Parasites (e.g. bacteria, ectoparasites, viruses) can markedly reduce host fitness by continually draining resources [1]. Thus, hosts are favoured by selection to develop physiological, behavioural or immunological responses to combat parasite pressures [2]. Parasites can also be transmitted to host offspring, and selection should favour maternal responses that extend protection against parasites to eggs, embryos or chicks [1]. One line of maternal defence involves the deposition of non-genetic resources into eggs (e.g. antibodies, nutrients), where they influence strongly offspring health, growth, survival and phenotype [3]. However, as the deposition of these egg resources can be costly, females are faced with trading-off offspring quality against their own survival and reproductive prospects throughout life [4].

In birds, females deposit immune factors in the yolk and albumen to protect embryos with poorly developed immune systems from harmful pathogens and to enhance their development [5], [6], [7], [8], [9]. For example, lysozyme in egg albumen is a major component of maternal innate immunity in birds [10], [11]. This family of enzymes acts by digesting peptidoglycans of bacterial cell walls and thus protecting the embryo from harmful bacteria that penetrate the eggshell [11], [12]. In addition, females allocate varying amounts of carotenoids into yolk, which can quench free radicals produced during early growth and allow proper functioning of the embryonic immune system [13], [14], [8]. Carotenoids also can boost maternal health status [15]. As these compounds cannot be manufactured de novo by birds and must be extracted from the diet [16], mothers face allocation trade-offs between self-maintenance and deposition in eggs.

Another critical trait by which females influence offspring survival is egg size. Several studies demonstrate that egg size is related to hatching success [17], [18], early nestling survival [19], nestling condition [20] or fledgling survival [21]. For example, in mallards (Anas platyrhynchos), young hatched from larger eggs are better able to survive during the few days immediately after hatching [19], [22], [23].

Among the pathogens to which eggs or newly hatched animals may be exposed, there is the community of bacteria on the feathers of parents [24], [25], [26] or in the nests [27], [28] that can be transmitted to chicks or eggs by contact. These bacteria can influence bird metabolism and immune status, since during maintenance behaviour, birds ingest feather-associated bacteria that are then found in their digestive tracts [29]. Such maintenance behaviour includes preening (or direct application of preen oil to feathers using the bill), which is thought to constitute a first line of defence against parasites [30], [31], [32], [33]. Preen oil (also known as urogygial gland secretions) is composed of a mixture of aliphatic monoester waxes, fatty acids and monohydroxy wax-alcohols [34], and there are at least three modes by which preen oil could influence feather-degrading bacteria (FDB; see [26]). First, preen oil may simply form a physical barrier that prevents FDB from reaching the feather surface [33]. Second, chemicals in preen oil could have anti-microbial actions [32], [35]. Third, antibiotic-producing symbiotic bacteria could be cultivated within the uropygial gland and preened onto feathers [36]. In preliminary studies, we found that deprivation of preen gland access in mallards led to a significant change in the structure of plumage bacterial communities (Giraudeau et al. unpublished data). In addition, we found that surgical removal of the preen gland in adult house sparrows (Passer domesticus) led to an increase in the abundance of non-feather degrading bacteria on the plumage (Czirjak et al. unpublished data). These results show that bacterial communities change when birds do not have access to their preen gland.

Despite the potential importance of preen oil for limiting feather-associated bacterial growth, the effect of preening behavior on immune status and breeding decisions has been seldom considered to date. The goal of this study was thus to examine, for the first time, how preening behavior (and its expected disease-protecting effects) during the reproductive period influences female immunoresponsiveness and parental investment decisions at the level of the egg. We used an experimental approach by blocking preen gland access [37] in one group of captive female mallards, while another group acted as unmanipulated, control animals. An assumption of our study is that having restricted access to the preen gland elevated maternal exposure to and accumulation of environmental microorganisms [32] and altered female immunoresponsiveness. We measured lysozyme concentration in maternal plasma to assess if our treatment induced an innate constitutive immune response (i.e. higher lysozyme concentration [38], [39], [11]). In addition, we measured egg size, clutch size, yolk carotenoid concentration and albumen lysozyme concentration as indices of maternal investment. Because we also expected that chicks raised by mothers with no access to their preen gland would also suffer increased microbial exposure, we predicted that experimental females would lay eggs with higher carotenoid and lysozyme concentrations in order to increase embryo immunological protection [40]. In addition, we expected that experimental females would produce lighter eggs compared to control females due to the energetic costs of mounting an immune response [41]. Finally, differential allocation of lysozyme along the laying sequence was recently shown in the yellow-legged gull (Larus michahellis, [42]). The authors proposed that females deposit higher concentrations of lysozyme in the last egg because transmission of bacteria between eggs and from the mother may be more intense for the last egg. Thus, we predicted an increase of albumen lysozyme concentration with laying order.

We found that blocking preen gland access in breeding female mallards modified several aspects of their condition and health. First, our treatment increased maternal lysozyme concentration in plasma and thereby constitutive immune status. Lysozymes are a major component of innate antibacterial immunity [10], [11], acting by digesting peptidoglycans of bacterial cell walls [11]. Thus, higher plasma levels of this antimicrobial enzyme in experimental females is consistent with the fact that they were exposed to more bacteria during egg laying [38], [44], [11]. As preen oil constitutes the first line of defence against potential environmental plumage parasites and particularly feather-associated bacteria [30], [32], [33], we argue that experimental females faced higher densities of detrimental bacteria on their plumage and probably ingested them while grooming [29]. In a previous experiment on mice, inoculation of pathogenic bacteria (flagellin from Salmonella enterica serovar Enteritidis) led to an elevation of plasma lysozyme [44]. For the moment, these kinds of experiment seem to be lacking in birds [11] and our study provides, for the first time, information about plasma lysozyme concentration in birds subjected to a microbe-rich environment.

We also found that experimental females lost more body mass and were in worse body condition during the study than controls females. Bacterial exposure and mounting an immune response may have significant effects on energy intake and metabolism, such as an increased utilization of glucose by peripheral tissues [45]. Consistent with our results, previous studies in poultry and mammals have shown that stimulation of the immune system with LPS results in an acute reduction of body weight gain and feed intake [46], [47], [48], [41]. An alternative hypothesis could be that females with the device that did not allow them access to the preen gland were more stressed than females that had access to their preen gland. However, we have three lines of evidence that work counter to this argument. First, in a preliminary study we did not find any significant difference in time devoted to several behaviours (e.g. grooming, feeding, walking, sleeping, bathing, courtship) between birds equipped with an APM or with a control-preening mechanism (same mechanism as the APM but without the small tube that prevents bill/preen gland contact; [37]). Second, we did not find any effect of our treatment on circulating carotenoids levels, a variable that can be affected by stressful conditions [49], [50]. Third, we found increased, not decreased, deposition of yolk carotenoids by experimental females compared to control ones (see more below). Taken together, these results suggest that blocking preen gland access does not increase stress in mallards, but we cannot rule out this hypothesis. Unfortunately, we do not have blood samples available to test this idea, but in future work we suggest that measuring blood or feces corticosterone levels would be a useful addition to this line of work.

In addition to the somatic effects on mothers, we found that restricting maternal access to the preen gland altered egg investment. Experimental females deposited more carotenoids and produced lighter eggs than control females. The fact that eggs were lighter is consistent with the idea that these females were in worse condition and health and thus could devote fewer resources to eggs, as has been shown previously [51], [52], [53]. Dufva (1996) [54], for example, showed that female great tits (Parus major) infected with Trypanosoma blood parasites laid smaller eggs than non-infected females. However, the shift in carotenoid allocation differs from the results of Saino et al. (2002a) [13], who found that wild female barn swallows (Hirundo rustica) whose immune system had been experimentally challenged with an antigen deposited lower lutein concentrations in their eggs than control females. If we assume that microbe exposure was also higher in eggs and chicks of females that had no access to their preen gland, then it is possible that elevated carotenoid levels worked to boost immunity [13], [8] in comparatively pathogen-challenged offspring. It is interesting that another health-related metric – albumen lysozyme concentration – was not altered by the preen gland treatment, suggesting a quite carotenoid-specific mechanism here. Small sample size could have limited our statistical power in this analysis as well. However, it should be noted that the lysozyme patterns we obtained were very consistent and similar in relation to the laying sequence. The precise carotenoid allocation mechanism at work is far from clear; carotenoids probably were not a limiting resource in our captive study (i.e. birds were fed ad libitum, carotenoid levels did not decline with laying order), and treated females did not show higher circulating plasma carotenoid levels compared to controls.

In conclusion, we found evidence that breeding female mallards are somatically affected by having no access to their preen gland and also modulate their reproductive investment accordingly. Further studies should examine the proximate mechanisms and fitness consequences of maternal deposition of immune compounds in eggs as a function of pathogen exposure.