Light increases the rate of embryonic development: implications for latitudinal trends in incubation period

Authors


Correspondence author. E-mail: caren.cooper@cornell.edu

Summary

1. In wild birds, incubation period shortens and the general pace of life quickens with distance from the equator. Temperature and various biotic factors, including adult behaviours, cannot fully account for longer incubation periods of equatorial birds and only explain some of the variation between tropical and temperate life histories. Here we consider the role of differences in light in driving variation in incubation period. In poultry, incubation periods can be experimentally shortened by exposing eggs to light. The positive influence of light on embryonic growth, called photoacceleration, can begin within hours after an egg is laid.

2. We artificially incubated house sparrow (Passer domesticus) eggs under photoperiods similar to those found at temperate (18Light : 6Dark) and tropical (12L : 12D) latitudes. We also measured embryonic metabolic rate during light and dark phases.

3. Eggs of house sparrows collected from the wild developed more rapidly under ‘temperate’ than ‘tropical’ photoperiods and had higher metabolic rates during phases of light exposure than during phases of darkness. Metabolic rates during light phases were high enough to account for a 1 day difference in incubation periods between temperate and tropical birds.

4. Based on a synthesis of photoacceleration studies on domesticated galliformes and our experimental results on a wild passerine, we provide the first support for the testable hypothesis that differences in photoperiod may influence variation in the rate of embryonic development across latitudes in birds.

Introduction

Organisms exhibit a myriad of behavioural, physiological, and anatomical adaptations that occur in limited sets of combinations (i.e. life history strategies). These traits tend to co-vary and form recurring associations, forming a life-history axis characterized as a continuum from a fast to slow pace of life (Ricklefs & Wikelski 2002). In general, pace of life increases with latitude. For example, bird species that breed at temperate latitudes typically have higher metabolic rates (Wiersma et al. 2007), lay more eggs per nest (Moreau 1944; Lack & Moreau 1965), have shorter incubation periods (Ricklefs 1969), and higher nestling growth rates (Ricklefs 1976) than in closely related tropical species.

The metabolic pattern of avian development is unique among vertebrates. Effectively ectotherms early in development, embryos depend on external sources of heat to grow until development has advanced enough to establish homeothermy, which occurs several days after hatching in altricial birds. Temperature, atmospheric gases, humidity, and light influence the rate and quality of embryo growth (Shutze et al. 1962; Romanoff & Romanoff 1967). Surprisingly, patterns of female nest attendance, and resulting incubation temperatures, fail to fully account for latitudinal trends for incubation periods of birds (Ricklefs 1969; Tieleman, Williams & Ricklefs 2004; Robinson et al. 2008), particularly when accounting for phylogenetic constraints (Martin et al. 2007).

Lines of latitude are merely human constructs that describe a suite of selective pressures actually responsible for latitudinal variation in life-history traits. Many studies focus on biotic factors as potential drivers of latitudinal patterns, such as the ultimate and proximate roles of food limitation (Lack 1968), predation on offspring (Skutch 1949, 1985; Martin et al. 2000), and rates of adult mortality (Karr et al. 1990; Johnston et al. 1997; Brawn et al. 1999). Incubation period is a critical variable to consider in life-history studies because variation in developmental rates has cascading consequences that influence expression of other traits via carry-over effects throughout an organism’s entire lifetime (Blount et al. 2006). Adult attendance and incubation temperature have long been considered the primary factors affecting latitudinal variation in incubation period. Yet recent studies have shown conflicting results, which may reflect a diversity of selective pressures influencing this trait (Ricklefs 1969; Tieleman, Williams & Ricklefs 2004; Robinson et al. 2008; but see Martin et al. 2007). Research on biotic factors may produce varying results because the biotic factors may not vary as consistently with latitude as abiotic factors. The contribution of abiotic factors, such as those associated with variation in sunlight, is relatively unexplored. To our knowledge, no studies of wild birds have looked at the role of the abiotic factor sunlight on latitudinal differences in incubation period. We investigated light because photoperiod is a patently reliable component of the environment varying predictably with latitude, and it is well-established that light increases embryonic development rates in domesticated species.

How light affects embryonic development: insights from domesticated species

In the mid-1960s the poultry industry discovered that light intensity above a certain threshold accelerated embryonic development (photoacceleration in sensuShutze et al. 1962; Isakson, Huffman & Siegel 1970), prompting experiments to evaluate possible mechanisms.

There are marked differences between the development rates of chicken embryos incubated under continuous high-intensity light (1100–3000 lx) and those incubated under continuous darkness. Embryos exposed to light hatched approximately 1 day earlier than those receiving no, or only a short pulse of, daily light (Lauber 1975; Ghatpande, Ghatpande & Khan 1995). Photoperiod also affects embryonic growth and development. Poultry embryos exposed to continuous light produced the shortest incubation periods (Siegel et al. 1969), while those exposed to 12-h light-dark cycles took 0 to 5 h longer to hatch (Walter & Voitle 1972, 1973; Rozenboim et al. 2004). The rate and mechanism of photoacceleration varies with stage of embryo development. Light stimulation produces the fastest rate of embryonic development during early incubation and a less rapid rate during the middle of incubation. Light stimulation during the last week of incubation did not have any apparent effect on development (for an overview see Siegel et al. 1969). In experiments comparing 0 and 24-h light treatments, differences in embryonic development can often be detected after as few as 10 h (Siegel et al. 1969).

The physiological mechanisms by which light stimulates embryonic growth and development differ prior to and after formation and maturation of retinal photoreceptors, the hypothalamic pacemaker and the pineal gland – the primary components of the avian circadian system (reviewed in Dawson et al. 2001). Light appears to stimulate mitosis in neural crest mesoderm during the first 2 days of chicken embryo development, accelerating closure of the neural tube (embryonic d1 or stage 7 of the Hamburger–Hamilton (H–H) classification of avian embryonic development) and subsequent somite development (Isakson, Huffman & Siegel 1970). This is consistent with observations that high light intensity increases embryonic cell proliferation (Ghatpande, Ghatpande & Khan 1995). Recently, Halevy et al. (2006) demonstrated that mesodermal differentiation into myoblast cells occurred earlier (embryonic d5) in embryos exposed to green light (560 nm). This process is regulated by a known family of transcription factors (MyoD) and may be triggered by photic cues from retinal or pineal photoreceptors acting on the neuroendocrine system (Halevy et al. 2006). There is also evidence that when light penetrates to the cellular level, as would be possible in early stage avian embryos, it can directly activate cytochromes in the mitochondrial electron transport chain and stimulate cellular metabolism (Karu 1988). Work with mammalian cells in the 1980s provided evidence that visible light can regulate cellular metabolism by way of cAMP, subsequently leading to the initiation of DNA synthesis (Karu 1988). Thus, light may influence gene expression during early development (d1 to d5 in chickens embryos), thereby accelerating the growth process (Shafey 2004). After the formation of the pineal gland (H-H stage 17, d3 in chickens), entrainment of the avian embryo to photoperiod is mediated by production of melatonin (Hill et al. 2004). Light reduces melatonin production (Akasaka et al. 1995) while increased melatonin production occurs at night (Dawson & Van’t Hof 2002). As little as 1 h of light decreases melatonin production in embryonic chickens (Zeman et al. 1999); and even very low light intensity (e.g. as low as 0·15 lx) can reduce melatonin production in mammals (Lynch, Deng & Wurtman 1984). Similarly, low light intensity (10 lx) can entrain embryonic starlings, Sturnus vulgus (Gwinner, Zeman & Klaassen 1997). Note that natural light levels in North American nest boxes often exceed this level during the breeding season (183 ± 196 lx, range 5·3–650 lx; M. Voss, unpublished data). Light also influences the activity of clock genes found in some avian tissues (e.g. multiple regions of the brain, the ovary; Nakao et al. 2007); however little is known about the embryonic ontogeny of avian circadian clocks (Okabayashi et al. 2003). What is known suggests that avian circadian clocks mature late in the embryonic chicken (d16 in the suprachiasmatic nucleus; d18 in pineal gland Okabayashi et al. 2003; d19 retinal photoreceptors Wai et al. 2006), when photoacceleration is no longer significant (Siegel et al. 1969). Regardless of the mechanism of origin, prenatal entrainment to photoperiod does not usually persist as chicks tend to re-entrain to the ambient photoperiod during the first few days after hatch in experimentally manipulated studies (Zeman et al. 1999).

Finally, light exposure can cause changes in embryonic metabolic rate (Preda et al. 1962). For example, the metabolic rate of pigeon (Columba livia) embryos is greater in light than dark (Prinzinger & Hinninger 1992). Heart rate also increases in in vitro embryo hearts exposed to light (Gimeno, Roberts & Webb 1967) and in in vivo embryos during daylight hours (Moriya et al. 1999).

In summary, experimental work in domesticated species reveals several key conclusions: (i) light accelerates embryonic development, in part by increasing metabolic activity, (ii) the extent of photoacceleration is affected by the photoperiod, and (iii) the rate of photoacceleration varies with the stage of embryo development, primarily due to differences in the mechanism by which light affects development. Prior to pineal gland formation, light affects specific developmental pathways such as neural crest mitosis. After pineal gland formation, light acts primarily by modifying melatonin synthesis. Taken as a whole, extensive studies in domesticated species reveal that variation in photoperiod can have critical and important effects on the rate and route of development of embryonic birds.

Objectives

We expect sensitivity to light to be an evolutionarily conserved trait, not one expressed only in domesticated lines of birds. Indeed, given that selection for increasing developmental speed to maximize production of industrial goods has presumably been quite strong, it is possible that the magnitude of the response could be greater in wild birds.

We tested for the positive effects of light on house sparrow (Passer domesticus) development using two experiments: (i) comparing direct measures of embryonic metabolic rates in artificial incubators under sequential light and dark phases, and (ii) comparing total development time of artificially incubated embryos raised under two different photocycles. We hypothesized that embryo metabolism would be lower during dark phases and higher during light phases, mimicking circadian rhythms in metabolic rate. We hypothesized that longer daily exposure to light would result in shorter embryonic development. We present results from the first photoacceleration experiment conducted on a wild bird species, the house sparrow. Then we explore the possible implications of photoacceleration on trends in incubation periods (seasonal, latitudinal, between cavity, enclosed- and open-cup nesters) and variation in egg coloration. We suggest that temperature and light, received by embryos as a consequence of adult incubation rhythms and the abiotic environment, act as epigenetic effects (Nichelmann, Hochel & Tzschentke 1999; Nichelmann 2004) that influence the rate of development and imprint a pace of life on the developing bird.

Materials and methods

Study species

House sparrows, introduced from Europe to North America in the 1850s, are currently widely distributed across North America. North American house sparrows lay 1–8 eggs per clutch (mode = 5 eggs), rear as many as four broods per season, and initiate up to eight attempts per season (Lowther & Cink 2006). Although only females develop a brood patch, males assist in incubation. In the wild, the incubation period ranges from 10 to 14 days, with an average of 11 days from last egg to first hatch (Lowther & Cink 2006). In artificial incubators, Wetherbee & Wetherbee (1961) reported their longest incubation as 282 h (11·75 days).

Embryonic metabolic rate

We collected house sparrow eggs (n = 52 eggs from 13 clutches) from natural nests found in nest boxes in Erie, PA on the day the 4th egg was laid. We pre-warmed and calibrated both of our incubators (36·09C ± 2·86 SD; still air Hovabator window models) prior to inserting eggs. Incubators contained an automatic egg turner that turned eggs 0·25 revolutions per hour and with an average relative humidity of 73·3% ± 5·3%. Because the egg turner was designed for larger eggs, we put a square of chemwipe under each egg before positioning it in the automatic turner to prevent them from slipping through the egg cups. Light within the incubator was limited to the output from a Zilla® mini reptile UVB full spectrum compact fluorescent light (1155·09 ± 613·52 lx) set to an 18L : 6D photoperiod, phase-shifted so light phase began at noon (EST). A 20cc syringe fitted with rubber stoppers and tubing was used to make a small respirometer chamber that fit inside the incubator. We pumped room air through Ascarite© filled CO2 scrubbers into ultra-low permeability Tedlar© gas bags in order to standardize ambient CO2 levels. We then pumped the CO2-free air from the gas bag at 400 mL min−1 through the respirometer chamber to Quibit© gas analyzers (CO2 and O2). We measured CO2 production for each egg under dark and light conditions within 4 h of each other and paired the observations for individual eggs. Metabolism measurements were made for each egg over several days of incubation. We weighed and candled eggs daily to check for adequate development. We excluded from the analysis any eggs that failed to develop or died early in the incubation period. We computed metabolic rate in μL CO2 min−1 egg−1 from CO2 measurement adjusted for flow rate.

Embryonic development time

We gathered house sparrow eggs in Ithaca, NY to simultaneously test the prediction that 12L : 12D photoperiods would lead to longer incubation periods than eggs incubated under 18L : 6D photoperiods. In Ithaca, NY, we used a total of four Hovabator still air window incubators, two with egg turners for incubating eggs and two with screens on the floor for hatching. We refer to the former as egg-rotating incubators and the latter as hatching incubators. The two egg-rotating incubators were kept at 37·22C ± 0·72 SD with average relative humidity of 69·94% ± 5·60 SD and 37·23C ± 0·49 SD with average relative humidity of 68·53% ± 12·65 SD. The two hatching incubators were kept at slightly lower temperatures and higher relative humidity (35·56C ± 0·23 SD with average relative humidity of 84·57% ± 2·74 SD and 36·39C ± 0·51 SD with average relative humidity of 81·56% ± 7·13 SD). Each incubator contained two windows, one of which we covered with electrical tape and one of which we fitted with a light fixture containing a full-spectrum florescent light (Zilla® mini reptile UVB full spectrum compact fluorescent). We used electrical tape to adhere the light to the window and to light-proof the edges. The light produced 1203·59 ± 974·36 lx at egg height. Since we left small holes open for air circulation, we surrounded each incubator with black-out curtains to prevent ambient light from reaching the eggs. Thus, eggs experienced light only from the light fixtures attached to the windows. Desired humidity levels were reached by maintaining a specific surface area of water in each incubator. For first broods, one set of egg-rotation and hatching incubators were set at 12L : 12D, the equatorial photoperiod, and one set of egg-rotation and hatching incubators were set at 18L : 6D, a photoperiod common to temperate latitudes during the breeding season. The temperature, humidity, and air flow levels in incubators are difficult to control precisely. For this reason, other studies (e.g. Lauber 1975) kept control and illuminated eggs in the same incubators and applied light exposure via optical fibres placed into a small hole in the eggshell. Instead, we repeated the experiment with second broods after switching which incubators were assigned as 18L : 6D and 12L : 12D treatments. Both treatments began the light phase at 6:00 EST.

We monitored house sparrows nests in boxes several times a week from late April through June 2009. We collected eggs between 9:00 and 11:00. We numbered each egg with a permanent marker, weighed it to the nearest 0·01 g on portable electronic balance, and alternated the assignment to 18L : 6D and 12L : 12D treatments for each egg collected from different nests. We transported eggs between the field and incubators in a compartmentalized sewing box filled with bird seed, with each egg’s air sac facing up (Kuehler et al. 1993). We weighed and candled eggs daily during incubation. Eggs were typically transferred to hatching incubators on day 9 when candling indicated normal growth and several days of consistent embryo movement. After transfer to hatching incubators, eggs were rotated and checked for signs of pipping twice daily. Seven of the 88 eggs cracked during handling and were excluded from the experiment.

Results

Embryonic metabolic rate

The variance around each treatment mean was approximately equal and the distributions were normal. The mean metabolic rate was 1·30 ± 0·57 μL CO2 min−1 egg−1 (mean ± 95% C.I.) for the dark treatment and 1·92 ± 0·73 μL CO2 min−1 egg−1 for the light treatment. We used a general linear mixed model with egg ID as a random effect to estimate the difference in light and dark phase metabolic rates (sas version 9.1, SAS 2002). The positive least square mean estimates of the difference in light-to-dark phase metabolic rate indicated that light phase metabolic rate was greater than dark phase metabolic rate throughout egg development (Fig. 1).

Figure 1.

 Embryonic metabolic rate measured on 52 eggs collected from nests of wild House Sparrows. (a) Least square mean estimate of embryonic metabolic rate (μL CO2 min−1 egg−1) was higher during light phases than dark phases (bars with 95% confidence intervals) from measurements across embryo development. (b) The average difference in embryonic metabolic rate between light and dark phases was always greater than zero.

Extrapolating additive effects

We used the average light and dark metabolic rates to estimate the difference in incubation period for 18L : 6D and 12L : 12D photoperiods. These photoperiods represent approximate temperate and tropical daylengths, respectively. With the temperate house sparrow incubation period of 11 days as a point of reference, we assumed that a typical temperate day during the incubation period had 1080 light minutes and 360 dark minutes. If total daily CO2 production consists of 1080 min of light × 1·92 μL CO2 min−1 egg−1 and 360 min of dark × 1·30 CO2 μL−1 min−1 egg−1 then the metabolic rate to support daily development would be 2548 μL CO2 day−1 egg−1; total metabolism for 11 days of development would produce 28024 μL COegg−1. We also calculated the daily CO2 production under a tropical photoperiod (720 min light × 1·92 μL CO2 min−1 egg−1 and 720 mins dark × 1·30 μL CO2 min−1 egg−1 = 2324 μL CO2 day−1 egg−1). If we assume that 28024 μL COegg−1 over 11 days is the baseline metabolic rate required to produce a normal embryo under the temperate photoperiod, then 12·10 days would be necessary under the tropical photoperiod to reach the same point. The difference in daylight hours should result in tropical incubation periods only 1·10 days longer, about 10%, than those under temperate photoperiods.

Embryonic development time

We considered alpha ≤0·05 to be statistically significant. We included eggs for which the laydate accuracy was high, i.e. the difference between the estimated minimum and maximum laydate was zero (n = 23) or 1 (n = 13) or 2 (n = 11). Of these 47 eggs, 24 were in the 18L : 6D treatment (n = 14 in incubator 1, n = 10 in incubator 2) and 23 were in the 12L : 12D treatment (n = 10 in incubator 1, n = 13 in incubator 2). Thirteen eggs, distributed approximately equal across treatments, failed to pip or hatch. Including all 47 eggs in a general linear mixed (GLM; Proc Mixed in sas) model of treatment with nest origin as a random effect, indicated no difference in the initial mass of eggs entering each treatment (GLM, d.f. = 39, = 0·67). Including the 34 eggs that survived to pip in a mixed model of treatment, incubator, and their interaction, with nest origin as a random effect, eggs in 18L : 6D treatment had significantly (GLM, d.f. = 23, = 0·007) shorter incubation periods (12·7 days ± 0·26 SD) than eggs in the 12L : 12D treatment (13·7 days ± 0·26 SD; Fig. 2). Incubation periods also differed between incubators, but the interaction between incubators and treatment was not significant (GLM, d.f. = 23, = 0·73). Results were similar after removing the potential outlier (GLM, d.f. = 22, P = 0·03 with shorter incubation periods (12·7 days ± 0·21) in 18L : 6D than in the 12L : 12D treatments (13·4 days ± 0·22), P = 0·04 for incubator effect, and P = 0·8 for the treatment-incubator interaction). The results confirm our prediction regarding photoacceleration: eggs incubated under the longer photoperiod hatched about 1 day earlier than eggs incubated under the shorter photoperiod.

Figure 2.

 Incubation period (number of day from laydate to pipping) measured on 34 eggs collected from nests of wild House Sparrows (open circles) in two incubators under two photoperiod treatments, with the least square mean estimates from a mixed model (black circles). The incubation period for eggs incubated under temperate (18L : 6D) photoperiod was shorter than for eggs incubated under tropical (12L : 12D) photoperiod, even with outlier (inc 1, tropical) removed.

Discussion and synthesis

The combined body of evidence, from research on precocial domesticated birds and our experimental evidence on an altricial wild bird species, suggests that light can accelerate embryonic development within a species. Embryos raised in common garden conditions showed differences in metabolic activity and development period based purely on differences in light. The differences we report in incubation period in house sparrows under 12L : 12D (tropical) vs. 18L : 6D (temperate) photoperiods are consistent with the hypothesis that photoacceleration may be an underlying mechanism contributing to latitudinal gradients in incubation periods. The effective biological relevance of sunlight on embryonic development in the wild may depend on (i) the relative importance of circadian rhythms compared to additive effects, (ii) interactions with other influential features of the embryonic environment, such as temperature, habitat, and parental behaviour that influences how photoperiod translates into the amount and quality of light received by embryos, and (iii) the importance of phylogenetic constraints and local adaptations. Temperature, in particular, likely plays an important role in driving incubation period, as experimental work has revealed the direct effect of temperature on development periods (Martin et al. 2007; Ardia et al. 2009; Ardia, Perez & Clotfelter 2010), but differences in incubation periods are not explained only by temperature (Robinson et al. 2008).

Additive or circadian

Light could have an additive effect, with more total light resulting in incrementally more rapid development. This effect would produce a latitudinal gradient in incubation periods where tropical species have incubation periods 1–2 days longer, on average, than eggs exposed to long photoperiods of temperate latitudes. Our results reveal an approximate 10% difference in incubation period produced from a 6-h daily difference in light exposure to embryos of one species in an experimental environment. Phylogenetically controlled comparisons of incubation periods between temperate and tropical species exhibiting approximately 2-h day−1 differences in daylength show greater differences in incubation period, ranging from 30% to 50% (Martin et al. 2007; Martin & Schwabl 2008). The 10% difference we measured indicates that light, via photoacceleration, could explain a portion of the variation in incubation period across latitudes. Our values are more comparable to seasonal declines in incubation period, reported to be about 1 day between spring and summer for Eastern Bluebirds (Cooper et al. 2005). Additional studies that combine measurements of incubation period across latitudes with experimental manipulation of light exposure to eggs collected across latitudes will be needed to assess the relative contribution of photoacceleration compared with other contributing factors.

Although our results only permit us to infer summative effects of light, the potential importance of the circadian aspects of photocycles is also ecologically relevant because physiological processes can be influenced by pulses of light. After development of the pineal gland, pulses of light during adult recesses may register as complete photoperiods to embryos. It is widely established that brief periods of light to signal the beginning of the day and another to signal the end of the day, called skeleton photoperiods or intermittent lighting, can be interpreted by birds as a ‘subjective day’ (Bacon 1984; Rowland 1985) and entrain physiological processes (Slaugh et al. 1988). It is conceivable that the incubation patterns of birds could create a skeleton photoperiod by exposing eggs to brief pulses of light, essentially defining a subjective day. These could then entrain metabolic cycles that are higher during the ‘light’ phase than the ‘dark’ phase. Most evidence suggests that activity periods of wild birds are tightly linked to the local photoperiod (Daan & Aschoff 1975). Therefore, dawn and dusk pulses of light during adult recesses may be enough to synchronize embryonic biological rhythms to the local photoperiod. In adults, experimental exposure to skeleton photoperiods can entrain daily rhythms in metabolism (Wikelski et al. 2008) and while to the best of our knowledge this has not been directly tested in embryos, light pulses during daylight hours may be sufficient to entrain metabolic rhythms (Styrsky, Berthold & Robinson 2004).

Interactions with other features

Under natural incubation conditions, eggs are contact-incubated by adults and therefore experience conditions greatly different from conditions in experimental incubators. Specifically, in the wild, eggs are unlikely to receive heat generated by contact incubation at the same time as receiving light, though there may be situations where ambient conditions provide equivalent heat or in which light reaches eggs through the sides of nests while an adult is incubating. Most eggs experience their longest exposures to sunlight during the laying period (females typically lay one egg per day), when eggs receive little parental attendance. Presumably, light cannot enhance development outside the embryonic thermoneutral zone, although we are not aware of studies that examine this possibility. A possible scenario may be that light stimulates protein production (Shafey 2004; Halevy et al. 2006), and when heat resumes more proteins are then available as molecular switches to turn on metabolic pathways. Furthermore, some species may initiate incubation before all eggs in a clutch are laid when adults sleep on nests at night (Rompré & Robinson 2008). This transfer of heat at night, while leaving eggs mostly unattended during the day, may launch developmental processes in the embryo and also allow embryos to entrain to photoperiod during the first days prior to full onset of diurnal incubation.

Inferring the ecological significance of lab-based photoacceleration is complicated by many other factors. Many eggs are in nests located in areas of low light intensities, such as forests or in cavities. In full sunlight, light intensity is over 10 000 lx, cloudy days about 1000 lx, nest boxes can range from 5 to over 500 lx, and natural cavities may receive even less light. Interestingly, cavity nesting species have longer incubation periods than open nesting species (Martin 1995). Yet, in some ways, extrapolating photacceleration results from domesticated to wild birds is highly conservative because eggshell density, which is much lower in wild birds, plays a role in the amount of light that reaches the blastodisc (Shafey 2004). There is some evidence that passerine eggs could be affected by even lower light intensities than reported for domesticated birds (Rahn & Paganelie 1989). The condition of reduced eggshell density allowing increasing light to penetrate shells and the potential for increased sensitivity to light may allow (or have been selected to) link the physiological mechanisms underlying photoacceleration with pulsed and reduced light cues. Furthermore, egg colour could influence the extent to which photoacceleration is important in natural settings. The thickness of eggshell and degree of pigmentation may preferentially filter some wavelengths of light (Romanoff & Romanoff 1949), reducing the efficiency of the photoacceleration mechanism. A recent poultry study (Shafey et al. 2004) showed that eggshells are particularly efficient at blocking UV wavelengths, and that the brown eggs of one breed are more effective than the white eggs of another breed in blocking light transmittance across the spectrum. The eggs used in our experiment varied in colour from white to blue to grey, but we did not quantify the degree of pigmentation or opacity of each egg. Future research is needed to examine the role of photoacceleration in explaining the longer incubation periods of cavity nesting birds relative to open nesting birds, and the potential to account for variation in egg pigmentation.

Phylogenetic constraints and local adaptations

Some species may show no differences in incubation period across a broad geographic range, suggesting a lack of response to light and other environmental features. For example, the incubation period of white-crowned sparrow subspecies (Zonotrichia sp.) appears phylogenetically constrained as it does not vary among subspecies, with altitude, nor with latitude (Morton 1976; King & Hubbard 1981; Carey et al. 1982). Nevertheless, the incubation period of white-crowned sparrows does decrease between spring and summer (Mead & Morton 1985). Because photoacceleration is a proximate mechanism contributing to the rate of development, selection pressures may result in adaptations that enhance or diminish photoacceleration. For example, the degree of eggshell pigmentation has been shown to mediate the amount of light reaching the blastodisc during early avian development (Shafey et al. 2004). Ghatpande, Ghatpande & Khan (1995) demonstrated that the amount of light reaching the embryo determines the extent to which photoacceleration enhances development. Species with heavily pigmented eggshells, such as the white-crowned sparrow, may therefore not experience photoacceleration to the extent of other species.

Pigmented eggshells represent one example of adaptations that may augment photoacceleration. A variety of adaptations, such as reduced shell pigmentation, increasingly exposed nesting sites, or possibly variation in key isozymes responsible for energy metabolism, may play a role in the greater differences in natural incubation period between temperate and tropical species, even when differences in photoperiod length are shorter than those in our experiment. For example, Schaefer et al. (2004) measured 35–40% longer incubation periods in Sylvia species in tropical Africa (14·5–15 days) compared with many temperate-zone Sylvia species in Europe (10–12 days, averaging around 11 days). Since our study only included house sparrows living in temperate regions (presumably with many local adaptations to temperate conditions), we cannot make greater inferences about how the embryonic development of house sparrows residing in the tropics may or may not respond to light.

Although we provide experimental evidence for photoaccelerated development in a Passerine species, the broader influence of light exposure on avian development and metabolism still requires additional investigation. Our synthetic review, coupled with limited experimental evidence from a wild species, suggest some major latitudinal differences in life-history traits could be explained, in part, by variation in basic environmental features such as latitudinal differences in quantity and quality of light.

Acknowledgements

We received support from the U. S. National Science Foundation IRCEB (grant 0212587 to WDR).

Ancillary