Recently-adopted foraging strategies constrain early chick development in a coastal breeding gull

Field study

We studied the development of free-ranging chicks in a mixed breeding colony of Lesser Black-backed Gulls and Herring Gulls (Larus argentatus) at the outer port of Zeebrugge (Belgium, 51°20′53″N 3°10′20″E). The colony counted 1458 Lesser Black-backed Gull pairs in 2016 and 1326 pairs in 2017 (Stienen et al., 2017; Stienen et al., 2018). Lesser Black-backed Gulls exhibit a limited sexual dimorphism (males being on average larger than females but with a large overlap) and show bi-parental care. The species lays two or three eggs during May and June, of which the first two are largest (Verboven et al., 2005). At the onset of breeding in 2016 and 2017, during egg laying, 26 and six Lesser Black-backed Gull nests, respectively, were haphazardly selected for monitoring and enclosed with chicken wire. Sample size was reduced in 2017 to avoid further disturbance of the colony after a year with high nest predator pressure. Nests were visited every first, third and fifth weekday, until all remaining chicks had reached an age of 30 days after the hatching date, thus covering the periods from 27 May to 19 July 2016 and from 7 June to 14 July 2017.

On the estimated day of hatching, eggs of monitored nests were substituted by 2 first- or second-laid eggs of equal developmental stage (= pipping eggs), each originating from a separate nest. This resulted in a standardized clutch size of 2 and ensured hatching synchrony, thus preventing survival differences due to hatching asynchrony within and among clutches and ruling out effects of parental genetic quality on chick development. Upon hatching, nestlings were individually marked with colored tape, and down feathers were collected for molecular sexing. During each visit, chick body mass (to the nearest g) and head length (to the nearest mm) were measured. At 30 days of age, the right innermost primary feather (P1) of each monitored chick was plucked and stored for isotope analysis. In total, 15 female and 22 male chicks were sampled in 2016, and 5 females and 5 males were sampled in 2017.

Aviary experiment

To avoid unnecessary disturbance of the Zeebrugge colony, Lesser Black-backed Gull eggs for the aviary experiment were collected from a nearby colony in the port of Ostend, Belgium (51°13′15″N2°56′27″E) on 29 May 2016. We consistently collected one of the two largest eggs (i.e., first or second laid egg) from haphazardly selected 3-egg clutches. Chicks hatching from larger eggs have a higher chance of successfully fledging at 30–40 days after hatching (Del Hoyo et al., 2018). To avoid laying date effects and ensure synchronized hatching, only pipping eggs were sampled. All collected eggs were placed in an incubator (temperature 37.5 °C, humidity 62%) in a nearby aviary facility of Ghent University hosted in the Ostend Bird Rescue Center (VOC Oostende) until they were fully hatched. Of 34 collected eggs, 32 hatched between 30 May and 4 June 2016, 20 hatchlings (10 females and 10 males) were assigned to the current study. At hatching, chicks were randomly assigned to a diet of either ground whole adult chicken (Gallus gallus domesticus) or ground whole Sprat (Sprattus sprattus), with 10 chicks per diet treatment. These food types were chosen based on their macronutrient composition, which was representative of the food items found in regurgitates of chicks at the colony of Zeebrugge (see Table A1). We compensated for the loss of vitamins C and B1 due to prolonged deep-freeze storage by adding them to the thawed food. Food items were thawed at room temperature 6 to 8 h before feeding the chicks. On the first day after hatching, food was offered at five different times using tweezers. Afterwards, and during the remaining duration of the experiment, food was available ad libitum on a plate, and refreshed every 3 h between 9 AM and 9 PM. Chicks were measured every 5 days from the day of hatching until 40 days of age. We measured body mass (to the nearest g), total head length (head and bill, to the nearest 0.1 mm) and length of the external part of the P1 feather (measured to the nearest mm with a digital caliper). At 40 days of age, the P1 feather of all 20 chicks was plucked and stored for stable isotope analysis.

Sexing

All field and aviary chicks were molecularly sexed using DNA samples extracted from down feathers. A polymerase chain reaction (PCR) was performed using Fridolfsson & Ellegren (1999) 2550F/2718R primers, and sex was subsequently determined by electrophoresis of these samples.

Diet assessment

We estimated the terrestrial versus marine component in the diet of field chicks by means of carbon and nitrogen stable isotope analysis of feather samples, a technique that has proven efficient in assigning the proportion of assimilated diet components to a restricted number of sources in gulls, with marine food being characterized by higher values of δ¹³C and δ¹⁵N than terrestrial food (Moreno et al., 2010; Steenweg, Ronconi & Leonard, 2011; Weiser & Powell, 2011). First, we assessed the local diet of Lesser Black-backed Gulls based on an analysis of pellets and regurgitations of individuals breeding at Zeebrugge (Research Institute for Nature and Forest, 2006–2017; See Fig. SA1), supplemented by literature data (Camphuysen, 2011; Garthe et al., 2016). Based on this information, we collected and analyzed three samples of locally sourced swimming crabs (Subfam. Polybiinae), chicken meat, fish (Cod Gadus morhua, Sole Solea solea, Plaice Pleuronectes platessa), earthworms (Fam. Lumbricidae) and fried potatoes (Solanum tuberosum). All food samples were freeze dried, ground and subjected to accelerated solvent lipid extraction as described in Bodin et al. (2009). Second, we cut P1 feathers of free-ranging chicks into 3 sections, each corresponding to a 10-day period. The length of feather sections in field chicks was estimated as the average length of P1 feathers in aviary chicks at 10, 20 and 30 days after hatching (values in Fig. SA1). Only feather vellum was used in the analysis of stable isotope ratios. All feather sections were cleaned for 5 min in an ultrasonic bath, left 12 h in a 2:1 chloroform –methanol wash, and oven-dried at 50 °C for 24 h. After this, food and feather samples were finely cut and placed in tin cups. Third, isotopic ratios were obtained by mass spectrometry at the Department of Applied Analytical and Physical Chemistry of Ghent University. Isotope ratios are reported in per mil (‰) using standard delta notation: ${\displaystyle \delta =\left[\left(\frac{Rsample}{Rstandard}\right)-1\right]\times 1000}$ where R represents the 13C/12C or 15N/14N ratio. Standards were Vienna Pee Dee Belemnite for carbon or air N₂ for nitrogen, respectively. Fourth, stable isotope signatures in feathers were corrected for tissue fractionation by means of trophic enrichment factors (TEFs). TEFs were calculated for carbon and nitrogen (δ¹³C and δ¹⁵N) between fish and chicken fed to aviary chicks, and the P1 feathers of these chicks, as in Hobson & Clark (1992), following the formula: δX = δa − δd, where δa = stable isotope composition of feather vellum tissue and δd = stable isotope composition of the diet. TEF for a given isotope was then estimated as the average of individual δX values for that isotope (see Table SA2 for the obtained values).

Next, the proportion of terrestrial food in the chicks’ diet was estimated based on these carbon and nitrogen isotopic ratios and TEFs using a Bayesian stable-isotope mixing model (package MixSIAR, (Stock et al., 2018; R Core Team, 2018). Models were fitted using the Markov-chain Monte-Carlo algorithm, simulating 3 chains over 1000000 time steps. Model convergence was assessed by means of the Gelman–Rubin (Gelman & Rubin, 1992) and Geweke (1992) diagnostics.

Data analysis

The estimated proportion of terrestrial food in the diet of field chicks was compared between three age periods (0–10, 10–20, and 20–30 days after hatching) and between sexes by means of a beta regression with identity link (Ferrari & Cribari-Neto, 2004) using R package betareg (Cribari-Neto & Zeileis, 2010). Significance of effects of the tested variables and their interaction were assessed with an analysis of deviance using Chi-squared tests. Following Camphuysen (2013), a 3-parameter logistic growth curve was fitted to the body mass and total head length of each chick, from both field and aviary, approximating by least-squares the parameters of the logistic function: ${\displaystyle y=\frac{a}{1+b{e}^{-kt}}}$ where y is the body mass (g) or head length (mm), a is the corresponding upper asymptote, b the body mass or head length at the point of inflection, k is the growth rate (days⁻¹) and t is the number of days since hatching. The obtained upper asymptote and growth rate values were analysed by multiple regression with sex and the estimated proportion of terrestrial food in diet (field chicks) or diet treatment (aviary chicks), as well as their interactions, as explanatory variables. Significance of effects of the explanatory variables and their interactions were assessed by means of F-tests. Body condition was approximated using residual body mass, i.e., the residual values of a linear regression of body mass on total head length (Reist, 1985). Residual body mass values were averaged per individual within each age period. These were analysed by means of linear mixed effects models, using R package nlme (Pinheiro et al., 2018), with sex, age period and the proportion of terrestrial food in diet (field chicks) or the diet treatment (aviary chicks), and all their possible two- and three-way interactions, as explanatory variables. Chick identity was included as a random intercept.

Models were fitted using the Restricted Maximum Likelihood (REML) estimation to reduce bias in parameter estimations (Harville, 1977). Model residual normality and homoscedasticity were assessed respectively by means of a Shapiro–Wilk normality test (Shapiro & Wilk, 1965) and Breusch-Pagan test (Breusch & Pagan, 1979). Significance levels of all tests were set at 5%. Only significant interactions were retained, while main effects were always retained to avoid parameter estimation bias (Whittingham et al., 2006).

Ethical approval

All applicable international, national, and institutional guidelines for the care and use of animals were followed during the aviary experiment (Ghent University Ethical Committee, project EC number 2015-017) as well as in the field study (University of Antwerp Ethical Committee for Animal Experiments, project EC number 2013-73). Additionally, all procedures performed in the aviary were in accordance with the regulations of the institution at which the studies were conducted (VOC Oostende).

Variation in diet composition of field chicks

Estimated proportions of terrestrial food in the chick diet ranged from 4% to 80% (Fig. 1). The estimated proportions were concentrated at the more marine values (mean 30%, median 24% terrestrial component), implying that most field chicks were raised on a predominantly marine diet. The proportion of terrestrial food in the chick diet did not vary significantly in relation to chick age or sex (Table 1).

Effects of diet composition on chick development

Asymptotic body mass of both field and aviary chicks was significantly related to their diet composition (Table 2). In field chicks, asymptotic body mass was negatively related to the proportion of terrestrial food (Fig. 2A), while aviary chicks reached a higher body mass when raised on a terrestrial diet (Fig. 2B). Males attained a higher body mass than females in both field and aviary chicks, with similar effect sizes in both environments. Growth rates in field chicks (Fig. 2C) did not vary with diet or sex. In contrast, aviary chicks gained body mass faster when raised on a terrestrial diet (Fig. 2D). Male aviary chicks showed a trend toward faster body mass gain than females, which was however not found to be statistically significant (Table 2).

Asymptotic head length of field chicks was inversely related to the proportion of terrestrial food in their diet, while in aviary chicks, diet treatment had no effect (Table 2). Male chicks attained a larger head length than females in both the field and aviary (Figs. 3A, 3B). Growth rates for head length in field and aviary chicks were not significantly related to their diet (Table 2), and sex differences in growth rates of head length were not statistically significant in the field (Fig. 3C) or in the aviary (Fig. 3D).

Effects of diet composition on body condition

For field chicks, the retained model for residual body mass contained an interaction between the proportion of terrestrial food in the diet and growth period (Table 3). During the first 10 days of growth (Fig. 4A), residual body mass did not vary with the proportion of terrestrial food, while in the second and third 10-day period (Figs. 4B, 4C), it decreased with increasing proportion of terrestrial food. In aviary chicks, males showed higher residual body mass than females (Figs. 4D, 4E, 4F), but no evidence was found for diet or age effects.