Plumage coloration differs between offspring raised in natural cavities and nestboxes

ABSTRACT Most of our knowledge of secondary hole-nesting birds comes from populations breeding in human-provided nestboxes, yet these might differ from natural cavities in a number of parameters, including internal dimensions or microclimate, leading to differences in reproductive ecology. Here, we demonstrate differences in plumage coloration, an important visual signal of individual quality, in Blue Tit (Cyanistes caeruleus) and Great Tit (Parus major) nestlings raised in natural cavities and nestboxes. For this study, we collected feather samples over two breeding seasons and applied reflectance spectrophotometry to calculate color metrics. Carotenoid chroma was higher in nestbox-reared Blue Tits, whereas brightness was higher in nestbox-reared Great Tits (with a similar tendency for Blue Tits). The latter result might be explained by the better nutritional condition of Great Tit nestlings raised in nestboxes. Furthermore, we found no evidence for preference of adults expressing more elaborate plumage towards a specific cavity type in either species. Consequently, we assume that differences in nestling plumage reported here are driven by rearing conditions (nestboxes vs. natural cavities) and not by differences in plumage-based parental quality. Our study adds to the growing body of evidence confirming that anthropogenic environmental modifiers, such as nestboxes, might influence avian physiology and the resulting phenotype. LAY SUMMARY Natural cavities and nestboxes differ in many ways, including dimensions and microclimatic conditions, which can have consequences for predation risk and avian reproductive success. Currently, knowledge on plumage color of hole-nesters comes solely from studies on populations breeding in nestboxes or aviaries. So far, no study tested whether such results hold for birds breeding in natural cavities. To fill this gap, we examined the impact of rearing cavity type—natural cavities vs. artificial cavities (nestboxes) on carotenoid-based coloration of Blue Tit (Cyanistes caeruleus) and Great Tit (Parus major) nestlings. We found that BlueTit nestlings expressed plumage with more saturated color in nestboxes, whereas in GreatTits nestlings raised in nestboxes expressed brighter plumage. Our results add to the growing body of evidence that differences between natural cavities and nestboxes can bear consequences for nestling physiology and phenotype. RÉSUMÉ La plupart de nos connaissances sur les oiseaux cavicoles secondaires proviennent de populations se reproduisant dans des nichoirs artificiels, mais ceux-ci peuvent différer des cavités naturelles d'un certain nombre de paramètres, incluant les dimensions internes ou le microclimat, ce qui entraîne des différences dans l'écologie de la reproduction. Dans cette étude, nous démontrons qu'il existe des différences dans la coloration du plumage, un signal visuel important de la qualité individuelle, chez les oisillons de Cyanistes caeruleus et Parus major élevés dans des cavités naturelles et dans des nichoirs. Nous avons collecté des échantillons de plumes pendant deux saisons de reproduction et nous avons appliqué la spectrophotométrie de réflectance pour calculer les paramètres de la couleur. La saturation des couleurs des caroténoïdes était plus élevée chez les oisillons de C. caeruleus élevés dans des nichoirs, alors que la brillance était plus élevée chez les oisillons de P. major élevés dans des nichoirs (avec une tendance similaire pour C. caeruleus). Ce dernier résultat pourrait s'expliquer par le meilleur état nutritionnel des oisillons de P. major élevés dans des nichoirs. En outre, nous n'avons trouvé aucune preuve de la préférence des adultes au plumage plus élaboré pour un type de cavité spécifique chez l'une ou l'autre de ces espèces. Par conséquent, nous supposons que les différences de plumage chez les oisillons sont dues aux conditions d'élevage (nichoirs vs cavités naturelles) et non à des différences de qualité parentale basées sur le plumage. Notre étude s'ajoute au nombre croissant de preuves confirmant que les modificateurs environnementaux anthropogéniques, tels que les nichoirs, peuvent influencer la physiologie aviaire et le phénotype qui en résulte.


LAY SUMMARY
•Natural cavities and nestboxes differ in many ways, including dimensions and microclimatic conditions, which can have consequences for predation risk and avian reproductive success.•Currently, knowledge on plumage color of hole-nesters comes solely from studies on populations breeding in nestboxes or aviaries.So far, no study tested whether such results hold for birds breeding in natural cavities.To fill this gap, we examined the impact of rearing cavity typenatural cavities vs. artificial cavities (nestboxes) on carotenoid-based coloration of Blue Tit (Cyanistes caeruleus) and Great Tit (Parus major) nestlings.•We found that Blue Tit nestlings expressed plumage with more saturated color in nestboxes, whereas in Great Tits nestlings raised in nestboxes expressed brighter plumage.•Our results add to the growing body of evidence that differences between natural cavities and nestboxes can bear consequences for nestling physiology and phenotype.

INTRODUCTION
Anthropogenic activities can lead to the disappearance or deterioration of natural habitats (Chamberlain et al. 2009).The shortage of old forest stands, leading to a scarcity of natural cavities they offer, is often mitigated by providing nestboxes, readily used by some birds, such as Great Tits (Parus major) and Blue Tits (Cyanistes caeruleus) (Mänd et al. 2005).For researchers, nestbox provisioning can facilitate the acquisition of larger sample sizes (Wesołowski 2011, Maziarz et al. 2017, Sudyka et al. 2022).However, the reproductive ecology of birds using nestboxes differs from those using natural cavities in some essential parameters, including earlier breeding dates, accelerated onset of incubation (Fargallo et al. 2001, Czeszczewik 2004, Sudyka et al. 2022; but see Mitrus 2003), and altered gut microbiota composition (Maraci et al. 2022).These deviations may stem from lower humidity (except for woodcrete boxes in the presence of thermoregulating nestlings) and less stable microclimatic conditions in nestboxes when compared to natural nesting sites (i.e., higher daily average temperature amplitude and worse insulation properties; Maziarz et al. 2017, Sudyka et al. 2023).Relative to natural cavities, birds breeding in nestboxes may also experience higher ectoparasite infestation rates (Fargallo et al. 2001, Wesołowski andStańska 2001) but lower predation risk (Kuitunen and Aleknonis 1992, Purcell et al. 1997, Czeszczewik 2004, Maziarz et al. 2017).Consequently, conclusions stemming from studies of nestboxbreeding birds whose evolutionary history is associated with natural cavities may have limited relevance.The problem of extrapolating results from nestbox studies has been pointed out in the past (Møller 1989, Robertson and Rendell 1990, Lambrechts et al. 2010, Wesołowski 2011, Sudyka et al. 2023), which inspired comparative research on birds breeding in both artificial and natural breeding sites (Purcell et al. 1997, Czeszczewik 2004, Norris et al. 2018, Sudyka et al. 2022).However, a topic not being examined in this comparative context so far is offspring plumage coloration, which represents a significant knowledge gap, given its postulated signaling function and behavioral consequences (Morales andVelando 2018, García-Campa et al. 2023).
To our knowledge, all studies on coloration of hole-nesters were done on populations breeding in nestboxes or birds kept in aviaries.The signaling function of juvenile coloration is still debated, yet there is increasing evidence of its role in parent-offspring communication in both nesting and post-fledging periods.Carotenoid coloration is the result of the amount and composition of pigments deposited in the growing feather and the quality of keratin, which is the main feather building material (Shawkey and Hill 2005).For example, in Blue Tits, nestling body mass is positively related to multiple components of carotenoid-based breast coloration: UV chroma (Jacot and Kempenaers 2007, Morales and Velando 2018, García-Campa et al. 2023), total brightness (García-Campa et al. 2023), and carotenoid chroma (Johnsen et al. 2003).Consequently, plumage color might signal nestling quality and thus modulate parental provisioning effort.In line with this, Morales and Velando (2018) demonstrated that experimental reduction of the UV component of Blue Tit offspring coloration modified both nestling begging and paternal feeding behavior.Moreover, García-Campa et al. (2021) found that lutein-supplemented females fed more intensively UV-reduced nestlings compared to controls, which suggests that parental nutritional condition might impact the interpretation of offspring color signal.
To date, most work on juvenile coloration comes from nestbox populations of two parids, Blue Tits and Great Tits.They commonly breed in nestboxes, tolerate human presence, and their overall numbers in Europe are slowly increasing (BirdLife International, 2024), which makes them important model species to study life-history traits and plumage coloration.The yellow coloration of Blue Tit and Great Tit offspring breast, belly, and cheeks is based on two carotenoids-lutein and zeaxanthin (Partali et al. 1987, Isaksson et al. 2008).Because carotenoids cannot be synthesized de novo by birds (Olson and Owens 1998), their main sources for nestlings are maternal deposition in egg yolk and carotenoid-rich food provided by the parents (Fitze et al. 2003a, Biard et al. 2005, Isaksson et al. 2006).The amount of pigment received from parents is modulated by environmental carotenoid availability, clutch size, and parental provisioning effort.The main source of carotenoids in the diet of Blue and Great tit nestlings are caterpillars, but their abundance changes non-linearly throughout the breeding season.Thus, it is crucial to synchronize nestling growth period with the peak of food availability (Wesołowski and Rowiński 2014).In particular, carotenoid-rich food sources within the first 6 days after hatching are crucial for color expression (Fitze et al. 2003b).In line with this, laying date may correlate with nestling coloration in several European parid populations (Slagsvold and Lifjeld 1985, Arriero and Fargallo 2006, Pagani-Núñez et al. 2014, Janas et al. 2020; but see Biard et al. 2017).Furthermore, the amount of food provisioned per nestling might be related to color phenotypes.For example, in Great Tits, nestlings raised in experimentally reduced broods expressed more chromatic (Tschirren et al. 2003) or brighter (Jacot et al. 2010, Matrková andRemeš 2012) breast plumage coloration.Similarly, Blue Tits from reduced broods developed plumage with higher carotenoid and UV chroma (Jacot and Kempenaers 2007).
In adults, carotenoid-based coloration is thought to be a condition-dependent trait (Griggio et al. 2009, Doutrelant et al. 2012).For example, in Great Tits, breast plumage coloration was positively related to body condition (Galván 2010, Hegyi et al. 2015; but see Isaksson et al. 2008) and feather growth rates (Senar et al. 2003).Moreover, in Blue Tits, breast plumage coloration was negatively related to molt rate (Ferns andHinsley 2008, Griggio et al. 2009) and haemosporidian infection (del Cerro et al. 2010).As honest signal, carotenoidbased coloration might be a good predictor of parental quality as it was reported to correlate with nestling provisioning rates and the proportion of fledged young (García-Navas et al. 2012).For example, in Blue Tit immune-challenged females, breast plumage coloration was positively related to the allocation of carotenoids to egg yolk (Midamegbe et al. 2013).Furthermore, Doutrelant et al. (2008) found significant positive relationships between female breast plumage coloration, clutch size, and fledging success.In addition, saturation (i.e., chroma) of feathers of Great Tit fathers was positively related to that of their fostered nestlings, which underline the role of paternal feeding in the first period after hatching (Fitze et al. 2003b).Taken together, parental quality, signaled by carotenoid-based traits, should be taken into account when analyzing sources of variation in nestling plumage.
Here, we examined the association between the type of cavity nestlings were reared in-natural cavities vs. nestboxes-and offspring plumage coloration, specifically carotenoid chroma and brightness of Blue Tit and Great Tit nestling breast feathers.Carotenoid chroma reflects the amount of pigment deposited in the feather (Andersson and Prager 2006), whereas brightness is related to the quality of the reflective keratin structure and might depend on the degree of feather abrasion and the trace amounts of melanin (Shawkey andHill 2005, Isaksson et al. 2008).Here, we used feather coloration data from Blue Tits and Great Tits breeding over two years in an ecologically homogeneous old-growth forest.One plot was supplied with nestboxes, whereas the other, closely located, offered only natural cavities as breeding sites.This setup allowed us to test for consequences of breeding in human-provided nests.In the same study site, Sudyka et al. (2022) demonstrated that fledgling number and fledging success were lower for Blue Tits breeding in nestboxes, an effect absent in Great Tits.Based on this differential sensitivity to cavity type, we hypothesized that nestling plumage coloration might depend on rearing cavity type in a speciesspecific manner.Specifically, if carotenoid-based coloration is condition-dependent, we predicted that Blue Tit nestlings would develop feathers with reduced carotenoid chroma and brightness in nestboxes compared to natural cavities.In contrast in Great Tits, since these tend to have higher body mass in nestboxes than in natural cavities (Sudyka et al. 2022), we predicted an increase of brightness and carotenoid chroma metrics in nestboxes.To ensure that potential differences in nestling color metrics were not simply a by-product of parental preference and competition for a particular cavity type, we also examined adult plumage coloration (breast carotenoid chroma and brightness, plus crown brightness and UV chroma in Blue Tits) as a proxy of parental quality in relation to the cavity type where they bred.

Study Site and Sampling
The study was conducted during two consecutive field seasons (2018)(2019) in Bielany Forest (~150 ha), situated in the northern part of the city of Warsaw.The area is a remnant of the Mazovian primeval forest and due to its natural value, it is protected under the Natura 2000 network (PLH140041, Special Area of Conservation, 129.84 ha) and is under reserve protection as the Nature Reserve Bielany Forest (130.35 ha).We collected data within two study plots, hereafter referred to as "natural cavity" and "nestbox" plots, with edges separated by a minimum of 200 m.Such study design was to avoid a non-random cavity type choice by parents due to the interand intra-specific competition for a particular cavity type (Lõhmus and Remm 2005).Alternatively, we would risk that different subsets of birds would breed in nest boxes and natural cavities if these were offered on the same plot (to avoid this bias we excluded from the current analysis the few nests in natural cavities that were located in the nestbox plot).As a result, it would be impossible to disentangle between parental quality and the impact of nest type.Importantly, other similar studies applied spatial separation of nest boxes and natural cavities areas for these exact reasons (Mitrus 2003, Czeszczewik 2004, Norris et al. 2018).Nevertheless, having two separate plots might be associated with some differences in habitat quality.To specifically confirm lack of this potential confounder, Sudyka et al. (2022) found no difference in food availability between study plots in both field seasons, assessed by the widely used method of frass fall collection (Wesołowski and Rowiński 2014).Ambient temperature and humidity, as well as anthropogenic factors (noise and air pollution, measured as parts per million [PM] 2.5 concentration) were found to be uniform between study plots (Sudyka et al. 2022).Moreover, in terms of phytosociology, both plots are an oak-hornbeam forest (Tilio-Carpinetum), with some of the oak stands reaching 300-400 years, with a drier variant of Tilio-Carpinetum typicum covering the whole nestbox site and half of the natural cavity site and a fertile variant of Tilio-Carpinetum dominating the north-eastern part of the natural cavity plot (Pawłat-Zawrzykraj et al. 2021).Overall, due to the small spatial scale, our study site encompassing both plots offered a homogenous habitat that was uniformly used by the tits as confirmed by the similar breeding densities of both species (Di Lecce et al. 2023).The natural cavity plot (52°17ʹ32.2″N,20°57ʹ34.0″E)consisted of 50 ha of monitored area (with 30 ha of core area with most intense nest searches) and was characterized by a very high availability of tree hollows, as only ~21% of cavities occupied in 2019 were used in previous breeding seasons.The nestbox plot (52°17ʹ52.1″N,20°57ʹ09.7″E)had an area of 15 ha and consisted of 65 nestboxes (woodcrete Schwegler 1b, Supplementary Material Figure 1) hung every 50 m, at the mean height of 2.91 m.Schwegler nestboxes have an entrance hole of 32 mm of diameter, 12 cm of internal diameter, and were 24 cm high (Sudyka et al. 2023).For details on differences in microclimatic conditions between natural cavities and nestboxes, refer to Sudyka et al. (2023).
Weekly nestbox monitoring and nest searches in the natural cavity plot started at the beginning of April in both studied seasons.Natural cavity searches were carried out daily for the first 2-3 weeks by 3-7 experienced field assistants using direct observations.After this period, they were continued at lower intensity throughout the season to detect nests built by the parents that lost their first brood (for more details, see the supplementary information in Sudyka et al. 2022).Nests were monitored to record laying date, clutch size, incubation start, hatching date, number of fledglings, and fledging date.On the 14th day after hatching (hatching day = day 0), we measured nestlings body mass and tarsus length and plucked ~12 feathers from both ventral feather tracts at the shoulder joint height.Feathers were immediately stuck in one bundle to a black sheet of paper covered with double adhesive tape, ensuring that the background did not show through from underneath, and transported to the lab for measurements.Sex of nestlings was determined using a machine-learning population assignment approach, using single-nucleotide polymorphisms (SNPs) that differed between known females and males (Foll andGaggiotti 2008, Trenkel et al. 2020), as part of another project conducted in Bielany Forest (Di Lecce et al. 2023 for details).Adults were caught with mist nets or custom-made ratchet traps ~14th day after nestlings hatching.The traps were installed inside the nestbox door and closed when an entering bird moved a wire located a few centimeters below the entrance.Age of adults, categorized as birds in their second calendar year (i.e., "first year breeding bird") or after second calendar year (later referred to as "older"), was assigned according to the molt limit between the primary and greater coverts, while sex was discriminated based on the presence of the brood patch, as eggs are incubated only by females in both Blue Tits and Great Tits (Svensson 1994).We measured body mass, tarsus length, and wing length as well as took feather samples: crown and breast feathers from Blue Tits and breast feathers from Great Tits.Feather samples were stuck to and preserved on a black paper covered with a double adhesive, transparent tape, similarly to feathers from nestlings.

Plumage Reflectance Measurements
We measured reflectance in laboratory conditions, using a spectrophotometer USB400 (Ocean Optics, Orlando, FL) with 300-700 nm range, a xenon pulsed light source and a probe (7 × 400 μm) with a 2 mm high collar made from opaque photographic film.Before the onset of measurements and every 20 min, we took a reference scan of the white standard (Spectralon).From each sample we took 10 reads with the probe held perpendicular to the feather surface, as in Janas et al. (2018).Raw spectra were averaged and processed in the pavo package (Maia et al. 2019) in R. The raw reflectance was measured with 0.201 nm bin width, but it was interpolated to 1 nm bins, while converted into rspec object in the pavo.We calculated total brightness, quantified as a sum of reflectance values over all wavelengths (hereafter brightness), and two measures of chroma: carotenoid chroma and UV chroma (Montgomerie 2009).Carotenoid chroma is the relative reflectance in the region of the spectrum with maximum carotenoid absorbance (R λ450 ), thus it constitutes the best approximation of carotenoid content in the feathers (Örnborg 2002(Örnborg , Peters et al. 2004)).UV chroma was calculated as a total reflectance in between 320 and 400 nm divided by total reflectance (brightness).Carotenoid chroma and UV chroma of feathers with carotenoid-based coloration (i.e., breast feathers) are strongly negatively correlated, and both are independent of brightness (Supplementary Material Figure 2).To characterize crown coloration of adult Blue Tits, we calculated brightness and UV chroma.

Statistical Analysis
To examine the sources of variation in nestling body mass, we applied linear mixed effect models with body mass as response variable, cavity type, laying date (expressed as the number of days from April 1), clutch size, nestlings' sex and year fitted as predictors and a random term of nest ID (a unique nest ID with a year code).Initial models tested for interactions between cavity type and all fixed factors, which were sequentially removed from the model if not significant.
To test for differences in color parameters of Blue Tit and Great Tit nestlings raised in natural cavities and nestboxes, we fitted general linear mixed models (using the lmer function, package lmerTest, version 3.1-3; Kuznetsova et al. 2017), separately for each species, with brightness and carotenoid chroma as response variables in separate models.To account for the impact of common rearing environment and for the cases when nestlings were weighted on other than the 14th day, we fitted a unique nest ID (nest ID with a year code, hereafter: nest ID) and Time shift D14, respectively, as random terms.Full models included the focal cavity type (natural cavities or nestboxes), parental color metrics (to quantify parentoffspring correlations), parental age (categorized as 2nd year of life and older; Svensson 1994), as adults plumage shows age dichromatism (Evans et al. 2010), laying date (to account for seasonal variability in carotenoid availability; Slagsvold and Lifjeld 1985), clutch size (to account for between-siblings competition for food), nestlings sex (coloration is sex-specific early in life; Johnsen et al. 2003), nestlings mass (to assess if coloration is individual condition-dependent; García-Campa et al. 2023), and year of study (on our study site there were substantial differences between years in weather and food availability, with 2019 being a harsh year; Sudyka et al. 2022).We also fitted two-way interactions of all these factors with cavity type to see how rearing environment modulates links of these factors with coloration metrics.In the final models, we always retained the factors of largest biological relevance and in our study focus: cavity type, parental color metrics, nestling sex, and year of study.Parental age, laying date, clutch size, and nestling body mass (together with random term of Time shift D14) were removed from the models in a backward elimination procedure if their importance was not statistically supported (Quinn and Keough 2022; Supplementary Material Table 3 for results of models before reduction).To check the model's assumptions and to detect potential influential data points we assessed the diagnostic plots: residuals vs fitted values and influence plots with studentized residuals plotted against hat values and found no major deviations from the normal distribution.We controlled for multicollinearity of predictors in final models by checking the variance inflation factor (VIF), calculated using the function vif in the package metafor, according to the formula: following (Zuur et al. 2010).In all cases, it was lower than the threshold value of 2, adopted in this study, so we could retain all factors of interest in the models.All color variables (together with clutch size and laying date) were scaled to zero mean and unit standard deviation.Backward elimination in model selection was used here to specifically test our main hypothesis, that is plumage differences between the two types of cavities.To provide data driven support for our results we used the package MuMIn v. 1.47.5 (Bartoń, 2023) for model selection based on Akaike information criterion (AIC c ).We generated 4 TABLE 1. Models testing the effects of cavity type (natural cavity set as reference value), parental plumage characteristics (breast plumage brightness and carotenoid chroma), nestling sex and year of study on plumage color traits of Blue Tit and Great Tit nestlings.All color variables, clutch size and laying date (expressed as number of days from April 1) were scaled to zero mean and unit SD.Due to high model Akaike uncertainty (all Akaike weights [w i ] < 0.8), we averaged all models with cumulative w i < 0.95 (Symonds and Moussalli 2011).We report full averaged estimates, standard errors (SE) and confidence interval (CI) in Supplementary Material Table 4.The main results of averaged models concerning differences in plumage coloration in relation to cavity type corroborate those obtained using the backward elimination method (Table 1).
To examine potential differences in color traits of adult individuals breeding in natural cavities and nestboxes, we fitted separate general linear mixed models, treating brightness and carotenoid chroma of breast feathers and brightness and UV chroma for Blue Tit crown as response variables.The models included the categorical variables of cavity type (natural cavity or nestbox), sex, age (individuals in their second year or older), and study year (2018 or 2019) alongside the random term of nest ID.In the initial models, we tested for 3-way interactions between cavity type, sex and age; 2-way interactions between cavity type and year, sex and age; as well as interactions between age and sex.Whenever nonsignificant (P > 0.05), interactions were sequentially removed from the models.
To test if differences in coloration are noticeable for birds and thus might have consequences for parents-offspring signaling, we applied the receptor-noise limited discrimination model (Vorobyev andOsorio 1998, Vorobyev et al. 1998).Specifically, we assessed values of Just Noticeable Differences (hereafter JNDs) for both chromatic (ΔS) and achromatic contrasts (ΔL); for detailed description of methods and results, Supplementary Material.
All statistical analysis were performed in R (R Core Team 2021).

RESULTS
In total, we analyzed 467 Blue Tit nestlings, with feather samples available for 453 individuals, from 32 natural cavities and 32 nestboxes, and 410 Great Tit nestlings, with feather samples available for 390 individuals, from 32 natural cavities and 26 nestboxes (Supplementary Material Table 1).Great Tit nestling body mass showed a non-significant tendency to be higher in nestboxes (estimate = 0.62, P = 0.1; Supplementary Material Table 2 and Figure 3), was higher in males (estimate = 0.49, P < 0.001) and negatively related to clutch size (estimate = -0.48,P = 0.01) (Supplementary Material Table 2).In Blue Tits, nestling body mass was higher in males (estimate = 0.57, P < 0.001) and there was a statistically significant interaction between cavity type and year, with birds raised in natural cavities being heavier in 2019 (estimate = -0.73,P = 0.04; Supplementary Material Table 2 and Figure 3).
We found significant differences in nestling plumage coloration depending on whether they developed in nestboxes or in natural cavities.Carotenoid chroma was higher in Blue Tit nestlings reared in nestboxes relative to those reared in natural cavities (Table 1, Figure 1B), and was negatively related to laying date and clutch size (Table 1, Figure 2).There was also a significant positive relationship between nestling and father carotenoid chroma (Table 1).Carotenoid chroma, but not brightness, was also significantly higher in male nestlings (Table 1).
In Great Tit nestlings, brightness was related to cavity type, with nestlings from nestboxes being brighter than those reared in natural cavities.An interaction between cavity type and fathers' brightness was significant, but driven mainly by the negative association between father-offspring brightness in natural cavities (Table 1, Figure 3).Brightness was also negatively associated with laying date (Table 1).In contrast to Blue Tits, the expression of carotenoid chroma did not differ between Great Tit nestlings raised in natural cavities and nestboxes (Table 1, Figure 1D) and showed a positive association with clutch size.
We found no differences in adult Blue Tit crown brightness, breast plumage brightness, and carotenoid chroma when breeding in natural cavities and nestboxes (Table 2, Supplementary Material Figure 4).Notably, there was a significant interaction between cavity type and age in adult Blue Tits crown UV chroma, indicating that in natural cavities older individuals tended to have higher UV chroma than first year parents, whereas in nestboxes there were no age-related differences (Table 2).Similarly, parental Great Tit breast plumage brightness did not differ depending on cavity type (Supplementary Material Figure 5), but there was a significant interaction between cavity type and year of study for carotenoid chroma (Table 2).The visual JNDs values for both chromatic (ΔS) and achromatic contrasts (ΔL) were all lower than the threshold (JND > 1), indicating that the observed differences in plumage coloration of nestlings may not perceivable by the birds (Supplementary Material Figure 6).

DISCUSSION
This study revealed that the type of cavity where nestlings were raised-nestboxes or natural cavities-influenced juvenile plumage coloration, although the underlying patterns of color variation associated with cavity type are complex.In Blue Tits, we showed that carotenoid chroma was higher in nestbox-reared nestlings than in natural cavity-reared ones.In Great Tits, breast plumage brightness was higher in nestlings from nestboxes relative to those from natural cavities.Importantly, in Great Tits reared in natural cavities, there was a negative relationship between plumage brightness of nestlings and fathers, while an analogous relationship was absent in nestboxes.Although there is a reported general preference of Great Tits towards nestboxes over natural cavities (Lõhmus and Remm 2005), we found no evidence for preference of more ornamented adults toward a specific type of cavity in either species (possibly due to the study design adopting spatial separation of cavity types and thereby allowing to avoid this confounding effect).Consequently, we assume that differences in nestling plumage coloration according to rearing cavity type reported here are not simply driven by differences in plumage-related parental quality.We also demonstrated negative correlations of all color traits (except brightness in Great Tit nestlings) with laying date, which indicates that environmental carotenoid availability is indeed impaired with the season progression.
Contrary to our expectations, we found carotenoid chroma of Blue Tit nestlings raised in nestboxes was higher than   1).Dark green points represent nestlings from natural cavities and grey points denote nestlings raised in nestboxes.Plot shows raw data points with regression lines and grey shades indicate 95% CIs.Interaction between clutch size and cavity type was nonsignificant but we present it to facilitate discussing the underlying causes of the detected differences in nestling carotenoid chroma.TABLE 2. Summary statistics of linear mixed effect models examining the relationship between adult Blue Tit and Great Tit plumage characteristics and cavity type (natural cavity was set as reference value to the nestbox effect), year of study, age (first year breeding bird or older), and sex.Interactions between cavity type and other main factors were kept only if significant.All color variables were scaled to zero mean and unit SD.  1) and clutch size (Figure 2).The negative relationship between carotenoid chroma and clutch size is in line with higher chroma found in Blue Tits from experimentally reduced broods (Jacot and Kempenaers 2007).It suggests that while caring for smaller broods, Blue Tit parents may provide more carotenoid-rich food to each nestling.However, in our study, clutch size tended to be higher in nestboxes (the effect was significant only in 2018), and there was no interaction between clutch size and cavity type, which rules out clutch size as the cause.Although carotenoid chroma was negatively related to laying date and the start of incubation was accelerated in nestboxes, hatching date did not differ in relation to cavity type in both Blue Tits and Great Tits (Sudyka et al. 2022).Moreover, as nestling body mass did not vary in relation to cavity type, higher chroma in nestbox-reared birds cannot be explained by differences in nutritional condition (Supplementary Material Table 3 and Figure 3).It is thus possible that our result may be driven by differences in incubation conditions possibly due to varying microclimate between the two cavity types (Sudyka et al. 2023), affecting the processing of carotenoids deposited by mothers to egg yolk.Other processes occurring later in nestling development might also explain the detected difference.We can speculate that, for example, carotenoid metabolic processes might depend on nest microclimate or individual microbiomes, which have been shown to differ between the two cavity types (Maraci et al. 2022, Sudyka et al. 2023).Indeed, post-hatching nesting environments resulted in fewer Blue Tit nestlings fledging from nestboxes compared to natural cavities on our study site (Sudyka et al. 2022).
In Great Tit nestlings, plumage brightness, but not carotenoid chroma, was associated with cavity type (Table 1).Importantly, there was an interaction between cavity type and paternal plumage brightness (Table 1), with a significant negative relationship between nestlings and paternal brightness in natural cavities and no such relation in nestboxes (Figure 3).This negative association between father-offspring brightness in natural cavities (Figure 3) possibly stems from higher variance in plumage brightness of nestlings raised in natural nesting holes.Given that conditions in natural cavities are characterized by a much greater among-nest range of variability than in nestboxes, this pattern supports conditiondependence of feather brightness.In carotenoid-based plumage, brightness is considered to reflect the structural component (i.e., the quality of feather microstructure and its building material keratin; Shawkey andHill 2005, Jacot et al. 2010).Although two studies found no connection between Great Tit breast feathers microstructure (e.g., quantified as density of barbs and barbules, barb cortex and pith area) and brightness (Galván 2011, Gamero et al. 2015), the key role is likely to be played by the quality of keratin, determining its reflective properties.In line with this, a recent study by García-Campa et al. (2023) showed that at both within and between-nest level brightness is related to nestling body mass, further corroborating its condition-dependence.Although in our study brightness was not directly associated with nestling body mass (see Supplementary Material Figure 7), the latter tended to be higher in Great Tits reared in nestboxes vs. natural cavities, suggesting a potential role of nutritional condition.
In contrast to Blue Tits, in Great Tits neither carotenoid chroma nor parameters of reproductive success (Sudyka et al. 2022) differed between birds reared in natural cavities and nestboxes.We presume that this species-specific sensitivity to nesting conditions may stem from the disparities in reproductive ecology (e.g., the fact that Blue Tits, having lower body size than Great Tits, take care of larger broods; Dhondt et al. 1990).This potentially may entail higher within-brood competition that includes carotenoid-rich food resources or larger sensitivity to the unstable microclimate inherent to nestboxes (Sudyka et al. 2022).Overall, these findings add to the species-specific pattern showing consequences of nestbox breeding mostly in Blue Tits and not in Great Tits (Sudyka et al. 2022a, Di Lecce et al 2023).The variation detected in reflectance-based metrics is important in the context of physiological responses to developmental conditions that produce varying outcomes on nestling phenotype between the two types of nesting cavities.However, this variation does not appear to have clear signaling consequences (as quantified by the JND analysis; Supplementary Material Figure 7).
To conclude, our results add to the growing body of evidence highlighting the fact that differences between natural cavities and nestboxes can influence nestling phenotypic traits such as feather coloration.In our study, this is manifested in terms of higher carotenoid chroma in Blue Tit nestlings and higher brightness in Great Tit nestlings reared in nestboxes when compared to those reared in natural cavities.As such, the detected differences in plumage coloration between cavity types are important indicators of (anthropogenic) environmental modifiers to avian physiology.We acknowledge that due to the difficulty of sampling in natural cavities, the research conducted on birds breeding in nestboxes will remain the main source of knowledge on the reproductive biology of secondary cavity nesters.Yet, given the differences in nestling coloration reported in offspring reared in nestboxes as opposed to natural cavities, we advocate for further research on the constraints imposed by artificial nestboxes to avian biology, with particular focus on the long-term physiological consequences of these differences in terms of survival and future reproductive success.

FIGURE 1 .
FIGURE 1. Differences in coloration of Blue Tit (A, B) and Great Tit (C, D) nestlings.Plotted on raw data, horizontal bars denote data median, box edges represent 25%-75% quartiles, whiskers indicate 1.5 IQR and points above and below whiskers denote values outside of 1.5 IQR values.Analyses included random term of nest ID to avoid pseudo-replication.Asterisks indicate significance levels for "cavity type" factor in models presented in Table 1.*P < 0.05, **P < 0.01.
Table shows estimates, standard errors (SE), degrees of freedom (df), t-values, and P-values.Significant (P < 0.05) associations are marked in bold.P ≤ 0.1 is given in italics.*P < 0.05, **P < 0.01, ***P < 0.001.brightness and carotenoid chroma of Blue Tit and Great Tit nestlings fitted as response variables.Each global model included cavity type, parental plumage characteristics (breast feathers plumage brightness and carotenoid chroma), parental age (first year breeding bird or older), nestling sex, body mass, and year of study.The factor cavity type, as focal to our study, was kept in all models.

Table 1 .
*P < 0.05, **P < 0.01.Relationship between nestling and paternal brightness (Table1).Dark green points and line represent nestlings from natural cavities and grey points and line denote nestlings raised in nestboxes.Plot shows raw data points with regression lines and grey shades indicate 95% CIs.Relationship between carotenoid chroma of Blue Tit nestlings and clutch size (Table Table shows estimates, standard errors (SE), degrees of freedom (df), t-values, and P-values.Significant associations (P < 0.05) are marked in bold.P ≤ 0.1 are given in italics, *P < 0.05, **P < 0.01, ***P < 0.001.