96

Constant and Cycling Incubation Temperatures Have Long-Term Effects

on the Morphology and Metabolic Rate of Japanese Quail*

Noah Ben-Ezra certain range to ensure proper embryonic development (Du-

Gary Burness† Rant et al. 2013; McClintock et al. 2014). Deviations from
Department of Biology, Trent University, Peterborough, optimal temperatures have been shown experimentally to have
Ontario K9L 0G2, Canada wide-ranging effects on subsequent morphological and phys-
iological traits (Hepp et al. 2015). For example, in both pre-
Accepted 7/4/2016; Electronically Published 9/8/2016 cocial and altricial bird species, suboptimal incubation tem-
peratures result in decreased offspring mass and/or size (wood
Dryad data: http://dx.doi.org/10.5061/dyad.76qt1. ducks, Aix sponsa [Hepp and Kennamer 2012; DuRant et al.
2013]; tree swallows, Tachycineta bicolor [Ardia et al. 2010]; blue
tits, Cyanistes caeruleus [Nord and Nilsson 2011]). Suboptimal in-
cubation temperatures can also affect growth rates (DuRant et al.
ABSTRACT
2010, 2013) and locomotor performance (Hopkins et al. 2011).
Incubation temperature can have profound effects on growth Although suboptimal incubation temperatures may have negative
and development of embryos and young birds. However, few implications for postfledging survival (Berntsen and Bech 2015)
studies have examined the role that cycling incubation temper- and fitness (Hepp and Kennamer 2012; Hepp et al. 2015), effects
ature may play in phenotypic variation and whether these effects of incubation temperature have, for the most part, rarely been
persist to adulthood. We incubated Japanese quail eggs at control examined beyond the juvenile stage (but see Wada et al. 2015)
temperatures (37.57C), at low temperatures (36.07C), and under yet may be an important source of adult phenotypic variation.
a cyclical treatment that maintained the same average tempera- Suboptimal incubation temperatures can also have negative
ture as the low treatment (36.07C) with high temperatures that metabolic effects on offspring. When Zebra finch (Taeniopygia
were the same as the control (37.57C) and low temperatures that guttata) eggs were periodically cooled, 12-d-old embryos dis-
still allowed for development of the embryo (28.07C). Individuals played reduced mass and an increased metabolic rate, suggest-
in the low treatment group were smaller in mass and size than ing less efficient development (Olson et al. 2006). Additionally,
individuals in the control group but had an increased basal met- low incubation temperatures have been shown to increase rest-
abolic rate relative to individuals in the cyclical treatment group. ing metabolic rate in 14–15-d-old blue tit fledglings (Nord and
Temperature cycling offset the effects of low incubation tem- Nilsson 2011) and transiently in 25-d-old female zebra finch
peratures on metabolic rate and embryonic development but not juveniles (Wada et al. 2015). This negative effect on metabolic
the effects on adult mass and size. Although Japanese quail are rate may be more pronounced—and long-lasting—in precocial
sexually size dimorphic, with females larger than males, we could species that undergo a greater amount of development in the
detect no evidence of sex-specific sensitivity to suboptimal incu- egg.
bation temperatures. These results highlight the importance of Studies investigating the influence of incubation temperature
incubation temperature and pattern as sources of morphological on posthatching development have largely focused on effects of
and physiological variation of adult birds. constant temperatures. However, in many natural systems cy-
clical incubation temperatures are common, as the brooding
Keywords: growth, sexual size dimorphism, epigenetics, de- parent leaves the nest periodically to either feed or avoid dep-
velopment, avian, metabolism. redation (Weathers and Sullivan 1989). Since the cessation of
incubation will cause a drop in egg temperature, cyclical tem-
perature change may have a similar effect on developing em-
bryos as constant low incubation temperatures (Olson et al.
Introduction
2006). For example, when Japanese quail (Coturnix japonica)
Egg incubation temperature is highly regulated in oviporous eggs were exposed to 8-h interruptions to incubation, the length
species (e.g., birds, reptiles) and must be maintained within a of incubation increased by ∼30% (Callebaut 1990). In wood ducks,
extending interruptions to incubation has also been shown to
*This paper was submitted in response to a call for papers for a Focused Issue increase the length of incubation (Carter et al. 2014). Whether
on “Early-Life Effects on the Adult Phenotype: A Comparative Perspective.” any of these presumed negative effects on embryos persist into
†Corresponding author; e-mail: garyburness@trentu.ca.
adulthood is not well known, and to date no study of which we
Physiological and Biochemical Zoology 90(1):96–105. 2017. q 2016 by The
are aware has examined whether cyclical incubation contributes
University of Chicago. All rights reserved. 1522-2152/2017/9001-5153$15.00. to phenotype variation independently of reduced temperature.
DOI: 10.1086/688383 Previous studies of cyclical incubation (e.g., Olson et al. 2006,

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Effects of Incubation Temperature on Adult Phenotypes 97

2008) compared treatments that were both cycling and at a re- low (36.07C, 70 eggs), and cyclical (mean p 36.07C, low p 28.07C,
duced average temperature with constant high-temperature treat- high p 37.57C [see below], 56 eggs). Sample sizes were unequal
ments. As such, the potential effects of cyclical incubation have to account for predicted differences in hatching success in the
not been examined in isolation. For example, it is possible that second trial (see below). We chose 37.57C for the control because
an incubation pattern that more closely mimics natural fluctua- this is the most commonly used temperature in commercial/
tions may provide developmental benefits to embryos, as reported recreational quail farming, while 36.07C is at the lower end of the
previously in some reptile species (reviewed in Bowden et al. range that is used (Callebaut 1990). The cyclical treatment was
2014). Whether these potential benefits exist in birds, however, designed to maintain an average temperature of 36.07C, but,
is not known. It is therefore important to consider temperature unlike the low-temperature treatment, the temperature cycled
cycling when exploring the effects of incubation temperature on over 3-h periods. For 127 min of each 3-h period the incubator
an individual’s phenotype. was maintained at 37.57C. The incubator was then shut off by a
In some species, males and females differ predictably in body digital timer, and the temperature decreased toward ambient for
size (sexual size dimorphism), and during development the two the remaining 53 min, eventually reaching 28.07C before the timer
sexes may respond differently to environmental stressors. The di- turned the incubator back on and the temperature rose back to
rectionality of this effect (if any) may result from a combination 37.57C. It should be noted that we measured air temperature in
of factors, including nutritional requirements, hormonal differ- the incubators, not internal egg temperature. While this is com-
ences, genotype, and clutch size (Jones et al. 2009). In most avian monly done (e.g., Nord and Nilsson 2011; Wada et al. 2015), eggs
species, males are larger than females, so it is often difficult to in the cyclical treatment may therefore have experienced slightly
disentangle the effects of being male from that of the necessity of different temperatures than were recorded due to thermal iner-
attaining a larger adult size (e.g., Chin et al. 2013). In zebra finches, tia. The cooling/warming sequence was established via prelimi-
however, males and females are similarly sized, yet when incu- nary trials to ensure that the incubation temperature was on
bated at suboptimal temperatures males displayed reduced growth average the same as that of the low-temperature treatment while
relative to females (Wada et al. 2015). This suggests that males maintaining the same maximum temperature as the control (i.e.,
may be more sensitive to incubation temperature than females, 37.57C). Preliminary temperature trials to establish appropriate
independent of body size. Japanese quail display reverse sexual cycling periods were conducted using empty incubators. Be-
size dimorphism, with females larger than males. This species thus cause of later concern that the presence of eggs may have influ-
presents an opportunity to further examine sex-specific effects of enced the cooling/warming sequence, we retested the cycling tem-
incubation temperature by uncoupling the effects of being male peratures with and without eggs. We estimate that the presence
from the necessity of reaching a larger adult size. of eggs increased the average temperature of the cycling incu-
In the current study, we tested the effect that constant in- bator by up to 0.37C. As a result, the average temperature of our
cubation temperature and temperature cycling has on the event- cycling treatment may not have perfectly mimicked the low in-
ual size and metabolic rate of adult Japanese quail. We hypoth- cubation temperature. Although we argue that the effect was
esized that the temperature experienced by an individual as an likely minimal (see “Discussion”), our results are interpreted in
embryo would have short- and long-term effects on its pheno- light of this.
type. Specifically, we predicted that (1) individuals incubated at The humidity of incubators was maintained at 60% 5 5%,
low temperatures would be lighter and smaller at hatch and at and all temperatures were measured 50.17C using a digital ther-
adulthood, (2) low incubation temperatures would result in adults mometer (BAT-12; Physiotemp). Because egg turning occurred
with increased mass-adjusted basal metabolic rates (BMRs) rela- via an automatic turning plate once every 3 h, each treatment
tive to the controls, (3) cyclical incubation would offset some of had similar egg-turning bouts. Egg turning was stopped on in-
the negative effects of low incubation temperature, and (4) there cubation day 13 (onset of incubation p incubation day 0). Start-
would be sex-specific effects of incubation temperature, although ing on incubation day 14, we increased the humidity to 75% 5
we could not predict directionality a priori. 5% and started checking eggs for movement and pipping (break-
ing through the shell). In the cyclical treatment, any potential
change in humidity during off-bouts was within the range of error
Material and Methods of the humidity control.
Once pipping began, eggs were checked at least every 7 h
Husbandry and Treatments
around the clock, and hatch time was recorded. Because of the
Fertilized Japanese quail eggs were transported from Cro Quail degree of wetness of the chicks and our previous experience with
Farms in St. Anns, Ontario, Canada, to Trent University by car hatchling Japanese quail, it was possible to estimate age (in hours
at room temperature (∼3 h). Eggs were then weighed, numbered posthatch) to within 2 h. When discovered, hatchlings were
using a nontoxic marker, and placed randomly into each of the marked on the leg(s) with a nontoxic colored marker for iden-
three digital incubators (R-com 20; MX-20). Although artificial tification and moved to a different incubator to dry until 10 h after
convection air incubation may have different implications for estimated hatch time (377 5 17C and 60% 5 10% humidity;
embryo development than parental contact incubation, it is com- Hovabator model 1583). At 10 h of age (based on the estimated
monly used in incubation temperature studies. The incubators hatch time, above), individuals were weighed and marked with
were maintained at different temperatures: control (37.57C, 48 eggs), acrylic paint on the head to indicate hatch day.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
98 N. Ben-Ezra and G. Burness

Posthatch husbandry followed Chin et al. (2013). Briefly, at bers, reweighed, and returned to their cages. Mass during res-
10 h of age hatchlings were moved to a metal brooder (357 5 pirometry was calculated as the average of the two masses. Due
27C, 40% 5 10% humidity). When chicks were 5 d old (hatch to a scale malfunction, three birds from the second trial were
day p day 0), they were removed from the brooder and moved missing prerespirometry masses; these masses were estimated
in groups of five or six to wire cages (45 cm # 45 cm # 95 cm) using a linear regression (r 2 p 0.97, F1, 72 p 2,656.14, P ! 0.001,
lined with paper towels and randomized between two envi- n p 74) based on the reduction in mass of all other birds dur-
ronmental chambers. The environmental chambers were main- ing respirometry. On average, individuals lost 2.5% of their ini-
tained at 307 5 17C for 2 wk, at which point the temperature was tial mass during a respirometry trial.
decreased by 3.57C every 6–7 d until reaching a final temperature The respirometry protocol followed Chin et al. (2013). Briefly,
of 247C. On day 10, they were separated into groups of three or ambient air was scrubbed of CO2 and water using three columns
four, and on day 25 they were housed individually. Males and of Drierite, one column of soda lime, and one column of Ascarite
females, sexed visually by plumage on day 25, were kept at sep- II, connected in series with Bev-A-Line tubing (Cole-Palmer).
arate ends of the environmental chamber to minimize aggres- The column of Ascarite II also contained layers of Drierite at the
sion between males in adjacent cages. As individuals were sexed top and bottom of the column to absorb any additional water
visually, sex ratios were not obtained at hatch. Individuals were produced during the absorption of CO2. Air then flowed through
fed MasterFeed Turkey Starter (24% protein) until day 24, and the multiplexor (model TR-RM; Sable Systems) into the respi-
Mazuri Adult Breeder Diet (21% protein) was blended in after rometry chambers or a piece of Bev-A-Line tubing when mea-
day 24 until given exclusively on day 37 5 2 d. A light cycle suring baselines. Flow rate was determined by mass flow con-
of 14L∶10D was used throughout the experiment. Water and trollers (model 840; Sierra Instruments) and was set to 1,800 mL/
food was supplied ad lib. from when chicks were moved to the min. On exiting the chamber, air was scrubbed of water vapor
brooder until the end of the study. To ensure adequate sample using magnesium perchlorate. A subsampler pump (TR-SS3; Sa-
sizes, the experiment was replicated, separated by a 1-mo pe- ble Systems) drew in air at 150 mL/min into the carbon dioxide
riod. A total of 97 chicks hatched across the entire study (con- analyzer (CA-10; Sable Systems) followed by the oxygen analyzer
trol p 29, low p 28, cyclical p 40), and a total of 88 birds (FC-10a; Sable Systems). Before each trial, we calibrated the CO2
survived to adulthood (control p 26, low p 25, cyclical p 37). analyzer by spanning it to 0% and 1.00% using dry ambient air
Individuals were considered to be adults by day 55 (Chin et al. and compressed gas, respectively. We calibrated the O2 analyzer
2013). All animal use was approved by the Animal Care Com- to 20.95% using ambient air. ExpeData software (ver. 1.7.30) was


mittee at Trent University, in accordance with the Canadian used to record and analyze the data. To calculate VO2 , the lowest
Council on Animal Care (protocol 13027-23094). continuous 5-min period of oxygen consumption was selected
for each individual, and the average carbon dioxide emission and


oxygen consumption from this period were recorded. VO2 was
Morphological Measurements
calculated using equation (10.6) in Lighton (2008) and multiplied


Starting at hatch (day 0) and then on days 3, 5, 10, 20, 30, 40, by 20.08 to convert VO2 (mL O2) to metabolic rate (J; Schmidt-
and 55 posthatch, we measured body mass using a digital pan Nielson 1990). Raw data are available in the Dryad Digital Re-
balance (50.01 g), and as an index of structural size we mea- pository (http://dx.doi.org/10.5061/dryad.76qt1; Ben-Ezra and
sured the distance from the back of the head to the front of the Burness 2016).
bill using digital calipers (50.01 mm). This metric is com-
monly used as an index of body size in birds (e.g., Ardia et al.
Statistics
2010). Head-bill measurements were taken in triplicate and
averaged. All statistical analyses were performed in JMP 11.0 (SAS), and
statistical significance was claimed at P ! 0.05. No data were
transformed.
Adult BMR
To analyze initial egg mass, we used a general linear model
As an index of adult BMR (day 601), we measured the noc- (GLM) with trial number, treatment, and the treatment # trial


turnal oxygen consumption rate (VO2 ) of 77 (control p 25, number interaction. Eggs in the first trial were ∼5% lighter than
low p 29, cyclical p 23) of 88 individuals surviving to adulthood eggs in the second trial (least squares mean 5 1 SEM: trial 1,
using flowthrough respirometry. Food was removed at 0730– 13.24 5 0.12 g; trial 2, 13.91 5 0.12 g; trial number: F1, 168 p
0830 hours on the day of measurement. At 1600–1800 hours, two 14.69, P p 0.002) but did not differ between treatments (treat-
or three quail were removed from their cages, transported to the ment: F2, 168 p 0.98, P p 0.379). By chance, because eggs were
respirometry room (∼30 m) in a breathable cloth bag, weighed distributed randomly, there was a treatment # trial number in-
(5 0.01 g), and placed in one of three 4-L paint-can respirometry teraction (F2, 168 p 3.41, P p 0.035), with eggs allocated to the low
chambers, painted flat gray inside. The chambers were placed in temperature incubator of trial 2 being heavier than eggs in all
an incubator (Thermo Low Temperature Incubator model 915; treatments of trial 1 (Tukey HSD, P ! 0.05) but not differing from
Fisher Scientific) set at 247C (within the Japanese quail ther- those of the other treatments in trial 2 (Tukey HSD, NS). When
moneutral zone; Ben-Hamo et al. 2010). At 0800–0900 hours the split by trial number, we confirmed that egg mass did not differ
following morning, individuals were removed from the cham- among treatments in trial 1 (treatment: F2, 83 p 0.33, P p 0.717).

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Effects of Incubation Temperature on Adult Phenotypes 99

In trial 2, egg mass differed among treatments (F2, 85 p 4.49, P p

0.014), with eggs in the low treatment group being heavier than
those in the control treatment group (least squares mean 5 1 SEM:
control, 13.57 5 0.22 g; low, 14.38 5 0.18 g; Tukey HSD, P !
0.05); there were not other significant differences (Tukey HSD,
NS). Trial number and the interaction between trial number and
treatment were included in all subsequent analyses to examine
the influence of egg distribution.
To analyze the mass and size of hatchlings and adults, we used
GLMs. We started with a full model, including sex, treatment
(incubation temperature), trial number, the interaction between
sex and treatment, and the interaction between trial number
and treatment. Sex, treatment, and the sex # treatment inter-
action were retained in all tests regardless of significance due to
a priori interest. All other main effects and interactions were
Figure 1. Hatching success and length of incubation (least squares
excluded using backward stepwise elimination if P values were
mean 5 1 SEM) for Japanese quail eggs in control (37.57C), cyclical
10.10, starting with the least significant interactions followed (mean: 36.07C; range: 28.07–37.57C), and low (36.07C) temperature treat-
by main effects. To include the main effect of sex, individuals ment groups. Sample sizes for length of incubation are in parentheses;
that hatched but died before day 25 (when sexing became pos- error bars for the cyclical treatment are shown but are hidden behind
sible; total, n p 9; each treatment, n p 3) were excluded from the symbol. Number of eggs incubated: control, n p 48; cyclical, n p 56;
low, n p 70. Differing letters denote P ! 0.05 for a Tukey HSD.
all posthatch analyses.
To analyze hatching success and sex ratio, we used a GLM
with a binomial distribution and logit function using treatment of treatment remained the same, as did all subsequent com-
as the main effect. Trial number was initially included in these parisons described above (P ! 0.01 for all), so trials were kept
models but was removed as it was nonsignificant (P 1 0.1). To pooled. Variance in hatch day also differed among treatment
test for an effect of treatment and trial number on the length of groups (Levene, F2, 94 p 11.06, P ! 0.001). The cyclical treat-
incubation, we used ANOVA, with a Welch’s ANOVA for anal- ment showed significantly less variance than the control (Levene,
ysis of treatment due to lack of homogeneity of variance. Dunn- F1, 67 p 20.48, Dunn-Šidák-adjusted P p 0.003) and the low-
Šidák-adjusted P values were used for post hoc analyses of length temperature treatment (Levene, F1, 66 p 20.32, Dunn-Šidák-adjusted
of incubation and hatching success. Because it was not always P p 0.002). The control and low-temperature treatments did not
possible to match a hatchling to the egg from which it came, we differ from each other (Levene, F1, 55 p 0.04, P p 0.851).
could not compare these two metrics.


For the analysis of BMR (as indexed by VO2 ), we used a GLM
Hatching Success Was Affected by Temperature and Cycling
with body mass at respirometry included as a covariate. Sex,
treatment, and sex # treatment were again retained regardless Incubation temperature significantly affected hatching success
of statistical significance. Interaction terms with respirometry (GLM, x 2 p 13.24, P p 0.001, n p 174; fig. 1). Hatching suc-
mass (respirometry mass # sex, respirometry mass # treatment) cess was significantly higher in the cyclical treatment group
as well as trial number, trial number # treatment, bird age at (71%) than in the low treatment group (40%; Dunn-Šidák-
respirometry, and chamber number were initially included. Bird adjusted P p 0.001). The control treatment (60%) did not dif-
age (range: 60–94 d posthatch) and chamber number did not fer from the low-temperature treatment (Dunn-Šidák-adjusted
affect BMR (P 1 0.1 for both) and were removed from the final Pp0.057)or the cyclical treatment (Dunn-Šidák-adjusted P p
model. 0.236). Mortality rates were similar across treatments (table 1).

Incubation Temperature and Cycling Had Short-Term

Results
Effects on Hatchling Size but Not Mass
Length of Incubation Was Affected
Incubation temperature did not affect hatchling body mass
by Temperature and Cycling
(least squares mean 5 1 SEM: control, 9.35 5 0.21 g; cyclical,
Incubation temperature significantly affected how long it took 9.38 5 0.18 g; low, 9.57 5 0.22 g; treatment: F2, 79 p 0.32, P p
for individuals to hatch (Welch’s ANOVA, F2, 42 p 109.70, P ! 0.731), but hatchlings in the second trial were heavier than
0.001; fig. 1). Individuals in the low treatment group took signif- those in the first (least squares mean 5 1 SEM: trial 1, 9.10 5
icantly longer than individuals in the cyclical treatment group 0.16 g; trial 2, 9.76 5 0.17 g; trial number: F1, 79 p 7.91, P p
(Welch’s ANOVA, F1, 31 p 102.53, Dunn-Šidák-adjusted P p 0.006), consistent with their heavier egg mass. The interaction
0.003), which in turn took significantly longer than individuals between trial number and treatment was not significant (F2, 79 p
in the control treatment group (Welch’s ANOVA, F1, 33 p 104.74, 3.00, P p 0.055) but approached significance due to the initial
Dunn-Šidák-adjusted P p 0.002). When split by trial, the effect difference in egg mass (see “Material and Methods”).

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
100 N. Ben-Ezra and G. Burness

Table 1: Hatching success of Japanese quail eggs (n p 174) Adult BMR Was Affected by Incubation
and survival of hatchlings Temperature and Cycling
Temperature treatment Incubation temperature affected adult BMR (treatment: F2, 67 p
Variable Control Cyclical Low 6.48, P p 0.003; fig. 3). Birds in the low treatment group had a
higher BMR than those in the cyclical treatment group (Tukey
Eggs incubated 48 56 70 HSD, P ! 0.05); control birds fell between those in the other
Eggs hatched 29 40 28 treatment groups but did not attain a significant difference from
Individuals surviving to either (Tukey HSD, NS). Individuals in the second trial had higher
sexing (day 25) 26 37 25 BMRs than individuals in the first (least squares mean 5 1 SEM:
Posthatch survival (%) 89.7 92.5 89.3 trial 1, 145.34 5 4.81 kJ/day; trial 2, 163.60 5 4.54 kJ/day; trial:
Note. Survival was 100% between day 25 and the end of the study. F1, 67 p 7.89, P p 0.007). As expected, there was a positive effect
of body mass on BMR (P ! 0.05; table 2). Adult mass-adjusted
BMR did not differ between males and females, nor was there
Incubation temperature significantly affected hatchling head-
an interaction between sex and treatment (P 1 0.1 for both). The
bill length (least squares mean 5 1 SEM: control, 22.98 50.14 mm;
interaction between respirometry mass and treatment approached
cyclical, 22.45 5 0.12 mm; low, 23.01 5 0.15 mm; treatment:
significance (P p 0.069; table 2) but did not attain significance
F2, 82 p 5.80, P p 0.044). Hatchlings in the control and low
even when the nonsignificant sex # treatment interaction term
treatment groups had greater head-bill lengths than those in the
was removed. Due to a priori interest (see “Material and Methods”),
cyclical treatment group (Tukey HSD, P ! 0.05 for both) but did
not differ from each other (Tukey HSD, NS). At hatch, males
and females did not differ in body masses or head-bill lengths,
nor was there a sex-specific effect of incubation temperature
(sex # treatment, P 1 0.10).

Adult Mass and Size Were Affected by Incubation

Temperature but Not Cycling
Incubation temperature had significant long-lasting effects on
adult body mass (least squares mean 5 1 SEM: control, 254.96 5
4.81 g; cyclical, 240.13 5 4.10 g; low, 239.64 5 4.94 g; treat-
ment: F2, 82 p 3.42, P p 0.038; fig. 2A), although a post hoc Tukey
HSD test was unable to identify where this difference lay (control
vs. low, P p 0.055; control vs. cyclical, P p 0.073; cyclical vs. low,
P p 0.99). This likely was the result of the borderline significant
sex-specific effect (sex # treatment: F2, 82 p 2.69, P p 0.074). In
fact, less conservative post hoc Student’s t-tests showed that con-
trol birds were heavier than those in both the cyclical (Dunn-
Šidák-adjusted P p 0.031) and low (Dunn-Šidák-adjusted P p
0.029) treatment groups, while birds in the low temperature and
cyclical groups did not differ from each other (Dunn-Šidák-adjusted
P p 0.470).
Incubation temperature also affected adult size (least squares
mean 5 1 SEM: control, 43.55 5 0.21 mm; cyclical, 42.53 5
0.18 mm; low, 42.60 5 0.21 mm; treatment: F2, 82 p 8.07, P p
0.001; fig. 2B). Control birds were larger in size (as head-bill
distance) than birds in the cyclical and low treatment groups
(Tukey TSD, P ! 0.05 for both), which did not differ from each
other (Tukey HSD, NS). Females were heavier than males at
adulthood (least squares mean 5 1 SEM: female, 268.86 5 3.79 g;
male, 220.96 5 3.77 g; sex: F1, 82 p 80.26, P ! 0.001; fig. A2A) and Figure 2. Japanese quail adult (day 55) measurements for body mass
had greater head-bill lengths (least squares mean 5 1 SEM: (A) and head-bill length (B) for individuals incubated as embryos at
female, 43.16 5 0.16 mm; male, 42.63 5 0.16 mm; sex: F1, 82 p control (37.57C), cyclical (mean: 36.07C; range: 28.07–37.57C), and
5.33, P p 0.023; fig. A2B). There was no sex-specific effect low (36.07C) environmental temperatures, with sexes pooled. Plotted
values are least squares means 5 1 SEM from a general linear model.
of treatment on adult size (sex # treatment: F2, 82 p 2.02, P p Sample sizes are in parentheses; the plus sign in A denotes P ! 0.10
0.1395), meaning that males and females were affected similarly for a Tukey HSD (see “Results” for details), and the asterisk in B
by incubation temperature. denotes P ! 0.05 for a Tukey HSD.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Effects of Incubation Temperature on Adult Phenotypes 101

were 184.1 5 1.1, 173.3 5 1.0, and 187.4 5 1.1 degree-days,

respectively (mean 5 SE). Although birds in the control and low
treatment groups hatched at similar degree-days, the develop-
ment of the birds in the cyclical treatment group was accelerated.
Even if the mean temperature of the cyclical treatment was ele-
vated by up to 0.37C (see “Material and Methods”), embryonic
development was still faster than that for the other treatments
(176.8 5 0.9 degree-days). This suggests that the acceleration
and deceleration of developmental rate that occurs when incuba-
tion temperature is fluctuating can alter the relationship between
developmental rate and temperature, as reported previously for
a variety of insect species (Wu et al. 2015).

Hatching Success
Figure 3. Basal metabolic rate of adult Japanese quail, incubated as em- Hatching success was reduced by low temperature but not by
bryos at control (37.57C), cyclical (mean: 36.07C; range: 28.07–37.57C), cycling. In fact, the cyclical treatment had the highest hatch-
and low (36.07C) environmental temperatures. Plotted values are least
ing success at 71%, although this was not significantly different
squares means 5 1 SEM from a general linear model, with body mass
included as a covariate. Sample sizes are in parentheses; different letters from that for the control treatment (60%), and both were simi-
denote P ! 0.05 for a Tukey HSD. lar to previously reported hatching successes in Japanese quail
(Copur et al. 2010; Javurkova et al. 2015). This result mirrors
we retained the nonsignificant sex # treatment interaction in what was found with even longer cycling periods (8 h off-bouts)
the final model. of incubation in Japanese quail (Callebaut 1990). In that study,
however, average incubation temperature was not controlled
Discussion for. Our results reveal that cycling can mitigate the reduced hatch-
ing success normally seen with constant low incubation tempe-
Length of Incubation
rature (e.g., blue tits [Nord and Nilsson 2011], wood ducks [Hepp
Incubation was extended for the low treatment (∼18.6 d) rela- et al. 2006]). This further suggests that high on-bout tempera-
tive to the control (∼16.0 d), with the cyclical treatment (∼17.3 d) tures during cyclical incubation have more of an influence on
displaying an intermediate duration as well as a reduced vari- hatching success than the average temperature. Eggs in the cy-
ance in hatch time. This suggests that cyclical temperatures during clical treatment group spent only 53 min per 3-h cycle at the
incubation may counteract some of the negative impact of con- control temperature (37.57C), but hatching success was similarly
stant low incubation temperatures. A decrease in average tem- high. The significantly reduced hatching success for the low tem-
perature is widely known to increase incubation period (Eiby perature treatment also indicates that even with the same aver-
and Booth 2009; Ardia et al. 2010; Nord and Nilsson 2011; Du- age temperature as the cyclical treatment, maintaining incuba-
Rant et al. 2012; Kolackova et al. 2015). Our results, however, tion at a constant 36.07C, without 1.57C higher on-bouts, has
demonstrate that fluctuating incubation temperatures can lead a dramatic negative influence on embryo development and egg
to more rapid hatching. Although we recognize that the tem- hatchability.
perature of our cycling treatment may have been up to 0.37C
above the low treatment temperature, we do not think this was
Hatchling Morphology
the cause of differences in hatching time, as explained below.
Instead, the decrease in length of incubation of the cyclical treat- Incubation temperature did not affect hatchling body mass, al-
ment compared with the low treatment more likely reflects though cycling did affect size, as hatchlings in the cyclical treat-
accelerated embryonic development that may occur during the
higher on-bout temperatures (Olson et al. 2006).
Because of the temperature sensitivity of development in birds, Table 2: Results of a general linear model for factors
incubation period has been expressed previously in degree-days influencing basal metabolic rate in adult Japanese quail
(e.g., Olson et al. 2006; Wada et al. 2015). This assumes that high Main effect df F P
temperatures linearly increase the rate of development, so that
developmental stage can be predicted on the basis of incubation Sex 1, 67 .43 .514
temperature (Olson et al. 2006; Koch 2015). We calculated degree- Treatment 2, 67 6.48 .003
days as the mean incubation temperature above physiological Sex # treatment 2, 67 1.33 .270
zero (when development does not occur) multiplied by the num- Trial 1, 67 7.89 .007
ber of days (e.g., Matsuzawa et al. 2002). Using 26.07C as phys- Respirometry mass 1, 67 7.29 .009
iological zero for Japanese quail (Conway and Martin 2000; Lacin Respirometry mass # treatment 2, 67 2.78 .069
et al. 2008), hatching times of control, cyclical, and low treatments Note. Numbers in boldface type indicate significant main effects.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
102 N. Ben-Ezra and G. Burness

ment group had shorter head-bill lengths than individuals in zebra finches but not females (Wada et al. 2015). Chin et al.
both the low and the control treatment group. In contrast, when (2013) demonstrated that female Japanese quail, as the larger
zebra finch eggs were incubated with a temperature treatment sex, were more negatively affected by early food restriction.
that cycled (and was also cooler on average), embryos were Meanwhile, in a previous study we did not detect any sex-
structurally larger (in longitudinal measurements) relative to specific effects in Japanese quail when altering rearing tem-
control embryos that were incubated at a warmer and constant perature (Burness et al. 2013). In the current study, we found
temperature (Olson et al. 2008). Although this is opposite to that both sexes responded equally to low incubation temper-
our finding, zebra finches are an altricial species, while Japa- ature, showing a reduction in size and body mass at adulthood.
nese quail are precocial. These differing developmental strat- The difficulty in detecting a consistent pattern even within the
egies may therefore lead to an altered response to suboptimal same species likely reflects the complex suite of interactions that
growth conditions, especially if growth metrics are being dif- influences sex-specific effects (Jones et al. 2009). Our results
ferentially prioritized (Olson et al. 2008). It is also possible that suggest, however, that incubation temperatures do not appear
the smaller structural size that we detected at hatch reflects a to influence male and female Japanese quail differently.
trade-off between development of adequate structural size and
duration of incubation. Hatchlings from the cyclical incuba-
Adult Metabolic Rate
tion treatment were structurally the smallest and in terms of
degree-days hatched the earliest, indicating an acceleration of As predicted, birds in the low treatment group had a higher
development (Wu et al. 2015). BMR than birds in the cyclical treatment group. However, the
BMR of the birds in the low and cyclical treatment groups did
not differ from that of the birds in the control treatment group.
Adult Morphology
Effects of incubation temperature on metabolism have been de-
Although a number of studies have looked at the effects of low tected previously in altricial species. For example, experimentally
temperature on hatchlings and subsequent development (re- reduced incubation temperatures resulted in an elevated meta-
viewed in DuRant et al. 2013), seldom have individuals been bolic rate in fledgling blue tits (Nord and Nilsson 2011). Sim-
measured through to adulthood (but see Berntsen and Bech ilarly, there was also a transient increase in BMR in female zebra
2015; Wada et al 2015). In the current study, we demonstrated finches as a result of low incubation temperature, although this
that incubation temperature can have long-term effects on an effect was no longer present by adulthood and was never seen
individual’s body mass and size. As predicted, adults in the in males (Wada et al. 2015). Our results suggest that depending
control treatment group (37.57C) were heavier and larger than on developmental strategy, there may be long-lasting metabolic
individuals in the low treatment group (36.07C). This implies effects of incubation temperature pattern. Japanese quail are pre-
that the effects of incubation temperature may be far reaching. cocial and undergo a greater proportion of their development
It has previously been shown that low incubation temperature within the egg than do altricial species, such as blue tits and zebra
can lead to decreased offspring mass and survival (e.g., wood finches. As such, the effects we detected may differ among spe-
ducks; Hepp and Kennamer 2012) as well as decreased recruit- cies coincident with life-history strategies and developmental
ment (e.g., wood ducks; Hepp and Kennamer 2012), while patterns.
reduced mass alone can increase time to sexual maturity (e.g., A plausible mechanistic explanation for the increased BMR
lesser scaups, Aythya affinis; Dawson and Clark 2000). Incu- in adults in the low incubation treatment group relative to the
bation temperature may therefore play a significant role in the cyclical treatment group may be the efficiency of embryonic de-
survival and reproduction of offspring. velopment (e.g., Nord and Nilsson 2011). When zebra finch em-
Although cycling offset several of the negative effects of a low bryos were exposed to periodic cooling, they showed elevated
average incubation temperature, it appeared to have no in- metabolic rates and decreased yolk reserves, indicating that de-
fluence on adult body mass and head-bill length. This finding is velopmental efficiency was being impaired (Olson et al. 2006). If
contrary to our prediction that cycling would offset the long- such inefficiency is maintained during postnatal growth, it may
term negative effects of low incubation temperature on the contribute to the phenotypic variation we detected in BMR at
size and mass of adults. These results show that regardless of adulthood. Furthermore, the results of the current study sug-
whether incubation temperature was constant or cyclical, low gest that it is incubation temperature alone, not a cyclical pat-
mean incubation temperatures negatively affect adult mor- tern, that elevates BMR. This distinction was not made by Olson
phology. This is consistent with data on the influence of incu- et al. (2006), as they compared a constant high temperature treat-
bation temperature on the growth rates of wood ducks from ment with a cyclical low temperature treatment.
hatch to 9–12 d (DuRant et al. 2010, 2013). An alternative mechanism that could explain the variation we
Female quail grew at faster rates and were both heavier and found in adult BMR is that embryonic survival differed between
structurally larger at adulthood than males, as shown previously treatments. Hatching success was decreased for the low treat-
(e.g., Chin et al. 2013). Contrary to our prediction, however, ment (40%) relative to the cyclical treatment (71%). If embryos
there was no observed sex specificity in how individuals re- with high metabolic rates experienced increased survival at low
sponded to the treatments. It has been found that suboptimal temperatures and their relatively high metabolic rates persisted
incubation temperatures result in lower body masses of male to adulthood, this could explain the difference we found. Because

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Effects of Incubation Temperature on Adult Phenotypes 103

we did not measure embryonic metabolic rate, however, we do out the study. A. Rooke, Burness laboratory members, and three
not know whether this may have occurred. anonymous reviewers made valuable comments that improved
the clarity of the manuscript. Funding was provided by a Nat-
ural Sciences and Engineering Research Council (NSERC) Dis-
Cyclical Temperature Treatment covery grant and grants from the Canadian Foundation for
Innovation and the Ontario Innovation Trust to G.B. as well as
As previously mentioned (see “Material and Methods”), the
an NSERC–Undergraduate Student Research Award to N.B. All
mean temperature of the cyclical treatment may have been
experiments were approved by the Trent University Animal Care
underestimated by as much as 0.37C. Although individuals in
Committee and were conducted in accordance with current Ca-
this treatment group hatched at an intermediate time between
nadian law.
those in the low (36.07C) and control (37.57C) treatment groups,
there were no other measures for which they were an interme-
diate. Long-term morphology measurements showed that birds APPENDIX
in the cyclical treatment group were just as negatively affected as
those in the low treatment group, while their hatching success
and BMR were similar to the control birds. These results are in-
consistent with what would be expected of an intermediate tem-
perature treatment. Furthermore, the small potential increase of
0.37C (vs. 1.57C between the low and control treatments) sug-
gests that the marked differences found between the cyclical and
control treatments (including length of incubation) were a result
of cycling alone, as originally intended.

Summary and Conclusions

In summary, our hypothesis that there would be short- and long-
term effects of incubation temperature and cycling on offspring
phenotype was supported. As predicted, low incubation tem-
perature increased the time it took for individuals to hatch and
decreased hatching success. It did not, however, have the pre-
dicted effect on hatchlings, with birds in the low and control
treatment groups hatching at similar mass and size. At adult-
hood, birds in the low treatment group were lighter and smaller
than birds in the control treatment group and had elevated
mass-adjusted BMR, indicating long-term impacts of incuba-
tion temperature. Consistent with our prediction, the cyclical
treatment ameliorated some of the negative effects of low in-
cubation temperature. This was seen in improved hatching suc-
cess for the cyclical treatment (same as control), duration of in-
cubation (intermediate between low and control), and adult BMR,
where birds in the cyclical treatment group had a lower BMR than
birds in the low treatment group. Measures of body mass and size
in the cyclical treatment group showed no improvement relative
to birds in the low treatment group. Contrary to our prediction,
males and females were equally affected by temperature treat-
ment. Future studies using a similar technique of isolating cyclical
incubation from a change in average temperature would provide
valuable information regarding the potential benefits of cyclical
incubation.

Acknowledgments
Figure A1. Japanese quail growth curves for body mass (A) and head-
We thank M. Olson for assisting with various aspects of this bill length (B) for control, cyclical, and low treatment groups, with
project as well as J. Allen and the Trent University Animal Care sexes pooled. Plotted values are means 5 1 SEM; error bars are in-
staff for housing, feeding, and caring for the quail through- cluded, but some are hidden behind symbols.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
104 N. Ben-Ezra and G. Burness

Berntsen H.H. and C. Bech. 2015. Incubation temperature in-

fluences survival in a small passerine bird. J Avian Biol 46:1–5.
Bowden R.M., A.W. Carter, and R.T. Paitz. 2014. Constancy in
an inconsistent world: moving beyond constant tempera-
tures in the study of reptilian incubation. Integr Comp Bio
54:830–840.
Burness G., J.R. Huard, E. Malcolm, and G.J. Tattersall. 2013.
Post-hatch heat warms adult beaks: irreversible physiolog-
ical plasticity in Japanese quail. Proc R Soc B 280:20131436.
doi:10.1098/rspb.2013.1436.
Callebaut M.E. 1990. Hatching of Japanese-quail chicks (Co-
turnix coturnix japonica) following long, daily cyclical inter-
ruptions of their incubation. Poult Sci 69:2241–2243.
Carter A.W., W.A. Hopkins, I.T. Moore, and S.E. DuRant.
2014. Influence of incubation recess patterns on incubation
period and hatchling traits in wood ducks Aix sponsa. J Avian
Biol 45:273–279.
Chin E.H., A.L. Storm-Suke, R.J. Kelly, and G. Burness. 2013.
Catch-up growth in Japanese quail (Coturnix japonica): re-
lationships with food intake, metabolic rate and sex. J Comp
Physiol B 183:821–831.
Conway C.J. and T.E. Martin. 2000. Effects of ambient temper-
ature on avian incubation behavior. Behav Ecol 11:178–188.
Copur G., M. Baylan, and S. Canogullari. 2010. Egg weight but not
egg shape index, determines the hatchability in Japanese quail
(Coturnix coturnix japonica). J Anim Vet Adv 9:1890–1895.
Dawson R. and R.G. Clark. 2000. Effects of hatching date and
egg size and growth, recruitment, and adult size of lesser
scaup. Condor 102:930–935.
DuRant S.E., G.R. Hepp, I.T. Moore, B.C. Hopkins, and W.A.
Hopkins. 2010. Slight differences in incubation temperature
affect early growth and stress endocrinology of wood duck
(Aix sponsa) ducklings. J Exp Biol 213:45–51.
DuRant S.E., W.A. Hopkins, G.R. Hepp, and J.R. Walters. 2013.
Figure A2. Japanese quail growth curves for body mass (A) and head- Ecological, evolutionary and conservation implications of
bill length (B) for males and females, with experimental treatments incubation temperature-dependent phenotypes in birds. Biol
pooled. Plotted values are means 5 1 SEM; error bars are included
Rev 88:229–509.
but are hidden behind symbols.
DuRant S.E., W.A. Hopkins, A.F. Wilson, and G.R. Hepp. 2012.
Incubation temperature affects the metabolic cost of thermo-
regulation in a young precocial bird. Funct Ecol 26:416–422.
Literature Cited Eiby Y.A. and D.T. Booth. 2009. The effects of incubation tem-
perature on the morphology and composition of Australian
Ardia D.R., J.H. Perez, and E.D. Clotfelter. 2010. Experimental Brush-turkey (Alectura lathami) chicks. J Comp Physiol B
cooling during incubation leads to reduced innate immunity 179:875–882.
and body condition in nestling tree swallows. Proc R Soc B Hepp G.R., S.E. DuRant, and W.A. Hopkins. 2015. Influence
277:1881–1888. of incubation temperature on offspring phenotype and fit-
Ben-Ezra N. and G. Burness. 2016. Data from: Constant and ness in birds. Pp. 171–178 in D. Deeming and S.J. Reynolds,
cycling incubation temperatures have long-term effects on eds. Nests, eggs, and incubation: new ideas about avian re-
the morphology and metabolic rate of Japanese quail. Physiol production. Oxford University Press, Oxford.
Biochem Zool, Dryad Digital Repository, http://dx.doi.org Hepp G.R. and R.A. Kennamer. 2012. Warm is better: incubation
/10.5061/dryad.76qt1. temperature influences apparent survival and recruitment of
Ben-Hamo M., B. Pinshow, M.D. McCue, S.R. McWilliams, and wood ducks (Aix sponsa). PLoS ONE 7:e47777. doi:10.1371/
U. Bauchinger. 2010. Fasting triggers hypothermia, and am- journal.pone.0047777.
bient temperature modulates its depth in Japanese quail (Co- Hepp G.R., R.A. Kennamer, and M.H. Johnson. 2006. Maternal
turnix japonica). Comp Biochem Physiol A 156:84–91. doi:10 effects in wood ducks: incubation period and neonate phe-
.1016/j.cbpa.2009.12.020. notype. Funct Ecol 20:307–314.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 Effects of Incubation Temperature on Adult Phenotypes 105

Hopkins B.C., S.E. DuRant, G.R. Hepp, and W.A. Hopkins. retta caretta) pre-emergent hatchlings in Minabe, Japan.
2011. Incubation temperature influences locomotor per- Mar Biol 140:639–646.
formance in young wood ducks (Aix sponsa). J Exp Zool A McClintock M.E., G.R. Hepp, and R.A. Kennamer. 2014. Plas-
315:374–279. ticity of incubation behaviors helps wood ducks (Aix sponsa)
Javurkova V., E. Krkavcova, J. Kreisinger, P. Hyrsl, and L. maintain an optimal thermal environment for developing
Hyankova. 2015. Effects of experimentally increased in ovo embryos. Auk 131:672–680.
lysozyme on egg hatchability, chicks complement activ- Nord A. and J.Å. Nilsson. 2011. Incubation temperature af-
ity, and phenotype in a precocial bird. J Exp Zool A 323:497– fects growth and energy metabolism in blue tit nestlings. Am
505. Nat 178:639–651.
Jones K.S., S. Nakagawa, and B.C. Sheldon. 2009. Environ- Olson C.R., C.M. Vleck, and D.C. Adams. 2008. Decoupling
mental sensitivity in relation to size and sex in birds: meta- morphological development from growth in periodically
regression analysis. Am Nat 174:122–133. cooled zebra finch embryos. J Morphol 269:875–883.
Koch K. 2015. Influence of temperature and photoperiod on Olson C.R., C.M. Vleck, and D. Vleck. 2006. Periodic cooling
embryonic development in the dragonfly Sympetrum stri- of bird eggs reduces embryonic growth efficiency. Physiol
olatum (Odondata: Libellulidae). Physiol Entomol 40:90– Biochem Zool 79:927–936.
101. Schmidt-Nielson K. 1990. Animal physiology: adaptation and
Kolackova M., L. Prokupkova, I. Albrecht, and D. Horak. environment. 4th ed. Cambridge University Press, Cambridge.
2015. Incubation temperature influences trade-off between Wada H., B. Kriengwatana, N. Allen, K.L. Schmidt, K.K. Soma,
structural size and energy reserves in mallard hatchlings. and S.A. MacDougall-Shackleton. 2015. Transient and per-
Physiol Biochem Zool 88:1–10. manent effects of suboptimal incubation temperatures on
Lacin E., O. Coban, and N. Sabuncuoglu. 2008. Effects of egg growth, metabolic rate, immune function and adrenocortical
storage material and storage period on hatchability in Jap- responses in zebra finches. J Exp Biol 218:2847–2855.
anese quail. Asian-Australas J Anim Sci 21:1183–1188. Weathers W.W. and K.A. Sullivan. 1989. Juvenile foraging pro-
Lighton J.R.B. 2008. Measuring metabolic rates: a manual for ficiency, parent effort, and avian reproductive success. Ecol
scientists. Oxford University Press, New York. Monogr 59:223–246.
Matsuzawa Y., K. Sato, W. Sakamoto, and K.A. Bjorndal. 2002. Wu T.H., S.F. Shiao, and T. Okuyama. 2015. Development of
Seasonal fluctuations in sand temperature: effects on the in- insects under fluctuating temperature: a review and case study.
cubation period and mortality of loggerhead sea turtle (Ca- J App Entomol 139:592–599.

This content downloaded from 045.029.215.030 on January 25, 2020 14:15:12 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).