Introduction

Breeding safely is the biggest challenge for forest birds (reviewed by Wesołowski and Tomiałojć 2005). Predation is usually the main reason for nest loss and adult death (e.g. Nice 1957; Ricklefs 1969; Wesołowski and Tomiałojć 2005), but is better avoided by the birds occupying the safest nest sites. Tree cavities provide the most secure places for breeding and they are also usually superabundant in natural forests (review in Wesołowski 2007). However, numerous bird species fail to utilize this ecological opportunity. The question of why more bird species do not nest in tree cavities was posed by Lack (1954) and Nice (1957) more than half a century ago, and still awaits an adequate answer.

The existence of constraints that are impassable for some species could be an explanation (Wesołowski 2007). Successful breeding in tree cavities requires a cavity-nesting bird to deal with several problems that are nonexistent or of minor importance in other types of nest sites. The major issues of cavity nesting could include: how to keep the nest contents dry, how to provide sufficient air exchange, and how to tend broods in very dark places. Inability to tackle any of these problems would prevent the birds from taking advantage of the security provided by cavities. One can envisage that individual bird species have differed in their ability to cope with these challenges, and only a fraction of them have possessed the anatomical or physiological traits that were necessary for the evolution of the cavity-nesting habit (Wesołowski 2007).

There are reasons to expect that light conditions within tree cavities might constrain both the evolution of the cavity-nesting habit and the nest site choice within these species. Tree cavities are rather dark places in which the amount of light declines rapidly with increasing distance from the entrance (Wesołowski and Maziarz 2012). Therefore, to use such sites, a bird has to be able to operate in low light conditions, either by being able to see details at low illumination or by using other sensory modalities (reviews in Land and Nilsson 2002; Martin and Osorio 2008). The use of olfaction, touch and/or sound to orientate in darkness is observed in some birds (review in Martin 2011), but has not been demonstrated in tree-cavity nesters. Therefore, to guide behaviour within nest cavities, it seems highly probable that cavity-nesting birds must rely on vision. Changes in parental feeding behaviour in reaction to experimental manipulation of the chicks’ appearance seem to support this; for example in Great Tits Parus major (Heeb et al. 2003; Wiebe and Slagsvold 2009, 2012), Northern Flickers Colaptes auratus and Pied Flycatchers Ficedula hypoleuca (Wiebe and Slagsvold 2009, 2012).

If birds do rely on visual cues within the nest cavity (and no visual system can operate in complete darkness; Land and Nilsson 2002), there must be some limiting illumination level below which they cannot see, i.e. some nest cavities might be too dark to be usable. Additionally, a diurnally active adult bird, when feeding its young in a cavity, has to be able to rapidly adjust its sight to the light conditions at the nest by switching between the well-lit environment outside to the dim light inside the cavity and back again, often within a few seconds. The ability to cope with rapidly fluctuating light levels through, for example, the use of a pupil with a high dynamic range, may have evolved only in cavity nesters, but this supposition has not been tested (Wesołowski and Maziarz 2012).

There is limited data on illuminance in nest cavities (summarized in Wesołowski and Maziarz 2012). Our previous study of illumination in the nest cavities of Marsh Tits Poecile palustris and Great Tits breeding in the primeval forest of Białowieża National Park (Poland) showed that the birds were able to feed young in dim light when illuminance fell within the mesopic–scotopic range of vertebrate eyes, equivalent to moonlight–twilight in open habitats (Wesołowski and Maziarz 2012). This work also suggested that the patterns of nest-site use could be affected by the requirement for sufficient illumination within the cavity. However, we do not know how important the problem of “having enough light” is for other cavity-nesting birds, or how the birds deal with it within specific cavities.

To check how general this constraint could be, we measured illumination in cavities of another obligatory cavity nester, the Collared Flycatcher Ficedula albicollis (Old World Flycatchers, Muscicapidae), breeding in the same forest. We chose this species for two reasons: (1) the cavity-nesting habit of flycatchers has evolved independently from that of the tits (von Haartman 1957), and (2) Collared Flycatchers usually use tree cavities of different dimensions than the tits in BNP (Wesołowski 1996; Walankiewicz et al. 2007; Wesołowski and Maziarz 2012). We believe that examining the illumination within flycatcher nest cavities would not only enhance our understanding of the flycatchers’ behaviour but also, by comparison with the tits, provide a greater insight into the evolution of cavity nesting in general.

In the primeval forest habitat of Białowieża National Park (Eastern Poland), Collared Flycatchers have a wide spectrum of tree cavities to choose from (Tomiałojć et al. 1984; Walankiewicz and Mitrus 1997; Czeszczewik et al. 1999, 2012; Wesołowski 2003). In such conditions, they breed mostly in nonexcavated cavities in living trees (Walankiewicz 1991; Walankiewicz et al. 2007). The position and dimensions of the cavities used by this species overlap partially with those used by Pied Flycatchers, Marsh Tits, or Great Tits (Walankiewicz 1991; Wesołowski 1989, 2007; Czeszczewik and Walankiewicz 2003), but, as the cavities are superabundant (see above), Collared Flycatchers are not constrained by interspecific competition for nest sites. By building bulky nests in cavities, the birds may better prevent soaking of the nest contents by water accumulating within the cavity (Wesołowski et al. 2002), and adjust the distance of the nest from the cavity entrance to avoid predation (Wesołowski and Maziarz 2012).

We took advantage of intensive studies of Collared Flycatchers carried out in Białowieża National Park (Walankiewicz et al. 2007; Czeszczewik et al. 2012) to measure illuminance inside nest cavities used by this species. Here, we present the results and compare them with the values obtained in cavities utilized by tits (Wesołowski and Maziarz 2012). We discuss the implications of the findings for understanding adaptations to cavity nesting in birds.

Study area and methods

The study was carried out in the strictly protected part of the Białowieża National Park (Eastern Poland, ca. 52°40′N, 23°50′E), within three large, permanent study plots (C, M, W) located in oak–hornbeam–lime old-growth forest of primeval origin (see Wesołowski et al. 2010 for detailed descriptions and photos). Tree cavities are superabundant here, and the cavity-nesting birds have a wide spectrum of nesting places to choose from (Wesołowski 1996, 2007).

Illuminance was recorded in Collared Flycatcher nest cavities in 2013, in an identical manner to that described for tits in Wesołowski and Maziarz (2012). For measurements we used Konica Minolta T-10 M meters with mini receptor heads (measuring range 0.01–299,000 lx). The receptor heads were connected by a flexible cable to the main device and were small enough (16 mm diameter) to pass through even the narrowest cavity openings. This feature allowed us to put only the sensors inside cavities (Fig. 1). All measurements were performed by well-trained observers, the same as in the previous study (Wesołowski and Maziarz 2012).

Fig. 1
figure 1

Changes in illuminance with increasing cavity depth, recorded in the artificial Collared Flycatcher nest cavity. Ambient illumination was that of a sunny sky at midday (illuminance ca. 66,000 lx in the open). The dimensions of the artificial cavity corresponded to median values of tree cavities used by Collared Flycatchers in Białowieża National Park, Poland (top right, Walankiewicz et al. 2007; see “Methods” for details)

Full size image

We measured illuminance in nest cavities containing 8–12 day-old nestlings. To record the minimum possible light level at which parents were able to feed their young, we began observations before sunrise and took measurements at each nest cavity at the time when parents commenced feeding (for a detailed description, see Wesołowski and Maziarz 2012). For safety reasons, we took measurements only in cavities accessible from a ladder, up to ca. 6 m above the ground, which fell into the lower range of Collared Flycatcher nest-cavity heights in the study area (median 8 m; Walankiewicz et al. 2007).

During each visit we took the measurements: (1) with the receptor head inserted into the nest cavity, placed centrally just above the nestlings, with the receptor window faced upwards; (2) with the receptor head situated at the cavity entrance, held in the plane of the cavity opening, which was sideward or upward; (3) on the forest floor under a gap in the forest canopy nearest to the cavity tree, with the receptor window held facing upwards. The gap had to be big enough so that full daylight could reach the forest floor. The measurements were always taken in this order, and all illuminance values were recorded in the normal fast mode of the light meter.

At each nest cavity, we also gathered additional information which could have an influence on the collected values. These were the time (hour:min) of the adults’ arrival at the nest, the time of the first feeding (the first time the adult entered the nest cavity in the morning), the times when measurements were taken inside and outside of the nest cavity; weather conditions (clear sky or cloudy); the cavity depth (vertical distance between the lower edge of the cavity entrance and the top of the nest material); the sensor surface depth (the vertical distance between the lower edge of the entrance and the top of the sensor).

To minimize the duration of the morning visit to each nest, we measured the physical dimensions of each cavity a few days after the young had fledged. These dimensions were: cavity diameter (least/greatest dimension of the cavity interior cross-section; measurements taken perpendicular to each other in a plane at the nest rim level); entrance diameter (least/greatest dimension of the cavity entrance cross-section; measurements taken at the narrowest place in the entrance plane), and the entrance orientation (sideward/upwards). We used a collapsible ruler and a flashlight bulb fixed to a flexible wire to perform measurements (further details in Wesołowski 1996).

To gather information on the possible variation in illuminance with cavity depth, we made an artificial cavity in which dimensions corresponded to the median values of Collared Flycatcher nest cavities in the Białowieża National Park (Walankiewicz et al. 2007). The artificial cavity was a 60-cm-long cylinder made of thick, dark brown, matt cardboard that was closed at both ends with opaque lids. It was constructed in the same way as the artificial tit cavities in the previous study (see Wesołowski and Maziarz 2012; Fig. 1 for details). The internal diameter of the Collared Flycatcher model cavity was 10 cm; the entrance was 5 cm wide and 8 cm high. The artificial cavity was fixed vertically to a tree trunk in a position typical of a Collared Flycatcher nest cavity. To register the maximum light intensity inside the cavity, the measurements were taken at the highest ambient light level, at midday on a sunny, cloudless day. The illuminance inside the cavity was recorded at 10-cm intervals beneath the entrance, down to 50 cm in three sequences in immediate succession. The values of three measurements at 20 cm below the entrance differed by up to 61 %, but by only 10–20 % at 30–50 cm depth. The means of three measurements were used in the analyses.

Results

Illuminance was measured in the nest cavities of 20 Collared Flycatcher pairs. The median entrance area of those cavities was 21.2 cm2 and the median cavity depth was 13 cm (Table 1). Birds tended to nest deeper in cavities with larger and upward-facing openings (Fig. 2). The median distance to the nest in the latter (18 cm, n = 10) was larger than that in cavities with sideward openings (12 cm, n = 10, Mann–Whitney test, Z = −1.8, P = 0.070). All cavities in which nests were further than 18 cm from the entrance had upward-facing openings (Fig. 2).

Table 1 Measurements of Collared Flycatcher nest cavities in Białowieża National Park
Full size table
Fig. 2
figure 2

The relationship between entrance area and distance from the bottom of the opening to the nest rim in cavities of Collared Flycatchers. The orientation of opening of each cavity, i.e. sideward (diamonds) or upward (triangles), is also shown

Full size image

During the observation period (between June 4 and 15; Table 1), sunrise was at or around 04:00 local time (UTC + 2 h zone). The adults appeared at the nest cavity between 31 and 16 min before sunrise (no roosting females were found at any of the nest cavities), and started to feed their young from between 22 min before and 16 min after sunrise (Table 1). During cloudless weather (35 % of all observations), the birds usually arrived 8 min earlier (median = 29 min before sunrise) than on overcast days (Mann–Whitney test, Z = −2.2, P = 0.030). However, the birds started to feed their young only 4 min earlier on cloudless days (median = 12 min before sunrise) than on overcast days (Mann–Whitney test, Z = −1.0, P = 0.322). The time when feeding was commenced was not related to any of the measured cavity dimensions (r s = −0.30 to 0.28, P = 0.21–0.94).

Illuminance at the level of the nestlings’ heads was measured 1–13 min (median = 8 min) after the observation of the first feeding, when it varied between 0.01 and 0.4 lx (median = 0.05 lx). Usually only ca. 4 % of the incoming light (measured at the cavity entrance) reached the nest (Table 1). The nest illumination was not related to the cavity depth, nor to the entrance size (r s = −0.37 and 0.16, P = 0.113 and 0.501, respectively). The illuminance at the entrance of cavities with upward-facing openings (median = 2.04 lx, n = 10) was almost twice that at the entrances of cavities with sideward openings (median = 1.14 lx; Mann–Whitney test, Z = −1.6, P = 0.112). However, as the former cavities tended to be deeper (see above), the illuminance at nest level was the same in both types of nest cavities (median ca. 0.05 lx).

The illuminance in the artificial cavity measured at a sunny midday, when the ambient light reached its maximum level in the open (ca. 66,000 lx), declined with distance from the entrance (Fig. 1). It dropped from 1.3 lx at 10 cm below the entrance to 0.2 lx at 50 cm. Only ca. 0.002 % of the ambient light (measured in a forest gap) reached 10 cm below the entrance within cavities.

Discussion

The range of illuminance values found in Collared Flycatcher cavities at the level of the nest in this study (0.01–0.4 lx) is within the lower value of nest illuminance recorded in Great and Marsh Tits (0.03–3 lx; Wesołowski and Maziarz 2012). As we measured illuminance with some delay from the time of the first feeding (1–13 min, median = 8 min), the actual amount of light in the nest cavities at the time of the first feeding is likely to be lower than our data indicate; the amount of light reaching the nest was most probably further reduced because a fraction of the incoming light was blocked by the body of the parent entering the cavity. Therefore, it is likely that the illuminance of nests early in the morning could be lowered by approximately 0.01 lx, and remained below 0.1 lx for most of the day, similar to the amount of light reaching the nest in the tits’ cavities (Wesołowski and Maziarz 2012). Hence, these two groups of birds, despite evolving the cavity-nesting habit independently (von Haartman 1957) and occupying different tree cavities (Wesołowski 1996; Walankiewicz et al. 2007; Wesołowski and Maziarz 2012), appear to have converged on a similar lowest usable light level.

The range of light levels experienced by Collared Flycatchers in nest cavities early in the morning corresponded to the illumination that would occur outside the cavity during night time when illumination is provided by a new moon and up to a half-phase moon (under cloudless conditions; Martin 1983). Yet, at midday, the nest cavities were also dark, and still within the range of naturally occurring night conditions, although at the upper levels produced on open ground under a full moon in a clear sky (Martin 1983). This places the illuminance values recorded in the Collared Flycatcher nest cavities within the mesopic–scotopic range. In the mesopic range, colour vision is highly impaired or lost (Lind and Kelber 2009); in the scotopic range of light levels, colour vision does not occur (Martin 1983). Therefore, it is unlikely that flycatchers would be able to employ colour vision inside cavities. Our data on illuminance in the nest cavities of Collared Flycatchers and the tits (Wesołowski and Maziarz 2012) lend empirical support to critical reviews of hypotheses which propose that the colour of eggs and the gapes of nestlings have a signalling function in cavity-nesting birds (Reynolds et al. 2009; Wiebe and Slagsvold 2009; Holveck et al. 2010).

Because illuminance declines rapidly with depth inside the cavity (Fig. 1), it is possible that birds can control the amount of light reaching the nests by placing them at different distances below the entrance (i.e. by building up a nest to an appropriate depth). In BNP, Collared Flycatchers situated their nests relatively close to the cavity entrances (median depth = 13 cm; Table 1). This is within the reach of larger predators, such as Pine Martens Martes martes, which are unable to squeeze through the entrance and enter the cavity but can still insert a limb to reach the nest (Wesołowski 2002). By placing the nests only a few cm further below the entrance, the birds could lessen that threat substantially, and they could completely avoid this type of predation by situating them 20 cm below the entrance (Löhrl 1987; Wesołowski 2002). Also, locating broods further below the entrance could make it more difficult for a predator to detect by inspecting the cavity interior from the entrance—and this could further impair nest detection by predators (Wesołowski and Rowiński 2012). However, despite the clear, selective, anti-predator advantage of placing a nest further below the entrance, Collared Flycatchers rarely placed their nests far enough below the entrance to avoid the probability of predation (Fig. 2). This suggests that deeper placement of a nest within the cavity could be costly. The costs might include a higher risk of nest flooding (Wesołowski et al. 2002) and inadequate nest illumination for the care of young. The latter possibility is supported by our data showing that in cavities with smaller and sideward-facing openings, nests tended to be situated closer to the entrance, while all the nests placed at distances greater than 18 cm were found in cavities with larger and upward-facing openings (Fig. 2). This indicates that there is a trade off between safety and illumination: to achieve an adequate nest illumination, the birds put their nests closer to the entrance than would be predicted if the main driver of nest position was safety from predation (see also Löhrl 1977).

It is noteworthy that Collared Flycatchers started to feed their young before sunrise—about half an hour earlier and at about tenfold lower ambient illuminance than recorded in tits (Wesołowski and Maziarz 2012). This suggests that Collared Flycatchers could be under an additional selective pressure that forces them to accelerate the commencement of feeding their young each day. Foraging and feeding young at dawn requires an ability to locate food sources at low light levels (Kacelnik 1979). Hence, it is possible that Collared Flycatchers are more capable of hunting for insects at lower ambient light levels, and begin feeding young earlier than the tits.

Although the birds appeared at cavities later on cloudy mornings, they started to feed their young at a similar time as on cloudless mornings. There was a 40-min spread in the timing of the first feeding by the parents in different nest cavities; however, this did not seem to be due to variation in nest illuminance among nest cavities. Despite substantial variation in cavity dimensions, young were fed when the nests were illuminated to a similar extent. Thus, it seems reasonable to assume that the light level at which feeding commenced was close to the minimum illumination at which Collared Flycatchers could tend the young efficiently.

Apart from being able to see in dim light, the visual systems of diurnally active cavity nesters have to cope with the rapid changes of light levels that occur when entering and leaving the nest cavity (3–5 log units; Cassey 2009). Cavity-nesting birds can operate and feed their young when faced with almost instant changes in light conditions (from a bright environment outside to the darkness inside cavities, and reverse), but how they do this remains unclear. At least in part, the mechanism could involve the use of a highly dynamic entrance pupil (Martin 1999; Lind and Kelber 2009), or keeping the retina permanently dark-adapted by reducing the pupil to a very small aperture at high light levels and opening the pupil when entering the nest cavity (Reynolds et al. 2009). Nevertheless, the pupil constriction could compensate for the only 300-fold range—at most—of retinal illumination when the pupil is closed to a pinhole aperture, and it will probably be much less than this, as pinhole-aperture pupils have not been recorded in any species of passerine bird (Martin 1999). So, perhaps, there might be some other—as yet unknown—mechanisms of very rapid light/dark adaptation in cavity-nesting species. A thorough study aimed at finding and understanding these mechanisms would certainly be worthwhile.

Summing up

The results of this Collared Flycatcher study confirm previous observations carried out on Great and Marsh Tits (Wesołowski and Maziarz 2012): species occupying different tree cavities, whose cavity-nesting habit evolved independently from flycatchers. They provide greater evidence that tree cavities used by nesting birds can indeed be very dark places, in general. To use them, the birds have to be able to see in dim light, to switch back and forth rapidly between a dark cavity interior and well-lit ambient environment, and maintain adequate spatial resolution in both environments. Thus, some species might be prevented from nesting in cavities because of their inability to meet these requirements. Our data also demonstrate that, under many natural light levels, there is unlikely to be enough light in nest cavities to enable birds to use colour vision. Finally, they indicate that lighting requirements can affect nest-site use by cavity nesters. We conclude that nest cavities pose real sensory challenges for the bird species which use them. Our observations of two taxa of birds, in which cavity nesting evolved independently, yield similar results. Therefore, we tentatively propose that similar constraints with respect to light levels have limited the ways in which cavities are used by other diurnal birds breeding in enclosed spaces (e.g. tree cavities or burrows). Exploration of birds’ visual performance in relation to the light levels experienced in tree cavities, and the light level changes experienced on entering and leaving tree cavities, would provide important insights into physiological and behavioural aspects of the sensory ecology of birds.