Thermal Environment of the Nest During Development of Tree Swallow (Tachycineta bicolor) Chicks

845 The Auk 119(3):845 851, 2002 Thermal Environment of the Nest During Development of Tree Swallow (Tachycineta bicolor) Chicks SUSAN B. CHAPLIN, 1 MORA L. CERVENKA, 2 AND ALISON C. MICKELSON 3 Department of Biology, University of St. Thomas, St. Paul, Minnesota 55105, USA ABSTRACT. We continuously recorded temperatures of the nest cup (T n )andair(t a )justoutsidethe nest box throughout development of Tree Swallow (Tachycineta bicolor)chicksineast-centralminnesota. Brood size was manipulated (three or six chicks per nest) to study the effect of number of chicks in the nest on its thermal environment. From day 0 to day 4, T n paralleled T a,approximately2 7Chigher,with nocturnal T n maintained at 24C and daytime T n maintained at 26C. From day 6 to day, T n was relatively constant throughout the day, maintained at 32C (day)and29c (night).thet n became increasingly independent of T a,asdeterminedbyregression analysis of T n versus T a with age. Nocturnal T n of larger clutches (six chicks) was 2C warmerthan smaller clutches (three chicks), and T n of larger clutches exhibited greater independence from T a at 1 E-mail: sbchaplin@stthomas.edu 2 Present address: School of Dentistry, University of Minnesota, St. Paul, Minnesota 55414, USA. 3 Present address: School of Nursing, University of Minnesota, St. Paul, Minnesota 55414, USA. day 10 than in smaller clutches. The occurrence of nest homeothermy at day 10 in six-chick nests correlates with near-maximal body mass of chicks and completion of feather insulation. Small (three-chick) clutches showed greater dependence of T n on T a at 10 days of age than large (six-chick) clutches; we propose that development of thermoregulatory capacity may have proceeded more slowly in chicks from those nests. RESUMEN. Registramos continuamente la temperatura de la copa del nido (T n )ydelaire(t a )inmediatamente exterior a la caja de nidificación durante el período de desarrollo de polluelos de Tachycineta bicolor en el centro-este de Minnesota. Manipulamos el tamaño de la nidada (tresoseispolluelos por nido) para estudiar el efecto del número de polluelos en el nido sobre el ambiente térmico de éste. Desde el día 1 al día 4, T n fluctuó enformaparalela con T a,perofueaproximadamente2 7Cmayor; la T n nocturna se mantuvo a 24C mientrasquela diaria a 26C. Desde el día 6 hasta el día, T n fue relativamente constante a través del día, manteniéndose a 32C duranteeldía y a 29C durantelanoche.

846 Notes [Auk, Vol. 119 Mediante un análisisde regresióndet n versus T a determinamos que la independencia entre T n y T a fue incrementando con el tiempo. La T n nocturna de las nidadas grandes (seis polluelos) fue 2C más alta que la de las nidadas pequeñas (tres polluelos), y al décimo día la T n de las nidadas grandes presentó una mayor independencia de la T a que la de las nidadas pequeñas. La presencia de homeotermia en el décimo día de los nidos con seis polluelos se correlacionó con la masa corporal máxima de los polluelos y con la finalización del aislamiento por plumas. Con 10 días de edad, T n mostró unamayordependencia de la T a en las nidadas pequeñas (tres polluelos)que en las nidadas grandes (seis polluelos). Proponemos que el desarrollo de la capacidad termoregulatoria podría haber procedido más lentamente en polluelos de nidadas más pequeñas. Tree Swallows (Tachycineta bicolor)beginbreeding in April in Minnesota and the first young hatch approximately 1 June, when air temperatures at night may range from 5 15C. On cool days and during inclement weather, average daytime air temperatures at that latitude may only be 15 20C, posing significant risk of cooling to newly hatched chicks without an attendant brooding parent. Perhaps because of the risk of hypothermia associated with early breeding at northern latitudes, Tree Swallows line their cavity nests of dry grass with feathers (Robertson et al. 1992). The importance of that feather insulation was documented by Lombardo et al. (1995), who found that Tree Swallow chicks from nests with artificially reduced feather insulation had lower growth rates and were smaller and less mature at day (approximately two-thirds of the nestling period). Few measurements of nest thermal environment during posthatch chick development have been reported. Calder (1971) recorded temperatures of artificial eggs inserted into two Calliope Hummingbird (Stellula calliope) nestsbycontinuousrecording during incubation, brooding of young chicks, and late postnatal development. However, most studies have focused on nest or egg temperatures during laying and incubation periods (Haftorn 1978, Cooper and Afton 1981, Zann and Rossetto 1991, Smith and Montgomerie 1992, Flint and Maccluskie 1995, Wilson and Verbeek 1995). This study reports results of continuous recording of nest cup and air temperature (at the surface of the nest box) throughout Tree Swallow nestling development and as a function of clutch size. On the basis of the premise that nest temperature is critical to nestling development, several questions have been investigated with these data. (1) How is nest temperature influenced by the ambient environment (air temperature, variation in air temperature, or time of day)? (2) Does the number of chicks in the nest affect the thermal environment of the nest during development? (3) How is nest temperature influenced by growth and maturation of chicks that is, does the pattern of nest temperature change daily during the nestling period, specifically with changes in nestling biomass and development of endothermy? Methods. Tree Swallows were studied at the Twin Cities Army Ammunition Plant, New Brighton, Minnesota, from 1995 through 1997. Over 400 bluebird (Sialia sialis) nestboxeshadbeeninstalledonthe 2,400 acre site during the previous six years. Pairs of nests in close proximity and with similar hatch dates were selected for continuous temperature monitoring. Clutch size was reduced to three young in one nest, and was held at six in the other. Data loggers (Model ML-1, Mini-Mitter Co., Bend, Oregon) were installed in six pairs of nests to record air and nest temperatures throughout chick development. Prior to their installation in nest boxes, probes for the data loggers were calibrated in a water bath against a reference thermometer over a temperature range of 20 40C, were programmed to log temperatures 10 times an hour throughout the day, and synchronized with real time. Response time of the external nest probe was 1 s.usingfemalenesting behavior as a guide to approximate hatch time (S. B. Chaplin pers. obs. during breeding seasons 1993 and 1994), the day before predicted hatch time, we drilled aholeinthebackoftheboxandthreadedasingle thermocouple probe through the grass cup and into the bottom of the nest, where it rested slightly above the feathered nest cup lining and in the middle of the clutch of eggs or hatched young. The entry point was taped so that the probe was securely positioned (i.e. would not move when the eggs or chicks moved), and the data logger (with its own internal temperature sensor for monitoring air temperature) was secured to the outside of the nest box with duct tape. The logger was left in place until the chicks fledged (17 18 days in those nests), and the position of the nest probe with respect to the chicks was checked every other day when the chicks were weighed. Logger data (date, time, nest and air temperatures) was downloaded to a computer (using Mini-Mitter DA- TACAN software) as a DOS text file, which was then imported into an EXCEL file for graphical representation of daily changes in nest and air temperature for each nest, and then imported into STATVIEW and MINI-TAB for statistical analysis. To evaluate the effect of air temperatures on nest temperature throughout development, mean nest and air temperatures were computed for two periods only for each nest: 0000 0400 CST (a stable period of nocturnal air temperature) and 0600 00 (representing daytime air temperatures that had the greatest daily variation). Daytime and evening data (00 0000) were not used in this analysis because loggers were often in direct sun during part of that time and did not provide accurate air temperature measurement. Mean nest and air temperatures of the two selected periods and their coefficients of variation (CV)

847 were used in a repeated measures ANOVA, in which the effect of clutch size, time of day, and age on nest temperature (T n )orcoefficientofvariationofnest temperature (CVT n ) were analyzed. Because of a highly significant correlation of air temperature and nest temperature (P 0.0001), analysis of covariance (ANCOVA) was also performed with air temperature as a covariable. Results. The pattern of nestling development in Tree Swallow chicks in central Minnesota is similar to that reported by Zach and Mayoh (1982) and Robertson et al. (1992). At hatch, six chicks filled approximately one third of the volume of the nest cup whereas three chicks occupied far less volume. By day 6, six chicks completely filled the nest cup, whereas three chicks only completely filled the volume of the nest cup by day 8. Rapid increases in body mass occurred up to day 10 in three-clutch nests, when peak body mass of 22 g was achieved (Table 1). Chicks in six-clutch nests lagged slightly behind, reaching peak body mass of 22 g on day (Table 1). Chicks in three-clutch nests were significantly heavier than those from six-clutch nests on day 6, 10, and prior to fledging on day 16 (P 0.01 by unpaired t-test, Table 1). However, there was no difference in chronology of morphological development between the two groups of chicks. Feathers erupted from spinal tracts first, beginning on day 6, and the nestling was completely feathered by day to 14. Primary feather growth began on day 7 8, feathers completely erupted from their sheaths by day 14, and continued to elongate until chicks fledged. Chicks became active in the nest beginning on day 10 and jumped to the nest hole to be fed when parents arrived by day. They jumped or flew from the nest box spontaneously when disturbed, beginning on day 16, and chicks in each nest fledged naturally shortly thereafter (17 or 18 days in all nests). Figure 1 (A D) illustrates the pattern of daily nest and air temperature (T n and T a,respectively)change for representative days in one nest of six chicks. Mean nocturnal T a increased throughout the nestling period in early to mid June from a mean of 11 to 20C, whereas mean daytime air temperatures (0600 00 hours) did not change appreciably throughout the nestling period (mean 22C). Early in chick development (day 0 4), T n paralleled T a,but 2 7C higher,suchthatnocturnal T n was maintained at 23.5 1.0 (SE) C anddaytimet n reached only 26.1 0.9 (SE) C (Fig.1A,B).However, from day 6 to day, T n was relatively constant throughout the day, maintained at a mean of 31.9 0.7 (SE) C during the day and 29.7 0.8 (SE) C atnight(fig.1c, D). Late in development (day 16 18), daytime T n was only 4C greaterthant a (mean of all nests 25.3 1.5 [SE] C) and nocturnal T n was actually lower than at any period other than day 0 4 (24.6 1.6 [SE] C). During nestling development, mean T n was most significantly correlated (P 0.0001) with mean T a (r TABLE 1. Growth of three- and six-clutch Tree Swallow chicks in boxes with data loggers (1995 1997). Data are means one SE; n the number of chicks measured. Age (days) 0 2 4 6 8 10 14 16 Three-clutch Mass (g) (SE) 1.44 (0.11) 4.06 (0.17) 7.80 (0.48) 13.67 (0.30) 18.87 (0.22) 22.14 (0.43) 22.33 (0.92) 20.58 (1.6) 22.70 (1.1) n 9 18 9 15 9 Six-clutch Mass (g) (SE) 1.46 (0.19) 3.78 (0.18) 8.55 (0.37).17 (0.49) 17.72 (0.71) 20.06 (0.40) 22.46 (0.35) 21.18 (0.41) 20.70 (0.26) n 24 24 24 0.54) and maximum T a (r 0.53) and less significantly related to minimum T a (r 0.33, P 0.0009) or the coefficient of variation in T a (r 0.23, P 0.00). Mean daytime (0600 00) and nocturnal (0000 0400) T n increased from hatch day to day 10 (Figs. 2 and 3), then decreased slightly until fledging. Both daytime and nocturnal T n means stabilized at 30C (see above) from day 6 to day of nestling life. After day, the mean daytime T n actually decreased (Fig. 2) as chicks became more active and were less likely to sit near the probe in the bottom of the nest cup. In anestedanalysisofcovariance(ancova)witht a as the covariable, chick age was the most significant factor explaining variation in T n (P 0.0001), and clutch size was just barely significant (P 0.042; see discussion below). Neither time of day nor the interaction of any of the factors was significant in the ANCOVA. As an indicator of influence of chick activity on T n, we found that the nest was significantly warmer with chicks in the nest than when it was empty (postfledge day 18 compared to day 16, just prior to fledging), using a paired t-test (P 0.0001). In contrast, nest temperature of newly hatched chicks (day 0) was not significantly different than incubation T n (one day prior to hatch), using a paired t-test (P 0.65). Number of young in the nest did have a significant effect on T n, as determined by the ANCOVA described above (P 0.042), largely due to differences in nocturnal regulation of nest temperature between three- and six-clutch nests. Daytime T n of three- and six-clutch nests was remarkably similar throughout nestling life (Fig. 2). However, the mean nocturnal T n for three-clutch nests averaged 2.1C coolerthan sixclutch nests over the entire nestling period (Fig. 3). In addition, mean nocturnal T n of three-clutch nests never reached the six-clutch daytime average of 30C, and was only as high as 28C onday6and8ofnestling life. Thus, a six-clutch nest presented a warmer thermal environment at night, and was not as highly

848 Notes [Auk, Vol. 119 FIG. 1. Nest(opencircles)andair(closedcircles)temperaturesovera24hperiodduringselecteddays of nestling life in a six-chick nest of Tree Swallows hatched 31 May 1997. Time is CDT; sunrise and sunset are indicated by arrows on the abscissa and increased a total 10 min over the 10 days in this record. correlated with T a (days 0 10) as the nest environment of the three-clutch chicks was (r for six-clutch nest 0.32, P 0.0019 ; r for three-clutch nest 0.60, P 0.0001). The larger number of chicks in a six-clutch nest maintained a 10 13C nocturnaldifferential between T n and T a, whereas three-clutch nests maintained only 8 11C differential during days 4 10 of nestling life. Regulation of nest temperature is defined by maintenance of a relatively stable T n,independentof variation in T a.althought n is strongly correlated with T a over the course of the entire nestling period, that relationship changes as a function of chick age and clutch size. Linear regressions of T n and daytime T a (0600 00) split by age and clutch size show similar patterns of increasing y-intercept (predicted nest temperature at T a of 0C) and decreasing slope (i.e. less change in T n with change in T a )aschicksgetolder and larger (Fig. 4). Whereas the slopes of T n versus T a are greater than 0.6 for chicks on days 0 2, slopes of the regressions approach 0 in older chicks 8 to 10 days of age. That pattern disappears after 10 days of age when chicks become active in the nest and the nest probe no longer reflects the temperature of huddled chicks. Although the pattern of decreasing slopes of T n versus T a regressions with age is similar in threeand six-clutch nests, the y-intercepts of the T n versus T a regressions at each age are significantly greater in the six-clutch nests, meaning nests were warmer at all T a (t-test comparison of intercepts, P 0.05; Fig. 4). In addition, slope of T n versus T a regressions were 0.6 for three-clutch chicks up to day 8, whereas the slope dropped to 0.3 by day 4 in six-clutch chicks (Fig. 4). Slopes of the T n versus T a regressions for sixchick nests were significantly less than those for three-chick nests on day 4, 6, and 8, meaning sixchick T n exhibited greater independence from T a (ttest comparison of slopes, P 0.05). Discussion. Creation of a favorable and stable thermal environment for developing eggs and chicks is integral to the reproductive success of avian species (Ricklefs 1974, O Connor 1984). Assuming food is not limited, chicks will grow and mature fastest at a particular optimal temperature, specific to that species. From continuous records of nest temperature taken during the Tree Swallow nestling period, we have demonstrated that mean nest temperature during both day and night was maintained within 2C of30cduringdays6 14ofthenestlingperiod

849 FIG. 2. Meandaytime(0600 00)temperatures of nest (closed symbols) and air (open symbols) at three-chick (squares) and six-chick (circles) Tree Swallow nests as a function of age. Bars indicate one standard error above or below the mean. FIG.4. Nesttemperatureasafunctionofmorning air temperature (0600 00) during development of Tree Swallow chicks in six-clutch (top) versus threeclutch (bottom) nests. Data are from six nests of each clutch size throughout their development, with days 2(circle),6(square),and10(triangle)represented. Regressions are best-fit linear relationships. The line of equality y x is represented for comparison. FIG. 3. Meannocturnal(0000 0400)temperatures of nest (closed symbols) and air (open symbols) at three-chick (squares) and six-chick (circles) Tree Swallows nests as a function of age. Bars indicate one standard error above or below the mean. (Figs. 2 and 3). Early in postnatal development (day 0today6,orupto11gbodymass),broodingby parents helps raise the temperature of the nest (Dunn 1979, Marsh 1980, Robertson et al. 1992). Maintenance of a nocturnal nest thermal environment 30C wasfirstachievedatday6oftreeswallow nestling life, the age when body mass of the chicks completely fills the nest cup. From ages 6 to 14 days, chicks were brooded infrequently during the daytime (S. B. Chaplin pers. obs.), but six-chick nests nevertheless maintained stable nest temperatures for much of the day and night. Only when chicks were completely feathered and more active in the nest (from day 14 on) did the nest thermal environment again fluctuate with daily changes in air temperature. Several studies of the thermal environment of the nest have been conducted during the egg-laying and incubation phases, but comparative data on the nestling thermal environment of altricial species are lacking. Nest temperatures during incubation in several avian species measured 30 35C, somewhat higher than the posthatching nest temperatures recorded in this study (see Haftorn 1978, Smith and Montgomerie 1992, Blem and Blem 1994, Brown 1994, Wilson and Verbeek 1995). Calder (1971) implanted thermocouples in fake eggs added to Calliope Hummingbird nests and measured mean egg temperatures of 30 35C during incubation and young chick brooding stages. Thus, the microenvironment of the nest cup surrounding eggs (other studies) and developing chicks (this study) appears to be regulated at a level of approximately 30 35C in several avian species.

850 Notes [Auk, Vol. 119 In this study, we have also demonstrated a significant effect of clutch size on nest temperature. The mean nocturnal T n of three-clutch nests was 2Clower on average than that of six-clutch nests over the entire nestling period (Figs. 2 and 3), and the difference between nocturnal nest and air temperatures was also 2C lessduringdays4 10ofnestlinglife in three-clutch nests than in six-clutch nests. However, a 2C coolernesttemperature did not adversely affect growth of three-clutch chicks which, in this study, were significantly heavier than six-clutch chicks at 10 days (2.1 g) and prior to fledging (2.0 g) (Table 1). From that, one might conclude that nest temperature or its regulation really is not very critical to chick development. However, if regulation of stable nest temperature is reflective of the thermoregulatory status of the chicks, then maturation of thermoregulatory development of chicks may have proceeded more slowly in three-clutch chicks, in spite of their rapid accumulation of biomass. Support for that hypothesis comes from the observation that the slopes of the regressions of T n versus T a remained constant (0.6) between day 0 and day 8 in three-clutch nests, whereas the slopes of those regressions steadily decreased (less dependence of T n on T a )overthesameperiodinsix-clutchnests(0.52 to 0.26) (Fig. 4). Further corroborative evidence of the effect of the thermal environment on development of Tree Swallow chicks is presented by Lombardo et al. (1995) who showed that chicks from nests with reduced feather insulation were smaller and less mature at day than chicks from normally insulated nests in the absence of any increased ectoparasite load. Several other studies have documented the effect of clutch size and food quality and quantity on growth rate and development of Tree Swallows chicks (Paynter 1954, DeSteven 1980, Quinney et al. 1986, Smith and Montgomerie 1992). In those studies, larger clutches generally developed faster unless there was demonstrable food limitation. Dunn (1979) argued that larger clutches allow earlier attainment of effective homeothermy, the ability of individual chicks to regulate body temperature within 75% of the adult level: from four days in broods of six to seven individuals, to six days in a clutch of three chicks, to eight days in a clutch of one chick. Thus, the allocation of energy between growth and maturation of tissues during development may be different in chicks reared in different sized clutches and deserves further study. Acknowledgments. This study was conducted at the Twin Cities Army Ammunition Plant (TCAAP), New Brighton, Minnesota, with the permission of the U.S. Army and Alliant Technology. We are grateful to Craig Andreson for establishment and maintenance of the bluebird boxes on the TCAAP property. LITERATURE CITED BLEM, C.R.,AND L. B. BLEM. 1994.Compositionand microclimate of Prothonotary Warbler nests. Auk 111:197 200. BROWN, C.R.1994.Nestmicroclimate,eggtemperature, egg water loss, and eggshell conductance in Cape Weavers, Ploceus capensis. Ostrich65:26 31. CALDER, W.A.1971.Temperaturerelationshipsand the nesting of the Calliope Hummingbird. Condor 73:314 321. COOPER, J.A., AND A. D. AFTON. 1981.Amultiple sensor system for monitoring avian nesting behavior. Wilson Bulletin 93:325 333. DESTEVEN, D. 1980. Clutch size, breeding success, and parental survival in the Tree Swallow (Iridoprocne bicolor). Evolution 34:278 291. DUNN, E.H.1979.Ageofeffectivehomeothermyin nestling Tree Swallows according to brood size. Wilson Bulletin 91:455 457. FLINT, P.L.,AND M. C. MACCLUSKIE. 1995.Adevice for simultaneously measuring nest attendance and nest temperature in waterfowl. Journal of Field Ornithology 66:515 521. HAFTORN, S.1978.Egglayingandregulationofegg temperature during incubation in the Goldcrest (Regulus regulus). Ornis Scandinavica 9:2 21. LOMBARDO, M.P.,R.M.BOSMAN, C.A.FARO, S.G. HOUTTEMAN, AND T. S. KLUISZA. 1995.Effectof feathers as nest insulation on incubation behavior and reproductive performance of Tree Swallows (Tachycineta bicolor). Auk 1:973 981. MARSH, R.L.1980.Developmentofthermoregulation in nestling Tree Swallows. Condor 82:46 63. O CONNOR, R.J.1984.TheGrowthandDevelopment of Birds. J. Wiley and Sons, New York. PAYNTER, R. A., JR. 1954. Interrelations between clutch size, brood size, prefledging survival, and weight in Kent Island Tree Swallows. Bird- Banding 25:102 131. QUINNEY, T.E.,D.J.T.HUSSELL, AND C. D. ANKNEY. 1986. Sources of variation in growth of Tree Swallows. Auk 103:389 400. RICKLEFS, R.E.1974.Energeticsofreproductionin birds. Pages 152 297 in Avian Energetics (R. A. Paynter, Jr., Ed.). Nuttall Ornithological Club Publication, no. 15. ROBERTSON, R.J.,B.J.STUTCHBURY, AND R. R. CO- HEN. 1992. TreeSwallow, Tachycineta bicolor. In The Birds of North America, no. 11 (A. Poole, P. Stettenheim, and F. Gill, Eds.). Academy of Natural Sciences, Philadelphia, and American Ornithologists Union, Washington, D.C. SMITH, H.G.,AND R. MONTGOMERIE. 1992.Maleincubation in Barn Swallows: The influence of nest temperature and sexual selection. Condor 94: 750 759.

851 WILSON, S.F.,AND N. A. M. VERBEEK. 1995.Patterns of Wood Duck nest temperature during egg-laying and incubation. Condor 97:963 969. ZACH, R.,AND K. R. MAYOH. 1982.Weightandfeather growth of nestling Tree Swallows. Canadian Journal of Zoology 60:1080 1090. ZANN, R.,AND M. ROSSETTO. 1991.ZebraFinchincubation: Brood patch, egg temperature, and thermal properties of the nest. Emu 91:107 0. Received 26 June 2001, accepted 9 February 2002. Associate Editor: C. Blem