1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

Studies of avian species have shown that maternal effects mediated by the transfer of egg hormones can profoundly affect offspring phenotype and fitness. We previously demonstrated that the injection of a physiological amount of testosterone (T) in the eggs of ring-necked pheasants (Phasianus colchicus) disrupted the covariation among male morphological traits at sexual maturity and positively affected male mating success. Here, we investigate whether egg T exposure affected adult male circulating T levels at the onset of the breeding season (reflecting gonadal maturation), and the relationship between circulating T and male traits. Egg T exposure did not affect pre-mating plasma T. T levels were not associated with the expression of secondary sexual and non-sexual traits or socio-sexual behaviour (social rank, overall fighting ability and mating success). However, wattle brightness decreased with increasing circulating T in males hatched from T-eggs (T-males) but not among control males. In dyadic encounters during the peak mating period, control males with higher pre-mating T levels had higher chances of being dominant over other control males. However, higher pre-mating T levels did not predict success in male-male competition in encounters involving T-males. We suggest that the long-term effects of egg T on male phenotype do not originate from differential gonadal maturation according to egg T treatment. Rather, prenatal androgens may have priming effects on functioning of target tissues, translating into differential phenotypic effects according to androgen exposure during embryonic development.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

Maternal effects, which occur whenever mothers influence the development and phenotype of the progeny via maternal care and/or pre- and post-zygotic transmission of resources, constitute a fundamental source of phenotypic variation (Mousseau & Fox 1998; Wolf et al. 1998; Badyaev & Uller 2009; Wolf & Wade 2009). Moreover, they represent an important source of transgenerational phenotypic plasticity, allowing information about the environment experienced by the mother to be translated into adaptive phenotypic variation of the offspring (Badyaev & Uller 2009; Wolf & Wade 2009; Ho & Burggren 2010). Avian species constitute a highly suitable model to investigate the fitness consequences of maternal effects, because their cleidoic eggs contain a cocktail of substances of maternal origin, including hormones, vitamins and immune factors (Royle et al. 2001; Grindstaff et al. 2006; Groothuis et al. 2006; Rubolini et al. 2011), which have been shown to have both short- and long-term consequences for progeny development, phenotypic traits and fitness (Gil 2003; Groothuis et al. 2005; Groothuis & Schwabl 2008).

Maternal hormones, mostly steroids, vary markedly among clutches and between individual eggs in a clutch, often in a predictable fashion (Schwabl 1993; Royle et al. 2001; Groothuis et al. 2005; Groothuis & Schwabl 2008; Love et al. 2008; Rubolini et al. 2011). Such variation has paved the way for the experimental investigation of hormone-mediated maternal effects on offspring phenotype and fitness (Groothuis et al. 2005). Experimental manipulations are generally performed by increasing the hormonal egg content within the physiological limits via egg hormone injections, thus mimicking increased maternal transfer, and subsequently measuring the effects on progeny morphological and behavioural traits (Groothuis et al. 2005; Groothuis & Schwabl 2008). During ontogeny, exposure to steroids exerts ‘pleiotropic’ effects, by affecting different developmental pathways involving, for example, sexual, muscular and skeletal development, as well as metabolism and the immune system (see comprehensive reviews in Groothuis et al. 2005; Groothuis & Schwabl 2008). Early effects due to hormonal priming of target tissues may be long-lasting and contribute to shaping the adult phenotype (Groothuis et al. 2005; Groothuis & Schwabl 2008). Indeed, long-term studies showed that egg androgens affected male morphological and secondary sexual traits (Strasser & Schwabl 2004; Eising et al. 2006; Rubolini et al. 2006; Riedstra et al. 2013), female reproduction (Rubolini et al. 2007), as well as socio-sexual behaviour (Strasser & Schwabl 2004; Eising et al. 2006; Partecke & Schwabl 2008; Bonisoli-Alquati et al. 2011a; Schweitzer et al. 2013) and personality traits (Tobler & Sandell 2007; Ruuskanen & Laaksonen 2010). These effects, however, were often inconsistent across studies and species (Müller et al. 2008; Müller & Eens 2009; Bonisoli-Alquati et al. 2011b; Ruuskanen et al. 2012).

The specific mechanisms of action underlying long-lasting phenotypic effects of prenatal androgens are not well understood and are likely to be the result of both ‘organizational’ and ‘activational’ effects (Carere & Balthazart 2007; Groothuis & Schwabl 2008; Navara & Mendonca 2008). Indeed, steroid exposure during critical periods of embryonic development can cause permanent, irreversible modifications of the phenotype that last through adulthood, a so-called ‘organizational’ effect (Arnold & Breedlove 1985; Adkins-Regan 2007). Later on, at puberty or after sexual maturity, endogenously released steroids can (reversibly) stimulate behavioural traits, including those whose neural substrates were permanently affected by early exposure to maternal hormones, a so-called ‘activational’ effect (Arnold & Breedlove 1985; Adkins-Regan 2007).

For example, egg androgens may affect the development of the hypothalamo/pituitary/gonadal (HPG) axis, and thus, patterns of gonadal maturation and hormone secretion during adult life. Variation in circulating androgens in relation to early androgen exposure could alter adult socio-sexual behaviour and the expression of androgen-dependent secondary sexual traits. However, evidence that early androgen exposure affects functioning of the HPG axis is limited to the post-fledging stage, that is, well before sexual maturation. Elevated egg androgens caused higher plasma androgen levels in starling (Sturnus unicolor) nestlings close to fledging (Müller et al. 2007), but reduced plasma T in feral fowl (Gallus g. domesticus) chicks (Pfannkuche et al. 2011). Moreover, quail (Coturnix japonica) chicks hatched from T-injected eggs showed a tendency to excrete more T metabolites than chicks hatched from control eggs (Daisley et al. 2005). Studies analysing circulating hormones in sexually mature adult birds did not disclose any effect of elevated egg androgens (Partecke & Schwabl 2008; Riedstra et al. 2013; Schweitzer et al. 2013). However, early androgen exposure may affect adult behaviour and phenotype without affecting circulating androgen levels. This could be the case because early effects of in ovo androgens may affect the sensitivity of target tissues later in life (Carere & Balthazart 2007; Groothuis & Schwabl 2008), implying that similar adult circulating androgen levels can have differential persistent effects on androgen-dependent morphological traits and/or transient effects on behaviour according to early androgen exposure.

An alternative to the traditional ‘organizational/activational’ view of the long-term effects of in ovo hormonal exposure is that long-lasting effects of prenatal androgens represent indirect, cascading effects of early activational effects (Carere & Balthazart 2007). For example, early social experiences (involving, e.g., sibling interactions) affected by prenatal androgen exposure can ‘prime’ adult behaviour, independently of direct hormonal actions during adulthood (reviewed in Carere & Balthazart 2007).

In this study, we focus on pre-mating circulating T levels of adult male ring-necked pheasants (Phasianus colchicus) that were prenatally exposed to high T. Natural egg T levels were increased within the physiological limits by injecting in the egg albumen a T solution to mimic increased transfer of maternal egg T, and the morphology and behaviour of males hatching from T-injected eggs (T-males hereafter) were compared with those of males originating from control eggs (control males hereafter) (Bonisoli-Alquati et al. 2011a,b; Baratti et al. 2012).

We have previously shown on the same set of males that egg T affected the expression of wattle red colouration (T-males having less red, more orange wattles than controls), disrupted the covariation among several secondary sexual traits and between wattle colour and cell-mediated immune response (Bonisoli-Alquati et al. 2011b). It also affected male mating success, with T-males obtaining more copulations than control males, specifically with control females (Bonisoli-Alquati et al. 2011a). However, egg T exposure did not affect male social rank or overall success in male-male competition (Bonisoli-Alquati et al. 2011a).

The first aim of this study was thus to investigate whether circulating T levels at the onset of the breeding season (during a period of social stability; see 'Methods') differed between T- and control males: higher T levels in T-males at this time may suggest a permanent, organizational effect of maternal T on the HPG axis and/or gonadal maturation, with T-males maturing earlier than controls. Secondly, we investigated whether pre-mating T levels covaried with the concomitant expression of both sexual and nonsexual traits according to egg T treatment. A covariation between T levels and expression of male morphological traits might be expected, because in pheasants, the expression of important male ornaments, such as the size of the wattle (but not spur length), is dependent on circulating T (Briganti et al. 1999). Moreover, T may be immunosuppressive (Duffy et al. 2000; Roberts et al. 2004), and we therefore expected a negative covariation between circulating T and intensity of the cell-mediated immune response, as assessed by the standard phytohaemagglutinin skin testing technique (Lochmiller et al. 1993; Saino et al. 1997). Thirdly, we investigated whether pre-mating T levels predicted the odds of a male being dominant in subsequent dyadic aggressive encounters occurring during peak mating activity, and whether this relationship was dependent on egg T treatment. Our general prediction was that males with higher pre-mating T levels were more likely dominant in subsequent male-male combats. Indeed, circulating androgens predict aggressiveness in intrasexual conflicts in Vertebrates (Adkins-Regan 2005), and T implants increased the frequency of aggressive interactions in pheasants, with T-implanted males achieving higher rank (Briganti et al. 1999). Although we previously failed to detect a relationship between success in intrasexual conflicts and egg T treatment in the same set of males (Bonisoli-Alquati et al. 2011a), we argue that this could be due to the indirect effects of interindividual variation in T levels in modulating aggressive behaviour. The investigation of the effects of plasma T levels on aggressiveness in relation to egg T treatment may help elucidate the mechanisms behind the previously documented effects of egg T on male socio-sexual behaviour.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

Study Species

The ring-necked pheasant is a highly dimorphic galliform bird species, in which males are larger than females and characterized by conspicuous multiple secondary sexual traits, including elongated tail feathers, ear tufts, long spurs and periorbital wattles (Mateos 1998). During the mating season, males acquire and defend mating territories (Ridley & Hill 1987) via male-male aggressive interactions (threats, direct attacks and long-lasting fights) and territorial displays (erecting the wattle and ear tufts, the so-called ‘wattle display’) (Mateos & Carranza 1997). Position in the dominance hierarchy covaried with morphology and secondary sexual traits in some studies (body mass and tail length: von Schantz et al. 1989; spur length: Göransson et al. 1990; Hillgarth 1990; ear tuft length and wattle size: Mateos & Carranza 1997; Papeschi & Dessì-Fulgheri 2003), although these associations were not observed in our stock of males (Bonisoli-Alquati et al. 2011a). Previous correlational and experimental studies demonstrated that circulating T levels predict wattle size and dominance rank (Briganti et al. 1999; Papeschi et al. 2000).

Experimental Manipulation of Egg T and Subject Housing

The study was conducted in captivity at a game farm. Full details of experimental procedures are extensively reported in previously published articles (Bonisoli-Alquati et al. 2011a,b; Baratti et al. 2012) and are therefore only briefly summarized here. We purchased 500 freshly laid, unincubated eggs, randomly chosen from a very large breeding stock (L'envol de Retz, Machecoul, France: Although maternal and paternal identity of individual eggs was not known, the experimental eggs were randomly sampled from a very large pool of eggs, and the 90 males included in the present study were randomly chosen among those that reached sexual maturity (see below and Baratti et al. 2012). It is thus very unlikely that shared parentage among males confounded our results. Half of the eggs (T-eggs) were inoculated with a solution of T in ethyl alcohol dissolved in distilled water, while the other half, serving as controls, were inoculated with a solution of ethanol in distilled water.

The concentration of T in the albumen of a sample of pheasant eggs was 25.8 pg/g (22.7 s.d.) (Bonisoli-Alquati et al. 2011a). We aimed at increasing the mean egg albumen T concentration by 2 s.d. (Bonisoli-Alquati et al. 2011a,b). To this end, we injected 477 pg of T in each egg, assuming an average albumen mass of 10.5 g (Bonisoli-Alquati et al. 2011a). All the eggs were incubated in a professional incubator. Hatching success did not differ between treatments (control eggs: 80.4%; T-eggs: 82.0%) (Bonisoli-Alquati et al. 2011a,b), and chicks were maintained in captivity in the same aviary (12 × 8 × 3.5 m indoor; approximately 20 × 10 × 2.5 m outdoor). When 200 d old, the two sexes were separated in two different aviaries (each of the same size as the chick aviary) in mixed-treatment groups. All individuals were individually marked with unique combinations of colour rings on the legs.

Measurement of Phenotypic Traits and Blood Sampling

At sexual maturation, when development of the adult plumage was completed (270 d of age; pheasants become sexually mature in the spring season after the one of hatching; Cramp 1998), we measured male phenotypic traits and took a blood sample. Males were captured sequentially in the aviary using a special trapping chamber in batches of 5–10 individuals at a time (to minimize stress to the birds that were not being measured). We recorded a suite of sexual and nonsexual traits on 92 males (43 control and 49 T-males), including cell-mediated immunity (CMI) through the phytohaemagglutinin skin test (Lochmiller et al. 1993).

We measured tarsus length, body mass, spur and right ear tuft length, wattle size and colour, as detailed in Bonisoli-Alquati et al. (2011a,b). Wattle colour was measured using a portable spectrometer (Avantes AvaSpec 2048, the Netherlands) connected to a dual deuterium-halogen light source (Avantes DH-2000) [see Bonisoli-Alquati et al. (2011b) for details about frequency of calibration, position of measurements and number of repeated measurements for each individual]. Reflectance spectra were analysed using segment classification [see Armenta et al. (2008) for details]. As measures of wattle coloration, we calculated hue, chroma and brightness (Endler 1990) in the visible range (400–700 nm), because Galliformes are relatively insensitive to UV wavelengths (Ödeen & Håstad 2006). We derived colour measurements from reflectance spectra by means of ad hoc implemented routines in Microsoft Excel.

Cell-mediated immunity was estimated by subcutaneously injecting 0.2 mg of phytohaemagglutinin (PHA) dissolved in 0.05 ml phosphate-buffered saline (PBS) in the right wing web, while the left wing web was injected with PBS only to serve as a control. Cutaneous thickness at the injection sites was measured 24-h post-injection with a pressure-sensitive micrometer, assuming that the difference in thickness between the PHA-injected wing web and the sham-injected one estimates the intensity of cell-mediated immune response, with higher values implying a stronger immune response (Lochmiller et al. 1993; Saino et al. 1997).

After recording morphological measurements and before the PHA-PBS injection, we collected a blood sample (approximately 2 ml, kept into heparinized 2.5 ml vials) from the brachial vein using 2.5-ml disposable syringes. We later used these samples for the determination of plasma T levels. All blood samples and phenotypic measurements were taken on the same day in all individuals, mostly during morning hours. To minimize the effect of handling stress on blood parameters, individuals of a given batch (5–10 birds) were all measured and blood sampled before capturing, measuring and blood sampling those of the subsequent batch. Blood was sampled shortly before the peak territorial season, after all males had developed full nuptial plumage and during a period of gonadal maturation, when plasma T levels are expected to be highly variable among individuals (Goymann et al. 2007; Kempenaers et al. 2008), with the aim of detecting possible ‘organizational’ effects of prenatal T exposure on functioning and/or maturation of the HPG axis. However, we used the information on pre-mating T levels to test whether they predicted the outcome of intra- and intersexual interactions during the peak mating activity (when males were introduced into the aviary containing females), a period of high social instability and likely of highly fluctuating T levels (Wingfield et al. 1990).

Behavioural Observations

Starting from approximately 2 weeks after blood sampling, from the second half of March (i.e. during the peak territorial season; Cramp 1998), 42 males were randomly assigned to three groups (n = 14 per group, including seven control and seven T-males). We sequentially transferred each group into an aviary containing 108 females (i.e. a 1:8 male-to-female ratio), and left the males of a group and the females free to interact for 7 d. The 1:8 male-to-female ratio was similar to that observed in territories maintained by dominant males in wild populations (Johnsgard 1999) and is in accordance with game pheasant farming practices. We conducted behavioural observations for five consecutive days, starting 2 d after we transferred the males to the aviary, to allow acclimatization of males and establishment of social hierarchies. Each day, two observers conducted observations for six consecutive hours, from 7:00 a.m. to 1:00 p.m. We scored all interactions occurring between pairs of males, by noting the identity of both interacting birds and by classifying males as either the winner or the loser in each dyadic interaction. Winners and losers are easily recognized, as winners initiate the contests by displacing, threatening, pecking and/or chasing losers, which then retreat or escape (Mateos & Carranza 1997). On no occasion did the two observers disagree when classifying the outcome of interactions. An interaction was considered to have terminated when the winner stopped pecking at and chasing the loser.

We observed 1767 male–male interactions, which we used to calculate male social rank, as indexed by the David's score (Gammel et al. 2003) (see details in Bonisoli-Alquati et al. 2011a). This index is relatively insensitive to minor deviations from linearity in the hierarchy rank and is considered more reliable than other dominance indices (Gammel et al. 2003). Rank was expressed as the relative rank within each group, with rank 1 assigned to the highest ranking male and rank 14 assigned to the lowest ranking male within each group. Besides social rank, we additionally calculated the total number of wins by an individual as a global index of his fighting ability and success in male-male combats.

We also identified the participants in all copulations. We considered only the copulations where the female presented herself in a squatting position to the approaching male (i.e. when the female was performing a copulation acceptance behaviour; Mateos 1998). As in previous studies, we considered the number of copulations as an index of female preference, and thus male sexual attractiveness (Bonisoli-Alquati et al. 2011a; Baratti et al. 2012). Hereafter, we will use the total number of copulations obtained by a male and the total number of partners with which we documented at least one copulation as proxies for its mating success.

Plasma Hormone Assay

Vials with blood samples were kept cool and centrifuged within a few hours after sampling. Plasma was separated and kept at −20°C until assay. We could not obtain blood samples for two individuals, and one outlier individual with a very high value of plasma T (>7000 pg/ml) was excluded from all analyses, so the final sample size was 89 subjects (42 control and 47 T-males) for the analyses of the covariation of circulating T with phenotypic traits, and 41 subjects (21 control and 20 T-males) for behavioural analyses. Plasma T concentration was determined by direct radioimmunoassay (RIA) as described by Goymann et al. (2006). Briefly, an aliquot of 50 μl of plasma was extracted twice with dichloromethane after addition of tritiated T (Perkin Elmer, Rodgau, Germany). The organic phase was then separated from the aqueous phase by snap-freezing on a dry ice/ethanol mixture. Dried extracts were resuspended in PBS with 1% gelatine (PBSG), and an aliquot was transferred to scintillation vials to measure individual recoveries, which resulted to be 80% (12% s.d.). The RIA was conducted in a single run on duplicate aliquotes of the extracted samples using T antiserum (T3-125, Esoterix Endocrinology, Calabasas, CA, USA). After addition of tritiated T, the bound and free fractions were separated by addition of dextran-coated charcoal in PBSG assay buffer and centrifugation. The supernatant was decanted into scintillation vials and counted. Calculations were conducted with Immunofit 3.0 (Beckman Inc. Fullerton, CA), using a four-parameter logistic curve fit. The lower detection limit of the assay was determined by the 95% confidence intervals (c.i.) for the zero standard and was set at 7.8 pg/tube. The intra-assay coefficient of variation was 11%. Because the T antibody we used shows a significant cross-reaction with 5α-dihydrotestosterone (DHT) (44%), the reported concentration of T may include a fraction of DHT. We note that this is unlikely to affect our results concerning T levels as most studies have shown that DHT levels usually correlate well with T levels, although at much lower concentrations (approximately 5–10 times less) (e.g. Wingfield & Farner 1978, 1993; Fusani et al. 2003).

Statistical Analyses

Plasma T levels were log10-transformed before statistical analyses because of the highly skewed distribution, with a few high values (as it is often the case in studies of circulating T levels; Papeschi et al. 2000; Kempenaers et al. 2008) whose disproportionate leverage could bias any parametric statistic.

We first investigated whether pre-mating T levels predicted the concomitant expression of phenotypic traits [non-sexual traits: tarsus length, body mass, CMI; sexual traits: ear tuft length, spur length, wattle area, wattle hue, chroma and brightness (the latter three only in the visible range)] and socio-sexual behaviour during the subsequent peak mating period (number of wins in male–male interactions, dominance rank, total number of copulations and number of female partners) in linear models where we included egg T treatment (0 = control, 1 = T-males) and the interaction between circulating T levels and egg T treatment as predictors.

Secondly, we investigated whether pre-mating T levels predicted the outcome of male-male dyadic aggressive encounters during the subsequent peak mating activity period according to their respective egg treatment. To this aim, we first calculated for each dyad of interacting males (n = 219 male dyads) the number of wins for each member of the pair. We defined as dominant the male that won more than 50% of the aggressive interactions (‘dominant male’ hereafter), while the other contestant was defined as the ‘subordinate male’. Four dyads in which the two males had an identical proportion of wins were excluded from all analyses, leading to a final sample size of 215 male dyads. As in most cases (195 of 215 dyads) a male either won or lost all interactions against any given opponent, defining dominant males using more restrictive cut-offs (e.g. 80% of interactions won) did not alter our conclusions (details not shown for brevity). We ran a conditional logistic regression analysis for matched-pairs, case–control, datasets (Breslow & Day 1980; Hosmer & Lemeshow 1989), where the odds of a male being the dominant (code 1; the case) or subordinate (code 0; the control) individual in each dyad was the binary dependent variable. Fitting a conditional logistic regression model is identical to fitting a logistic regression to a dataset with a constant response (e.g. the case, 1), where the fitted fuction is constrained to pass through the origin and where each of the j predictors (x·j) is expressed as the difference (d·j) between the value of the case and the control for each i-th case–control matched pair (i.e. dij xij1xij0) (Breslow & Day 1980). We tested whether the effect of pre-mating T levels varied according to a three-level factor accounting for egg treatment of the two males in each dyad (encounter category: CC = control male vs. control male, n = 54 dyads; CT = control male vs. T-male, n = 118 dyads; TT = T-male vs. T-male, n = 43 dyads). As in conditional logistic regression models the fitted functions are constrained to pass through the origin, we could not include in the model the main effect of encounter category. In essence, our model tests for the data being fitted best by three different logistic functions representing the relationship between the odds of being dominant and circulating T (one for each level of encounter category) or by a unique logistic function for all levels of encounter category. In this latter case, the odds of being dominant vary according to circulating T but would be independent of encounter category. To avoid biases in significance tests and parameter estimates due to repeated testing of the same males in the different dyads, we assessed the p-value of the interaction term with a randomization test and obtained confidence intervals of parameter estimates by bootstrapping (see below for details of bootstrapping). The randomization test we adopted to determine the significance of the interaction term was performed as follows. First, we computed the likelihood ratio (LR = difference in −2 × log-likelihoods) between the model including and the one excluding the interaction term (LR of the model fitted to observed data, LRobs hereafter). Second, at each run, we randomly reassigned egg treatments of individual males (i.e. each male was randomly assigned to T or control group) and this random assignment was maintained for all interactions involving a particular male in a given randomization. Factor encounter category was thus recomputed at each run according to random reassignment of egg treatments of individual males, and a model including observed circulating T and its interaction term with this randomly generated encounter category was fitted to the data. LR between this model and the one excluding the interaction term was noted (LR of the models fitted to randomly generated data, LRrand hereafter). The rationale behind reshuffling male egg treatments is that we aimed at testing whether different encounter categories led to differences in the relationship between the odds of being dominant vs. circulating T or whether this relationship was the same for all encounter categories. The process was repeated 5000 times, and the significance of the interaction term was calculated as the probability of obtaining a value more extreme than the LRobs value over the distribution of all LRrand values.

Conditional logistic regression was run using the clogit function of the R package survival (vers. 2.37-4), with dyad specified as a ‘stratum’ variable (the stratum variable accounts for the paired nature of the dataset). Standard errors and confidence intervals (95% c.i.) of parameter estimates for each encounter category were obtained by bootstrapping (with replacement) male dyads 5000 times. Both standard errors and 95% confidence intervals were calculated based on the adjusted bootstrap percentile (BCa method), as implemented in the function of the R package boot (vers. 1.3-7). This method of calculating bootstrap c.i. is appropriate when the distribution of bootstrap estimates is skewed (DiCiccio & Efron 1996), as it was the case for some of the estimated parameters (details not shown). All analyses were run using R software (vers. 2.15.3) (R Core Team 2013).

Ethical Note

The study was carried out with the approval of the Ethical Committee of the former Department of Biology (now Department of Biosciences), University of Milan, Italy. The aviary was large enough for individuals to withdraw, and no wounds or injuries took place during aggressive encounters. Signs of stress (such as feather erection and open beak), which could occur during aggressive encounters, rapidly disappeared afterwards.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

Egg Treatment, Pre-Mating Plasma T Levels, Morphology and Socio-Sexual Behaviour

Plasma T levels of control and T-males at the onset of the breeding season were highly variable in both groups and did not statistically differ [control males: 506.6 (658.8 s.d.) pg/ml, n = 42; T-males: 404.9 (405.0 s.d.) pg/ml, n = 47; t-test on log10-transformed values: t87 = 0.88, p = 0.38]. Variances in T levels were also similar between groups (Levene's test, p = 0.29). T levels did not predict the expression of both nonsexual and sexual traits, with the exception of wattle brightness, that significantly and negatively covaried with T levels in T-males (Table 1, Fig. 1). Moreover, we detected a statistically significant (p = 0.048) interaction between egg treatment and T levels on wattle area (Table 1, Fig. 1): however, this interaction term is difficult to interpret because the estimated parameters for both groups did not statistically differ from 0 (see footnotes to Table 1). After controlling for circulating T levels, egg T treatment significantly predicted wattle hue, as shown in a previous analysis of the same dataset, with T-males showing a higher wattle hue (i.e. a less red wattle) than controls (see Bonisoli-Alquati et al. 2011b). T levels did not predict overall success in intrasexual competition or mating success, as measured by the total number of interactions won, dominance rank or number of copulations obtained (Table 1).

Table 1. Parameter estimates (s.e.) of linear models testing the effects of circulating T level, egg T treatment (0 = control males, 1 = T-males) and their interaction on the expression of non-sexual traits, sexual traits, and measures of agonistic and sexual behaviour of male ring-necked pheasants. Parameter estimates of main effects refer to models excluding the non-significant interaction terms, except for wattle area and brightness (see Table footnotes)
 T levelEgg treatmentT level × egg treatment
  1. * = p < 0.05; ** = p < 0.01.

  2. a

    Estimates refer to values centred around their mean value.

  3. b

    Model estimates: control males = −0.34 (0.29), t = 1.16, p = 0.25; T-males = 0.42 (0.24), t = 1.74, p = 0.086.

  4. c

    Model estimates: control males = 0.04 (0.05), t = 0.95, p = 0.35; T-males = −0.10 (0.04), t = 2.55, p = 0.013.

  5. d

    Test performed on log10(x + 1)-transformed values.

Nonsexual traits (n = 89)
Tarsus length (mm)−0.13 (0.66)0.35 (0.67)−0.69 (1.34)
Body mass (g)17.02 (28.19)36.67 (28.78)33.98 (57.51)
CMI (mm)−0.10 (0.12)−0.16 (0.12)−0.40 (0.23)
Sexual traits (n = 89)
Ear tuft length (mm)0.36 (0.50)−0.15 (0.51)0.34 (1.21)
Spur length (mm)0.34 (0.40)−0.29 (0.41)0.33 (0.81)
Wattle area (cm2)a0.06 (0.19)−0.02 (0.19)0.76 (0.37) *b
Wattle hue (×10)−0.12 (0.07)0.24 (0.07)**−0.17 (0.13)
Wattle chroma (×10)0.05 (0.08)−0.13 (0.08)0.19 (0.16)
Wattle brightness (×10)a−0.03 (0.03)−0.01 (0.01)−0.14 (0.06)*c
Agonistic and sexual behaviour (n = 41)
Dominance rank−0.66 (1.19)−1.01 (1.27)2.89 (2.42)
Number of winsd−0.06 (0.20)0.12 (0.21)−0.52 (0.41)
Number of copulationsd−0.08 (0.13)0.22 (0.13)−0.31 (0.25)
Number of female partners−0.71 (2.37)2.64 (2.51)−5.62 (4.81)

Figure 1. Relationship between (a) wattle area or (b) wattle brightness and circulating T levels among T (continuous line, filled circles) and control males (broken line, open circles). The lines represent linear regressions (see footnotes to Table 1 for statistics).

Download figure to PowerPoint

Plasma T Levels and Male Dominance in Dyadic Encounters

In a first descriptive analysis, we observed that in dyadic encounters involving control males (CC dyads), individuals with relatively higher pre-mating T levels than their opponents, as judged by the signed difference in circulating T levels between contestants, were the dominant ones in 67% of cases (36 of 54 dyads), a proportion significantly larger than 50% (binomial test, p = 0.020). On the other hand, in CT encounters, males with higher T were dominant in 48% of cases (57 of 118 dyads), and figures were similar for TT encounters (47%, 20 of 43 dyads).

Consistently with the above analysis, the conditional logistic regression indicated that the effect of pre-mating T levels on the odds of a male being dominant in dyadic encounters occurring during peak mating activity varied according to encounter category (encounter category × circulating T interaction, LRobs = 10.96, d.f. = 2, p-value from the randomization test = 0.017). This was the case because T levels significantly and positively predicted the odds of a male being dominant in dyadic encounters involving control males (estimate: 1.94, 95% c.i.: [0.47, 2.68]), but not in those involving at least one T-male (CT encounters, estimate: −0.02, 95% c.i.: [−0.71, 0.62]; TT encounters, estimate: 0.10, 95% c.i.: [−1.12, 0.92]) (Fig. 2). The effect of circulating T depended on the encounter category also when we included in separate models phenotypic traits potentially affecting male dominance (both nonsexual and sexual traits, excluding CMI; p-values from randomization tests always < 0.034, n = 8 models; see list of traits in Table 1). All results obtained by resampling and bootstrapping were entirely consistent with the outcome of conditional regression models fitted to the original dataset (details not shown for brevity).


Figure 2. Odds of a male being dominant in dyadic encounters in relation to the difference in pre-mating T levels between opponents [calculated as: log10(winner T level) – log10(loser T level)] according to encounter category (CC = control male vs. control male; CT = control male vs. T-male; TT = T-male vs. T-male). Lines show predicted values from the conditional logistic regression model, with standard errors derived from the bootstrap procedure (see 'Methods'). Lines are traced only within the actual range of variation of the predictor for each level of encounter category. Symbols represent the proportion of wins for each 8-quantile of the frequency distributions of differences in pre-mating T levels of each encounter category, plotted at the mean value of differences in pre-mating T for each 8-quantile. Bars represent standard errors for proportions (score method with continuity correction; Newcombe 1998). Note that lines and symbols are symmetrical with respect to the origin due to the nature of the conditional logistic regression model (see Statistical analyses). Standard errors are shown instead of confidence interval (reported in the 'Results') for clarity of representation.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

In this study of captive male pheasants, we showed that: (1) individuals that were exposed to elevated T in ovo did not differ from controls in their pre-mating plasma T levels; (2) pre-mating T levels did not predict the concomitant expression of morphological and immunological traits, nor behavioural traits (mating success and dominance rank) during peak mating period; the single exception was wattle brightness, that decreased with increasing T levels among T-males; (3) pre-mating T levels predicted the odds of a male being dominant during the subsequent peak mating period only in dyadic encounters involving control males, whereas this was not the case for encounters involving at least one T-male.

Prenatal T Exposure and Pre-Mating Circulating T Levels

Our findings argue against the hypothesis that the higher mating success of T-males we documented earlier (Bonisoli-Alquati et al. 2011a) originate from differences in functioning or maturation of the HPG axis, as circulating T levels shortly before the peak mating activity, although being highly variable likely due to intraindividual variation in gonadal state (as previously shown in pheasants; Papeschi et al. 2000), were not significantly different between the two groups of males. Rather, endogenous T levels of males during peak mating activity may have had different activational effects on socio-sexual behaviour (and thus mating success) depending on in ovo exposure to elevated androgens, which may have affected target tissues response/sensitivity. However, it should be noted that we measured T levels when males were kept in a unisexual group, separated from females. Hence, we cannot rule out the possibility that egg T affected T levels in the completely different socio-sexual context of the subsequent peak-mating period, when males were facing competition for access to females.

Recent studies carried out in the phylogenetically related feral fowl demonstrated that high egg T affected the density of brain androgen receptors and circulating T levels in the young, but not adult circulating T (Pfannkuche et al. 2011; Riedstra et al. 2013). Remarkably, prenatal T exposure feminized (rather than masculinized) male fowl phenotype, resulting in more female-like T levels, competitive behaviour of chicks, and expression of a secondary sexual trait (comb colour), thereby possibly reducing male sexual attractiveness (Riedstra et al. 2013). These findings are at odds with the positive effects of egg T on male mating success we observed in this stock of pheasants (Bonisoli-Alquati et al. 2011a), although we also documented less red wattles in T-males (Bonisoli-Alquati et al. 2011b). Such discrepancies are difficult to interpret and may be due to several reasons, including species-specific differences in the effects of prenatal androgens or differences in egg injection protocols, which may cause differences in timing of embryonic exposure to androgens. For example, Riedstra et al. (2013) (and most previous studies; see Groothuis et al. 2005) injected T in the egg yolk, while here we injected T in the albumen, and the T concentrations in the two matrices differ markedly (Rubolini et al. 2006; Gilbert et al. 2007; Bonisoli-Alquati et al. 2011a). Once in the egg, androgens injected in the yolk and albumen may differ in the timing of embryonic uptake or metabolism by the developing embryos (von Engelhardt et al. 2009). Therefore, yolk and albumen steroids may differently interact with the processes of gonadal, HPG axis or brain formation, leading to variable short- and long-term effects according to experimental injection procedure (Carere & Balthazart 2007).

Combined Effects of Prenatal T Exposure and Pre-Mating T Levels on Morphology and Social Dominance

We previously reported that elevated egg T caused a disruption of the covariation between male sexual and nonsexual traits (including CMI) in the same set of males studied here (Bonisoli-Alquati et al. 2011b). In this study, we moved a step further by analysing whether plasma T levels affected the concomitant expression of male sexual and non-sexual traits in relation to in ovo T injection. We found that the expression of most male traits was independent of pre-mating T levels, a result similar to that reported in wild pheasants, where only winter (but not pre-breeding) levels positively predicted wattle size (Papeschi et al. 2000). However, we detected a negative covariation between wattle brightness and circulating T among T-males, which was mostly due to a few T-males with T levels below the detection limit (see Fig. 1b), and a weak (p = 0.09) positive correlation between wattle area and T levels among T-males, resulting in a statistically significant interaction between T levels and prenatal T treatment on wattle area. Though wattle brightness or size do not seem to play a role in female mate choice or intrasexual competition in our pheasant stock, wattle size was previously experimentally shown to be T-dependent (Briganti et al. 1999; Papeschi et al. 2000). It is therefore plausible that the expression of wattle characteristics, including aspects of its coloration, is affected by pre-mating T levels, and more strongly so in T-males than controls. On the whole, these findings suggest for the first time that prenatal T may interact with circulating T levels in determining the concomitant expression of a T-dependent morphological trait.

The lack of statistically significant correlations between T levels and wattle size or immune response might be due to benign rearing conditions, with ad libitum food availability, reducing the costs (including immunosuppression) of maintaining high T levels. Accordingly, winter androgen levels covaried with wattle size in wild (Papeschi et al. 2000), but not captive, pheasants (F. Dessì-Fulgheri et al., unpubl. data).

The finding that pre-mating T levels positively predicted dominance among control males is in line with our expectations and consistent with the large body of literature documenting enhanced aggressiveness and higher success in intrasexual competition in individuals with higher circulating T (Adkins-Regan 2005; Kempenaers et al. 2008). On the other hand, the observation that higher circulating T levels did not predict the odds of a male being dominant in all encounters involving at least one T-male, suggesting that prenatal T exposure could alter the expression of androgen-dependent aggressive behaviour later in life, may have different explanations. For example, prenatal T exposure may have increased intraindividual variability in brain target tissues sensitivity/characteristics, decoupling any relationship between pre-mating T levels and aggressiveness during the peak mating period. In turn, aggressiveness of T-males might be more tightly related to either the information content of the combined expression of multiple traits (Rubolini et al. 2006; Bonisoli-Alquati et al. 2011b) or to the expression of specific aggressive displays (not resulting in overt aggression), than to pre-mating T levels. Alternatively, exposure to elevated egg T might have affected dominance relationships before behavioural observations took place. These early experiences could have translated into adulthood, priming subsequent socio-sexual behaviour, including intrasexual conflicts (see discussion in Carere & Balthazart 2007).

To our knowledge, only a single previous study of house sparrows (Passer domesticus) (Partecke & Schwabl 2008) analysed whether the combined effects of increased egg T and circulating hormone levels predicted competitive and sexual behaviour. In that study, there were no differences in circulating hormones between birds from different egg treatments, and no differential effects of circulating hormones on socio-sexual behaviour according to egg T. Similarly to pheasants, sparrows exposed to elevated egg T expressed a higher frequency of aggressive, dominance and sexual behaviours (Strasser & Schwabl 2004; Partecke & Schwabl 2008).

Concluding Remarks

This study indicates that the effects of egg T on morphology and socio-sexual behaviour we documented here and in previous studies were not mediated by variation in pre-mating T levels. The lack of differences in adult T levels according to prenatal T exposure we observed here is consistent with previous studies and corroborates the idea that the effects of early prenatal hormone exposure tend to fade in the course of ontogeny, as it is commonly the case for other maternal effects (Wilson & Festa-Bianchet 2009; Krist 2011). Thus, long-term effects of elevated egg T on phenotypic traits and sexual attractiveness might be due to early priming effects, resulting in different phenotypic effects later in life according to prenatal androgen exposure.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited

We are sincerely grateful to G. P. Oldani for invaluable technical support and suggestions. We thank A. Matteo, B. Leoni, U. Oldani, P. Usorini, S. Bocchi and A. T. Gerevini for help with data collection. We also would like to thank W. Goymann for making available his laboratory for the plasma hormone assays. Finally, we thank two anonymous reviewers for constructive criticism and valuable suggestions.

Literature Cited

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Literature Cited
  • Adkins-Regan, E. 2005: Hormones and Animal Social Behaviour. Princeton Univ. Press, Princeton, NJ.
  • Adkins-Regan, E. 2007: Hormones and the development of sex differences in behavior. J. Ornithol. 148, 1726.
  • Armenta, J. K., Dunn, P. O. & Whittingham, L. A. 2008: Quantifying avian sexual dichromatism: a comparison of methods. J. Exp. Biol. 211, 24232430.
  • Arnold, A. P. & Breedlove, S. M. 1985: Organizational and activational effects of sex steroids on brain and behavior: a reanalysis. Horm. Behav. 19, 469498.
  • Badyaev, A. V. & Uller, T. 2009: Parental effects in ecology and evolution: mechanisms, processes and implications. Phil. Trans. R. Soc. Lond. B 364, 11691177.
  • Baratti, M., Dessi-Fulgheri, F., Ambrosini, R., Bonisoli-Alquati, A., Caprioli, M., Goti, E., Matteo, A., Monnanni, R., Ragionieri, L., Ristori, E., Romano, M., Rubolini, D., Scialpi, A. & Saino, N. 2012: MHC genotype predicts mate choice in the ring-necked pheasant Phasianus colchicus. J. Evol. Biol. 25, 15311542.
  • Bonisoli-Alquati, A., Matteo, A., Ambrosini, R., Rubolini, D., Romano, M., Caprioli, M., Dessi-Fulgheri, F., Baratti, M. & Saino, N. 2011a: Effects of egg testosterone on female mate choice and male sexual behavior in the pheasant. Horm. Behav. 59, 7582.
  • Bonisoli-Alquati, A., Rubolini, D., Caprioli, M., Ambrosini, R., Romano, M. & Saino, N. 2011b: Egg testosterone affects wattle color and trait covariation in the ring-necked pheasant. Behav. Ecol. Sociobiol. 65, 17791790.
  • Breslow, N. E. & Day, N. E. 1980: Statistical Methods in Cancer Research, Volume I: The Analysis of Case-Control Studies. International Agency for Research on Cancer, Lyon.
  • Briganti, F., Papeschi, A., Mugnai, T. & Dessì-Fulgheri, F. 1999: Effect of testosterone on male traits and behaviour in juvenile pheasants. Ethol. Ecol. Evol. 11, 171178.
  • Carere, C. & Balthazart, J. 2007: Sexual versus individual differentiation: the controversial role of avian maternal hormones. Trends Endocrinol. Metab. 18, 7380.
  • Cramp, S. 1998: The Complete Birds of the Western Palearctic on CD-ROM. Oxford Univ. Press, Oxford.
  • Daisley, N. J., Bromundt, V., Möstl, E. & Kotrschal, K. 2005: Enhanced yolk testosterone influences behavioral phenotype independent of sex in Japanese quail chicks Coturnix japonica. Horm. Behav. 47, 185194.
  • DiCiccio, T. J. & Efron, B. 1996: Bootstrap confidence intervals (with Discussion). Stat. Sci. 11, 189228.
  • Duffy, D. L., Bentley, G. E., Drazen, D. L. & Ball, G. F. 2000: Effects of testosterone on cell-mediated and humoral immunity in non-breeding adult European starlings. Behav. Ecol. 11, 654662.
  • Eising, C. M., Müller, W. & Groothuis, T. G. G. 2006: Avian mothers create different phenotypes by hormone deposition in their eggs. Biol. Lett. 2, 2022.
  • Endler, J. A. 1990: On the measurement and classification of color in studies of animal color patterns. Biol. J. Linn. Soc. 41, 315352.
  • von Engelhardt, N., Henriksen, R. & Groothuis, T. G. G. 2009: Steroids in chicken egg yolk: metabolism and uptake during early embryonic development. Gen. Comp. Endocrinol. 163, 175183.
  • Fusani, L., Van't Hof, T. & Hutchison, J. B. 2003: Season-related changes in circulating androgen, brain aromatase, and perch-calling in male ring doves. Gen. Comp. Endocrinol. 130, 142147.
  • Gammel, M. P., de Vries, H., Jennings, D. J., Carlin, C. M. & Hayden, T. J. 2003: David's score: a more appropriate dominance ranking method than Clutton-Brock et al'.s index. Anim. Behav. 66, 601605.
  • Gil, D. 2003: Golden eggs: maternal manipulation of offspring phenotype by egg androgen in birds. Ardeola 50, 281294.
  • Gilbert, L., Bulmer, E., Arnold, K. E. & Graves, J. A. 2007: Yolk androgens and embryo sex: maternal effects or confounding factors? Horm. Behav. 51, 231238.
  • Göransson, G., von Schantz, T., Fröberg, I., Helgée, A. & Wittzell, H. 1990: Male characteristics, viability and harem size in the pheasant (Phasianus colchicus). Anim. Behav. 40, 89104.
  • Goymann, W., Trappschuh, M., Jensen, W. & Schwabl, I. 2006: Low ambient temperature increases food intake and dropping production, leading to incorrect estimates of hormone metabolite concentrations in European stonechats. Horm. Behav. 49, 644653.
  • Goymann, W., Landys, M. M. & Wingfield, J. C. 2007: Distinguishing seasonal androgen responses from male-male androgen responsiveness-revisiting the Challenge Hypothesis. Horm. Behav. 51, 463476.
  • Grindstaff, J. L., Hasselquist, D., Nilsson, J. K., Sandell, M., Smith, H. G. & Stjernman, M. 2006: Transgenerational priming of immunity: maternal exposure to a bacterial antigen enhances offspring humoral immunity. Proc. Roy. Soc. B 273, 25512557.
  • Groothuis, T. G. G. & Schwabl, H. 2008: Hormone-mediated maternal effects in birds: mechanisms matter but what do we know of them? Phil. Trans. R. Soc. Lond. B 363, 16471661.
  • Groothuis, T. G. G., Müller, W., von Engelhardt, N., Carere, C. & Eising, C. M. 2005: Maternal hormones as a tool to adjust offspring phenotype in avian species. Neurosci. Biobehav. Rev. 29, 329352.
  • Groothuis, T. G. G., Eising, C. M., Blount, J. D., Surai, P., Apanius, V., Dijkstra, C. & Müller, W. 2006: Multiple pathways of maternal effects in black-headed gull eggs: constraint and adaptive compensatory adjustment. J. Evol. Biol. 19, 13041313.
  • Hillgarth, N. 1990: Pheasant spurs out of fashion. Nature 345, 119120.
  • Ho, D. H. & Burggren, W. W. 2010: Epigenetics and transgenerational transfer: a physiological perspective. J. Exp. Biol. 213, 316.
  • Hosmer, D. W. & Lemeshow, S. 1989: Applied Logistic Regression. Wiley, New York, NY.
  • Johnsgard, P. A. 1999: The Pheasants of the World: Biology and Natural History. Smithsonian Institution Press, Washington, DC.
  • Kempenaers, B., Peters, A. & Foerster, K. 2008: Sources of individual variation in plasma testosterone levels. Phil. Trans. R. Soc. Lond. B 363, 17111723.
  • Krist, M. 2011: Egg size and offspring quality: a meta-analysis in birds. Biol. Rev. 86, 692716.
  • Lochmiller, R. L., Vestey, M. R. & Boran, J. C. 1993: Relationship between protein nutritional status and immunocompetence in Northern Bobwhite chicks. Auk 110, 503510.
  • Love, O. P., Wynne-Edwards, K. E., Bond, L. & Williams, T. D. 2008: Determinants of within- and among-clutch variation in yolk corticosterone in the European starling. Horm. Behav. 53, 104111.
  • Mateos, C. 1998: Sexual selection in the ring-necked pheasant: a review. Ethol. Ecol. Evol. 10, 313332.
  • Mateos, C. & Carranza, J. 1997: Signals in intra-sexual competition between ring-necked pheasant males. Anim. Behav. 53, 471485.
  • Mousseau, T. A. & Fox, C. W. 1998: The adaptive significance of maternal effects. Trends Ecol. Evol. 13, 403407.
  • Müller, W. & Eens, M. 2009: Elevated yolk androgen levels and the expression of multiple sexually selected male characters. Horm. Behav. 55, 175181.
  • Müller, W., Deptuch, K., Lopez-Rull, I. & Gil, D. 2007: Elevated yolk androgen levels benefit offspring development in a between-clutch context. Behav. Ecol. 18, 929936.
  • Müller, W., Vergauwen, J. & Eens, M. 2008: Yolk testosterone, postnatal growth and song in male canaries. Horm. Behav. 54, 125133.
  • Navara, K. J. & Mendonca, M. T. 2008: Yolk androgens as pleiotropic mediators of physiological processes: a mechanistic review. Comp. Biochem. Physiol. A 150, 378386.
  • Newcombe, R. G. 1998: Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat. Med. 17, 857872.
  • Ödeen, A. & Håstad, O. 2006: Complex distribution of avian color vision systems revealed by sequencing the SWS1 opsin from total DNA. Mol. Biol. Evol. 20, 855861.
  • Papeschi, A. & Dessì-Fulgheri, F. 2003: Multiple ornaments are positively related to male survival in the common pheasant. Anim. Behav. 65, 143147.
  • Papeschi, A., Briganti, F. & Dessì-Fulgheri, F. 2000: Winter androgen levels and wattle size in male common pheasants. Condor 102, 193197.
  • Partecke, J. & Schwabl, H. 2008: Organizational effects of maternal testosterone on reproductive behavior of adult house sparrows. Dev. Neurobiol. 68, 15381548.
  • Pfannkuche, K. A., Gahr, M., Weites, I. M., Riedstra, B., Wolf, C. & Groothuis, T. G. G. 2011: Examining a pathway for hormone mediated maternal effects–yolk testosterone affects androgen receptor expression and endogenous testosterone production in young chicks (Gallus gallus domesticus). Gen. Comp. Endocrinol. 172, 487493.
  • R Core Team 2013: R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna.
  • Ridley, M. & Hill, A. D. 1987: Social organization in the pheasant (Phasianus colchicus): harem formation, mate selection and the role of mate guarding. J. Zool. 211, 619630.
  • Riedstra, B., Pfannkuche, K. A. & Groothuis, T. G. G. 2013: Increased exposure to yolk testosterone has feminizing effects in chickens, Gallus gallus domesticus. Anim. Behav. 85, 701708.
  • Roberts, M. L., Buchanan, K. L. & Evans, M. R. 2004: Testing the immunocompetence handicap hypothesis: a review of the evidence. Anim. Behav. 68, 227239.
  • Royle, N. J., Surai, P. F. & Hartley, I. R. 2001: Maternally derived androgens and antioxidants in bird eggs: complementary but opposing effects? Behav. Ecol. 12, 381385.
  • Rubolini, D., Romano, M., Martinelli, R., Leoni, B. & Saino, N. 2006: Effects of prenatal yolk androgens on armaments and ornaments of the ring-necked pheasant. Behav. Ecol. Sociobiol. 59, 549560.
  • Rubolini, D., Martinelli, R., von Engelhardt, N., Romano, M., Groothuis, T. G. G., Fasola, M. & Saino, N. 2007: Consequences of prenatal androgen exposure for the reproductive performance of female pheasants (Phasianus colchicus). Proc. Roy. Soc. B 274, 137142.
  • Rubolini, D., Romano, M., Navara, K. J., Karadas, F., Ambrosini, R., Caprioli, M. & Saino, N. 2011: Maternal effects mediated by egg quality in the Yellow-legged Gull Larus michahellis in relation to laying order and embryo sex. Front. Zool. 8, 24.
  • Ruuskanen, S. & Laaksonen, T. 2010: Yolk hormones have sex-specific long-term effects on behavior in the pied flycatcher (Ficedula hypoleuca). Horm. Behav. 57, 119127.
  • Ruuskanen, S., Lehikoinen, E., Nikinmaa, M., Siitari, H., Waser, W. & Laaksonen, T. 2012: Long-lasting effects of yolk androgens on phenotype in the pied flycatcher (Ficedula hypoleuca). Behav. Ecol. Sociobiol. 67, 361372.
  • Saino, N., Calza, S. & Møller, A. P. 1997: Immunocompetence of nestling barn swallows in relation to brood size and parental effort. J. Anim. Ecol. 66, 827836.
  • von Schantz, T., Göransson, G., Andersson, G., Fröberg, I., Grahn, M., Helgée, A. & Wittzell, H. 1989: Female choice selects for a viability-based trait in pheasants. Nature 337, 166169.
  • Schwabl, H. 1993: Yolk is a source of maternal testosterone for developing birds. Proc. Natl Acad. Sci. USA 90, 1144611450.
  • Schweitzer, C., Goldstein, M. H., Place, N. J. & Adkins-Regan, E. 2013: Long-lasting and sex-specific consequences of elevated egg yolk testosterone for social behavior in Japanese quail. Horm. Behav. 63, 8087.
  • Strasser, R. & Schwabl, H. 2004: Yolk testosterone organizes behavior and male plumage coloration in house sparrows (Passer domesticus). Behav. Ecol. Sociobiol. 56, 491497.
  • Tobler, M. & Sandell, M. 2007: Yolk testosterone modulates persistence of neophobic responses in adult zebra finches, Taeniopygia guttata. Horm. Behav. 52, 640645.
  • Wilson, A. J. & Festa-Bianchet, M. 2009: Maternal effects in wild ungulates. In: Maternal Effects in Mammals (Maestripieri, D. & Mateo, J. J., eds). Univ. of Chicago Press, Chicago, pp. 83103.
  • Wingfield, J. C. & Farner, D. S. 1978: The annual cycle of plasma IrLH and steroid hormones in feral populations of the white-crowned sparrow, Zonotrichia leucophrys gambelii. Biol. Reprod. 19, 10461056.
  • Wingfield, J. C. & Farner, D. S. 1993: Endocrinology of reproduction in wild species. In: Avian Biology Vol. 9 (Farner, D. S., King, J. R. & Parkes, K. C. eds.). Academic Press, New York, pp. 163327.
  • Wingfield, J. C., Hegner, R. E., Dufty, A. M., Ball, G. F., Naturalist, T. A. & Dec, N. 1990: The “Challenge Hypothesis”: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. Am. Nat. 136, 829846.
  • Wolf, J. B. & Wade, M. J. 2009: What are maternal effects (and what are they not)? Phil. Trans. R. Soc. Lond. B 364, 11071115.
  • Wolf, J. B., Brodie, E. D. I., Cheverud, J. M., Moore, A. J. & Wade, M. J. 1998: Evolutionary consequences of indirect genetic effects. Trends Ecol. Evol. 13, 6469.