FLIGHT COST OF A SMALL PASSERINE MEASURED USING DOUBLY LABELED WATER: IMPLICATIONS FOR ENERGETICS STUDIES

PAUL TATNER AND DAVID M. BRYANT Departmentof BiologicalScience, University of Stirling,Stirling FK9 4LA, UnitedKingdom

AI•STRACT.--Themetabolic cost of rest and activity (cm• COz'g •'h -•) was measuredfor the European Robin (Erithacusrubecula, 18.6 g) at temperatures(Ta, øC) from -15 to +300C. Regressionsexpressing these costsare: night resting (M•N) = 4.23 -- 0.0733Ta,day resting (Mro)= 5.10 - 0.0807Ta,hopping (Mhop) = 8.60 -- 0.1256T•. Daily energy expenditure (DEE) for robins held in an outdoor aviary was 64.8 kJ/day (SD = 9.2, n = 6), determined using the doubly labeled water technique.This was positively related to time spentin flight (t•r, h) suchthat DEE = 50.9 + 23.4t•y. Flight costfor robinswas estimatedas 25.6 kJ/h (SD = 5.0, n = 6). This flight costis about twice that predictedby variousallometric equations. Robin flights in the aviary were short (3 m) and brief (0.78 s), indicating a mean flight speed(3.85 m/s) that was lower than the theoretical minimum power velocity (5.86 m/s). The EuropeanRobin has a relatively high wing loading (0.263g/cm z) and aspectratio (7.33).In a small bird with flapping flight these characteristicsimply a high cost,particularly at low flight speeds. The high costof flight was offsetby its shortduration. During 30-min observationperiods, an averageof 100.2s was spent in flight, implying a sustainedenergy demandof only 3.04 kJ/h (2.7 x basalmetabolic rate). The exceptionallyhigh flight costreported here (23 x basal metabolicrate) may be typicalof short,brief aerialforays. Other doublylabeled water studies reveal a positivecorrelation between the time spentin flight and DEE, indicatingits domi- nant impacton energy turnover in somefree-living birds. Received17 April 1985,accepted 19 September1985.

THEaccuracy of time-activity laboratory(Mu- species(Utter and LeFebvre 1970, Hails 1979, gaasand King 1981)studies for determiningthe Flint and Nagy 1984, Westerterp and Drent daily energy expenditure (DEE) of free-living 1985). birds depends upon representative time bud- gets and realistic values for the energy costof METHODS activities (Weathers and Nagy 1980, Koplin et al. 1980,Williams and Nagy 1984, Bryant et al. Energeticcost of activity.--Robins from a wooded 1985). This was well illustrated by Weatherset area on Stirling University campus(56ø08'N, 3ø54'W) al. (1984) for low-cost activities that are pur- were caught at dusk between November and March during 1983-1985. Wing dimensionsand body mass suedover long periods,but it is also important were noted, and the birds were placedin a respirom- for high-costactivities that usuallyoccupy only eterovernight to determinemetabolic rate for a range a small proportionof the day (Mugaasand King of temperatures(-15•C to +30øC).Three to five rep- 1981). licates were performed at 5øC intervals, each over- This study is concerned with the determi- night trial maintaining one temperature(+ 0.5øC). nation of activity costsin a small, woodland The birdswere releasedapproximately 1 h after dawn, passefine,the European Robin (Erithacusrube- having completeda bout of hopping activity in the cula).It employsgas-analysis respirometry to spindle-mountedrespirometer. Robins were caught measureresting and some locomotorcosts. Time direct from the wild so that the respirometryresults would include the effects of acclimatization to sea- budgets and doubly labeled water measure- sonal and local conditions. ments of DEE for aviary robins were used in The respirometertrain was an open-flow system, combinationwith laboratoryestimates of non- equipped with an MSA infrared gas analyzer for flight activity to establishthe costof flight for monitoringCO• production,and a BeckmanOM2 po- a speciesthat exhibits a largely nonaerial ex- larographicoxygen analyzer. Carbon dioxide was re- istence. It contrastswith previous studies on moved from incurrent air using Carbasorb,and air this topic that have focusedon mainly aerial leaving the respirometer chamber was passed over

169 The Auk 103: 169-180. January 1986 170 TATNERAND BRYANT [Auk, Vol. 103

Drierite prior to gas analysis.The respirometerwas the dark period, during which robins were resting enclosedin a darkenedincubator. Each overnight run either in the aviary or in the respirometer;(2) the was divided into a series of 2-h sample periods, in- handling period, during which body mass was re- terspersedwith ambientair sampling.During the first corded, blood samples taken, and the birds then 2 h the bird settled down in a darkened incubator at placed in a cloth bag; and (3) the light period, when about the time it normally went to roost, and the the birds were active in the aviary. mean for this period was taken to indicate the cost Doublylabeled water technique.--Thebirds received of resting during the day (Mrr0. Metabolism usually intraperitonealinjections of 0.13 ml of 10 atom % D20 reached its lowest levels at about 0300. At dawn the and 0.22 ml of 20 atom % H2•80.Initial blood samples waking period was often characterizedby intermit- were taken 1 h later, after which one individual was tent activity, so this and the first period were exclud- placed in the respirometerand the others returned ed when calculating the average costof resting over- to the aviary.On the following morninga bloodsam- night (Mrs). After waking, the bird was induced to ple was obtained from the robin in the respirometer, hop by revolving the spindle-mountedrespirometer after which it was returned to the aviary. Blood sam- until a steadytrace was obtained from the gas ana- ples were obtained from all individuals 48 h after lyzers. The speedof rotation was selectedsuch that initial sampling. The sampleswere flame-sealedin the individual'sgait resembledthat of a foraging bird Vitrex microcapillaries(5 •1) for later analysis.A dif- (1.5-2.5 m/min). The energeticcost of ground activ- ferent individual was monitored in the respirometer ity (includingforaging) is estimatedin this work from on the secondnight. Natural enrichment levels of the relationship establishedduring hopping trials deuterium and oxygen-18 were determined from (Mhop).Individual responseto a rotatingchamber was other robins on the study site at the time of the ex- unpredictable,and it sometimeswas necessaryto re- periment (D = 152.82ppm, •O = 2,011.56ppm). ject trials in which birds employedbouts of hovering Blood samples were processedaccording to the or clung temporarily to the chamber wall. All gas methodgiven in Bryant et al. (1984) and Westerterp volumes were corrected to STP. and Bryant (1984). The theory of measuring energy Six robins were used to determine daily energy expenditure using doubly labeled water (D2•80) is expenditure using the doubly labeled water (D2•80) describedby Lifson et al. (1955), and assumptions technique(Lifson et al. 1955).They were held during associatedwith the method are outlined by Nagy July-August 1983 for 48 h in an outdoor aviary, as (1980). Body water content was taken as 65% (after groupsof 3 individuals. Different birds were placed Newton 1968) of the mean body mass,itself derived in the respirometeron successivenights to establish from measurementstaken at the beginning and end summertimevalues for night resting,day resting,and of a sampling period. hopping. Food [blowfly larvae (Calliphoraspp.) and cheese]was supplied with water ad libitum.The avi- RESULTS ary was partly exposedto the sky, but completely shielded from the wind by fencing and woodland. Respirometrymeasurements.--Figure 1 shows Time budgets.--In the aviary, the activity of each winter metabolism during night resting (M•), bird was recorded for 3-5, 30-min periods (n = 48) day resting (M•r0, and hopping (Mhop)as a per day. Individuals were recognizedby coloredring function of ambient temperature (T•). Predic- combinationsand by spotsof white typist'scorrec- tive equationswere producedusing regression tion fluid applied to the head or breastfeathers. analysis (Table 1). McNab (1980) noted that Activities were timed using a microprocessor-based regressioncoefficients derived in this manner data logger, with event resolutionlimited only by are poor representationsof minimal thermal the human responsetime (approximately0.2 s). Three classesof activity were recognized:resting, flying, conductanceif the regression lines do not ex- and ground activity. The latter categorycomprised trapolate to the animal's body temperature at principally hopping and pecking but also included zero metabolism.In this work the regressions occasionalpreening, bathing, and defecatingas these extrapolate to 57.7øC (M•), 63.2øC (M•o), and activitiesentail broadly similar energy demands.We 68.5øC(M•op), which are all much greaterthan designatedthese minimal energy-demandingactivi- the European Robin's proventricular tempera- ties as ground activity to simplify the analysis,and ture of 40.6øC(Udvardy 1953). For this reason becauseit is not yet possibleto distinguishbetween the mean conductance method advocated by the energy costsof these minor activities. Weathers McNab (1980) was investigated, assuming a et al. (1984) measured the energetic costs of alert perching, preening, eating, and hopping. The mean constantbody temperatureof 40.6øC.However, costsof eachactivity were ranked in the order men- agreementbetween this method and measured tioned, althoughthere were no statisticallysignifi- metabolismwas poor. This may have been due cant differences. to an incorrect assumption of constant body The 24-h day was divided into three periods:(1) temperature.Metabolic values reported here for January1986] Costof Flightin theRobin 171

RQ in summerand winter wasgreater over the night-restingphase (0.95 comparedwith 0.75; t = 5.24, P < 0.01) than the day-restingphase (0.88 comparedwith 0.75; t --- 3.27, P < 0.05). The apparent increasein metabolic rate indi- catedby the elevatedCO• productionof sum- mer robinswas reducedto 12%(Mrs) and 11% (M,o)when corresponding changes in RQwere taken into account. However, differences be- tween summer and winter resting metabolic rate at 25øCwere still statisticallysignificant Fig. 1. Robin winter metabolism as a function of (Table2). Summerpredictions using the winter ambienttemperature. MrN = night restingmetabo- regression equations therefore were incre- lism;MrD = dayresting metabolism, Mhop = hopping mentedby 12%for M•N and by 11%for M•v. metabolism.Regressions are given in Table 1. The energeticcost of hoppingat 25øCfor sum- merrobins was not significantlydifferent from MrN are nightly averages,and therefore above that predictedby the winter regressionequa- the lowestresting level (40%greater, Tatnet tion (Table 2). unpubl. data). Thus, the thermal conductance Timebudgets.--Activity timed using the data for Mr• doesnot representthe minimal value. loggerwas expressed in termsof the propor- These limitations, together with the extent of tion of the 30~rainsample devoted to resting, variabilityin measuredmetabolism (Fig. 1, Syx ground activity, or flying. Theseproportions in Table 1), renderthe use of regressionanal- were normalized using an arcsinetransforma- ysis most appropriate in the context of this tion (Zar 1974)prior to statisticalanalysis. The work. values for each individual were determined on Day restingwas significantlymore demand- a dailybasis and used to apportionthe daylight ing than night resting(t = 5.42, P < 0.05), re- periodspent in the aviary(Table 3). Analyses quiring 21% [100 x (4.23 - 5.10)/4.23] more of varianceon pooled data showed that the energy.A comparisonof the respectiveregres- threecategories of activityoccurred in similar sion coefficients(Table 1) indicatesthat this dif- proportionsduring eachhour of daylight,in- ferencedid notvary with ambienttemperature dicatingthat diel activityrhythms were absent. (t = 0.736, P > 0.05; Zar 1974). An increaseof Theaverage proportion of timespent in flight thismagnitude (20-25%) between the night and wasthe only time-budgetvariable that differed dayphases is a generalfeature of avianresting significantlybetween individuals (one-way metabolism(Aschoff and Pohl 1970b, Calder ANOVA, F = 5.96,P < 0.05).Flights were short, and King 1974). asthe birdstended to fly from one sideof the The respirometrytrials during summer, un- aviaryto the other (3 m), landingon the sides dertakenat a meanTa of 25øC,indicated a higher onlymomentarily before making a returnflight. CO2production (Table 2) than would be pre- Cumulative flight time as a function of the dicted from the winter metabolism relation- numberof flightsin a samplingperiod is shown ships for both Mr• and Mrv. These increases in Fig. 2. Becausethe slopeof the regression were32% for M• (2.41to 3.17cm 3 CO2-g -• .h-1) was significant(t = 16.97, P < 0.001), the du- and 24%for M,v (3.10to 3.85crn • CO•.g-•.h-•). ration of a single flight was taken as 0.78 s. Summerrobins also exhibited higher respira- Duringthis time the bird flew approximately tory quotients (RQs). The difference between 3 m, so we necessarilyconsidered the costof

TABLE1. Robinwinter metabolism asa functionof ambienttemperature (T,, øC).

COaproduction ra (cm•'g-•' h-•) Sy• ( x 100) n Restingnight metabolism (M,•) 4.23- 0.0733T• 0.52 81.2 32 Restingday metabolism (M,o) 5.10- 0.0807T, 0.65 77.4 32 Hoppingmetabolism (Mao•) 8.60- 0.1256T, 1.40 79.7 30 172 TATNERAND BRYANT [Auk, Vol. 103

TABLE 2. Robin metabolism in winter and summer at 25øC. '

Night resting (MrN) Day resting(MrD) Hopping activity (Mhop) e (SD) n e (SD) n e (SD) n Summer cm• CO2.g-'.h -• 3.17 (0.18) 6 3.85 (0.34) 6 4.81 (0.98) 4 RQ 0.95 (0.08) 6 0.88 (0.08) 6 0.94 (0.07) 4 J.g-•.h -• 69.55 (4.06) 6 89.63 (5.69) 6 106.45 (21.63) 4 Winter b cm3 CO2.g-•.h-• 2.41 (0.13) 31 3.10 (0.16) 31 5.46 (0.41) 29 RQ 0.75 (0.03) 19 0.75 (0.05) 19 0.82 (0.10) 17 J.g-' .h -• 63.46 (3.38) 31 81.70 (4.28) 31 134.38 (10.17) 30 Summer vs. winter Difference (J.g-X.h-•) 6.09 7.93 -27.93 t (df) ' 3.45** (6.41) 3.24** (6.14) 2.45 Ns (3.19) ' J.g-•.h -• calculatedusing standardheat equivalentsfor CO2 as follows (units J/cm3 CO•): 21.95 at respi- ratory quotient (RQ) = 0.95, 23.30 at RQ = 0.88, 22.13 at RQ = 0.94, 26.44 at RQ = 0.75, 24.62 at RQ = 0.82 (Brody 1945). • cm3 CO•.g-X.h-• calculatedfrom the regressionequations in Table 1; standarddeviations of predicted valueswere calculatedusing MINITAB regression(Ryan et al. 1976).RQ mean valuesare basedon pooled data collectedover the temperaturerange -15øC to +30øC. • ** = P < 0.01, NS = not significant. short, brief flights. This type of flight resem- were calculatedon a daily basisand then av- bles that performedby the EuropeanRobin in eraged (n = 2 days) to derive the figures given its natural habitat when it moves short dis- in Table 4. Energy costswere determined using tancesbetween a perch and the ground, or be- the resultsof the winter respirometry,but tak- tween bushes. ing into account the observed summer differ- Nonflight daily energy expenditure (DEE) ences(Table 2). The appropriateequations are was calculatedfor aviary robinsusing the time- given in Table 4. It was not possibleto measure activity laboratory (TAL) method (after Mu- flight costdirectly, so for the present,the pe- gaasand King 1981). For each individual, the riod spent in flight is assignedan energy ex- costsof the five categoriesof activity (Table 3) penditure equivalent to the day-resting level.

TABLE3. Activity budgetsfor aviary robins.

Daytime Average ground temperatureNight- Daytimerest activity Daytimeflying (øC) timerest % of No. of % of No. of % of No. of Distur-bance • Individual Day Day Night• (h) day hours day hours day hours (h) RB35 1 18.7 23.0 8.67 87.3 11.19 6.9 0.88 5.8 0.74 2.51 2 21.9 15.6 5.62 78.2 13.20 20.3 3.43 1.5 0.25 1.50 RB36 1 18.7 12.7 5.85 94.9 15.56 2.1 0.34 3.0 0.49 1.75 2 21.9 25.0 10.23 88.0 11.24 7.6 0.97 4.4 0.56 1.00 RB37 1 18.7 12.7 6.1 89.0 14.15 4.2 0.67 6.8 1.08 2.00 2 21.9 15.6 5.62 75.7 13.16 15.0 2.60 9.3 1.62 1.00 RB41 1 17.7 25.0 11.13 84.0 9.87 3.7 0.43 12.3 1.45 1.12 2 20.1 11.2 8.75 86.9 12.18 6.2 0.87 6.8 0.95 1.23 R1M2 1 17.7 14.4 8.45 99.1 14.88 0.8 0.12 0.1 0.02 0.53 2 20.1 25.0 10.33 98.7 12.46 1.1 0.14 0.2 0.03 1.05 RB43 1 17.7 14.4 8.28 88.3 13.20 5.2 0.78 6.5 0.97 0.77 2 20.1 11.2 8.04 87.0 13.01 8.5 1.27 4.5 0.67 1.01

Night temperaturesabove 22øC indicate that the bird spent this period in the respirometer. Disturbance = when birds were handled. January1986] Costof Flight in theRobin 173

tayis time spentin flight per day (h). When no aoo. o time was spentin flight the regressionpredict- ed DEE to be 50.9 kJ (SD = 4.3). This is 16% Z o oo o greaterthan the mean nonflight DEE of 44.0 kJ (SD = 2.3) estimated using the TAL method (Table 4). The regressionequation also predict- ß•- o ed that for every hour of the day that robins spent in flight there was an increment in DEE of 23.4 kJ (SD = 4.98, n = 6). At the same time, there was a corresponding reduction in the daytime nonflight DEE of 2.2 kJ [1/16 x (70 x 50.9/100)]. This is basedon an averageof Fig. 2. Robin flight activity in an aviary during 30% nonflight DEE devoted to nighttime rest time budgets of 30 min. Regressionslope = 0.777 (Table 4), occurringover an averageof 8 h (Ta- (SD =0.046, t = 16.97, P < 0.01), regressioncon- ble 3), during which time robins do not fly. stant = 8.35 (SD = 7.17, t = 1.16, NS); r 2 = 86.2, n = Thus, the flight costof aviary robins was 25.6 47. kJ/h. Flight cost can also be calculatedusing the Doublylabeled water results.--The average dai- TAL estimatesof nonflight DEE (Table 4). These ly metabolicrate for aviary robins was 5.46 cm3 estimatesdo not include an estimateof flight CO2'g-•'h-• (SD = 0.67, n = 6), ranging from cost,the period spentin flight having been al- 4.21 to 6.07 cm3 CO2.g-'.h -' (Table 5). At an RQ locateda costequivalent to day resting. Thus, of 0.75, this gives a daily energy expenditure the difference between the DLW estimate of of 64.8 kJ/day for an 18.6-g robin, which is DEE and the TAL estimatemay be regardedas 2.42 x BMR (where BMR = 2.26 cm3 CO2.g-•. the net energy devotedto flight (i.e. the ener- h -• for an 18.6-g passerine;Aschoff and Pohl geticdemand above the day-restinglevel). Fig- 1970a). Four overnight respirometry trials at ure 3B shows this difference as a function of 25øCwith D2'SO-loadedbirds gave an average the time spent in flight. Although the regres- 'metabolicrate of 3.33 (SD = 0.50) cm3 CO2.g-•. sioncoefficient was significantly different from h-', which was not significantly different from 0 (t = 4.05, P < 0.05), the constant was not (t = the 3.76 (SD = 0.27) cm3 CO2.g-•.h -• obtained 2.06, P > 0.05). Hence, this regressionpredict- simultaneouslyusing infrared gas analysis. ed a net energy requirementof only 20.50 kJ Costof fiight.--Daily energy expenditurede- when 1 h/day was spent in flight. The extra termined usingthe doubly labeledwater (DLW) amount required to convert this figure to gross technique was significantly related to time flight energyis the costof day restingfor 1 h spent in flight (Fig. 3A). This is describedby at the mean daytime temperatureduring the the relationship DEE = 50.9 + 23.4tay,where experiment (19.6øC),which is 3.52 kJ/h. Hence, DEE is daily energy expenditure (kJ/day) and a secondestimate of flight costis 24.02 kJ/h. It

TABLE4. Energeticcosts for aviary robins basedon two 1-day TAL predictions(kJ/day). a

Disturbance Ground Flight period Nonflight Bird Night resting cost Day resting activity at Mrr• level DEE RB35 10.46 3.26 22.10 5.59 0.92 42.33 RB36 12.71 2.47 27.06 1.86 1.05 45.15 RB37 10.84 2.67 27.13 4.60 2.65 47.89 RB41 15.37 1.95 21.18 1.80 2.35 42.65 RB42 13.92 1.29 26.26 0.36 0.05 41.88 RB43 14.11 1.43 24.43 2.76 1.54 44.27 Mean 12.90 2.18 24.69 2.83 1.43 44.03 (SD) (1.94) (0.76) (2.57) (1.94) (0.96) (2.26) aEnergy cost(J.g-•.h-•): night resting (Mr,): 111.84 - 1.938Ta,plus 12% for summermetabolism; day resting(M•v) = 134.84- 2.134T,,plus 11% for summermetabolism; ground activity (Mao,) = 211.73- 3.102T,. Disturbancecost (blood sampling, weighing, transit, etc.) taken as Mrv at 25øC,plus 12%. 174 TATNERAaND BRYANT [Auk, Vol. 103

TAi•LE5. Metabolic rate of aviary robins during gO' A summer determined using doubly labeled water o (D2•'O).

Metab- Body olism o o mass (cm3 CO2' DEE' DEE Bird (g) g-•. h-•) (kJ/day) ( x BMRb) RB35 17.9 6.07 70.32 2.67 RB36 19.7 5.50 68.83 2.48 RB37 19.5 6.02 74.68 2.70 RB41 18.4 5.55 65.07 2.46 0•5 1•.0 RB42 18.2 4.21 48.61 1.85 Time in flight Hrs/day RB43 17.8 5.39 61.07 2.36

Mean 18.6 5.46 64.76 2.42 (SD) (0.8) (0.67) (9.17) (0.31) ' Using RQ = 0.75, 1 cm3 CO2= 26.44J (Brody 1945). bPredicted from body mass,BMR = 1.11kJ/h (night phase;Aschoff and Pohl 1970a). may be argued that resting during the day in- cludesa high proportion of alert behavior that may have a higher costthan indicated by Mrs. An estimate of the alert cost was made using the upper confidencelimit of the MrD regres- Time in flight Hrs/day sion (Mrr• + tSyx,t at 95%),which is equivalent Fig. 3. Robin energy expenditureas a function of to an addition of 0.4 BMR (0.85 cm3 CO2-g-•. time spent in flight. (A) Regression slope • 23.4 h-•). Recalculating the nonflight TAL predic- ($D = 4.98, t • 4.69, P < 0.01), regressionconstant • tion of DEE gave a mean value of 53.98 kJ/day 50.9 (SD = 4.29, t = 11.85,P < 0.01);r • = 80.8, S = 5.46. (SD = 3.21), which is close to that predicted (B) Regressionslope = 20.5 (SD = 5.07, t = 4.05, P < above (50.9 kJ/day). Using these higher values 0.05),regression constant = 8.98 (SD = 4.37, t = 2.06, of nonflight DEE produced only a slight change P > 0.05); r• = 75.5, S = 5.55. Dlw = Doubly labeled water. in the predicted flight cost, which was 23.47 kJ/h (19.95 + 3.52). metabolismat low temperatures.Differences in DISCUSSION the slope of metabolic equations for summer and winter given by Gavrilov indicate a reduc- Winter/summer metabolism.--The winter me- tion in conductance of 7%, probably due to tabolism results may be compared with those thicker winter plumage. In Moscow, robin obtained by Gavrilov (in Kendeigh et al. 1977) plumage increasedfrom 7% of the body mass for Russian robins of slightly lower average (14.0 g) to 9.1% of the body mass (17.2 g) be- body mass (17.6 g). His equation for winter tween summer and winter (Shilov 1968). Sea- robins (night resting metabolism,cm 3 CO2-g-•- sonalshifts in thermoregulatorymetabolism are h -• = 5.84 - 0.1460T•) indicates a much greater well documented (Calder and King 1974) and increasein energy requirement as T• declines have been related to body mass (Kendeigh et than we found (Table 1). Possibly,this was due al. 1977). Regressioncoefficients for SMR (Ken- to an adjustment of the relationship under- deigh et al. 1977: 135) of an 18.6-g passerine, taken to ensure that metabolic rate was zero calculated for summer and winter, indicate a when the ambient and bird's body tempera- 12% increase in summer energy requirements tures were equal (see Kendeigh et al. 1977: ap- [i.e. winter temperature coefficient(b) = 1.490 pendix 5.1). A similar procedure was investi- kJ.bird-•-day-•.øC% compared with summer gated here (see Methods), and it producedthe temperature coefficient (b)= 1.676 kJ.bird-z. following equation for night resting metabo- day-•-*C-Z). Elevated resting metabolic rates lism: cm• CO•-g-•.h -• = 5.34 - 0.1315T•.Al- therefore are expected for summer robins, in though this is similar to Gavrilov's result, it accordance with the 11-12% increase we ob- was not acceptable because it overestimated served. January1986] Costof Flightin theRobin 175

The reasonfor the higher RQ valuesfor sum- T^BLE6. Wing characteristicsof small birds. mer robinsin this study is rather obscure.Work on substrate concentrations and turnover in Aspect ratio birds has demonstrated that in some winter- (wing- acclimatizedspecies there is a reduced reliance Body Wing Wing span2/ on carbohydrateas a metabolicfuel, although mass area loading wing this was observedonly for elevated rates of (g) (cm2) (g/cm •) area) thermogenesis(Marsh and Dawson 1982,Marsh Erithacus rubecula a 18.6 70.8 0.263 7.33 et al. 1984). Lanius excubitor b 31 144 0.215 -- The absenceof a significantmetabolic differ- Parusmajor b 14.5 62 0.234 -- Ernberiza ence between summer and winter hopping gubernatrix• 25.5 100 0.255 -- trials may be due to the small sample size and Passer domesticus b 28.3 76 0.372 -- high degree of inherent variability. Pohl and Spizellaa. arborea• 18 90 0.200 -- West (1973) reported a greater heat production Anthusspinoletta • 19 109 0.174 -- Parus ater' 9.1 53.8 0.169 6.02 during forced exercisefor Common Redpolls Parus rnontanus' 10.9 61.9 0.176 5.53 (Carduelisfiammea) in winter compared with Parus cHstatus c 12.2 60.2 0.203 5.75 Regulusregulus' 5.9 35.4 0.167 5.94 Daily energyexpenditure.--Robins in the avi- Certhiafamiliaris' 9.1 62.4 0.146 5.72 ary spent a high proportion of the daylight pe- Measured, n = 21. riod resting, with relatively little time devoted Greenewalt 1962. to activities akin to foraging (Table 3). These Norberg 1979. aspectsof their daily behaviormay be different from robins in their natural habitat at this time 1984). There is an implicit assumptionthat the of year, as East (1982) has shown that robins flight speed adopted coincides with the mini- forage for approximately 20% of the time at mum power speed(base of the U power curve). temperatures greater than 10øC.The propor- However, the energeticcost of flight is likely tion of time spent in flight also may be lower to reflect a compromise between flight speed than the free-living level. Walsberg'sequation and flight gait, both being influenced by the relating percentageof activeday spentin flight current demandson a bird's time and activity to body mass(Walsberg 1983: Eq. 11) predicts budget (Norberg 1981,Pyke 1981).There is also a figureof 7.6%(_+1 SD, 3.1-18.7%)for an 18.6-g likely to be an inverse relationship between bird, compared with a mean of 4.3% (+1 SD, the costassociated with the flight gait and the 1.1-9.2%) found here. So, it is likely that the time spent in the air. Hails (1979) showed that mean DEE estimate (DLW; 64.76 kJ/day) re- avian aerial insectivores,which spend most of portedin this studywill be low comparedwith their life in the air, exhibit flight coststhat are free-living individuals. This suggestionis in 49-73% lower than predicted on the basis of accordancewith a DEE prediction of 76.40 kJ/ body massfor nonaerial species.More recently, dayfor an 18.6-gnonbreeding passerine (Bryant a similar result was reported for a larger, non- et al. 1985), which is 18% above the value we passerineseabird (Flint and Nagy 1984). Our observed. study considersthe opposite end of the spec- Flight costs.--Birdflight is recognized as a trum in terms of adaptation to an aerial exis- costlybut highly efficientform of locomotion tence. For most of the year robins spend only (Tucker1970). Determination of flight costshas a small proportion of their time in flight. Thus, beenapproached using a variety of techniques, during 172 h of time budgeting this species, includingmaterial balance (Kespaik 1968, Lyu- East(1982) obtainedsuch small samplesizes for leeva1970), flight mechanics(Pennycuick 1968, birds in flight that it was necessaryto pool the 1975; Tucker 1968; Greenewalt 1975; Raynet information with other activities of short du- 1979a), and doubly labeled water (LeFebvre ration or infrequent occurrence.The European 1964, Utter and LeFebvre 1970, Hails 1979, Flint Robin's mode of foraging involves hopping, al- and Nagy 1984).The theoreticalapproach pre- though there are often short flights between dictsa U-shapedcurve relating power require- perchesand the ground (East 1980). The wing mentsto flight speed(Pennycuick 1975), but in loading and aspectratio of a robin are high studiesof DEE the costof flight usually is taken comparedwith most other birds with a similar as a multiple of BMR (seeWilliams and Nagy body massor wing area (Table 6). A high aspect 176 TATNERAND BRYANT [Auk, Vol. 103

TABLE7. Predictionsof EuropeanRobin flight cost.'

Flight cost Source (kJ/ h) Method Pennycuick 1975 (Eq. 25) 5.69 Body measurementsand flight speed Tucker 1973 (Eq. 49) 6.43 Body measurementsand flight speed Greenewalt 1975 • 5.61 Body measurementsand flight speed Hart and Berger1972 10.23 Body mass Kendeighet al. 1977 14.66 Body mass Hails 1979 11.90 Body mass ßRobin measurements (n = 21):body mass= 18.6g (0.182N), wingspan= 228 ram,wing area= 70.78cm 2, flight speed= 3.85 m/s, air density = 1.18kg/m 3 (Hodgroan1959). bEq. 23 (1.09 kJ/h) x 4 (muscularefficiency 25%) + SMR (=1.24 kJ/h; Lasiewskiand Dawson 1967). ratioindicates relatively long wings,which may ments.Teal (1969) provided data on the costof improve maneuverability, but this could be a brief flights (• = 68 s, SD = 42, n = 56) in small disadvantagein a small bird that exhibits flap- birds basedon direct measurementof CO2 pro- ping flight becauseof the increasedwing in- duction.His passefineresults are shownin Fig. ertia.These characteristics suggest that the flight 4, from which it can be determinedthat flight costof foraging robins may be high. cost(cm 3 CO2/h) = 53.70M•'%where M is body Predictionsof flight cost in the robin vary mass(g). This relationshippredicts that flight from 5.61 to 14.66 kJ/h (Table 7). The lowest costfor an 18.6-g robin will be 30.14 kJ/h (as- estimates,derived using aerodynamictheory suming RQ = 0.71; Torre-Bueno1977), rather (Pennycuick 1975), indicated a flight cost of higher than we observed. only 5.1 x BMR. This calculationis likely to The costof flight for robins in this study was have underestimatedinduced power for robins extremely high at 25.6 kJ/h [23 x BMR in this study, as the very short flight duration (night) = 1.11 kJ/h; Aschoffand Pohl 1970b]. meant that a proportion of the time was prob- This demand operates for only very short pe- ably very similar to hovering.Raynet (1979b: riods, however, as the average flight duration 751) noted that induced power is better pre- was 0.78 s. During the 30-min time-budgeting dicted at low flight speedsusing a model for periods cumulative flight time for robins av- hovering. The observedflight speed (3.85 m/ eraged 100.2 s. In this instance,712.5 J would s) was lower than the theoretical minimum be required for flight, plus 866.3 J for the re- power speed(5.86 m/s, Pennycuick1975). This maining period, assumingit was at the average would also increase observed flight cost rela- nonflight level (Table 4; 44.03 kJ/day or 0.51 tive to the theoretical prediction. Limitations W). Hence, the overall average sustaineden- associatedwith the modelsof flight mechanics ergy demandmay have been only 3.16 kJ/h employed here are discussedby Raynet (1982). (2.8 x BMR). The maximum cumulative time The flight cost predictions based on body spent flying in a 30-min bout was 320 s; even mass(Table 7) are much greater than thosede- when flying for three times as long, the sus- rived using theoreticalflight mechanics,prob- tainedenergy demand would not be exception- ably becausethe latter underestimatethe phys- al (6.06 kJ/h, 5.46 x BMR; see Westerterp and iological requirements of flight (Raynet 1982) Bryant 1984). and the increase in power required for low- Energydemand for a single flight (0.75 s) is speedflight in small birds (Tucker 1973:fig. 1, 5.3 W, of which 0.51 W may be accountedfor Westerterpand Drent 1985:fig. 4). Mechanical by resting metabolism,leaving 4.79 W associ- predictions of energy requirements for flight ated with flight activity. Maximum sustained do not make provision for the cost of take-off, output of striatedmuscle is estimatedto be 250 which is presumablygreater than that of sus- W/kg of muscletissue (Weis-Foghand Alex- tained flight and which assumesa high signif- ander 1977). In the EuropeanRobin the pec- icance in the present work. Allometric equa- toralis musclesweigh 3.0 g, implying a me- tions relating body mass to physiological chanical power output of up to 0.75 W. measuresof flight metabolismtend to utilize Assuming23% muscleefficiency (Pennycuick data based on protracted flight duration, and 1975),the maximumpower consumptionof the therefore may reflect minimum power require- flight muscleswill be 3.26 W. This leaves a January1986] Costof Flight in the Robin 177 shortfall of 1.53 W, which could be accounted for by the muscularactivity of the legs during take-off and landing. If this is correct,the en- ergy requirementsduring short forays can be .... õo• partitioned as 61% flight, 29% take-off and 1250 landing, and 10% resting metabolism. For sustainedflight, most allometric predic- ,• 625. tions are based on body massø'7•, which indi- cates that sustained power requirements in flight are a multiple of basalmetabolic rate (Hart 150' and Berger 1972).Hence, larger birds use pro- portionally less power in continuous flight. Theoreticalaerodynamic predictions indicate a more direct proportionality with body mass (M TMin Hart and Berger 1972, M •.•øin Raynet 1979a). These calculations are based on struc- tural and mechanical features and therefore are Body mass (g} more likely to reflect the ability for maximal power output, the duration of which probably Fig. 4. Metabolic rate (cm3 CO2/h) as a function is limited in the longer term by physiological of bodymass (g). Flight (circles)and resting(crosses) constraints,including substrate availability and valuesare passefine data from Teal(1969). Robin flight oxygen debt. Maximal power output in flight cost(-+1 SD) = r. Starlingflight cost(+1 SD) = s (af- ter Westerterpand Drent 1985). Regressionequations is thereforelikely to be greaterthan sustained are: Flight metabolism(Mar)= 53.70M•'ø3 (slope = output, and might be expectedto occurduring 1.03,SD = 0.05, t = 19.12,P < 0.001;log•0 constant = short, brief flights such as those documented 1.74, SD = 0.09, t = 20.01, P < 0.001;r 2 = 86.9, log•o by Teal (1969), Westerterpand Drent (1985:fig. S = 0.09). Resting metabolism (Mr,,t) = 16.98Mø'69 3), and the presentstudy. This is in accordance (slope= 0.69, SD = 0.14, t = 4.93, P < 0.001; log•o with the reanalysisof Teal's data, which indi- constant = 1.23, SD = 0.22, t = 5.45, P < 0.001; t'2 = cate power requirementsthat are proportional 55.1,log,o S = 0.14).Night basalmetabolism (BMR) = to body mass(M•'ø3; see Fig. 4.). 5.03Mø.726 (after Aschoffand Pohl 1970a).Body mass The high costof flight (23.4 kJ/h) reported (g) = M. here for the EuropeanRobin is a consequence of the short flight duration. It is thought to provide a realisticestimate for the costof non- Although accuratepredictions of flight and migratoryflights becausethe gait exhibitedwas other costsare important when constructing similar to that observed in the natural situa- TAL estimatesof DEE, they must be considered tion. As mentioned, structural aspects of the in conjunction with limitations of the time- tobin's wing and the slow flight speed indicate budgettechnique. It is noteworthythat in most that shortflights are likely to involve high cost. doubly labeled water studieswhere it was rel- Implicationsfor energeticsstudies.--The discus- atively easyto obtain an accurateestimate for sionof flight costindicates that attentionshould time spent in flight as a result of its conspicu- be paid to the foragingmode employedby the ousness or duration, this has correlated well specieswhen constructinga TAL budget. Thus, with the DEE (Table 8). There will be no vari- aerial-foragingspecies are adaptedfor low-cost ation in DEE due to the flight component,how- flights of 2.9-5.7 x BMR (Hails 1979) or 4.8 x ever, unlessthere is a significantdifference be- BMR (Flint and Nagy 1984),whereas foraging tween individuals in the amount of time they that involves many short flights may involve spend in flight. A priority in future studiesof costsas high as 21 x BMR (this study), or even this type should be to ensure that the flight higher becausethe costis not proportional to componentis accuratelysampled, even though basal metabolism (Fig. 4). Current allometric it may form only a smallproportion of the dai- flight equations (Hart and Berger 1972, Ken- ly activity. deigh et al. 1977,Hails 1979) employedto pre- Interest in the field of free-living energetics dict flight costsin TAL studiesgive estimates studieshas centered on obtaininga TAL model that are intermediate between these extremes. that will accuratelypredict DEE as revealedby 178 TATNERAND BRYANT [Auk, Vol. 103

TABLE8. Correlation between DEE and flight parameters.

Corre- lation Source Species (r) Flight parameter Flint and Nagy 1984 Sooty Tern (Sternafuscata) 0.92 Time in flight Bryant and Westerterp 1980 House Martin (Delichonurbica) 0.49 Percentageof period in flight Westerterpand Bryant 1984 Barn Swallow (Hirundo rustica) 0.85 Percentageof period in flight Bank Swallow (Ripariariparia) 0.82 Percentageof period in flight Westerterp and Drent 1985 EuropeanStarling (Sturnusvulgaris) 0.94 Time in flight Weathersand Nagy 1980 Phainopepla (Phainopeplanitens) None Time in flight Williams and Nagy 1984 Savannah Sparrow (Passerculussandwichensis) None Time in flight Weathers et al. 1984 Loggerhead Shrike (Laniusludovicianus) None Time in flight Nagy et al. 1984 JackassPenguin (Spheniscusdemersus) 0.88 Distance swum This study EuropeanRobin (Erithacusrubecula) 0.90 Time in flight

the doubly labeledwater method(Weathers and they concludedthat three of the TAL models Nagy 1980,Williams and Nagy 1984,Weathers examinedyielded mean estimatescomparable et al. 1984, Bryant et al. 1985). Weathers et al. to those given by the DLW method, but that (1984)emphasized the importanceof obtaining there was no correspondencein the variability. an accurateestimate of the operative tempera- When their DLW resultsare plotted as a func- ture (Te;Bakken 1976)and using measureden- tion of TAL predictions, there are no signifi- ergy equivalentsfor variousbehaviors. The lat- cant correlations (Table 9). This lack of corre- ter aspectwas endorsedhere by the comparison lation appearsto be due to a single individual of resting metabolismresults for Scottishand (No. 144), which was the least intensively ob- Russianrobins. Operative temperaturewas not served. Exclusion of this bird reveals an im- determined in the present study becausethe proved level of correlation between DLW and birds were sheltered from the effects of sun TAL results (Table 9), to the extent that the and wind. This may have affectedthe nonflight model of Holmes et al. (1979) provides a sig- estimatesof DEE (Table 4), but the similarity in nificantcorrelation. Although it is of question- predictionsof flight costindicate that this was able validity to have excluded individual 144, of minorsignificance. Williams and Nagy (1984) this reinterpretation does offer some encour- published an analysisof DEE for the Savannah agementfor accurateprediction of intraspecific Sparrow (Passerculussandwichensis) in which DEE variability using TAL methods.It also in- dicatesthe importanceof both increasingthe number of individuals sampled in each study TABLE 9. Correlation between DLW and TAL mea- suresof daily expenditure for the SavannahSpar- and obtainingan accuratemeasure of time spent row (after Williams and Nagy 1984). in brief but costlyactivities if the conclusions in this type of study are to withstand critical Coefficient of appraisal. determination (r 2) % ACKNOWLEDGMENTS Exclud- All ing bird We thank Professor Wilson, Dr. A. E. Fallick, and Model data No. 144 Mr. T. Donnelly of the ScottishUniversities Research Schwartz and Zimmerman 1971 0 0 & Reactor Centre for facilities and advice with the Utter and LeFebvre 1973 24 53 processingof doubly labeled water samples.Mr. A. Walsberg 1977 0 12 Annan and Mr. A. Michaels of the University of Stir- Kendeigh et al. 1977 0 0 ling MicroprocessorGroup manufacturedour data Holmes et al. 1979 11 89 a logger, for which we are most grateful. Elizabeth N. Walsberg 1980 19 60 Flint, R. E. Macmillan, and J. N. Mugaas made valu- Mugaas and King 1981 5 73 able comments on the manuscript. This work was Significant at 1%. undertaken on NERC grant No. GR3/4185A. January1986] Costof Flightin theRobin 179

LITERATURE CITED KESPAIK,J. !968. [Heat production and heat loss of swallows and martins during flight.] Eesti Nsv ASCHOI•I•,J., & H. POHI•. 1970a. Der Ruheumsatz von teaduste akadeemia toimetised. XVII Kaoide Bio- V/•geln als Funktion der Tageszeitund der K/•r- loogia No. 2: 179-190. pergr/•sse.J. Ornithol. 111: 38-47. KOPLIN, J. R., M. W. COLLOPY,A. R. BAMMAN, & H. ß& . 1970b. Rhythmic variationsin en- LEVENSON.!980. Energeticsof two wintering ergy metabolism.Federation Proc. 29: 1541-1552. raptors. Auk 97: 795-806. BAKKEN,G.S. 1976. A heat transferanalysis of an- LASIEWSKI, R. C., & W. R. DAWSONß 1967. A re-ex- imals:unifying conceptsand the applicationof amination of the relation between standard met- metabolismchamber data to field ecology. J. abolicrate and body weight in birds.Condor 69: Theoret. Biol. 60: 337-384. 13-23. BRODY,S. 1945. Bioenergeticsand growth. New LEFEBVRE,E.A. 1964. The use of D20 •s for measur- Yorkß Hafner Publ. Co. ing energy metabolism in Columbalivia at rest BRYANT,D. M., C. J. HAILS, & R. PRYS-JONES.1985. and in flight. Auk 8!: 403-4!6. Energy expenditure of free-living Dippers (Cin- LIFSON, N'., G. B. GORDON, & R. MCCLINTOCK. 1955. clus cinclus) in winter. Condor 87: 177-186. Measurementof total carbondioxide production , --, & P. TATNER. 1984. Reproductive by meansof D2•80.J. App1.Physiol. 7: 704-7!0. energeticsof two tropical bird species.Auk 101: 25-37. LYULEEVA,D.S. !970. Energy of flight in swallows and swifts. Dok. Akad. Nauk. SSR !90: !467-!469. ß •r K. R. WESTERTERP.1980. The energybud- MARSHßR. L., C. CAREYß& W. R. DAWSON. 1984. Sub- get of the House Martin (Delichonurbica). Ardea 68: 91-102. strate concentrationsand turnover of plasma CALDERßW. A., • J. R. KINGß 1974. Thermal and glucoseduring cold exposurein seasonallyac- climatized House Finches Carpodacusmexicanus. caloric relations of birds. Pp. 259-413 in Avian J. Comp. Physiol. B !54: 469-476. biology,vol. 2 (D. S. Farnerand J. R. King, Eds.). ß & W. g. DAWSON. 1982. Substrate metabo- New Yorkß Academic Press. EAST, M. 1980. Sex differences and the effect of tem- lism in seasonallyacclimatized American Gold- finches.Amer. J. Physiol.242: R563-R569. perature on the foraging behaviour of robins (Erithacusrubecula). Ibis 122: 517-520. MCNAB,B.K. !980. On estimatingthermal conduc- tance in endotherms. Physiol. Zool. 53: !45-!56. ß 1982. Time budgetingby EuropeanRobins MUGAAS,J. N., &. J. R. KINGß 1981. Annual variation Erithacus rubecula; inter- and intra-sexual com- of daily energy expenditureby the Black-billed parisonsduring autumnßwinter and early spring. Ornis Scandinavica 13: 85-93. Magpie: study of thermal and behavioural en- ergetics.Stud. Avian Biol. No. 5. FLINT,E. M., & K. A. 1N[AGY.1984. Flight energetics 1N[AGY,K.A. !980. CO2production in animals:anal- of free-living Sooty Terns. Auk 101: 288-294. ysis of potential errors in the doubly-labelled GREENEWALT,C.H. 1962. Dimensionalrelationships water method. Amer. J. Physiol. 238: 466-473. for flying animals.Smithsonian Misc. Coil. 144: ß W. R. SEIGFRIEDß•r R. P. WILSON. 1984. En- !-46. ergy utilisationof free-rangingJackass Penguins ß !975. The flight of birds. Trans.Amer. Phil. Spheniscusdemersus. Ecology 65: !648-!655. Soc. 65: !-65. NEWTON,I. 1968. The temperaturesßweights and HAILSßC.J. !979. A comparisonof flight energetics body compositionof molting bullfinches.Con- in hirundinesand other birds. Comp. Biochem. dor 70: 323-332. Physiol. 63A: 58!-585. NORBERG,R.A. !98!. Optimal flight speed in birds HART,J. S., & M. BERGERß1972. Energeticsßwater when feeding young. J. Anim. Ecol. 50: 473-477. economy and temperature regulation during NORBERG,U. M. 1979. Morphology of the wingsß flight. Proc.!Sth Intern. Ornithol. Congr.:189- legs and tail of three coniferous forest titsßthe !99. goldcrestand the treecreeperin relation to lo- HOOGMAN, C. D. (Ed.). 1959. Handbook of chemis- comotot pattern and feeding station selection. try and physics.Cleveland, Ohioß Chemical Rub- Phil. Trans. Royal Soc.London 287: 13!-!65. ber Publ. Co. PENNYCUICK,C. J. !968. A wind tunnel study of HOLMESßR. T., C. P. BLACK, & T. W. SHERRY. !979. gliding flight in the pigeon Columbalivia. J. Exp. Comparativepopulation bioenergeticsof three Biol. 49: 509-526. insectivorous passetines in a deciduous forest. ß !975. Mechanicsof flight. Pp. !-75 in Avian Condor 81: 9-20. biologyßvol. 5 (D. S. Farnerand J.R. King, Eds.). KENDEIGH, S.C., V. R. DOLNIK, & V. M. GAVRILOV. New Yorkß Academic Press. !977. Avian energetics.Pp. !27-201 in Graniv- POHLßH.ß •r G. C. WEST. 1973. Daily and seasonal orousbirds in ecosystems(J. Pinowski and S.C. variation in metabolicresponse to cold during Kendeigh, Eds.). I.P.B. Handbook No. 12. Cam- restand forcedexercise in the CommonRedpoll. bridgeßEngland, Cambridge Univ. Press. Comp. Biochem. Physiol. 45: 85!-867. 180 T^TNER^NIV BRYANT [Auk,Vol. 103

PYKE,G. H. 1981. Optimal travel speedsof animals. breeding season;estimates using D2•80 and time Amer. Natur. 118: 475-487. budget methods.Ecology 54: 597-604. RAYNER,J. M.V. 1979a. A new approach to animal W^LSBERG,G. E. 1977. Ecology and energeticsof flight mechanics.J. Exp. Biol. 80: 17-54. contrastingsocial systemsin Phainopeplanitens. ß 1979b. A vortextheory of animal flight. Part Univ. California Publ. Zool. 108: 1-63. 2, The forward flight of birds.J. Fluid Mech. 91: 1980. Energy expenditure in free-living 731-763. birds: patterns and diversity. Proc. 17th Intern. 1982. Avian flight energetics.Ann. Rev. Ornithol. Congr.: 300-305. Physiol. 44: 109-119. 1983. Avian ecologicalenergetics. Pp. 161- RYAN, A. R., JR., B. L. JOINER,& B. F. RYAN. 1976ß 220 in Avian biology, vol. 7 (D. S. Farner and J. MINITAB student handbook. Boston, Duxbury R. King, Eds.). New York, Academic Press. Press. WEATHERS,W. W., W. A. BUTTEMER,A.M. HAYWORTH, SCHWARTZ,R. L., & J. L. ZIMMERMAN. 1971. The time & K. A. NAGY. 1984. An evaluation of time- and energybudget of the male Dickcissel(Spiza budgetestimates of daily energyexpenditure in americana).Condor 73: 65-76. birds. Auk 101: 459-472. SHILOV,I.A. 1968. Heat regulation in birds. Spring- --, & K. A. N^GY. 1980. Simultaneousdoubly fieldßVirginiaß U.S. Dept. Commerce. labeled water (3HH•SO) and time-budget esti- TEAL,J. M. 1969. Direct measurementof COe pro- matesof daily energyexpenditure in Phainopepla duction during flight in small birds. Zoologica nitens. Auk 97: 861-867. 54: 17-23. WEIs-FOGH, T., & R. MCN. ALEXANDERß 1977. The TORRE-BuENO,J. R. 1977. Respirationduring flight sustainedpower output from striatedmuscle. Pp. in birds. In Respiratoryfunction in birdsßadult 511-525 in Scale effects in animal locomotion (T. and embryonic(J. Piiper, Ed.). BerlinßSpringer- J. Pedley, Ed.). London, AcademicPress. Verlag. WESTERTERP,K. R., •r D. g. BRYANTß1984. Energetics TUCKER,V.A. 1968. Respiratoryexchange and evap- of free existence in swallows and martins (Hir- orative water lossin the Budgerigar.J. Exp. Biol. undinidae) during breeding: a comparativestudy 48: 67-87. using doubly-labelledwater. Oecologia62: 376- 1970. Energetic cost of locomotionin ani- 381. mals. Comp. Blochem.Physiol. 34: 841-846. ß & g. H. DRENT. 1985. Energetic costsand ß 1973. Bird metabolismduring flight. J. Exp. energy saving mechanismsin parental care of BioL 58: 689-709. free-living passerinebirds asdetermined by the UDV^RDY, M.D. F. 1953. Contributions to the D2•80 method. Proc. 18th Intern. Ornithol. knowledge of body temperature in birds. Zool. Congr.: in press. Bidrag.Uppsala 30: 25-42. WILLS^MS,J. B., & K. A. N^GY. 1984. Daily energy UTTERßJ. M., & E. A. LEFEBvRE.1970. Energy expen- expenditure of SavannahSparrows; comparison diture for free-flight by the Purple Martin (Progne of time-energybudget and doubly-labeledesti- subis).Comp. Blochem.Physiol. 35: 713-719. mates. Auk 101: 221-229. ß & ---. 1973. Daily energy expenditure Z^R, J. H. 1974. Biostatisticalanalysis. Englewood of Purple Martins (Progne subis) during the Cliffs, New Jersey,Prentice Hall.

ERRATA

In the memorial to Edmund W. Mudge (1985, Auk 102: 869), "Edmund W. Mudge" should read "Edmund W. Mudge, Jr." There is an error in the article by SpencerG. Sealy(1985, Auk 102: 889-892). At the end of paragraph2 on page889, the sentencebeginning "In 1979a single 9-day-oldcuckoo gradually assumed an erectposture ß.." should read "In 1985 .... "