clutch completion when laying larger clutches and as ambi-ent temperatures increased. Incubation onset was correlated with patterns of hatching asynchrony at both sites; how-ever, brood reduction increased only in the suburbs, where nestling food is limiting, and only during the late nestling period. Hatching asynchrony may be an unintended conse-quence of beginning incubation early to minimize hatch-ing failure of early laid eggs. Food limitation in the sub-urbs appears to result in increased brood reduction in large clutches that hatch asynchronously. Therefore, site-specific rates of brood reduction may be a consequence of asyn-chronous hatching patterns that result from parental effort to minimize hatching failure in first-laid eggs. This illus-trates how anthropogenic change, such as urbanization, can lead to loss of fitness when animals use behavioral strate-gies intended to maximize fitness in natural landscapes.
Hatching failure has significant fitness consequences for birds as approximately 10 % of the eggs that are laid and survive incubation fail to hatch (Koenig 1982). Several social and ecological factors, such as egg size, clutch size, latitude, diet, nest type, and social organization may influence patterns of hatching failure (Koenig 1982). Hatching failure increases with increasing social organization and is higher at lower lati-tudes. The egg viability hypothesis posits an increased risk of hatching failure when eggs experience prolonged exposure to warm ambient temperatures prior to the onset of incubation, as might be experienced in large clutches and at lower lati-tudes (Stoleson and Beissinger 1999).
Abstract In birds, hatching failure is pervasive and incurs an energetic and reproductive cost to breeding individuals. The egg viability hypothesis posits that exposure to warm temperatures prior to incubation decreases viability of early laid eggs and predicts that females in warm environments minimize hatching failure by beginning incubation earlier in the laying period, laying smaller clutches, or both. How-ever, beginning incubation prior to clutch completion may incur a cost by increasing hatching asynchrony and possi-bly brood reduction. We examined whether Florida scrub jays (Aphelocoma coerulescens) began incubation earlier relative to clutch completion when laying larger clutches or when ambient temperatures increased, and whether variation in incubation onset influenced subsequent pat-terns of hatching asynchrony and brood reduction. We compared these patterns between a suburban and wildland site because site-specific differences in hatching failure match a priori predictions of the egg viability hypothesis. Females at both sites began incubation earlier relative to
Communicated by Indrikis Krams.
R. A. Aldredge · R. K. Boughton · R. Bowman Avian Ecology Program, Archbold Biological Station, Venus, FL, USA
R. A. Aldredge (*) Department of Biology, University of North Carolina-Chapel Hill, CB #3280 G33 Wilson Hall, Chapel Hill, NC 27599, USAe-mail: [email protected]
M. A. Rensel · S. J. Schoech Department of Biological Sciences, University of Memphis, Memphis, TN, USA
M. A. Rensel Department of Integrative Biology and Physiology, University of California, Los Angeles, Los Angeles, CA, USA
78 Oecologia (2014) 174:77–85
1 3
Birds typically lay one egg per day but the onset of incu-bation relative to clutch completion varies among species (Clark and Wilson 1981; Stoleson and Beissinger 1995). Some birds begin incubation upon clutch completion and their eggs typically hatch synchronously, within 24 h. Other birds begin incubation prior to clutch completion and their eggs hatch asynchronously, with a minimum of 24 h between the hatching of the first- and last-laid eggs. Within species, hatching tends to be more synchronous in smaller clutches and in clutches laid earlier in the season, a pattern predicted by the egg viability hypothesis (Hébert 2002).
Asynchronous hatching in birds often is explained as an adaptive strategy to increase the number or quality of young that survive to fledge (see review in Stoleson and Beissinger 1997). Most hypotheses focus on the roles of food availabil-ity and nest predation, especially during the nestling stage, as the selective forces that drive asynchronous hatching patterns (Lack 1954; Clark and Wilson 1981). An alterna-tive explanation is proposed by the egg viability hypothesis, which states that egg viability declines with increased pre-incubation exposure to temperatures high enough to promote embryonic development (24–27 °C; physiological zero) but below the temperature necessary for normal embryonic development (36–38 °C; normal incubation temperatures) (Arnold et al. 1987; Stoleson and Beissinger 1999). Beissinger et al. (2005) showed that 3 days of exposure to warm ambient temperatures appears to be a threshold above which hatching failure increases. Adults may begin incuba-tion prior to clutch completion to reduce exposure of eggs to ambient temperature and minimize hatching failure, but this behavior likely will incur a cost of increasing hatching asyn-chrony and brood reduction (Arnold et al. 1987).
If hatching asynchrony is large, the earliest hatched nestlings may be fed before their siblings hatch, leading to competitive size asymmetries among nestlings (Stoleson and Beissinger 1997), which could lead to brood reduction (the non-random loss of nestlings through starvation) when food is limiting. Initiating incubation early may be an adap-tive behavior to increase hatching asynchrony and thereby increase the rapidity of brood reduction when food is limit-ing, reducing the brood to a size suitable to the available resources (Lack 1968; Forbes et al. 2001; Valkama et al. 2002; Moreno-Rueda et al. 2007). In contrast, the egg via-bility hypothesis suggests that hatching asynchrony might occur as a consequence of minimizing hatching failure in first-laid eggs. Although the egg viability hypothesis and brood reduction hypotheses both predict that hatching asynchrony should increase with clutch size, the egg via-bility hypothesis exclusively predicts that hatching asyn-chrony should be greater in warmer environments.
We examined whether natural variation in incubation behavior in the Florida scrub jay (Aphelocoma coerules-cens) was driven by an effort to minimize hatching failure
in first-laid eggs, and whether differences in incubation pat-terns between a suburban and wildland population affected patterns of hatching asynchrony and brood reduction. Florida scrub jays in suburban habitats lay larger clutches and have higher rates of hatching failure than populations in wildlands (Bowman and Woolfenden 2001). A site com-parison that incorporates lay date reveals that suburban scrub jays lay larger clutches, on average, because breeding begins earlier in the season in the suburbs (Bowman et al. 1998). The median clutch size in the suburbs is four eggs, whereas in the wildlands it is three eggs. Because scrub jays begin incubation near clutch completion (Woolfenden and Fitzpatrick 1996), first-laid eggs in clutches of four or more eggs experience at least 3 days of exposure to ambi-ent temperature, likely leading to higher rates of hatching failure (see Beissinger et al. 2005). Consistent with pre-dictions of the egg viability hypothesis, hatching failure is higher in first-laid eggs in the suburbs than in first-laid eggs in the wildlands and hatching failure is higher in early laid eggs within a clutch than in late-laid eggs at both sites (Aldredge et al. 2012). Even though ambient temperatures exceed physiological zero for more hours of the day as the season advances, hatching failure does not increase in either site. Female scrub jays appear to begin incubation earlier in the laying sequence as the season advances, thus decreasing exposure of early laid eggs to warmer ambient temperatures.
We predicted that females at both sites would begin incubation at or near clutch completion in three-egg clutches and prior to clutch completion in four-egg clutches. We also hypothesized that variation in patterns of incubation onset would influence hatching asynchrony and brood reduction and that because suburban birds lay larger clutches, suburban females would begin incubation earlier and have greater hatching asynchrony. At both sites, we predicted that the clutches of females that initiate incuba-tion prior to clutch completion would hatch more asynchro-nously than those that begin incubation at or after clutch completion and that greater hatching asynchrony would increase the rate of brood reduction. If hatching asynchrony is an adaptive mechanism to match brood size with food availability, then brood reduction should occur early in the nestling period, but if asynchronous hatching patterns occur as a consequence of maintaining egg viability, brood reduction should occur later in the nestling period.
Materials and methods
Study organism and study sites
The Florida scrub jay is a relatively long-lived, perma-nently territorial species that breeds in cooperative family
79Oecologia (2014) 174:77–85
1 3
groups consisting of a breeding pair and from zero to six sexually mature non-breeding auxiliaries (often their off-spring from previous broods) (Woolfenden and Fitzpat-rick 1984; Schoech et al. 1996; Woolfenden and Fitzpat-rick 1996). Florida scrub jays lay clutches of between two and five eggs and begin incubation with the ultimate or penultimate egg (Woolfenden and Fitzpatrick 1996). Both incubation and nestling periods last approximately 18 days, but the incubation period is less variable than the nestling period. Both clutch size and nesting success tend to decline as the breeding season advances (Woolfenden and Fitzpatrick 1984; Bowman et al. 1998; Bowman and Woolfenden 2001). Our research was conducted on two populations of Florida scrub jays in south-central Flor-ida. One population in Placid Lakes Estates (27°15′N, 81°25′W) occurs in fragmented patches of oak scrub within a residential development. This suburban popula-tion has been studied intensively over the past 20 years (Bowman 1998; Bowman and Woolfenden 2001). Ambi-ent temperatures in the suburban site are 1–2 °C warmer than in the wildland site (Aldredge et al. 2012), a pat-tern consistent with the urban heat island effect (Fan and Sailor 2005). The wildland population occurs in the southern half of Archbold Biological Station, immediately adjacent to the population studied by Woolfenden and col-leagues over the past 43+ years (Woolfenden and Fitzpat-rick 1984) and 10–20 km south of the suburban site. This wildland scrub habitat is fire-maintained, occurs in large contiguous blocks, and has been studied intensively over the past 20+ years (Schoech et al. 1991; Mumme 1992; Rensel et al. 2011).
Onset of incubation
We found and monitored 350 nests during the 2005–2007 breeding seasons, from late February until early June. All nests were found prior to egg laying and monitored until the entire nest failed or young successfully fledged. Each nest was visited daily after the onset of egg laying and lay-ing order was identified by marking eggs with an indelible marker. Hatching success was recorded by visiting nests at least once daily from 17 days after clutch completion until all eggs were hatched or 48 h had elapsed since the last hatched egg (see “Hatching asynchrony and brood reduc-tion” section); each nestling was marked in the sequence in which it hatched (see below).
To quantify site-specific differences in daily patterns of ambient temperature we placed thermocouples (HOBO H8 Pro Series; Onset, Bourne, MA) in shaded housings within 2 m of sites known to have been selected by jays for nesting. Ambient temperature was recorded every 15 s and each thermocouple was moved to a new site every 6 days. Two thermocouples were used at each site and both
thermocouple temperatures were averaged to avoid the con-founding effect of variation in ambient temperature because of individual housing placement.
In addition to ambient temperatures, we recorded nest temperatures in 2006 and 2007. On the day the first egg was laid, thermoprobes (HOBO H12-001; Onset) were placed in each focal nest. We gently pushed the probes through the nest lining at the bottom of the nest until the sensor was just exposed at the interface of the eggs and lining. Thermoprobes recorded temperature every minute from laying of the first egg until 3 days after clutch completion. The actual onset of incubation was identified as the first day the temperature of the nest was greater than 31 °C throughout the night. In all nests examined, nighttime incubation continued after this first night, suggesting that scrub jays exhibit a ris-ing incubation temperature pattern and achieve full noc-turnal incubation within a few days of clutch initiation (Wang and Beissinger 2009). We could not characterize diurnal incubation behavior, even when full incubation began after clutch completion, because ambient temper-ature exceeded incubation temperature during the day. Artificial eggs, which could give a more accurate esti-mate of incubation temperature, were not used because Florida scrub jays reject artificial eggs (Fleischer and Woolfenden 2004) and because we sought only to be able to detect the onset of incubation rather than accu-rately measure nest temperatures.
Hatching asynchrony and brood reduction
Beginning 17 days after clutch completion, nests were checked four times a day (every 3–4 h) to determine the order and timing of hatching. Hatching pattern was recorded as the time, in 12-h increments, elapsed between the first- and last-hatched eggs. Nests in which all eggs hatched within one visit (i.e., one hatching interval) were recorded as a zero and the variable increased by one unit every 12 h until hatching was complete. All nestlings were marked within each nest according to hatch order and fol-lowed through the entire nestling period. Hatch order was identified by uniquely marking nestlings’ toenails with red nail polish and recording from which egg each nest-ling hatched, when known. Nests were revisited, nail pol-ish reapplied, and the amount of brood reduction recorded 5 days after the first egg hatched. At 11 days post-hatch, all nestlings were banded, morphological measurements taken, and any further brood reduction recorded. Five and 11 days post-hatch were used to characterize early and late in the nestling period, respectively, because these days occur before and after Florida scrub jays reach their maxi-mum growth rate at approximately 8 days after hatching (Woolfenden 1978).
80 Oecologia (2014) 174:77–85
1 3
As part of the long-term demographic work on both populations, nests were visited daily starting 17 days after the first egg hatched to record how many nestlings suc-cessfully fledged. We did not test the relationship between hatching synchrony and brood reduction when the young fledge (day 18) because of the error in quantifying brood reduction at this time. Nestlings are capable of leav-ing the nest 2–3 days prior to fledging, and searching for individual fledglings could prompt remaining nestlings to fledge prematurely, thereby reducing survivorship. Once some fledging had occurred, it was difficult to determine whether nestlings that disappeared had died prior to fledg-ing or had been depredated following departure from the nest.
Statistical analysis
All analyses were performed using R 2.10.1 (R Develop-ment Core Team 2009). Although clutch size in Florida scrub jays ranges from two to five eggs, both two- and five-egg clutches are uncommon, thus we used only three- and four-egg clutches in our analyses. Because the sample size for measuring the actual onset of incubation based on changes in nest temperature was small (n = 30; 22 subur-ban and eight wildland nests), we used the linear relation-ship between apparent incubation days and actual incu-bation onset from these 30 nests to estimate the onset of incubation for all nests for which we had apparent incu-bation days, but no temperature data from 2005–2007 (n = 182; see “Results” section). The apparent incubation period was defined as the number of days elapsed between the last-laid and the first-hatched eggs. The predicted onset of incubation was highly correlated with subsequent pat-terns of embryonic development within clutches (Aldredge et al. 2012), providing support that our estimates accurately quantified when females began incubation. Generalized estimating equations (GEEs) were used to determine which factors influenced (1) the frequency (proportion of nests experiencing brood reduction), and (2) the amount (num-ber of nestlings lost) of brood reduction that occurred by 5 or 11 days after hatching. Unlike generalized linear mixed effects models, GEEs give unbiased parameter estimates when using a nonlinear link function while controlling for correlation structure (Fieberg et al. 2009). The frequency of brood reduction was investigated using a binomial error distribution and the amount of brood reduction was inves-tigated using a Poisson error distribution. Individual nests were used as a grouping variable in the GEE analysis to account for the non-independence of observations at the same nest at 5 and 11 days post-hatch. For all other anal-yses, general linear models were used when data fit basic statistical assumptions. All values are reported as means (±1 SEM).
Results
Clutch size was larger in the suburbs than in the wildlands (z1,179 = 3.821, P < 0.001) and clutch size decreased simi-larly in both sites as the season progressed (z1,179 = −2.177, P = 0.029).
Onset of incubation
A significant positive relationship existed between actual incubation onset, determined by changes in nest tempera-ture, and number of apparent incubation days (t1,28 = 3.157, P = 0.004, R2 = 0.262; Fig. 1). In nests with an apparent incubation period of 18 days, actual incubation began 0.58 (±0.15) days prior to clutch completion (n = 30); in nests with a shorter apparent incubation period, incubation began earlier in the laying period, but eggs still required approxi-mately 18 days of incubation prior to hatching.
Females in both sites began incubation earlier in the laying period as the season progressed (model effect esti-mate ± SE: −0.004 ± 0.002, t1,177 = −2.348, P = 0.020). After accounting for this seasonal effect, suburban
Act
ual i
ncub
atio
n on
set
(day
s re
lativ
e to
clu
tch
com
plet
ion)
Apparent incubation days
−2.0
−1.5
−1.0
−0.5
0.0
0.5
16 17 18 19
Fig. 1 Female scrub jays began incubation earlier in the laying period as apparent incubation days decreased (n = 30; mean ± SE). Females whose nests had 18-day apparent incubation periods began incubation 0.58 days before clutch completion; scrub jay nests with a shorter apparent incubation period began incubation earlier in the laying period, but eggs still hatched approximately 18 days after incu-bation was initiated
81Oecologia (2014) 174:77–85
1 3
females began incubation earlier than wildland females (t1,177 = −2.853, P = 0.005), and females that laid four-egg clutches began incubation earlier than females that laid three-egg clutches (t1,177 = −3.558, P < 0.001). For nests that were initiated at the beginning of the breeding season (day-of-year = 60), wildland females began incubation 0.42 (±0.15) and 0.64 (±0.06) days before clutch comple-tion when laying three- and four-egg clutches, respectively. Suburban females began incubation 0.60 (±0.06) and 0.82 (±0.07) days before clutch completion when laying three- and four-egg clutches, respectively. By the end of the breeding season (day-of-year = 130), females at both sites began incubation 0.25 days earlier in the laying period for both clutch sizes.
Hatching asynchrony and brood reduction
Hatching patterns were determined for 107 nests during 2006 and 2007; fifty nests in the suburbs and 57 nests in the wildlands. Nests were considered synchronous if ≤24 h elapsed between the first- and last-hatched eggs and asyn-chronous if >24 h elapsed. A strong negative relationship existed between predicted incubation onset and hatching synchrony (t1,101 = −4.469, P < 0.001, R2 = 0.165; Fig. 2).
On average, females whose clutches hatched synchro-nously began incubation 0.39 (±0.11) days before clutch completion, whereas females with asynchronously hatch-ing clutches began incubation 0.80 (±0.13) days before clutch completion. Hatching asynchrony increased in both sites as the season progressed (model effect estimate ± SE: 1.616 ± 0.480, t1,101 = 3.042, P = 0.003). After accounting for this seasonal effect, hatching asynchrony was higher in the wildlands than in the suburbs (t1,101 = −3.209, P = 0.002), and an interaction existed between site and clutch size in explaining patterns of hatching asynchrony (t1,101 = 3.368, P = 0.001). Hatching was asynchronous in three- and four-egg clutches in the wildlands, but in the suburbs, three-egg clutches hatched synchronously while four-egg clutches hatched asynchronously (Fig. 3). For nests that were initiated at the beginning of the breeding season (day-of-year = 60), three- and four-egg clutches in the wildlands took 0.60 (±0.34) and 0.66 (±0.16) days, respectively, to hatch, but three- and four-egg clutches in the suburbs took 0.02 (±0.18) and 0.88 (±0.24) days,
Est
imat
ed in
cuba
tion
onse
t(d
ays
rela
tive
to c
lutc
h co
mpl
etio
n)
Hatching asynchrony (hours)
−0.50
−0.75
−1.00
−1.25
−1.50
0 12 24 36 48 60 72
Fig. 2 Hatching asynchrony increased as females began incubation before clutch completion (n = 182; mean ± SE). Females whose nests hatched synchronously began incubation 0.39 days before clutch completion, and hatching became increasingly asynchronous as females initiated incubation earlier in the laying period
3−egg clutches
4−egg clutches
Hat
chin
g as
ynch
rony
(ho
urs)
0
12
24
Wildlands Suburbs
Fig. 3 An interaction existed between site and clutch size in explain-ing seasonal patterns of hatching asynchrony (n = 107; mean ± SE). Hatching was similarly asynchronous in three- and four-egg clutches in the wildlands. In the suburbs, hatching was synchronous in three-egg clutches but asynchronous in four-egg clutches. Lines are based on parameter estimates from the general linear model examining how site and clutch size influence hatching asynchrony patterns. Site-spe-cific differences in hatching asynchrony at the beginning of the sea-son (day-of-year = 60) are shown. Hatching asynchrony increased as the season progressed but the relationship between site and clutch size did not change
82 Oecologia (2014) 174:77–85
1 3
respectively, to hatch. By the end of the breeding season (day-of-year = 130), three- and four-egg clutches at both sites took approximately 0.78 days longer to hatch than at the beginning of the breeding season.
The frequency of brood reduction was higher in the suburbs than in the wildlands (model effect estimate ± SE: 2.078 ± 0.582, Wald = 12.73, P < 0.001), higher 11 days than 5 days after eggs hatched (model effect esti-mate ± SE: 1.851 ± 0.589, Wald = 9.88, P = 0.002), and higher in asynchronous than in synchronous broods (model effect estimate ± SE: 1.900 ± 0.982, Wald = 3.74, P = 0.053). In addition, an interaction existed between the effects of site and hatching patterns on brood reduc-tion (model effect estimate ± SE: −2.360 ± 1.018, Wald = 5.37, P = 0.020). The odds that a nest experienced brood reduction was 6.37 times higher 11 days than 5 days after eggs hatched. The frequency of brood reduction was similar for synchronously and asynchronously hatched clutches in the wildlands and synchronously hatched clutches in the suburbs, but brood reduction increased
substantially in asynchronously hatched clutches in the suburbs (Fig. 4).
The within-nest intensity of brood reduction was higher in the suburbs than in the wildlands (model effect esti-mate ± SE: 1.177 ± 0.365, Wald = 10.38, P = 0.001; Fig. 5), more young were lost to brood reduction by 11 days than 5 days after eggs hatched (model effect esti-mate ± SE: 0.967 ± 0.321, Wald = 9.11, P = 0.003; Fig. 5), and an interaction existed between site and pres-ence of hatching asynchrony (model effect estimate ± SE: −1.362 ± 0.715, Wald = 3.63, P = 0.057). Again, the amount of brood reduction was similar for synchronously and asynchronously hatched clutches in the wildlands and synchronously hatched clutches in the suburbs, but the number of young lost to brood reduction increased in asyn-chronously hatched clutches in the suburbs.
Discussion
Greater hatching asynchrony in four-egg clutches in both sites could be explained by females beginning incubation
SynchronousAsynchronous
Fre
quen
cy o
f bro
od r
educ
tion
0.0
0.2
0.4
0.6
0.8
1.0
Wildlands Suburbs
Fig. 4 An interaction existed between site and hatching synchrony in explaining patterns of brood reduction. The frequency of brood reduction was similar for synchronously hatched nests in the sub-urbs and wildlands. Brood reduction was more likely in asynchro-nously hatched nests in the suburbs than in asynchronously hatched nests in the wildlands. Lines are based on parameter estimates from the generalized estimating equation examining how site and hatching synchrony influence the frequency of brood reduction. Site-specific differences in brood reduction 11 days after eggs hatched are shown. The frequency of brood reduction was lower for all site-specific hatching synchrony combinations 5 days after eggs hatched
Wildlands
Suburbs
Num
ber
of n
estli
ngs
lost
to b
rood
red
uctio
n
0.0
0.2
0.4
0.6
0.8
1.0
Day 5 Day 11
Fig. 5 More nestlings were lost to brood reduction by 11 days than 5 days after the first egg hatched, and brood reduction was greater in the suburbs than in the wildlands. No interaction existed between site and time after hatching. Lines are based on parameter estimates from the generalized estimating equation examining how site and days after hatching influence the amount of brood reduction. The amount of brood reduction in synchronously hatching nests in both sites is shown. More nestlings were lost to brood reduction in nests with asynchronous hatching
83Oecologia (2014) 174:77–85
1 3
earlier in the laying period to minimize hatching failure (as proposed by the egg viability hypothesis) or could occur to facilitate brood reduction when resource availability is low (Shawkey et al. 2004; as predicted by the brood reduction hypothesis). However, hatching became more asynchro-nous as the season advanced even though arthropod abun-dance typically increases in the wildlands (R. Bowman, unpublished data). Additionally, brood reduction in the suburbs increased during the nestling period. These pat-terns suggest that hatching asynchrony is not an adaptive mechanism to facilitate the rapidity of brood reduction but is an unintended consequence of maintaining the viability of first-laid eggs. Brood reduction continues to increase in suburban Florida scrub jay broods from 11 to 18 days after eggs hatch (R. Bowman, unpublished data), providing fur-ther support that brood reduction occurs late in the nestling period in the suburbs.
As predicted, the earlier a female began incubation rela-tive to clutch completion, the greater the degree of hatch-ing asynchrony. However, fewer than one-quarter of scrub jay females began incubation more than 1 day prior to clutch completion and most (>90 %) females began incuba-tion less than 1.5 days prior to clutch completion. Because hatching asynchrony can lead to increased brood reduction, females might be constrained in how early they can begin incubation and still maximize the number of young fledged (Stoleson and Beissinger 1997). Because suburban scrub jays have larger clutch sizes and are exposed to a warmer ambient temperature, they cannot begin incubation early enough to minimize hatching failure without incurring costs in increased hatching asynchrony and brood reduc-tion that the wildland population might not experience. In addition, the suburban environment also tends to be food limited for nestlings (Shawkey et al. 2004). Thus, suburban birds end up suffering even higher rates of hatching fail-ure (Aldredge et al. 2012) and brood reduction. Other spe-cies that move into urban areas when their preferred habitat decreases may face similar challenges raising nestlings in an environment where temperatures are warmer and arthro-pods are scarce. Future research also could examine demo-graphic, behavioral, or life history consequences of global climate change based on inferences from this urban heat island effect.
The overall pattern of greater hatching asynchrony in the wildlands than in the suburbs is the opposite of the pattern predicted by earlier incubation onset seen in the suburbs. Independent of the timing of incubation onset, variation in nest attentiveness during early brooding could explain the discrepancy between patterns of incubation onset and hatching asynchrony. Once some eggs have hatched, females may leave the nest more often or for longer peri-ods to acquire food for the newly hatched young and to replace endogenous reserves that were lost during the
incubation period (McMaster and Sealy 1999). Suburban females take fewer, shorter off-bouts during the incubation period because of predictable access to high-quality anthro-pogenic food (Aldredge et al. 2012), potentially reducing their hatching asynchrony, especially in three-egg clutches for which suburban females begin incubation near clutch completion. However, this hypothesis fails to explain why hatching asynchrony is so high in four-egg clutches in the suburbs. Additionally, the almost 14 h difference in hatch-ing time between three-egg clutches in the suburbs and wildlands seems too large to be explained only by differ-ences in early brooding behavior.
Our study draws attention to the importance of exam-ining alternative hypotheses. If our intent had been spe-cifically to test the brood reduction hypothesis, we might have concluded that asynchronous hatching patterns were an adaptive behavior to facilitate brood reduction in food-limited environments. However, our findings are consist-ent with the alternative hypothesis (i.e., the egg viability hypothesis) that hatching asynchrony occurs as a conse-quence of minimizing hatching failure in first-laid eggs in warm environments because (1) females appear to change incubation behavior to minimize exposure of eggs to warm ambient temperatures, (2) the timing of the onset of incu-bation influences hatching asynchrony patterns, and (3) hatching asynchrony increases brood reduction late in the nestling period. Several other alternative hypotheses exist (Stoleson and Beissinger 1995) and should be examined simultaneously (Stoleson and Beissinger 1997) to deter-mine whether egg viability can explain the evolution of hatching asynchrony in the Florida scrub jay, and in birds in general.
Birds face a tradeoff in laying and incubation behav-iors to minimize hatching failure without incurring costs of hatching asynchrony. The appropriate balance of these behaviors may depend on factors that influence hatching failure or brood reduction. Both seasonal and latitudinal gradients in avian clutch size are consistent with predic-tions of the egg viability hypothesis; in many passerines, clutch size decreases with advancing season and decreas-ing latitude (Lack 1968; Klomp 1970; Cooper et al. 2005). A smaller clutch size reduces exposure of early laid eggs to warm ambient temperatures. Florida scrub jays exhibit a seasonal reduction in clutch size, but females also initi-ate incubation earlier in the laying period. Therefore, our results suggest that a priori predictions of the egg viability hypothesis may explain seasonal trends in avian clutch size as females balance the tradeoff between minimizing hatch-ing failure in first-laid eggs and minimizing hatching asyn-chrony and brood reduction. Likewise, the egg viability hypothesis also may explain why avian clutch size declines from temperate to tropical regions (Stoleson and Beissinger 1999; Cook et al. 2003, 2005; Wang et al. 2011). Therefore,
84 Oecologia (2014) 174:77–85
1 3
future research should examine whether predictions of the egg viability hypothesis can explain the seasonal and latitudinal trends in avian clutch size. Through controlled experiments the relative importance of factors that regulate incubation onset should be tested by separating potentially correlated environmental effects, such as ambient tempera-ture, time of year, and resource availability.
Acknowledgments We thank Laura Kearns, Sonya LeClair, April Feswick, and Angela Tringali for field work assistance in the subur-ban site and the staff at Archbold Biological Station for logistical sup-port during the study. We thank the students, staff and faculty at the University of Memphis for collecting data in the wildland site, espe-cially Travis Wilcoxen and Gina Morgan. We thank one anonymous reviewer and S. R. Beissinger for helpful comments on the manu-script. Financial support was provided by the American Ornitholo-gists’ Union and the University of Central Florida. The long-term monitoring of suburban and wildland Florida scrub jays that allowed this work was supported by National Science Foundation awards to R. B. (IBN-03406328) and S. J. S. (IOS-0346328).
References
Aldredge RA, LeClair SC, Bowman R (2012) Site-specific dif-ferences in hatching failure between suburban and wildland Florida scrub-jays. J Avian Biol 43:369–375. doi:10.1111/ j.1600-048X.2012.05665.x
Arnold TW, Rohwer FC, Armstrong T (1987) Egg viability, nest predation, and the adaptive significance of clutch size in prairie ducks. Am Nat 130:643–653. doi:10.1086/284736
Beissinger SR, Cook MI, Arendt WJ (2005) The shelf life of bird eggs: testing egg viability using a tropical climate gradient. Ecol-ogy 86:2164–2175. doi:10.1890/04-1624
Bowman R (1998) Population dynamics, demography, and contribu-tions to metapopulation dynamics by suburban populations of the Florida scrub-jay. Non-game wildlife technical report no. 11. Florida Game and Fresh Water Fish Commission, Tallahassee, FL
Bowman R, Woolfenden GE (2001) Nest success and the timing of nest failure of Florida scrub-jays in suburban and wildland habitats. In: Marzluff JM, Bowman R, Donnelly R (eds) Avian ecology and conservation in an urbanizing world. Kluwer, Norwell, pp 383–402
Bowman R, Woolfenden GE, Fitzpatrick JW (1998) Timing of breed-ing and clutch size in the Florida scrub-jay (Aphelocoma coerule-scens). Ostrich 69:316–317
Clark AB, Wilson DS (1981) Avian breeding adaptations—hatch-ing asynchrony, brood reduction, and nest failure. Q Rev Biol 56:253–277. doi:10.1086/412316
Cook MI, Beissinger SR, Toranzos GA, Rodriguez RA, Arendt WJ (2003) Trans-shell infection by pathogenic micro-organisms reduces the shelf life of non-incubated bird’s eggs: a constraint on the onset of incubation? Proc R Soc Lond B 270:2233–2240. doi:10.1098/rspb.2003.2508
Cooper CB, Hochachka WM, Butcher G, Dhondt AA (2005) Seasonal and latitudinal trends in clutch size: thermal con-straints during laying and incubation. Ecology 86:2018–2031. doi:10.1890/03-8028
Fan HL, Sailor DJ (2005) Modeling the impacts of anthropogenic heating on the urban climate of Philadelphia: a comparison of implementation in two PBL schemes. Atmos Environ 39:73–84. doi:10.1016/j.atmosenv.2004.09.031
Fieberg J, Rieger RH, Zicus MC, Schildcrout JS (2009) Regression modeling of correlated data in ecology: subject-specific and
population averaged response patterns. J Appl Ecol 46:1018–1025. doi:10.1111/j.1365-2664.2009.01692.x
Fleischer TL, Woolfenden GE (2004) Florida scrub-jays eject foreign eggs added to their nests. J Field Ornithol 75:49–50
Forbes S, Glassey B, Thornton S, Earle L (2001) The secondary adjustment of clutch size in red-winged blackbirds (Agelaius phoeniceus). Behav Ecol Sociobiol 50:37–44. doi:10.1007/s002650100332
Hébert P (2002) Ecological factors affecting initiation of incubation behaviour. In: Deeming DC (ed) Avian incubation: behaviour, environment, and evolution. Oxford University Press, New York, pp 271–279
Klomp H (1970) The determination of clutch size in birds. A review. Ardea 58:1–124
Koenig WD (1982) Ecological and social factors affecting hatchabil-ity of eggs. Auk 99:526–536
Lack D (1954) The natural regulation of animal numbers. Oxford University Press, New York
Lack D (1968) Ecological adaptations for breeding in birds. Methuen, London
McMaster DG, Sealy SG (1999) Do brown-headed cowbird hatch-lings alter adult yellow warbler behavior during the hatching period? J Field Ornithol 70:365–373
Moreno-Rueda G, Soler M, Soler JJ, Martinez JG, Perez-Contreras T (2007) Rules of food allocation between nestlings of the black-billed magpie Pica pica, a species showing brood reduction. Ardeola 54:15–25
Rensel MA, Wilcoxen TE, Schoech SJ (2011) Corticosterone, brood size, and hatch order in free-living Florida scrub-jay (Aphelocoma coerulescens) nestlings. Gen Comp Endocrinol 171:197–202. doi:10.1016/j.ygcen.2011.01.011
Schoech SJ, Mumme RL, Moore MC (1991) Reproductive endocri-nology and mechanisms of breeding inhibition in cooperatively breeding Florida scrub jays (Aphelocoma c. coerulescens). Con-dor 93:354–364. doi:10.2307/1368951
Schoech SJ, Mumme RL, Wingfield JC (1996) Delayed breeding in the cooperatively breeding Florida scrub-jay (Aphelocoma coer-ulescens): inhibition or the absence of stimulation? Behav Ecol Sociobiol 39:77–90. doi:10.1007/s002650050269
Shawkey MD, Bowman R, Woolfenden GE (2004) Why is brood reduction in Florida scrub-jays higher in suburban than in wild-land habitats? Can J Zool 82:1427–1435. doi:10.1139/Z04-123
Stoleson SH, Beissinger SR (1995) Hatching asynchrony and the onset of incubation in birds revisited: when is the critical period? Curr Ornithol 12:191–271
Stoleson SH, Beissinger SR (1997) Hatching asynchrony, brood reduction, and food limitation in a Neotropical parrot. Ecol Mon-ogr 67:131–154. doi:10.1890/0012-9615(1997)067[0131:HABRAF]2.0.CO;2
Stoleson SH, Beissinger SR (1999) Egg viability as a constraint on hatching asynchrony at high ambient temperatures. J Anim Ecol 68:951–962. doi:10.1046/j.1365-2656.1999.00342.x
Valkama J, Korpimäki E, Holm A, Hakkarainen H (2002) Hatch-ing asynchrony and brood reduction in Tengmalm’s owl Aegolius funereus: the role of temporal and spatial variation in food abundance. Oecologia 133:334–341. doi:10.1007/s00442-002-1033-2
Wang JM, Beissinger SR (2009) Variation in the onset of incubation and its influence on avian hatching success and asynchrony. Anim Behav 78:601–613. doi:10.1016/j.anbehav.2009.05.022
Wang JM, Firestone MK, Beissinger SR (2011) Microbial and environmental effects on avian egg viability: do tropical mechanisms act in a temperate environment? Ecology 92: 1137–1145
Woolfenden GE (1978) Growth and survival of young Florida scrub jays. Wilson Bull 90:1–18
Woolfenden GE, Fitzpatrick JW (1984) The Florida scrub jay: demography of a cooperative-breeding bird. Princeton University Press, New Jersey
Woolfenden GE, Fitzpatrick JW (1996) Florida scrub-jay (Aphelo-coma coerulescens). In: Poole A, Gill F (eds) The birds of North America, no. 228. The Academy of Natural Sciences, Philadel-phia, The American Ornithologists’ Union, Washington, DC, pp 1–28
