The study of host—parasite interactions has been of primary interest for avian ecologists. However, bacteria have only recently received attention in this respect (Maul & Farris 2005, Benskin et al. 2009). Associations between birds and bacteria may involve pathogenic interactions, but also positive symbiotic interactions (Martín-Platero et al. 2006, Ruiz-Rodríguez et al. 2009). These interactions begin in the nest and may affect the growth and survival of altricial and semi-altricial nestlings (Potti et al. 2002, Moreno et al. 2003, Gonzâlez-Braojos et al. 2012). The early stage of the post-hatching period is critical for establishment of the gut microbial community. This process starts from a sterile gastrointestinal environment at the moment of hatching and continues towards establishing a relatively stable status as the nestling ages. Thus, Mills et al. (1999) reported that microorganisms colonize nestling cloacae shortly after hatching, suggesting the source of microbes to be adults, local food items or their local environment. Understanding the factors modulating bacterial abundances in nestling digestive tracts could improve our understanding of bird—bacteria interactions in the wild.

Several factors including climate, food, age and health state affect the composition of the gut microbiota of individual birds (Brittingham et al. 1988, Lombardo et al. 1996). Nutrient richness in the environment, humidity and temperature have been identified as important factors affecting growth in bacterial cultures (Ratkowsky et al. 1982, Madigan et al. 2006). Nutrient availability for digestive bacteria may change with seasonal variation in diet of both adults and nestling birds (Blanco et al. 2006, Novotny et al. 2007). Nutritional quality may also affect bacterial growth, especially for nestlings competing for parental food deliveries. Late-breeding pairs tend to offer less or poorer quality food to their nestlings (Naef-Daenzer et al. 2000, Rossmanith et al. 2007, Wilkin et al. 2009). Nestlings in larger broods may also suffer nutritionally from stronger competition with brood mates (Naguib et al. 2004, Pichorim & Monteiro 2008). Thus, breeding phenology and brood size could affect the growth of bacteria in digestive tracts through nestling nutrition in terms of quantity and quality. Gut mass declines in conditions of poor nutrition (Brzek & Konarzewski 2001), possibly driving higher competition among bacteria for space. To our knowledge, there is no published information about the associations of environmental factors with growth of bacteria in nestling digestive tracts. The only study, to our knowledge, relating gut bacterial growth to seasonal climate changes did not include nestlings (Janiga et al. 2007). Ambient temperature may affect bacterial growth through the thermoregulatory capacity of altricial nestlings, which improves with age until thermal independence from adult brooding behaviour (O'Connor 1984, Starck & Ricklefs 1998). If the body temperature of non-thermally independent offspring fluctuates more when the ambient temperature is low (Starck & Ricklefs 1998, Bize et al. 2007), bacterial growth might suffer accordingly. Thermally independent nestlings may offer more stable thermal regimes for gut bacteria. Rainfall may affect the foraging capacity of adults and thereby nestling nutritional condition (Rosa & Murphy 1994, Elliott et al. 2005, Spencer 2005, Geiser et al. 2008, Morrison et al. 2009, Arlettaz et al. 2010).

In this study we assessed whether environmental factors modulate the abundance of two types of gut bacteria (enterococci and Enterobacteriaceae) at two nestling ages in the Pied Flycatcher Ficedula hypoleuca. Enterococci are widely distributed in animal gastrointestinal tracts (Foulquié-Moreno et al. 2005) and may exist as commensal organisms of the alimentary tract of chickens (Klein 2003) and wild birds (e.g. Moreno et al. 2003). They have probiotic properties and are able to limit the colonization of the digestive tract by patho- genic bacteria (Mazur-Gonkowska et al. 2006). Moreno et al. (2003) found a significantly positive association between nestling mass shortly before fledging and the presence of Enterococcus faecium. Enterobacteriaceae are also common in the intestinal microflora of wild birds. Thus, Winsor et al. (1981) showed that the most prevalent intestinal bacteria of this group in Turkey Vultures Cathartes aura were Escherichia coli and Proteus mirabilis. Moreover, Enterobacteriaceae contribute to the digestion of food and play an important role in controlling other gut bacteria (Hudault et al. 2001, Reid et al. 2001). It has been shown that both enterococci and Enterobacteriaceae grow best at temperatures between 22 and 45°C (Ron 1975, Martínez et al. 2003, FoulquiéMoreno et al. 2005), so thermal fluctuations in thermally dependent nestlings during parental absences may affect bacterial growth conditions.

Accordingly, we hypothesized that: (1) later broods will have lower bacterial counts due to poor nestling nutrition, (2) larger broods will have lower bacterial counts due to poor nestling nutrition, (3) lower ambient temperatures will result in lower bacterial growth in non-thermally independent offspring (7 days) while having smaller effects in nestlings about to fledge (13 days), (4) higher rainfall will induce poorer bacterial growth due to restricted parental food deliveries to nestlings. Finally, (5) we looked for synergistic effects of rainfall and temperature in driving bacterial growth in guts of nestlings, as low nutrition may have especially strong effects when the costs of thermoregulation are high.

METHODS

The study was conducted during the 2009 breeding season in a deciduous forest of Pyrennean Oak Quercus pyrenaica at an elevation of 1200 m a.s.l. in Valsaín, Segovia province (40°54′N, 4°01′W), Spain. The local population of Pied Flycatchers breeds in nest-boxes and has been under study since 1991 (Sanz et al. 2003). Nest-boxes are cleaned every year after the breeding season. For the current study, nest-boxes were checked daily for nest-building activity, and the hatching dates and brood sizes were recorded.

The Pied Flycatcher is a small (12–13 g) passerine bird, which breeds in many forested areas of the Palaearctic region (Lundberg & Alatalo 1992). It only stays in European woodlands for the spring and summer, spending the rest of the year on migration or in the wintering areas in tropical West Africa. It breeds naturally in tree cavities, but if nest-boxes are provided, these are preferred over natural cavities. Egg-laying in the population under study typically begins in late May, and clutch sizes range from 4 to 7 eggs with a mode of 6 eggs (mean 5.5 ± 0.6). The female incubates alone and receives part of her food from her mate (Lundberg & Alatalo 1992). Young are brooded by the female only up to day 8 (hatching day = day 1) (Sanz & Moreno 1995). Both sexes feed the young. Young fledge within 14-16 days of hatching. This occurs in the second half of June in our study area (Moreno et al. 2001).

A sample of 54 broods of four to six chicks was used for this study. Of these nests, we obtained samples from two randomly selected chicks in 43 nests and one randomly selected chick in the remaining nests. Nestlings were measured and weighed at two ages (7 and 13 days).

RESULTS

On day 7 the loads of Enterobacteriaceae (r = 0.433; F_1,26 = 2.530; P = 0.011), but not of enterococci (r = 0.229; F_1,30 = 1.530; P = 0.126) were significantly correlated within broods. No significant within-brood similarity was found for loads on day 13 either for Enterobacteriaceae (r = 0.159; F_1,30 = 1.380; P = 0.192) or enterococci (r = 0.014; F_1,33 = 1.027; P = 0.468). Thus, we did not calculate intra-brood average bacterial counts.

For each bacterial type and nestling age, we present only the preferred model, as alternative models showed ΔAICc values higher than 2. The preferred model for Enterobacteriaceae on day 7 included hatching date, mean temperature, rainfall and the interactions between hatching date and rainfall and mean temperature and rainfall during the preceding period (Table 1). All variables and interactions included in the model with lowest AICc were significant (Table 2). There was a negative association of Enterobacteriaceae loads on day 7 with rainfall, but only for early broods (Figure 1). Likewise, rainfall showed a negative association with Enterobacteriaceae loads on day 7, but less so at low ambient temperatures (Table 2). The preferred model for Enterobacteriaceae on day 13 included only hatching date (Table 1), which showed a negative association with bacterial loads (Table 2, Figure 2).

Table 1.

Linear mixed models of bacterial loads (two bacterial types at two nestling ages) in relation to mean temperature in the preceding week (Temp), rainfall (Rain), hatching date (HD), brood size and their interactions. Nest was included as a random effect. Only the preferred models for each analysis are presented (no alternative models showed ΔAICc lower than 2). K = the number of parameters in each model; Weight = the probability that the model is the preferred model in the model set.

Table 2.

Parameter estimates of effects included in the preferred models (see Table 1) of Enterobacteriaceae and enterococci loads at two nestling ages. Effects are mean temperature in preceding week (Temp), rainfall (Rain), hatching date (HD) and interactions.

The preferred model for enterococci on day 7 included only mean temperature (Table 1), which showed a positive association with bacterial loads (Table 2). The model for enterococci on day 13 included only mean temperature (Table 1), which showed a positive association with bacterial loads (Table 2).

Figure 1.

Associations of Enterobacteriaceae loads on day 7 with rainfall for broods hatched before or after the median hatching date in the population. Lines represent the simple regression which is not corrected for other significant effects presented in Table 2. A single nestling per nest is presented to avoid pseudoreplication.

Figure 2.

Association of loads of Enterobacteriaceae on day 13 with hatching date. A single nestling per nest is presented to avoid pseudoreplication.

DISCUSSION

Our results showed that Enterobacteriaceae loads were lower in later-hatching nestlings. We found no association between brood size and loads of either bacterial type. Ambient temperature was positively correlated to Enterobacteriaceae loads, but only on day 7. Temperature was positively correlated with enterococci loads at two nestling ages. Higher rainfall resulted in lower Enterobacteriaceae loads on day 7.

Nestlings in late-hatched broods have fewer Enterobacteriaceae in their guts at both ages than earlyhatched nestlings. This effect may be due to seasonal changes in diet. For example, Waldenström et al. (2002) found that the prevalence of Campylobacter spp. in migrating birds was highly influenced by feeding habits. Also, Lombardo et al. (1996) suggested that different feeding habits might explain the greater prevalence of bacteria in insectivorous than in omnivorous birds. Blanco et al. (2006) found differences in composition, richness and prevalence of faecal microbiota associated with the diet of adult Red Kite Milvus milvus. Furthermore, they found that Klebsiella sp. showed a higher prevalence in January than in February, whereas Novotny et al. (2007) found that the occurrence of Yersinia enterocolitica in adult Alpine Accentors Prunella collaris was high in summer, especially during the nestling period. It is possible that latebreeding parents are less efficient at collecting prey, and that late nestlings may therefore be undernourished. Malnourished nestlings may have lower bacterial growth in their guts due to intestinal shrinkage (Brzek & Konarzewski 2001) or to lower nutrient input. However, the non-significant effect of brood size suggests that malnourishment induced by competition among siblings is less important than seasonal changes in diet. In contrast to Enterobacteriaceae, enterococci were not responsive to differences in hatching date, which suggests that they are relatively insensitive to nutritional effects.

As expected, loads of Enterobacteriaceae and enterococci of chicks at early nestling ages are positively associated with mean temperature, while in grown nestlings this association is only found for enterococci. At low ambient temperatures, poorly thermoregulating chicks (younger than 7 days) may lose relatively more heat, drop their body temperature and reduce their metabolism (Starck & Ricklefs 1998), thereby negatively affecting bacterial growth in the gut. Nestlings about to fledge maintain body temperature within a much smaller range than nestlings of 7 days (O'Connor 1984). The responsiveness of enterococci to temperature at late nestling ages indicates the extreme thermal sensitivity of these bacteria when compared with other gut bacteria.

Enterobacteriaceae loads showed a negative association with rainfall, but only in young nestlings. The negative effects of rainfall on day 7 on loads of Enterobacteriaceae could also be explained by poor thermoregulation as temperatures in the nest drop during rain showers. Rainfall also has a negative effect on the foraging capacity of adults (Radford et al. 2001, Geiser et al. 2008, Arlettaz et al. 2010). This could adversely affect nestling nutritional condition, so we would expect a negative effect of rainfall on the capacity to sustain large bacterial populations. By contrast, enterococci were not affected by rainfall at any nestling age. This suggests that enterococci are less responsive to rainfall-mediated nestling nutritional condition than Enterobacteriaceae.

Only early-hatched broods experienced nesting environments conductive to strong predicted effects of high rainfall on bacterial loads in nestling guts. In fact, loads of Enterobacteriaceae on day 7 in early-hatched broods, but not in late-hatched broods, showed a negative association with rainfall. This may be related to poor nestling nutrition due to low foraging capacity of adults during periods of high humidity. Finally, low temperatures and high rainfall may interact synergistically as predicted to induce thermoregulatory problems for small nestlings, thereby inducing reduced bacterial growth. In contrast, enterococci showed no response to rainfall at any ambient temperature.

To conclude, growth of important intestinal bacteria appears sensitive to seasonal and climatic factors, presumably mediated by nestling diet, thermoregulatory capacity and nutritional state. Different bacterial types vary in their responsiveness to environmental and seasonal variation.