King 1983), it is impossible to determine whether or not Ta influences attentiveness in- dependently of Te without experimentally un- coupling the two. We are ...
THE EFFECT OF EGG TEMPERATURE ON ATTENTIVENESS IN THE BELDING'S SAVANNAH SPARROW STEPHEND. DAVIS, JOSEPHB. WILLIAMS, WENDY J. ADAMS, AND STEPHANIE
L. BROWN
Natural Science Division,Pepperdine University,Malibu,California90265 USA
ABSTRACT.--We manipulatedegg temperatures in nestsof the Belding'sSavannahSparrow, (Passerculus sandwichensis beldingi), an open-nestingpasserineof which only the female incubates.Femalesincreasedtheir attentivehessby 22.0%when we cooledtheir eggsand decreased their attentivehess by 28.4%when we warmedtheir eggs.The rangeof adjustment
approximated the upperandlowerlimitsof attentivehess observed for ourpopulationduring onebreedingseason; adjustments wereprimarilymadeby changing the lengthof both attentiveand inattentiveperiods.Our resultsindicatethat thesebirdsadjusttheir attentive time in directresponseto fluctuations in egg temperature,independentlyof diurnal fluctuationsin air temperature. Theyalsosupportthe hypothesis thatattentivehess isultimately controlledby an internalrhythm and is only modifiedby, but not completelyexplainedby, responses to eggtemperature.Received 14September 1983,accepted 6 February 1984.
AVIAN egg temperature(Te) must be controlled within narrow limits to insure proper
embryologicaldevelopment(Lundy 1969,Romanoff and Romanoff 1972, Weinrich and Bak-
er 1978).Many speciesaccomplishthis by continuouslyincubatingtheir eggs;parentseither share in the responsibility of incubation, or a nonincubatingpartnerfeedsits incubatingmate (Drent 1975, Carey 1980). Other speciesengage in intermittent incubation by one parent. In this case, the maintenance of Te within narrow
limits is especially challenging. During daylight hours,the incubatingparent (usuallythe female) must compensatefor fluctuationsin environmental temperatureand also take sufficient time away from the nest to feed. This controlof T• is accomplishedprimarily through adjustmentsof attentiveness(Kendeigh 1952, White and Kinney 1974, Yom-Tov et al. 1978) and secondarilyby regulation of heat transfer
in attentiveness is to regulate T,, the most effective mode of regulation should be via the direct detection of and responseto T•. This is suggestedby a numberof investigators.For example, Drent (1972:259) suggestedthat the key stimulusto guide a parent in adjustingits incubationrhythm is T• measuredat the moment the parent returns to the nest after a recess. Experimentalevidence to supportthis hypothesis, however, is at best indirect.
Using hole nesters,von Haartman (1956) and Johnsonand Cowan (1974) artificially heated nest boxes in the field to determine
the influ-
ence of ambient temperatures at the nest on attentiveness. Von Haartman reported a decreasein attentiveness (sessionsat the nest) with
an increasein nest-boxtemperatureon the part of his Pied Flycatchers(Ficedulahypoleuca), but Johnson and Cowan found no response in CrestedMynas (Acridotheres cristatellus). Tewas throughthe brood patch (Drent et al. 1970, not measuredand therefore not separatedfrom
Haftorn 1982).
Many studiesdemonstratea correlationbe-
Ta of the nest-box complex in either of these experiments.
tween attentivenessand air temperature (Ta)
White and Kinney (1974) presentedindirect
(Kendeigh1952,White and Kinney 1974,Yom-
evidence that sensoryperception of Tethrough
Tov et al. 1978), but, because Ta and mean T,
the broodpatchof the incubatingfemaleis used
may vary in parallel (Huggins 1941, Yom-Tov
as a feedback
et al. 1978, Zerba and Morton 1983, Webb and
Working with Village Weaverbirds(Ploceuscucullatus),they found that anesthetationof the
King 1983), it is impossible to determine whether or not Ta influences attentiveness in-
dependentlyof Tewithout experimentallyuncouplingthe two.We arenot awareof anystudy in which
this has been done.
stimulus
to control attentiveness.
broodpatchof incubatingfemalesincreasedattentiveness.They speculatedthat the anesthetation causeda reduction of the frequency of
dischargeof sensoryreceptorslocatedin the
Becausethe primary function of adjustments skin of the brood patch,which encouragedat556
The Auk 101: 556-566. July 1984
July1984]
Effect ofEggTemperature onAttentivehess
tentiveness.They alsofound high coolingrates of eggswithin different neststo be associated with high attentiveness,implicatingan influence of Te.From a graphical analysisof attentiveness versus Ta, they found that a rectangular hyperbolic function fit their data best. This function predicted intercept Ta for zero attentiveness to be 37.0øC + 0.6 (95% confi-
dence interval), which correspondedcloselyto their mean maximum Te during the day (36.7øC + 1.5 SD) and during the night
557
permanentresidentof the coastalmarshesof Southern California and a speciesin which only the female incubates,as our study species.Our study site was a
saltmarshneighboringthe easternarm of PointMugu Lagoonat the Point Mugu Naval Air Station near Oxnard, California (34ø06'30"N, 119ø05'W). Male Sa-
vannah Sparrows established territories in early March, and breeding occurredfrom March to July. Femalesconstructednestsjust aboveground level in low-profile halophytic succulentsand grasses.The nestswere open but shadedfrom direct solar radia-
tion during mostof the day. The mean clutchsize was 3.2 eggsñ 0.05 (SE,n = 356), and the incubation (36.7øC+ 1.4 SD). From these results, White and period was 13.2 days ñ 0.3 (SE, n = 6). Kinney hypothesizedthat 37øCcorrespondsto Egg temperature.--Wemeasuredegg temperatures a "releasetemperature"of the egg, which the with 40-gaugethermocouplewire insertednear the incubating female detectsby the brood patch centerof one egg in a clutchand cementedin place asa cue to leave the nest. They alsoarguedthat with a dental resin (Caulk Grip Cement, L. D. Caulk the detection of Ta near 37øCpromotes cessa- Co., Div. Dentsply Intern. Inc., Milford, Delaware). tion of incubation. Temperatureswere continuously monitored by po-
Recently,Zerbaand Morton (1983)foundthis "releasetemperature"in the Mountain Whitecrowned Sparrow (Zonotrichia leucophrys oriantha)to be relatively constantat 38.0øC+ 1.5 SD throughout the day in an environment with large temperaturefluctuations.They concluded that incubatingfemalesmustbe acutelysen-
tentiometric recorder (Linear, Model 142, Linear In-
sitive to Te.
from
The purposeof our studywas to separatefor the first time the effects of Te from those of Ta
strumentsCorp., Irvine, California). We calibrated thermocoupleand recorderoutput eachday by comparisonwith a standardthermometercalibratedto the National Bureau of Standards.Temperature read-
ings were accurateto within _+0.2øC.Visual comparisons verified that chart recordings accurately representedthe time that a bird arrived and departed the nest to within
20 s.
Definitionsand computations.--We defined "attentiveness"
as the number
of total minutes
a bird in-
on attentivenessin a speciesdisplaying intermittent incubation by one parent. We did this
cubatedher eggsfor a given hour, "attentiveperiod" as the number of minutesspent on the eggsper visit by controllingTein neststhroughoutthe day, to the nest, and "inattentive period" as the number in situ, by either warming or coolingthe eggs of minutes spent off the eggs per recessfrom the nest. We computedmean T• from the chart recordand observingresponsesin attentivebehavior. ingsby meansof a digitizer (Hipad, HoustonInstruWe selected two treatment levels for our conment, Austin, Texas)and softwarethat computedthe trol of T•: one near the mean minimum Te, to mean Teper hour, per day period, and per night peapproximatethe normaltemperaturesensedby riod from 100 equally spacedtemperatureobserva-
an incubating parent at the moment of its re-
tions for each hour (one temperature measurement
turn to the nest after a recess, and the other
per 36 s).Therefore,a meanT, for a 24-hperiodrep-
near the mean maximum Te, to represent the resentedthe mean of 2,400temperatureobservations. normal temperatureat the moment when an We verified that the measurement and statistical incubation sessionstopped and recessbegan. computationof mean T, were accurateby comparing Therebywe obtaineddirectevidenceto testtwo them with manual measurementsmadedirectly from
hypotheses: (1) the hypothesisof Drent (1972) that birds use Te measured at the moment of return to the nest after a recessto adjust atten-
tiveness,and (2) the hypothesisof White and Kinney (1974)that birds usea "releasetemperature" of their eggsto cue departurefrom the nest.
MATERIALS
AND METHODS
the chart recordingsand with the SPSSstatistical
package(Nie et al. 1975).We definedthe meanmaximum Teper hour as the mean of the egg temperatures taken at the end of each attentive period for a
given hour.Similarly,we calculatedthe meanminimum T• per hour from the beginningof eachattentive period.Mean T, per hour was taken from the U.S. Navy WeatherStationlessthan 1 km from the
studysite.We alsorecordedair temperatures in the vegetationadjacentto nestswith a shieldedthermocouple.
Birdsand studyarea.--We chosethe Belding's Savannah Sparrow (Passerculus sandwichensis beldingi),a
Apparatusto manipulate eggtemperature.--We designedan apparatusto deliver heatedor cooledair
558
DAVISETAL.
to the bottom of a nest just beneath the eggs.It consisted of a 15-mm-diameter,3-m-long copper tube enclosed within an insulated 30-mm-diameter copper tube. A variable-speedfan gently forced air through the inner tube to the nestwhile polyethyleneglycolwascirculatedby a refrigeratedwater bath (Neslab, Model RTE-8, Neslab Instruments
Inc.,
Portsmouth,New Hampshire)through the outer tube to regulatethe temperatureof the air stream.We used thermocouplesto moniter air-streamand water-bath temperatures.We placedthe fan 20 m and the water bath 50 m from the nest to reduce motor
noise at the
nest.
Air velocities
delivered
to the bottom
of the nest
varied from 0 to 2 m/s while the bird was on the nest
(measuredwith a hot wire anemometer, Model HWA-
[Auk, Vol. 101
cooled,and left untreatedon alternatedaysfrom 0600 to 1800.We used data from the untreated day as our control
for
each nest.
We
varied
the order
of the
treatmentsamong neststo determine whether or not the sequenceof treatmentsaffected results. An example of digitized T, data for a typical nest on three consecutivedays of control, cool, and heat treatmentsis shownin Fig. 1. Becausethe incubating female aided the heating of eggs but hindered the coolingof them, Teon the heatedday is nearly constant, whereas T, on the cooled day is not. T, on the cooledday continuesto reflectan incubationpattern. Visual inspectionof this pattern indicatesthat attentive periodsare longer on the cooledday than on the control day. We maintained T• below the mean minimum T, of
103, Thermonetics Corp., San Diego, California). the control on cooleddaysand abovethe mean maxDuring cooling treatments,we manually increased imum T, of the controlon heateddays(Fig. 1). This velocitiesto 4 m/s when the bird was off the nest by allowed us to test our hypothesisthat T• directly afmeans of a variable rheostat attached to the fan mofectsattentive behavior,specificallythe T, at the motor. Smokeplacedin the air streamshowedthat most ment when the female returns to the nest for a session (mean minimum Te) and at the moment when of the air followed the path of least resistanceout of the bottom and sides of the nest and not between the she departs from the nest for a recess(mean maxieggsinto the nest cup. This finding was supported mum T•, "release temperature"). by the measurementof air velocities10 mm above the eggs of less than 0.1 m/s (below background RESULTS levels)at maximum fan speeds.Thermocouplesplaced in one typical nestcup indicateda maximumdifferNormal attentivepatternof the population.--Atence of 3.0øCbetween the ambient temperature and tentiveness at seven nests varied with the time the temperature10 mm above the eggsand no difof day on 22 days between 27 April and 10 July ferencebetweenthe two temperaturesat 50 mm above 1979 (Fig. 2A). A polynomial regressionof the the eggs.This indicated that Tewas controlledprimarily by conductiveheatexchangebetweenthe fine- data predicted a minimal attentivenessof 31.9 mesh nesting material and the undersurfaceof the min at 1257 (obtainedby setting the derivative eggsand not through forced convectionby air cur- of attentiveness with respect to time equal to rents. Becauseairflow into the nest cup was presum- zero). Mean attentivenessduring daylight hours ably even more restrictedwhen the bird was on the (0500-2000) was 39.3 min _+ 0.51 (SE, n = 383). nest (due to increased resistance to air flow and our
The mean number of attentive periodsper hour
reduction of fan speed),we assumedthat air flowing into the nest contributed little to the temperature experiencedby the female through the broodpatch. Experimental protocoL--Ourexperimentalobjectives were twofold. First, we attempted to describe the normal attentive pattern of our population under various conditions of weather and microhabitat by measuringTa, T•, and attentivenessthroughoutthe breeding seasonof 1979 (27 April to 10 July) at 7 nestsfor 22 days. We selectedthree different microhabitatswithin our study site, which offered different conditionsand vegetationaltypes (Salicorniavirginica,Frankenia grandifolia, and Monanthocloe littoralis) for nesting. Second,we used these resultsto design an experiment to manipulate Tein situ during the breeding seasonof 1980.We selectedfour representativenests, very similar in microhabitat(all in Salicornia vegeta-
was 3.4 + 0.08 (SE, n = 383).
tion and within
a 1-ha area), time of incubation
(within 20 days of each other, 8-28 June), and exposure to daily To.In each nest, eggs were heated,
Attentivepatternof manipulatednests.--Attentivenessat four nestson 15 daysbetween 8 and 26 June 1980 varied with mean hourly Ta(Fig. 2B). A linear regression of the data (y = 73.102 - 1.946x) predicted an x-intercept (extrapolation to 0 attentiveness) at 37.6øC+ 8.4, -7.8 (95% confidenceinterval), a temperature not significantly different from the mean maximum T• per hour of 38.3øC(P > 0.50). We manipulated T, independently of T• in thesesamefour nests,giving a total of 5 cooled, 5 heated, and 15 untreated (control) days (Fig. 3). Each treatment period was continuous for 12 h (0600-1800).
Except during the heat treatment at 1000 at nest C and at 0900-1500 at nest D, females generally increasedtheir attentivenesswhen eggs were cooled and decreased
attentiveness
when
July1984]
559
Effectof EggTemperature onAttentivehess
control J
I
I
I
I
I
/
O9
10
11
12
16
17
18
19
_cooled
i 20Fheated 0 /
•
06
07
0
• 14
15
40-
2O 0
13
HOUR
OF DAY
Fig. 1. Example of digitized egg-temperaturedata from a manipulated nest from 0600 to 1900 Pacific StandardTime (PST)on a control day (16/6/80), cooledday (17/6/80), and heatedday (18/6/80).
eggswere heated, relative to control days (Fig. 3). The aberrant increase in attentiveness be-
tween 0900 and 1500during the heat treatment at nestD coincidedwith the onsetof high winds (gustingto 46 km/h). We know that the female increased her attentivenessat the nest during the windy period, but we cannotbe certain that she was actually incubating her eggs,because T, was held constant during heat treatments (cf. Fig. 1). We also collected data on an "untreated" nest on this same day of high winds,
nificantly increase rates of water loss, and Walsberg(1983) demonstratedthat birds do not respond to manipulated nest humidities. Pos-
sibly,the femalewasattemptingto compensate for an increasein the cooling rate of eggs associated with high winds. We tested this by artificially ventilating three eggs in a nest, in situ, with a 10 m/s (36 km/h) air velocity. Cooling rates of the eggs (from 40.0øC down to 35.0øC,with a T, of 26.0øC)increased from 2.2øC/
cubationasindicatedby T, recordings(data not
min to 9.8øC/min. Cooling rate was not a factor on the heated day of Fig. 3D, however, because we continuously held eggs at or above 38øC.
shown).
We also observed
however, and found a dramatic increase in in-
The
reason
for this increase
in attentiveness
with the onsetof high winds is not clear. It is unlikely that the incubating female was attempting to prevent egg dessication.Tracy and Sotherland (1979) have shown that movement of air acrossthe surface of an egg will not sig-
that the female
had extreme
difficulty in flight navigationduring the windy period, and it may be that she remained on her nest to conserveenergy (cf. Tucker 1971). Figure 4A shows combined data for attentiveness and mean T, for heated, cooled, and
untreateddays.A Bartlett'stestof homogeneity
560
DAVIS ETAL.
[Auk, Vol. 101
60
50
40
30
•o
IO
A o O3
o5
o7
o9
il
13
15
t7
i9
21
Hourof day (PST) 6O
50 40 3O 2O
B
Meanair temperature(øC) Fig. 2. (A) Attentivehessduring daylight hours (0500-2000, PST) at seven nestsfor 22 daysand 383 h during the breeding seasonof 1979. Polynomial regression,y = 85.87 - 8.34 x 10 2x + 3.22 x 10 Sx2, F = 155.67,r2 = 0.45, n = 383. Dashedlines representa range of attentivehessextending 22.0%higher and 28.4% lower than the polynomial regression,from 0600 to 1800. (B) Attentivehessversus mean air temperature during daylight hours at four nestsfor 15 daysand 270 h, between 8 and 26 June 1980.Linear regression, y = 73.102 - 1.946x,r2 = 0.51, n = 270, P < 0.001.
of variances between the three treatments (So-
kal and Rohlf 1969: 370) indicated a significant heteroscedasticity (X2 = 73.23, P < 0.001). Therefore, we tested the significance of the treatments nonparametrically. A Mann-Whitney U-test indicated a significant increasein
attentiveness on cooled days (t = 22.50, P < 0.001) and significant decreasein attentiveness on heated days (t = 7.30, P < 0.001) relative to untreated (control) days. Furthermore, a Kruskal-Wallis
test for each of the 12 h of treatment
indicated that the treatments differed signifi-
July1984]
Effect ofEggTemperature onAttentiveness
30'
g 2o I0
!
'
,....,..
561
.
"¾..,.'
/ &..
• 60
+- :'"' - '¾? 'c-
'i
40
•
\ .j-.,
..
20
\ .:
\:
"•,'
:
I0
":
..•
03 05
0709
II
13 15 17 19 21
05 07 09
II
13 15 17 19 2t
Hourof day (PST) Fig. 3. Attentivenessversushour of the day for four birds, labeled A, B, C, and D, and three treatments, labeled heated (•), cooled(&), and untreated (•) days.
cantlyat all hours(rangefrom X• = 7.04to 16.67,
8.5øCabove the control on heated days. Fur-
P < 0.05 to P < 0.001). The mean T• was 30.6øC + 0.39 (SE, n = 60)
thermore, when our birds returned to the nest
for cooleddays,35.5øC+ 0.15 (n = 180) for control days, and 40.4øC_+0.17 (n = 60) for heated days (Fig. 4A). Egg temperaturesduring manipulation were within the range normally experienced by the incubating female and approximated the upper and lower temperature limits experiencedupon departure from (mean maximum T• per hour = 38.3øC + 0.14 SE, n = 178) and arrival at the nest (mean minimum T• per hour = 31.1øC _+ 0.24, n --- 178). The mean
T, of 40.4øCon heated days was close to the surfacetemperature of the brood patch of incubating females (40.2øC _+0.42 SE, n = 8).
On cooled days, the mean minimum T, per hour was 23.9øC _+ 0.59 (SE, n = 58), and the mean maximum T• was 33.3øC _+ 0.45 (n = 56). On heated days, the mean minimum T• was 39.6øC_+0.21 (n = 55), and the mean maximum
T• was 40.7øC- 0.17 (n = 54). Therefore, T• perceivedby the incubatingfemale at the moment of return to the nest after a recessaveraged 7.2øCbelow the control on cooled days and
on heated days they immediately experienced a T• at or above their "release temperature" of 38.3øC(mean maximum T, of the control); at no time, however, did they experience T• as high
as a "releasetemperature"on cooleddays. Because
the
treatments
were
on
different
days,our data from untreateddaysrepresented an adequate control only if T• was similar among days. Fortunately, ocean air currents ameliorated temperature fluctuations at our study site (Fig. 4B). A two-way analysisof variance, with treatmentsand hour of the day as covariates,indicated that T• did not differ significantly between days (F = 1.98, P > 0.1). Attentive and inattentive periods at control nests.--Out of 11 nests measured during the breeding seasonsof 1979 and 1980, correlation analysis (Pearson product-moment, Nie et al.
1975:280) of attentive and inattentive periods versus mean hourly T, between 0600 and 1800 indicated that there was a negative correlation between attentive period and T• (P -< 0.05) at 7 nests and a positive correlation between inat-
562
DAVIS ETAL.
[Auk,Vol. 101
25
6O 5O 4O
50
o
55
•
c:: 5O
40 • c:: 20
"•
45 •
I0 o
0:5
05
07
09
II
15
15
17
19
21
Hourof doy (PST) Fiõ.4. (A) Attentivehess versushourof the dayfor heated(I), cooled(&),anduntreated (O)days,with associated meaneggtemperatures on heated(....... ), cooled(.... ), and untreated(.... temperatureversushour of the day for heated(I), cooled(&), and untreated(O) days.
tentive period and T, (P -< 0.05) at 8 nests(data not shown).Combineddatafor all 11 nestsgave a regressionequation of y = 20.86 - 0.496x (n = 1,390, r = -0.13,
P < 0.001) for attentive
period versus T, and y = 3.38 + 0.257x (n = 1,409, r = 0.08, P < 0.002) for inattentive peri-
od versusT,. Consideringthe high variability in attentive and inattentive periodswithin and among individual birds in a population, as observed here and by others (Kendeigh 1952, Weeden 1966, Haftorn 1979), the significance of these correlationsis probably real. Attentiveand inattentiveperiodsat manipulated nests.--High variability in comparing single observationsof attentive and inattentive pe-
o)days.(B) Air
riods versus Tais typical, and grouping data to enhance patterns that are present is common (Kluijver 1950, Kendeigh 1952, Weeden 1966, Haftorn 1979).We have employedsucha method by grouping our data into Tarangesof 1.0øC intervals so that eachpoint representsthe mean attentive or inattentive period for our four ma-
nipulatednestsin eachtemperaturerange(Fig. 5).
On control days, the attentive period correlated negatively with Ta,whereasthe inattentive period displayeda biphasicrelationship (Fig. 5). At Tabelow 13.5øC,inattentiveperiods decreasedwith increasing T,, whereas at T, above13.5øCthe patternwas reversed.Solving
July1984]
EffectofEggTemperature onAttentivehess
563
In Fig. 6A, attentive periods are compared with Taon control, cooled, and heated days. A comparison of the regression equation for cooled dayswith that for control days indicated that their slopes were the same (t = 0.36,
40
35
P >
0.50 but their
elevations
were
different
(t = 4.43, P < 0.001). A similar comparisonbetween heated days and control days also indicated equal slopes(t = 0.29, P > 0.50) and un3 '- 25 equal elevations(t = 5.17, P < 0.001).Therefore, it appears that the length of attentive periods 'o 20 O increased on cooled days but decreased on ß ß 17 heated days. This suggeststhat birds adjusted 12. 15 attentive periods in a direct response to Teindependent of the influence of Ta.Furthermore, it appearsthat even on cooledand heated days the negative correlation between attentive 5 • & 43 period and T, persisted. 28 24 In Fig. 6B, inattentive periods are compared 0 ' with T, for control, cooled,and heated days.A 6• 1'2 1'4 16 1182•0 2•2 comparisonof the regressionequation for heatMean air temperature (%) ed days with that of control days (between T• Fig. 5. Mean attentive (I) and inattentive (,) pe- of 13øCto 22øC)indicated that their slopeswere riodsversusmeanair temperature(groupedinto 1.0øC the same (t = 0.09, P > 0.50) but their elevaintervals)for manipulatednestson untreateddays. tions were different (t = 9.00, P < 0.001). A
•
30
.=_
Numbersunder eachdata point representthe number of observations
for
each
mean
value.
Linear
regressionfor attentive period versusmean T,, y =
similar comparisonbetween cooled days and control days (between T• of 7øCto 14øC)indi-
catedequal slopes(t = 1.42, P > 0.10) and equal elevations (t = 1.06, P > 0.20). Femaleslengthtentive period versusmean T, from 7øCto 14øC,y = 24.08 - 1.32x (r: = 0.41, n = 8, 0.10 < P > 0.05); for ened inattentive periods on heated days relainattentive period versus mean T• from 13øCto 22øC, tive to control days, and a positive correlation between inattentive periods and Ta persisted. y = -6.85 + 0.97x (• = 0.63, n = 10, P < 0.01). In contrast,on cooled days inattentive periods were shortened,but only at Tagreaterthan 18øC. The reason for no shift in inattentiveness at Ta the equationsfor the point of intersectiongave below 18øCis not apparent. an inattentive period of 6.2 min and a Ta of 52.75 - 1.86x (r• = 0.72, n = 16, P < 0.001); for inat-
13.5øC.This biphasic pattern might be explained by the fact that T• below 13.5øCoccurred only during early morning hours when the birds were first coming off their nest after
DISCUSSION
Female Savannah Sparrows consistently responded to cooled eggs by increasing attena long night. Theselonger inattentive periods tiveness and to heated eggs by decreasingatduring early morning hours may be in re- tentiveness, relative to control days. sponseto an intensified hunger drive. Attentiveness on control and heated days Kendeigh (1952: 41) found a similar pattern tracked T• throughout the day, whereas attenin the House Wren (Troglodytes aedon)when he tivenesson cooled days followed T• only durplotted the daily average inattentive period ing the morning hours (Fig. 4A). This may reagainst daily average Ta.He found inattentive flect an accumulated energy drain on the periods to be negatively related to Ta below incubating parent on cooled days, which re13øCbut positively related to Taabove 13øC.He suits in an increasein foraging time (decrease attributed this increasein inattentive period in attentiveness)after midday. Haftorn (1982) with Ta below 13øC to an increase in the need shows that eggs artificially cooled under field for food and, thus, an increase in the time reconditions cause the body temperature of an quired to searchfor food on cool days. incubatingfemale to decline, and Biebach(1979)
564
D^vls•T ^[.
[Auk,Vol. 101
daysreducedattentivenessby 28.4%+ 2.31 (SE, n = 60), whereas a decrease in mean T• of 4.9øC 45
on cooled days increased attentiveness by 22.0% + 0.98 (n = 60). These adjustmentswere primarily made by decreasingor increasing both attentive and inattentive periods (Fig. 6). Our results provide the first direct evidence that T, influencesattentivenessindependently of Tain an open-nesting passerine.In addition, they indicatethat Teis not the only factorcon-
4O
35
trolling attentiveness. At no time did a bird
discontinueincubation in responseto heating or continuously incubate in responseto cooling of the eggs. With the Village Weaverbird, White and Kinney (1974) found intercept T• for zero attentivenessto be 37.0øC+ 0.6 (95% confidence interval) and the mean maximum Te to be 36.7øC+ 1.5 (SD). They hypothesizethat when an incubating female returns to the nest, she warms her eggs until a "release temperature" (ca.37øC)is reached;this is detectedby sensory receptorsin the brood patch,cueing departure. Our data do not support this hypothesis.Our correspondingintercept Ta for zero attentive-
E 20 •D
15
lO
nessand mean maximum T, were 37.6øC(+0.8) -r'
•D
and 38.3øC(_+1.9), respectively.In our heat treatment,we continuouslyheld Teat or above this "release temperature" (ca. 38øC),and yet
15
Q. •D
•. •o •D
•
5
0
6
I
I
I
I
[
[
i
I
8
10
12
14
16
18
20
22
Mean air temperature (•c)
Fig. 6. (A) Attentive periods and (B) inattentive periods versusmean air temperature (grouped into 1.0øCintervals) for manipulated nestson heated ([2), cooled (/x), and untreated (0) days. Numbers under eachdata point representthe number of observations for eachmean value. Numbers and linear regressions for control data are the sameas in Fig. 5. (A) Linear regressionfor attentive period versus mean T, on heateddays, y = 39.70 - 1.68x(r2 = 0.40, n = 10, P < 0.05) and cooled days, y = 58.29 - 1.62x, (r2 = 0.47, n = 10, P < 0.05). (B) Linear regression for inattentive period versus mean Ta on heated days, y = 1.27 + 1.00x, (r2 = 0.70, n = 12, P < 0.001) and cooled
days, y = 17.01 - 0.48x (r2 = 0.33, n = 11, 0.10 < P > 0.05).
the birds continued to incubate for some period, albeit 28.4%lessintensively (Fig. 4A). Conversely, on cooled days we continuously held Teseveral degreesbelow this "releasetemperature" of 38øC,yet the birds did not continuously incubate but simply increasedattentivenessby 22.0%.
Theseresultsare not surprisingwhen compared with thoseof experimentsin which Teis manipulated in dual-sex, continuous incubators. For example, in one experiment with the Ringed Turtle Dove (Streptopeliarisoria),where the normal range in Teis 38-39øC,Franks(1967) controlled T• at -3
to 8øC for 39 h and in
another experimentat 46øCfor 13 days;in both cases the
doves
continued
to incubate.
As in
our experiment, this indicates that sensing of T• by the brood patchof an incubatingfemale is not the immediate
cause of the incubation
rhythm. Yom-Tov et al. (1978) found in the Dead Sea
and Vleck (1981) show that eggs cooled in the
laboratory causeenergy requirementsfor incubating birds to increase. An increase in mean Te of 4.9øC on heated
Sparrow (Passermoabiticus) that, at Te and T• greaterthan 36øC(mean T• was 33.7øC,so 36øC is probably closeto the "releasetemperature" for this species,although this number is not
July1984]
EffectofEggTemperature onAttentivehess
given), attentivehess did not go to zero, as White and Kinney's model (1974) predicts,but abruptly increased.They attributed this increasein attentivehessby the female to a mechanism to avoid the overheating of eggs under conditions of high Taand suggestedthat this is probablyan adaptivebehavior commonin desert habitats, where Ta can reach 45øC or more
565
controlled by an internal rhythm (cf. Haftorn 1979, 1981). Therefore, attentive patterns can be modified by, but not completelyexplained by, responsesto Te or T•. Furthermore, it appears that adjustmentsto these external stimuli are limited to a defined range by this internal rhythm. For example, in our experiment, when we heated the eggs to maintain T• at a level that theoretically required no incubation for
at midday. When our birdswere presentedwith T, above their "release temperature" of 38øC, development (ca. 40øC), attentivehess did not attentivehess did not increase but diminished go to zero but was reducedby 28.4%.Similarly, by 28.4%. Unlike the Dead Sea Sparrow, how- when we cooled the eggs to maintain a T, that ever, our birds never experienceT• greaterthan theoreticallyrequired continuousincubationfor 37øC(de Violini 1975)during the breedingsea- development(ca. 30øC),attentivenessdid not son and therefore have probably not devel- becomecontinuousbut increasedby 22.0%.This oped a responseto avoid the overheating of suggeststhat the range delimited by the intertheir eggs. nal rhythm controlling attentivehess in our Kendeigh (1952: 87) proposed that the birds was from 28.4% below to 22.0% above the rhythm of incubation is causedby an alterna- mean attentive rhythm of our population. To tion between the drive to incubate and the drive test this hypothesis, we plotted this range to feed. He suggestedthat hunger contractions around the polynomial regressiondescribing in the proventriculus of the Song Sparrow the attentive patternsagainsttime of day for (Melospizamelodia)and HouseWren furnish the our population (Fig. 2A). We believe it is no primary stimulus to terminate the attentive pe- accident that this range closely approximates riod and initiate a searchfor food. It is implied the observedupper and lower attentive limits that once this appetite is satiated, the drive to of the population. Furthermore, we believe it incubate would take precedence,and the bird is no accident that shifts in attentive and inatwould return to the nest to begin another at- tentive periods in responseto manipulated T• tentive bout. If this were true in our experi- are made in such a way that slopesof period ment, one would expect the energy demand versus Taparallel control days (Fig. 6). (cf. Biebach 1979 and Vleck 1981), and there-
fore the hunger drive, to intensify on days CONCLUSIONS when eggs were cooled, leading to a shortening of attentive periods. This was not the case, It appearsthat female Savannah Sparrows as seen in Fig. 6A. Instead, attentive periods adjusttheir attentiverhythm in directresponse lengthenedon cooleddays,evidently in direct to egg temperature, independently of diurnal response to Te. Johnson and Cowan (1974) showed that the
Crested Myna, a tropical hole nester introduced into Vancouver,
British Columbia
from
SoutheastAsia in 1897, failed to adjust its attentive rhythm adequately to the 11øCcooler breeding climate in Vancouver. This resulted in poor nest attentivehess,low incubation temperatures, and a reduction of hatching success from 98% (Myna of West Bengal) to 61%. This reductionin hatching success was overcomeby heating nest boxesso they were not allowed to
go below 28øCor by allowing the closelyrelated Starling (Sturnusvulgaris,daylight attentivenessof 77% as opposedto 47% for the Myna) to incubate the eggs. Experiments such as this indicate to us that the attentive pattern of a speciesis ultimately
fluctuationsin air temperature.Theseadjustments are primarily made by changing the length of both attentive and inattentive periods. It alsoappearsthat this responseto egg temperature
does not cause the attentive
rhythm, becausewhen egg temperatureis held constantat or abovea "releasetemperature" of 38øCthe attentive rhythm continues.It appears that the rhythm of attentivehessis ultimately controlled by an internally regulated oscillation and that sensoryperception of egg temperature only modifies the rhythm of attentivehess. Furthermore,
the modification
of the
attentive rhythm in responseto egg temperature is limited to a certain range, which approximates the upper and lower attentive times observedfor the population under natural conditions.
566
DAVIS ETAL. ACKNOWLEDGMENTS
We thank P. Weldon and M. Moyher for helping with the constructionof the apparatusto manipulate egg temperature, S. Warford and J. Noli for developing computersoftware to digitize our chart recorder data, R. Dot
and the Commander of the U.S.
Naval Air Stationat Point Mugu for providing logistic support, and Pepperdine University for financial support of this study. We are grateful to S.C. Ken-
[Auk,Vol. 101
KENDEIGH, S.C.
deigh,R. Ricklefs,A. Place,F. Rohwer,J.Kingsolver, NIE, N.H., and G. Tallman for providing helpful commentson an earlier version of the manuscript.
1952.
Parental
care and its evolu-
tion in birds. Illinois Biol. Monogr. 22: 1-356. KLUIJVER, H. N. 1950. Daily routines of the Great Tit, Parusm. majorL. Ardea 38: 99-135. LUNDY,H. 1969. A review of the effectsof temperature, humidity, turning, and gaseousenvironment in the incubatoron the hatchability of the hen's egg. Pp. 143-176 in The fertility and hatchability of the hen's egg (T. C. Carter and B. M. Freeman,Eds.).Edinburgh, Oliver and Boyd. C. H. HULL, J. G. JENKINS,K. STEIN-
BRENNER, & D. IZ[.BENT. 1975. Statisticalpackage for the socialsciences.New York, McGrawHill.
LITERATURE CITED
]3IEBACH, H. 1979. Energetikdes Brutensbeim Star (Sturnesvulgaris).J. Ornithol. 120: 121-138.
CAREY,C. 1980. The ecologyof arian incubation. Bioscience 30: 819-824.
DRENT,R. H. 1972. Adaptive aspectsof the physiology of incubation. Pp. 255-280 in Proc. 15th Intern. Ornithol. Congr. (K. H. Voous,Ed.). The Hague, Leiden E. J. Brill. ß 1975. Incubation. Pp. 333-420 in Arian biology,vol. 5 (D. S. Farnerand J. R. King, Eds.). New York, Academic Press.
ßK. POSTUMA,& t. JOUSTRA.1970. The effect
of egg temperature on incubation behaviour in the herring gull. Behavior(suppl.) 17: 237-261. FRANKS, E. C. 1967. The responseof incubating Ringed Turtle Doves (Streptopelia risoria)to manipulatedegg temperatures.Condor 69: 268-276. VONHAARTMAN, L. 1956. Der einflussder temperatur auf den brutrhythmus experimentell nachgewiesen. Ornis Fennica 33: 100-107.
HAFTORN, S. 1979. Incubationandregulationof egg
ROMANOFF,A. L., & A. J. ROMANOFF. 1972. Patho-
genesis of the arian embryo. New York, John Wiley and Sons. SOKAL,R. E., & F. J. ROHLF. 1969. Biometry. San Francisco, W. H. Freeman.
TRACY,C. R., & P. R. SOTHERLAND.1979. Boundary layersof bird eggs:do they ever constitutea sig-
nificant barrier to water loss?Physiol.Zool. 52: 63-66.
TUCKER, V.A. 1971. Flight energeticsof birds.Amer. Zool.
11: 115-124.
DEVIOLINI,R. 1975. Climate summaryfor the Pacific Missile
Test Center.
Tech. Publ. TP-75-25.
VLECK,C. M. 1981. Energetic cost of incubation in the Zebra
Finch.
Condor
83: 229-237.
WALSBERG, G.E. 1983. A test for regulation of nest humidity in two bird species.Physiol. Zool. 56: 231-235.
WEBB,D. R., & J. R. KING. 1983. An analysisof the heat budgetsof the eggsand nest of the Whitecrowned Sparrow, Zonotrichialeucophrys, in relation to parental attentiveness.Physiol. Zool. 56: 493-505.
WEEDEN, J.S. 1966. Diurnal rhythm of attentiveness
temperature in the Willow Tit Parus montanus.
of incubatingfemale Tree Sparrows(Spitellaar-
Ornis
borea)at a northern latitude. Auk 83: 368-388.
Scandinavica
10: 220-234.
1981. Incubation rhythm in the great tit Parusmajor.Cinclus 4: 19-26. 1982. Regulation of body temperatureand
heat transfer to eggs during incubation.Ornis Scandinavica
13: 1-10.
HUGGINS, R.A. 1941. Eggtemperatures of wild birds under natural conditions.Ecology22: 148-157. JOHNSON,S. R., • I. M. COWANß 1974. Thermal ad-
WEINRICH, J.A., & J.R. BAKER.1978. Ad•lie Penguin (Pygoscelis adeliae)embryonic development at different temperatures. Auk 95: 569-576. WHITE, F. N., & J. L. KINNEY. 1974. Arian incubation.
Science
189: 107-115.
YOM-TOv, Y., A. AR, & H. MENDELSSOHN. 1978. In-
cubation behavior of the Dead Sea Sparrow. Condor
80: 340-343.
aptationas a factoraffectingcolonizingsuccess ZERBA,E., & g. L. MORTON. 1983. Dynamics of inof introduced Sturnidae (Aves) in North America. Can. J. Zool. 52: 1559-1576.
cubation in Mountain White-Crowned Sparrows. Condor
85: 1-11.
