The Condor 106:852±861 ᭧ The Cooper Ornithological Society 2004

THERMOREGULATION AND HABITAT PREFERENCE IN MOUNTAIN CHICKADEES AND TITMICE

SHELDON J. COOPER1 AND JAMES A. GESSAMAN Department of Biology, Utah State University, Logan, UT 84322-5305

Abstract. The Mountain Chickadee (Poecile gambeli) and the Juniper Titmouse (Baeo- lophus ridgwayi) are closely related, ecologically similar sympatric in portions of their range. However, Mountain Chickadees prefer higher altitude, cooler habitats than Juniper Titmice. We measured oxygen consumption, evaporative water loss, body temper- ature, and thermal conductance on seasonally acclimatized individuals to determine if ther- moregulatory differences correlate with habitat preference. The Mountain Chickadee's lower critical temperature was 4.2ЊC lower than the Juniper Titmouse's in summer and 2.4ЊC lower in winter. Thermal conductance decreased signi®cantly in winter relative to summer in Mountain Chickadees but not in Juniper Titmice. The Mountain Chickadee's upper critical temperature was 4.2ЊC lower than the Juniper Titmouse's in summer. Also in summer, Moun- tain Chickadees had signi®cantly higher body temperature above the upper critical temper- ature than Juniper Titmice, indicating less heat tolerance. The overall metabolic response to temperature in these two species suggests that physiology plays a role in maintaining their habitat segregation. Key words: ridgwayi, energy metabolism, evaporative water loss, habitat preference, oxygen consumption, Poecile gambeli, thermoregulation.

Termo-regulacioÂn y preferencia de haÂbitat en Poecile gambeli y Baeolophus ridgwayi Resumen. Las aves paserinas Poecile gambeli y Baeolophus ridgwayi, cercanamente emparentadas y ecoloÂgicamente similares, se distribuyen de modo simpaÂtrico en partes de sus rangos. Sin embrago, P. gambeli pre®ere ambientes maÂs elevados y frescos que B. ridgwayi. Medimos el consumo de oxõÂgeno, la peÂrdida de agua por evaporacioÂn, la tempe- ratura corporal y la conductancia teÂrmica en individuos aclimatados estacionalmente para determinar si las diferencias en termo-regulacioÂn se correlacionan con la preferencia de haÂbitat. La temperatura crõÂtica menor de P. gambeli fue 4.2ЊCmaÂs baja que la de B. ridgwayi en el verano y 2.4ЊCmaÂs baja en el invierno. La conductancia teÂrmica disminuyo signi®- cativamente en el invierno en relacioÂn al verano en P. gambeli pero no en B. ridgwayi.La temperatura crõÂtica mayor de P. gambeli fue 4.2ЊCmaÂs baja que la de B. ridgwayi en el verano. TambieÂn en el verano, P. gambeli tuvo una temperatura corporal signi®cativamente mayor, por arriba del lõÂmite superior de temperatura crõÂtica, que la de B. ridgwayi, indicando menor tolerancia al calor. La respuesta metaboÂlica global a la temperatura en estas dos especies sugiere que la ®siologõÂa juega un rol importante en mantener la segregacioÂn de sus ambientes.

INTRODUCTION son and Bartholomew 1968). However, Root Climate may in¯uence the biogeography, abun- (1988a, 1988b) provided data that indicate the dance, and habitat choice of physiologi- effect of climate on energy expenditure signi®- cally through its impact on energy and water cantly impacts species distribution and abun- balance, and/or ecologically through its impact dance. In addition, signi®cant energetic differ- on food availability and vegetation (Weathers ences are apparent in similar species from and van Riper 1982). For birds, the general dissimilar climates which may have an adaptive viewpoint is that climate limits birds through value (Weathers and van Riper 1982, Hayworth ecological and behavioral factors rather than by and Weathers 1984, Hinsley et al. 1993, Weath- physiological factors (Bartholomew 1958, Daw- ers and Greene 1998). The Mountain Chickadee (Poecile gambeli) and the Juniper Titmouse (Baeolophus ridgwayi) Manuscript received 10 March 2004; accepted 27 are small, mostly nonmigratory members of the July 2004. 1 Present address: Department of Biology and Mi- Paridae that occupy regions of western North crobiology, University of Wisconsin, Oshkosh, WI America (Fig. 1). The northern range extent of 54901-8640. E-mail: [email protected] the Juniper Titmouse may be in¯uenced by daily

[852] THERMOREGULATION IN CHICKADEES AND TITMICE 853

FIGURE 2. Annual temperature pro®les for Moun- tain Chickadee (higher elevation) and Juniper - mouse (lower elevation) study sites in northern Utah. (A) Mean daily maximum, (B) mean daily, (C) mean daily minimum, plotted for each month. Data are from FIGURE 1. Range of Mountain Chickadees and Ju- the Utah Climate Center, Logan, Utah, 1989±1994 (un- niper Titmice (McCallum et al. 1999, Cicero 2000). publ. data). energy expenditure especially during winter In the zone of elevational overlap in Utah (Cooper 2000). Throughout their range, Juniper where both species occupy pinyon-juniper Titmice prefer juniper or mixed pinyon-juniper woodlands, Mountain Chickadees prefer moister woodlands with mature that provide and cooler habitats than Juniper Titmice (SJC, cavities for nesting (Cicero 2000). The range of unpubl. data). For sites in northern Utah where Mountain Chickadees overlaps most of the range birds were captured for this study, mean daily of Juniper Titmice. However, chickadees gener- and mean daily minimum temperatures encoun- ally prefer pine and spruce-®r habitats (Dixon tered by the two species vary by as much as 4± 1961). Dixon (1961) stated that these two spe- 5ЊC seasonally (Fig. 2). In this study, we ex- cies are separated primarily by altitude in the amine the thermoregulatory abilities of these western United States. Mountain Chickadees species in relation to their respective environ- prefer elevations of 1800 to 3300 m (Bent 1946) ments. whereas Juniper Titmice prefer lower elevations Two-species comparisons, similar to the pre- ranging from around 700 to 2400 m (Cicero sent study have been criticized for statistical rea- 2000). Miller (1946) found that Mountain sons (Garland and Adolph 1994). However, Chickadees replaced Juniper Titmice above these studies are important tools in physiological 1980 m in the Grapevine Mountains along the ecology (Weathers and Greene 1998) and the -Nevada border and attributed these problems pointed out by Garland and Adolph habitat preferences to temperature zonation rath- (1994) can be addressed by performing meta- er than to vegetation community or structure. analyses on several two-species comparisons For example, patches of pinyon-juniper wood- (Gurevitch et al. 1992). lands above 1980 m were present at this site but were not occupied by Juniper Titmice. In addi- METHODS tion, Miller (1946) noted that Mountain Chick- Mountain Chickadees were captured in several adees occurred in several pinyon-juniper wood- locations within the Cache National Forest, lands in the western United States but generally Cache County, in northeastern Utah (41Њ52ЈN, not in the warmest locations examined, such as 111Њ30ЈW) at elevations of 2180 to 2250 m. Ju- the Kingston Mountains of California. niper Titmice were captured near Rosette, Box 854 SHELDON J. COOPER AND JAMES A. GESSAMAN

Elder County, in northwestern Utah (41Њ50ЈN, 3.0 cm above a 1-cm layer of paraf®n oil used 113Њ25ЈW) at an elevation of 1700 m. Mountain for the collection of fecal material. Oxygen con- Ç Chickadees and Juniper Titmice were captured sumption (VO2) was measured using open-cir- with mist nets in summer and winter of 1994, cuit respirometry with an Ametek Model S-3A 1995, and 1996. Mass at capture was measured oxygen analyzer (Pittsburgh, Pennsylvania). to the nearest 0.1 g with an Ohaus model CT- Dry, CO2-free air was drawn through the meta- 1200 (Pine Brook, New Jersey) portable elec- bolic chamber using a diaphragm pump. Outlet tronic balance. Following capture, birds were ¯ow rates of dry, CO2-free air were maintained transported to Logan, Utah, where they were by a Matheson precision rotameter (Model 604, housed individually in 0.3 ϫ 0.3 ϫ 0.3 m cages Montgomeryville, Pennsylvania) calibrated to placed in a 3 ϫ 3 ϫ 2.5 m temperature-con- Ϯ1.0% volumetrically (Brooks vol-u-meter, Hat- trolled environmental chamber. The environ- ®eld, Pennsylvania) and located downstream mental chamber was reprogrammed weekly to from the metabolic chamber. These ¯ow rates simulate the current photocycle and thermal re- yielded changes in oxygen content between in- gime of the study site. While in captivity, birds ¯ux and ef¯ux gas of 0.3% to 0.6% and main- were provided free access to water, grit, and tained oxygen content of ef¯ux gas above food (Tenebrio larvae and wild birdseed). Indi- 20.3%. Fractional concentration of oxygen in ef- viduals were tested within 1 week of capture ¯ux gas was determined from a 100-mL minϪ1 with birds tested from 17 May to 1 September subsample passed through the oxygen analyzer. designated ``summer birds,'' and those tested This subsample of ef¯ux gas was recorded every from 20 November to 1 March designated ``win- 15 sec using Datacan 5.0 data acquisition and ter birds.'' Summer titmice had an initial small analysis program (Sable Systems International, decrease (Ͻ5%) in mass following capture and Las Vegas, Nevada). Evaporative water loss was then maintained mass in captivity. All other determined over a 60-min timed interval by birds maintained mass while caged measuring the increase in mass of a downstream absorbent train containing Drierite (W.A. Ham- METABOLIC AND EVAPORATIVE WATER LOSS mond Drierite Company, Xenia, Ohio). All MEASUREMENTS weighings were made on an analytical balance Nighttime metabolic rate and evaporative water (Mettler H51AR, Columbus, Ohio). Weighings loss were measured for Mountain Chickadees were not recorded if any excreta were not com- and Juniper Titmice in both summer and winter. pletely covered by paraf®n oil. Measurements were made on individual birds Weathers et al. (1980) found that black paint using a 3.8-L metabolic chamber fashioned from used to achieve a high wall emissivity in the a paint can. The inside of the metabolic chamber metabolic chamber was hydroscopic, which was painted ¯at black to provide an emissivity leads to errors in evaporative water loss mea- near 1.0. Metabolic chamber temperature was surements. We determined the degree to which regulated within Ϯ0.5ЊC by placing it in a tem- our painted metabolic chambers were hydro- perature-controlled environmental chamber. scopic by placing a known mass of distilled wa- Metabolic chamber temperature was monitored ter in a container in the chambers at a range of continuously throughout each test with an Ome- temperatures similar to our experiments and ga (Stamford, Connecticut) thermocouple ther- measuring the increase in mass of a downstream mometer (Model Omni IIB, previously calibrat- absorbent train containing Drierite from the ed to a thermometer traceable to the U.S. Bureau evaporation of the water. We used the same ¯ow of Standards) attached to a 30-gauge copper- conditions as in the experiments. The mass loss constantan thermocouple inserted into the inlet of the water container was in close agreement port of the metabolic chamber. Metabolic rate of the mass gain in the Drierite absorbent train Ç was measured as oxygen consumption (VO2) (SJC, unpubl. data). Ç from 22:00 to 03:00 in summer and from 21:00 VO2 and evaporative water loss were mea- to 04:00 (MST) in winter. We fasted birds for at sured on individual birds exposed to a single least 4 hr prior to metabolic tests to insure post- ambient temperature (Ta) within a temperature absorptive conditions. Individuals were weighed range of Ϫ10ЊCto44ЊC. The order of tempera- and then placed inside the metabolic chamber tures selected was randomized. Each bird was where they perched on 1.0-cm wire mesh placed used only once during a 24-hr period and was THERMOREGULATION IN CHICKADEES AND TITMICE 855 tested no more than twice. No individuals were tersection of the regression line below and above tested twice within any one of the three temper- thermoneutrality, respectively, with a horizontal ature ranges of interest: below the lower critical line through mean BMR for each species and temperature, within the thermoneutral zone, or season. Published data for BMR (Cooper 2002), above the upper critical temperature. All indi- have been included to show the thermoneutral viduals were tested within 1 week of capture. At zone for each species in summer and winter. In the termination of each metabolic test, birds addition, the BMR data were analyzed for spe- were removed from the chamber and body tem- cies differences in addition to the seasonal dif- perature (Tb, to the nearest Ϯ0.1ЊC) was record- ferences reported in Cooper (2002). Data are re- ed with a 30-gauge copper-constantan thermo- ported as means Ϯ SE. We compared means ei- couple attached to an Omega Model HH25-TC ther using independent two-tailed t-tests or co- thermometer (previously calibrated to a ther- variance analysis (ANCOVA, with body mass as mometer traceable to the U.S. Bureau of Stan- a covariate). Regression lines were ®t by the dards). The thermocouple was inserted into the method of least squares. Slopes and intercepts cloaca to a depth where further insertion did not of regression lines were compared using AN- alter temperature reading (approximately 10±12 COVA. Statistical signi®cance was accepted at mm). Flow rates were maintained at 442±450 P Ͻ 0.05. All statistics were computed using mL minϪ1 for temperatures below 30ЊC and SPSS 6.1 (Norusis 1989). Due to the substantial 1096±1118 mL minϪ1 for temperatures above mass differences between species, all values of 30ЊC. These ¯ow rates maintained chamber dew heat production and evaporative water loss were point temperature below 12ЊC (Lasiewski et al. computed as mass-speci®c values. In addition, 1966). Individual birds were placed in the met- mass-independent values were calculated by di- abolic chamber for a total of 2 hr for tempera- viding rates of heat production and evaporative tures Յ30ЊC. Birds were allowed to equilibrate water loss by body mass raised to the 0.75 pow- Ç for the ®rst 60 min and VO2 was measured over er (Kleiber 1947, West et al. 1997, Weathers and the last 60 min of the trial. For metabolic trials Greene 1998). BMR and regression equations of Ç above 30ЊC individuals were placed in the cham- VO2 are also presented on a whole- basis ber for 60 min. The ®rst 10 min were used for because seasonal comparisons may be more ef- Ç equilibration (time needed for chamber to reach fective using per-bird VO2 (Dawson and Smith 99% equilibrium using equation of Lasiewski et 1986, Swanson 1991). Ç al. 1966) and VO2 was measured over the last 50 min of the trial. Oxygen consumption was RESULTS Ç calculated as steady state VO2 using Equation 4a BODY MASS of Withers (1977). All values were corrected for Mean body mass during nocturnal metabolic tri- standard temperature and pressure of dry gas. als for summer chickadees was 10.8 Ϯ 1.0g(n We calculated metabolic heat production assum- ϭ 29), which did not differ signi®cantly from ing 20.087 kJ of heat were produced per liter of the winter chickadee body mass of 11.1 Ϯ 1.0 oxygen consumed by fasting birds (Gessaman g(n ϭ 27, t54 ϭϪ1.1, P ϭ 0.28). Mean body and Nagy 1988). mass of titmice during nocturnal metabolic trials was signi®cantly higher in winter (17.2 Ϯ 1.1 g; STATISTICAL ANALYSES n ϭ 22 measurements of 16 titmice) compared Lower critical temperature (Tlc) and upper criti- to summer (16.0 Ϯ 1.0 g; n ϭ 33 measurements cal temperature (T ) were calculated using lo- uc of 23 titmice; t37 ϭϪ4.3, P Ͻ 0.001). cally weighted scatterplot smoothing (LO- WESS) regression followed by linear regression. METABOLIC RATE The LOWESS technique is useful because it Basal metabolism. BMR was 0.69 Ϯ 0.03 mL Ϫ1 Ϫ1 Ϫ1 makes no assumptions about the form of the un- O2 min (3.75 Ϯ 0.16 mL O2 g hr , n ϭ 14) derlying distribution (Cleveland 1985, Bakken et in summer chickadees and 0.81 Ϯ 0.04 mL O2 Ϫ1 Ϫ1 Ϫ1 al. 1991). LOWESS was used to determine the min (3.01 Ϯ 0.15 mL O2 g hr , n ϭ 16) in approximate Tlc and Tuc, and then metabolic rates summer titmice. BMR was signi®cantly higher between the two in¯ection points were pooled in summer chickadees compared to titmice (AN- and averaged as the basal metabolic rate (BMR). COVA: F1,27 ϭ 5.7, P ϭ 0.04). In winter, BMR Ϫ1 The Tlc and Tuc were then determined as the in- was 0.81 Ϯ 0.03 mL O2 min (4.36 Ϯ 0.15 mL 856 SHELDON J. COOPER AND JAMES A. GESSAMAN

The winter analyses below Tlc yielded the fol- lowing relations (Fig. 3B): Winter chickadees:

Ç Ϫ1 VO22 (mL O min ) ϭ 1.61 Ϫ 0.053Ta (3) (n ϭ 9, r2 ϭ 0.58, P Ͻ 0.01). Winter titmice:

Ç Ϫ1 VO22 (mL O min ) ϭ 2.03 Ϫ 0.053Ta (4) (n ϭ 10, r2 ϭ 0.71, P Ͻ 0.01). Interspeci®c comparisons of regression equa- tions for winter birds were not signi®cantly dif-

ferent in slopes (mass-speci®c: F1,15 ϭ 1.3, P ϭ 0.75 0.27; mass : F1,15 ϭ 0.7, P ϭ 0.43) or inter- 0.75 cepts on a mass basis (F1,16 ϭ 2.0, P ϭ 0.18). For chickadees, slopes were signi®cantly differ-

ent between seasons (mass-speci®c: F1,20 ϭ 8.0, FIGURE 3. Relationship between oxygen consump- 0.75 Ç P ϭ 0.02 ; mass : F1,20 ϭ 7.1, P ϭ 0.03) and tion (VO2) and ambient temperature (Ta) for (A) sum- mer-acclimatized and (B) winter-acclimatized Moun- intercepts were also signi®cantly different be- tain Chickadees and Juniper Titmice. Horizontal lines tween seasons (mass-speci®c: F1,21 ϭ 12.1, P ϭ 0.75 represent basal metabolic rate reported in Cooper 0.002; mass : F1,21 ϭ 11.2, P ϭ 0.003). For (2002). titmice, however, neither slopes (mass-speci®c: 0.75 F1,23 ϭ 0.0, P ϭ 1.0; mass : F1,23 ϭ 0.0, P ϭ 1.0) nor intercepts on a mass0.75 basis; F ϭ O gϪ1 hrϪ1, n ϭ 17) in chickadees and 0.99 Ϯ 1,24 2 3.2, P ϭ 0.09) were signi®cantly different be- 0.04 mL O minϪ1 (3.43 Ϯ 0.17 mL O gϪ1 hrϪ1, 2 2 tween seasons. The regression equations relating n ϭ 12) in titmice. BMR was signi®cantly high- heat production to T below thermoneutrality are er in winter chickadees compared to titmice a in Table 1. (ANCOVA: F1,26, ϭ 9.2 P Ͻ 0.01). Intraspeci®c comparisons show that winter birds had signi®- Lower critical temperature (Tlc) was calculat- cantly higher BMR compared to summer birds ed as the intersection of the regression line be- low thermoneutrality with a horizontal line (ANCOVA: chickadees, F1,28 ϭ 5.2, P ϭ 0.03; through mean BMR for each species and season, titmice, F1,25 ϭ 5.0, P ϭ 0.03). Below thermoneutrality. The summer analyses respectively. Tlc was 18.7ЊC in summer chicka- dees, 22.9ЊC in summer titmice, 14.7ЊC in winter below Tlc yielded the following relationships be- Ç chickadees, and 17.1 C in winter titmice. tween VO2 and Ta (Fig. 3A) Summer chicka- Њ dees: The mean nocturnal Tb within the thermoneu- tral zone and below the T of summer chicka- VOÇ (mL O minϪ1 ) ϭ 2.20 Ϫ 0.091T (1) lc 22 a dees was 35.5 Ϯ 1.8ЊC(n ϭ 18), which was not (n ϭ 15, r2 ϭ 0.71, P Ͻ 0.001). signi®cantly different from mean nocturnal Tb of summer titmice (35.9 Ϯ 2.6ЊC, n ϭ 24, t ϭϪ0.5, Summer titmice: P ϭ 0.60). The mean nocturnal Tb within the Ç Ϫ1 VO22 (mL O min ) ϭ 2.07 Ϫ 0.053Ta (2) thermoneutral zone and below Tlc of winter chickadees was 36.0 Ϯ 2.2ЊC(n ϭ 23), which 2 (n ϭ 17, r ϭ 0.75, P Ͻ 0.001). was not signi®cantly different than mean noc-

For interspeci®c comparisons of summer birds turnal Tb of winter titmice (35.3 Ϯ 2.0ЊC, n ϭ below thermoneutrality, the slopes of the two re- 18, t ϭ 1.1, P ϭ 0.29). gression lines were signi®cantly different (mass- Above thermoneutrality. The relationship be- 0.75 Ç speci®c: F1,28 ϭ 24.3, P Ͻ 0.001; mass : F1,28 tween VO2 and Ta above thermoneutrality is ϭ 7.5, P Ͻ 0.01) and the intercepts were signif- shown for summer and winter birds in Figure 3. icantly different (mass-speci®c: F1,29 ϭ 24.3, P The summer analyses above Tuc yielded the fol- 0.75 Ͻ 0.001; mass : F1,29 ϭ 14.2, P Ͻ 0.01). lowing relations: Summer chickadees: THERMOREGULATION IN CHICKADEES AND TITMICE 857

TABLE 1. Relationship of mass-speci®c (m W gϪ1) and mass-independent (m W gϪ0.75) heat production to ambient temperature (Ta, ЊC) outside the thermoneutral zone for seasonally acclimatized Mountain Chickadees and Juniper Titmice from northern Utah.

Regression Regression equation equation Species n (m W gϪ1) r2 P (m W gϪ0.75) r2 P Below thermoneutral zone Summer Chickadee 15 67.82±2.50Ta 0.65 Ͻ0.001 123.14±4.68Ta 0.67 Ͻ0.001 Titmouse 17 43.94±1.19Ta 0.80 Ͻ0.001 87.50±2.33Ta 0.79 Ͻ0.001 Winter Chickadee 9 48.82±1.66Ta 0.71 Ͻ0.01 88.93±3.01Ta 0.68 Ͻ0.01 Titmouse 10 39.49±1.19Ta 0.88 Ͻ0.001 80.42±2.34Ta 0.85 Ͻ0.01 Above thermoneutral zone Summer Chickadee 11 ±11.63 ϩ 1.04Ta 0.56 Ͻ0.01 ±21.35 ϩ 1.88Ta 0.56 Ͻ0.01 Titmouse 9 ±78.76 ϩ 2.66Ta 0.75 Ͻ0.01 ±163.89 ϩ 5.50Ta 0.73 Ͻ0.01

Ç Ϫ1 Ϫ1 Ϫ1 Ϫ1 VO22 (mL O min ) ϭϪ0.39 ϩ 0.034Ta (5) min (5.76 Ϯ 0.16 mL O2 g hr n ϭ 5) in chickadees and 1.08 Ϯ 0.18 mL O minϪ1 (3.96 2 2 (n ϭ 9, r ϭ 0.53, P Ͻ 0.01). Ϫ1 Ϫ1 Ç Ϯ 0.16 mL O2 g hr n ϭ 5) in titmice. VO2 Summer titmice: above 35ЊC was signi®cantly higher in winter chickadees compared to titmice (ANCOVA: F VOÇ (mL O minϪ1 ) 4.40 0.144T (6) 1,7 22 ϭϪ ϩ a ϭ 17.9, P Ͻ 0.01). The regression equations re- (n ϭ 11, r2 ϭ 0.64, P Ͻ 0.01). lating heat production to Ta above thermoneutra- lity are in Table 1. For interspeci®c comparisons of summer birds, Upper critical temperature (Tuc) in summer the slopes of the two regression lines were sig- birds was calculated as the intersection of the ni®cantly different (mass-speci®c: F1,16 ϭ 5.9, P regression line above thermoneutrality with a 0.75 ϭ 0.03; mass : F1,16 ϭ 7.1, P ϭ 0.02) but the horizontal line through mean BMR for each spe- intercepts were not signi®cantly different (mass- cies, respectively. T was 31.5ЊC in summer 0.75 uc speci®c: F1,17 ϭ 2.3, P ϭ 0.15; mass : F1,17 ϭ chickadees and 35.7ЊC in summer titmice. Tuc for 0.0, P ϭ 0.89). Interspeci®c comparisons of re- winter chickadees and titmice could only be es- gression equations for winter birds were not de- timated with LOWESS regression since metab- termined since heat production was not linearly olism above thermoneutrality was not a linear related to ambient temperature above 30ЊC. function of T (Fig. 3B). T was approximately However, the winter VÇ O value above ambient a uc 2 32ЊC for winter chickadees and 34ЊC for winter temperature of 35ЊC was 1.04 Ϯ 0.04 mL O 2 titmice. The pooled body temperatures for chickadees and titmice in summer and winter increased lin- early with increasing ambient temperature above 35ЊC (Fig. 4). The pooled data above 35ЊC yield- ed the following relationship:

Tba(ЊC) ϭ 26.30 ϩ 0.39T (7) (n ϭ 30, r2 ϭ 0.43, P Ͻ 0.01).

Mean body temperature above Tuc in summer chickadees was 42.9 Ϯ 0.6ЊC(n ϭ 11), which FIGURE 4. Relationship between body temperature was signi®cantly higher than summer titmice (Tb) and ambient temperature (Ta) above 35ЊC for Mountain Chickadees and Juniper Titmice in summer (41.4 Ϯ 0.3ЊC, n ϭ 9; t ϭ 2.2, P ϭ 0.04). Mean and winter. body temperature above 35ЊC in winter chicka- 858 SHELDON J. COOPER AND JAMES A. GESSAMAN

ent (mass-speci®c: F1,65 ϭ 1.3, P ϭ 0.25; 0.75 mass : F1,65 ϭ 0.0, P ϭ 0.98). The winter analyses of the relationship be-

tween log EWL and Ta yielded the following relations: Winter chickadees:

Ϫ1 Ϫ1 log EWL (mg g hr ) ϭ 0.645 ϩ 0.022Ta (10) (n ϭ 30, r2 ϭ 0.76, P Ͻ 0.001). Winter titmice:

Ϫ1 Ϫ1 log EWL (mg g hr ) ϭ 0.540 ϩ 0.022Ta (11) (n ϭ 17, r2 ϭ 0.94, P Ͻ 0.001). Interspeci®c comparisons of regression equa- tions for winter birds were not signi®cantly dif-

ferent in slopes (mass-speci®c: F1,43 ϭ 0.0, P ϭ FIGURE 5. Relationship between evaporative water 0.75 0.94; mass : F1,43 ϭ 0.0, P ϭ 0.86). The inter- loss and ambient temperature (Ta) for (A) summer-ac- climatized and (B) winter-acclimatized Mountain cepts were signi®cantly different on a mass-spe- Chickadees and Juniper Titmice. ci®c basis (F1,44 ϭ 6.4, P ϭ 0.02) but not on a 0.75 mass basis (F1,44 ϭ 2.0, P ϭ 0.17). For chick- adees, slopes were signi®cantly different be-

tween seasons (mass-speci®c: F1,59 ϭ 2.9, P ϭ dees was 41.2 Ϯ 0.2ЊC(n ϭ 5), which was not 0.03; mass0.75: F ϭ 2.7, P ϭ 0.03) and so were signi®cantly different than winter titmice (41.2 1,59 intercepts (mass-speci®c: F1,60 ϭ 5.0, P ϭ 0.03; Ϯ 0.5ЊC, n ϭ 5; t ϭ 1.8, P ϭ 0.12). 0.75 mass : F1,60 ϭ 4.9, P ϭ 0.03). For titmice, slopes were not signi®cantly different between EVAPORATIVE WATER LOSS seasons (mass-speci®c: F1,48 ϭ 0.6, P ϭ 0.46; 0.75 In summer, above 30ЊC the rate of evaporative mass : F1,48 ϭ 0.4, P ϭ 0.54) but intercepts water loss of both species increased exponen- were (mass-speci®c: F1,49 ϭ 13.4, P Ͻ 0.01; 0.75 tially (Fig. 5A), as is typical of endotherms. In mass : F1,49 ϭ 11.9, P Ͻ 0.01). winter, above 20ЊC the rate of evaporative water loss of both species increased exponentially THERMAL CONDUCTANCE (Fig. 5B). In order to compare the evaporative Overall thermal conductance below thermoneu- water loss response to varying temperature, the trality is equivalent to the slope of the line re- log EWL versus Ta for both species was deter- lating VÇ O to ambient temperature only if the mined. The summer analyses of the relationship 2 curve extrapolates to Tb at zero metabolism. between log EWL and Ta yielded the following Since the metabolic data from chickadees and relations: Summer chickadees: titmice did not conform to the Newton-Scholan- Ϫ1 Ϫ1 log EWL (mg g hr ) ϭ 0.846 ϩ 0.018Ta (8) der cooling model (Scholander et al. 1950) we calculated overall thermal conductance (Kes) for (n ϭ 33, r2 ϭ 0.73, P Ͻ 0.001). individuals using the equation of Bakken (1976) Summer titmice: Kes ϭ (M Ϫ E)/(TbaϪ T ) (12) Ϫ1 Ϫ1 log EWL (mg g hr ) ϭ 0.745 ϩ 0.020Ta (9) where M is metabolic rate and E is evaporative heat loss (assuming 2.43 J of heat for each mg (n ϭ 35, r2 ϭ 0.78, P Ͻ 0.001). of water evaporated). Thermal conductance be- low thermoneutrality was 1.57 Ϯ 0.13 mW gϪ1 For interspeci®c comparisons of summer ЊCϪ1 for summer chickadees (n ϭ 15), which birds, the slopes of the two regression lines were was signi®cantly higher than 1.02 Ϯ 0.07 mW Ϫ1 Ϫ1 not signi®cantly different (mass-speci®c: F1,64 ϭ g ЊC for summer titmice (n ϭ 17, t ϭ 3.9, P 0.75 1.0, P ϭ 0.33; mass : F1,64 ϭ 0.1, P ϭ 0.34) Ͻ 0.01). Thermal conductance for winter chick- and the intercepts were not signi®cantly differ- adees was 1.21 Ϯ 0.09 mW gϪ1 ЊCϪ1 (n ϭ 9), THERMOREGULATION IN CHICKADEES AND TITMICE 859 which was not signi®cantly different from win- ported for birds (Weathers 1981, Weathers and ter titmice (1.10 Ϯ 0.03 mW gϪ1 ЊCϪ1, n ϭ 10, van Riper 1982, Hayworth and Weathers 1984). t ϭ 1.6, P ϭ 0.14). Thermal conductance was For Juniper Titmice in northern Utah (this not signi®cantly different between summer and study), a Tuc of 35.7ЊC in summer is also lower winter titmice (t ϭ 1.1, P ϭ 0.28) but was sig- than several other passerines (see Weathers ni®cantly lower in winter chickadees compared 1981). In comparison, Juniper Titmice from to summer (t ϭϪ2.2, P ϭ 0.04). southeastern have a Tuc between 38ЊC and 40ЊC (Weathers and Greene 1998). Given DISCUSSION the very low Tuc of summer chickadees it is sur- Ç METABOLIC RESPONSE TO TEMPERATURE prising that the slope relating VO2 to Ta was ac- BELOW THERMONEUTRALITY tually less steep than titmice. This slope is re- Lower critical temperature in both species varied ferred to as the coef®cient of heat strain (Weath- with acclimatization state and was lowest in ers 1981) with steeper slopes re¯ecting lower winter. The Mountain Chickadee's Tlc was 4.2ЊC heat tolerance. The coef®cient of heat strain for lower in summer and 2.4ЊC lower in winter than summer chickadees is only 40% of the allome- the Juniper Titmouse's. In addition, the Tlc for tric prediction, while the coef®cient of heat Mountain Chickadees was 2.2ЊC and 6.4ЊC low- strain for summer titmice is 129% of the pre- er than predicted values based on body mass dicted value (Weathers 1981). The low coef®- (Weathers and van Riper 1982) for summer and cient of heat strain for summer chickadees is winter, respectively. The Tlc for Juniper Titmice very similar to values reported for Bridled Tit- was 3.4ЊC higher in summer and only 1.3ЊC low- mice (69% of predicted) and Juniper Titmice er than predicted values based on body mass (50% of predicted) from southeastern Arizona (Weathers and van Riper 1982). The slope and (Weathers and Greene 1998). intercept of the regression line relating metabol- In spite of their low coef®cient of heat strain ic rate to ambient temperature below thermo- in summer, Tb above the Tuc for Mountain Chick- neutrality varied seasonally in chickadees but adees was signi®cantly higher than Juniper Tit- not in titmice. This suggests that chickadees mice. Body temperature above the Tuc averaged have better insulation in winter, probably as re- 5.1ЊC higher than within the thermoneutral zone sult of increased plumage mass, but increased for summer chickadees but only 3.0ЊC for sum- plumage mass in titmice does not increase in- mer titmice. Increased levels of hyperthermia sulation (Cooper 2002). Minimal dry thermal permit chickadees to lose more heat by none- conductance below thermoneutrality also de- vaporative pathways than titmice (Weathers creased signi®cantly in winter chickadees but 1981). However, in spite of hyperthermia, chick- not in titmice. Overall insulative capacity in adees did not have any reduction in evaporative summer is greater in titmice than in chickadees water loss compared to titmice above Tuc indi- but in winter is not signi®cantly different be- cating more heat stress in chickadees above Tuc Ç tween the two species based on values of overall compared to titmice. In addition, VO2 above minimum thermal conductance. This suggests 35ЊC was signi®cantly higher in winter chicka- that seasonal changes in insulation are involved dees compared to winter titmice. Similar heat with winter acclimatization of the Mountain intolerance has been found in three Hawaiian Chickadee but not of the Juniper Titmouse. honeycreepers, the Palila (Loxioides bailleui; However, winter values of thermal conductance Weathers and van Riper 1982) and the Hawai`i exceeded those allometrically predicted for pas- and Kauai `Amakihi, (Hemignathus virens and serines (Aschoff 1981) by 26% for titmice and H. kauaiensis; MacMillen 1974), which are re- 13% for chickadees indicating that insulative stricted to cool, high forests or montane habitats. changes are probably not prominently involved with winter acclimatization in these two species. EVAPORATIVE WATER LOSS Evaporative water loss was not different be- METABOLIC RESPONSE TO TEMPERATURE tween summer chickadees and titmice. Both spe- ABOVE THERMONEUTRALITY cies had relatively high rates of evaporative wa- Mountain Chickadees appear to be markedly ter loss. For example, for summer birds evapo- heat intolerant. The Tuc for summer chickadees rative water loss above Tuc was approximately was only 31.5ЊC, which is one of the lowest re- 2.0ϫ higher in birds from this study compared 860 SHELDON J. COOPER AND JAMES A. GESSAMAN to Juniper Titmice from southeastern Arizona cadian phase. Comparative Biochemistry and (Weathers and Greene 1998). Whether this de- Physiology 69A:611±619. BAKKEN, G. S. 1976. A heat transfer analysis of ani- creased evaporative water loss in southern pop- mals: unifying concepts and the application of me- ulations that inhabit warmer, drier climates is a tabolism chamber data to ®eld ecology. Journal of common feature of the avian thermoregulatory Theoretical Biology 60:337±384. response to heat is uncertain. In winter, titmice BAKKEN, G. S. 1991. The effect of wind and air tem- had lower evaporative water loss than chicka- perature on metabolism and evaporative water loss rates of Dark-eyed Juncos, Junco hyemalis:a dees on a mass-speci®c basis but not on a mass- standard operative temperature scale. Physiologi- independent basis. Winter titmice and chicka- cal Zoology 64:1023±1049. dees had lower evaporative water loss over the BARTHOLOMEW, G. A. 1958. The role of physiology in entire range of temperatures tested compared to the distribution of terrestrial vertebrates, p. 81±95. In C. L. Hubbs [ED.], Zoogeography. American their summer counterparts. It is possible that cu- Association for the Advancement of Science, Pub- taneous water loss decreases in winter because lication Number 51. Washington, DC. increased plumage mass retards water-vapor dif- BENT, A. C. 1946. Life histories of North American fusion. However, for small birds, the total resis- jays, crows, and titmice. U.S. National Museum tance of plumage and the boundary layer of still Bulletin 191. CICERO, C. 2000. (Baeolophus inorna- air around the bird is generally quite low (Web- tus) and Juniper Titmouse (Baeolophus ridgwayi). ster et al. 1985). In A. Poole and F. Gill [EDS.], The birds of North Mountain Chickadees have greater cold tol- America, No. 485. The Birds of North Ameri- erance than Juniper Titmice as demonstrated by ca,Inc., Philadelphia, PA. a decreased lower critical temperature in both CLEVELAND, W. S. 1985. The elements of graphing data. Wadsworth, Monterey, CA. summer and winter and seasonal changes in COOPER, S. J. 2000. Seasonal energetics of Mountain thermal conductance. This could be advanta- Chickadees and Juniper Titmice. Condor 102: geous for chickadees that inhabit higher altitudes 635±644. since climate may be more unpredictable than COOPER, S. J. 2002. Seasonal metabolic acclimatization lower sites occupied by titmice. In addition, in Mountain Chickadees and Juniper Titmice. Physiological and Biochemical Zoology 75:386± chickadees have signi®cantly higher maximum 395. thermogenic capacity than titmice in both sum- DAWSON,W.R.,AND G. A. BARTHOLOMEW. 1968. Tem- mer and winter (Cooper 2002). On the other perature regulation and water economy of desert hand, Juniper Titmice appear to be better able to birds, p. 357±394. In G.W. Brown Jr. [ED.], Desert tolerate higher temperatures than Mountain biology. Academic Press, New York. DAWSON,W.R.,AND B. K. SMITH. 1986. Metabolic Chickadees. Summer titmice have a higher up- acclimatization in the American Gold®nch (Car- per critical temperature than summer chickadees duelis tristis), p. 427±434. In H. C. Heller, X. J. and also have lower body temperature under Musachia, and L. C. H. Wang [EDS.], Living in the heat stress. Thus, the overall metabolic response cold: physiological and biochemical adaptations. to temperature in these two species appears to Elsevier, Amsterdam. DIXON, K. L. 1961. Habitat distribution and niche re- play a role in maintaining their habitat prefer- lationships in North American species of Parus, ences. p. 179±216. In W. F. Blair [ED.], Vertebrate spe- ciation. University of Press, Austin, TX. ACKNOWLEDGMENTS GARLAND, T., JR., AND S. C. ADOLPH. 1994. Why not to do two-species studies: limitations on inferring Birds were captured under state and federal scienti®c adaptation. Physiological Zoology 67:797±828. collecting permits: UT 2COLL1401, USFWS PRT- GESSAMAN,J.A.,AND K. A. NAGY. 1988. Energy me- 779300. Thanks to Keith Dixon for providing infor- tabolism errors in gas-exchange conversion fac- mation on study sites and to Glenn Wilson for lab as- tors. Physiological Zoology 61:507±513. sistance. Kim Sullivan and two anonymous reviewers GUREVITCH, J., L. L. MORROW,A.WALLACE, AND J. S. provided many useful suggestions on an earlier draft WALSH. 1992. A meta-analysis of competition in of this manuscript. This study was supported in part ®eld experiments. American Naturalist 140:539± by research funding from the Ecology Center at Utah 572. State University to JAG and a North American Blue- bird Society student research grant and a Sigma Xi HAYWORTH,A.M.,AND W. W. W EATHERS. 1984. Tem- Grant-in-Aid of research to SJC. perature regulation and climatic adaptation in Black-billed and Yellow-billed Magpies. Condor LITERATURE CITED 86:19±26. HINSLEY, S. A., P. N. FERNS,D.H.THOMAS, AND B. ASCHOFF, J. 1981. Thermal conductance in mammals PINSHOW. 1993. Black-bellied Sandgrouse (Pte- and birds: its dependence on body size and cir- rocles orientalis) and Pin-tailed Sandgrouse (Pte- THERMOREGULATION IN CHICKADEES AND TITMICE 861

rocles alchata): closely related species with dif- tic and tropical mammals and birds. Biological fering bioenergetic adaptations to arid zones. Bulletin 99:237±258. Physiological Zoology 66:20±42. SWANSON, D. L. 1991. Seasonal adjustments in metab- KLEIBER, M. 1947. Body size and metabolic rate. Phys- olism and insulation in the Dark-eyed Junco. Con- iological Reviews 27:511±541 dor 93:538±545. LASIEWSKI, R. C., A. L. ACOSTA, AND M. H. BERNSTEIN. WEATHERS, W. W. 1981. Physiological thermoregula- 1966. Evaporative water loss in birdsÐI. Char- tion in heat-stressed birds: consequences of body acteristics of the open ¯ow method of determi- size. Physiological Zoology 54:345±361. nation, and their relation to estimates of thermo- WEATHERS,W.W.,AND E. GREENE. 1998. Thermoreg- regulatory ability. Comparative Biochemistry and ulatory responses of bridled and juniper titmice to Physiology 19:445±457. high temperature. Condor 100:365±372. MACMILLEN, R. E. 1974. Bioenergetics of Hawaiian WEATHERS, W. W., C. J. SHAPIRO, AND L. B. ASTHEI- honeycreepers: the Amakihi (Loxops virens) and MER. 1980. Metabolic responses of Cassin's Finch- the Anianiau (L. parva). Condor 76:62±69. es (Carpodacus cassinii) to temperature. Compar- MCCALLUM, D. A., R. GRUNDEL, AND D. L. DAHLSTEN. ative Biochemistry and Physiology 65A:235±238. 1999. Mountain Chickadee (Poecile gambeli). In WEATHERS,W.W.,AND C. VAN RIPER III. 1982. Tem- A. Poole and F. Gill [EDS.], The birds of North perature regulation in two endangered Hawaiian America, No. 453. The Birds of North Ameri- honeycreepers: the Palila (Psittirostra bailleui) ca,Inc., Philadelphia, PA. and the Laysan Finch (Psittirostra cantans). Auk MILLER, A. H. 1946. Vertebrate inhabitants of the pi- nÄon association in the Death Valley region. Ecol- 99:667±674. ogy 27:54±60. WEBSTER, M. D., G. S. CAMPBELL, AND J. R. KING. NORUSIS, M. J. 1989. SPSS for windows. SPSS Inc., 1985. Cutaneous resistance to water-vapor diffu- Chicago. sion in pigeons and the role of the plumage. Phys- ROOT, T. 1988a. Energy constraints on avian distribu- iological Zoology 58:58±70. tions and abundance. Ecology 69:330±339. WEST, G. B., J. H. BROWN, AND B. J. ENQUIST. 1997. ROOT, T. 1988b. Environmental factors associated with A general model for the origin of allometric scal- avian distributional boundaries. Journal of Bio- ing laws in biology. Science 276:122±126. Ç Ç geography 15:489±505. WITHERS, P. C. 1977. Measurement of VO2,VCO2, and SCHOLANDER, P. R., R. HOCK,V.WALTERS,F.JOHNSON, evaporative water loss with a ¯ow-through mask. AND L. IRVING. 1950. Heat regulation in some arc- Journal of Applied Physiology 42:120±123.