The growth rate of Black-tailed Gull chicks is negatively related to total mercury of female parents on Kabushima (Kabu Island), Japan

,


INTRODUCTION
Increases in anthropogenic mercury (Hg) emissions into the environment can lead to potential adverse health and reproductive effects on wildlife (Wolfe et al. 1998, Burger and Gochfeld 2002, Scheuhammer et al. 2007, Tan et al. 2009). As many seabirds are long-lived marine top predators, they accumulate high levels of Hg in their bodies from their prey (Wiener et al. 2002). High concentrations of Hg are known to have adverse effects on the reproductive success of seabirds (Burger and Gochfeld 2002). Even low levels of Hg contamination disrupt avian breeding and behavior through abnormal alteration of the endocrine system, causing some birds to skip breeding completely, or increase the number of incubation recesses, and decrease the number of chicks successfully raised (Tartu et al. 2013, Whitney and Cristol 2018, Hartman et al. 2019, Mills et al. 2020. Therefore, Hg at low levels may have lifetime fitness consequences. However, little is known about the effect of low Hg exposure on chick quality, which is affected by parental feeding behavior. To better understand the impact of such exposure, it is necessary to observe detailed parental breeding performance, such as chick growth, under field conditions. Marine organisms living in the higher Hg concentration region of the North Pacific Ocean are expected to experience elevated Hg contamination due to bioaccumulation and biomagnification of environmental Hg (Furness and Camphuysen 1997, Burger and Gochfeld 2002, Niizuma et al. 2021. Anthropogenic emissions of Hg in Asia have increased over the past several decades, and are still increasing, due to increased economic activity in East Asia and increased atmospheric Hg deposition in the Pacific Ocean (Pacyna et al. 2006, Driscoll et al. 2013, Lamborg et al. 2014, UNEP 2018. The surface water Hg concentration in the North Pacific Ocean is higher on the western side, especially off Japan's Sanriku Coast, located in northeastern Japan (Fig. 1), than in other waters (Sunderland et al. 2009). In fact, Black-tailed Gulls (Larus crassirostris) breeding in the colony on Kabushima (Kabu Island), at the northernmost point of the Sanriku Coast, accumulate total Hg in their body tissues at higher levels than other Laridae species (Goto et al. 2018). Our concern is that the current level of Hg contamination in Blacktailed Gulls could be negatively related to their breeding success, especially chick growth, through abnormal parental behavior due to mercury pollution.
In addition to the effects of Hg, individual differences in diets, which are assessed by nitrogen stable isotope values (δ 15 N), may also influence their breeding performance (González-Medina et al. 2018). δ 15 N represents the trophic level concentrated in both prey and biomarkers, allowing us to understand individual dietary differences (Wada et al. 1987, Hobson andWelch 1992). The effects of Hg on breeding performance, which is under the control of the parental diet, is of great interest.
During a two-year field study of Black-tailed Gulls, we investigated the effects of Hg concentrations in female parents on their breeding performance. We conducted the study over two years to account for environmental differences between years. In this study, we collected blood samples from breeding females to analyze Hg concentrations and δ 15 N ratios. We examined the effect of dietary differences on chick growth through the δ 15 N of their mothers. We also measured the clutch size, hatching success, fledgling success, and chick growth of each of the breeding females. We predicted that females with higher mercury concentrations would breed less successfully than those with lower concentrations, or that under current levels of Hg contamination of females, no negative effects of Hg on breeding success would be detected. We also predicted that chicks of female parents with higher Hg concentrations would grow more slowly than those of female parents with lower Hg concentrations.
In order to investigate the effect of Hg contamination on breeding success, we studied 16 nests in 2018 and 23 in 2019. We used snare traps to catch 16 incubating females in 2018 and 23 in 2019 from their nests and investigated their breeding success. As male and female Black-tailed Gulls have identical plumage, but differ in size (Chochi et al. 2002), the smaller females were identified in comparison with their larger mates during the period when both parents remained at their nest after egg-laying. In both 2018 and 2019, we were able to discern 10 identical breeding females (i.e., 10 nests of same female parent) and undertake a study on their reproductive success. We limited our pool of female parents to a minimum so as to minimize disturbance at the colony and to reduce any impact on their breeding attempts.
We took 1 ml blood samples from the wing veins of the gulls, then released them on their nests. We placed each blood sample in a heparinized tube. After centrifugation (5 min, 6200 rpm), we stored the blood cells and the serum in a 0.5 ml plastic screw cap vial with an O-ring (Asahi Techno Glass Co.) at -20 °C until the analysis.

Reproductive Success
Study plots were visited daily (between 8:00-12:00 h local time) from the start of laying on 28 April until the last chicks fledged on 30 June in 2018 and 5 July 2019, respectively. Study nests were indicated with pegs (30 cm tall, 5 cm wide), and eggs were marked using a waterproof felt-tip pen. Clutch size was defined as the number of eggs per nest. When egg remnants were found around nests or eggs disappeared, they were assumed to have been predated. Eggs that failed to hatch within seven days of the first chick in a clutch were recorded as unhatched. Hatching success was defined as the number of eggs hatched per nest. After hatching, a 30 cm wire mesh (φ 8 mm) was placed around each nest plot to prevent chicks from escaping. In 2019, chicks were ringed (with colored plastic rings) and weighed every five days from one day old to 30 days old, using a spring balance. Chick growth rate (g/day) was defined as the slope of the linear regression between 4 and 21 days of age because the chicks of large gull species grow almost linearly (Watanuki 1992). The chicks were assumed to fledge at 30 days of age. Fledging success was defined as the number of nests with one or more fledglings over the total number of nests studied.

Hg and Nitrogen Stable Isotope Analysis
Blood cells were freeze-dried to constant mass for 48 h and then ground to a fine powder using an electric crusher (TK-AM5, Titec) operating at liquid nitrogen temperature (Tomita et al. 2015). Concentrations of total mercury in dried blood cells were determined using a Direct Thermal Decomposition Mercury Analyzer (MA-3000; Nippon Instruments). Mercury atomic absorption standard solution (1005 mg/L, Kanto Kagaku Co., Ltd.) was diluted 1000 times and 100,000 times with L cysteine solution to prepare the standard solution. After preparation of the calibration standards using the standard solution, total mercury concentration was measured by thermal decomposition.
Dried blood cells were sealed in tin capsules for combustion. The nitrogen-stable isotopic values of dried blood cells were measured using a gas-source isotopic values mass spectrometer (ANCA-GLS and Hydra 20-20, Sercon Ltd., UK). The nitrogen isotope ratio of L-alanine M9R2064 (AZ100: δ 15 N-Air = 1.79 ± 0.2%) was used as a working standard. All samples were run in duplicate. Isotope ratios (δ 15 N) were shown as deviations from atmospheric nitrogen in parts per thousand (‰, Coplen 2011): where R is the isotopic ratio ( 15 N/ 14 N). Analytical reproducibility during the overall measurements was better than ±0.2%.

Statistical Analysis
All statistics were performed using R. 4.1.2 (R Development Core Team, Vienna, Austria). To consider whether to pool data, we initially ensured the inter-annual difference of total Hg concentration, δ 15 N, and reproductive parameters (clutch size, number of hatchlings per nest, and fledging success). We used generalized linear mixed models (GLMM, glmer, or lmer functions in the lme4 package; Bates et al. 2015) to test for between-year differences in total Hg concentration in blood, δ 15 N as the parental trophic level, clutch size, the number of hatchlings per nest, and fledging success, where years were explanatory variables, and individual identity was a random effect. We designated Gaussian (δ 15 N), using an identity link function, Gamma (Hg concentration) and Negative binomial (clutch size and the number of eggs hatching) using a log link function, and Binomial (fledging success: 0 = no fledgling, 1 = one or more chicks fledged) using a logit link function as the error distribution. In the GLMM for the number of hatchlings per nest, clutch size was offset after logarithmic transformation. For analyzing interannual differences, intra-individual changes were calculated separately from the overall dataset.
GLMMs were also used to investigate whether variation in reproductive parameters (clutch size, number of hatchlings per nest, and fledging success) was related to total Hg concentration in blood or δ 15 N, with each reproductive parameter designated as an explanatory variable, while individual identity and years were considered as random effects to account for the repeated identities of females between years. We used Gaussian error distribution (δ 15 N) with an identity link function and Gamma error distribution (Hg concentration) with a log link function.
To examine whether total Hg concentration in blood and/or δ 15 N influence offspring growth rates, we used a GLMM (Gaussian error distribution with identity link function) with chick growth rate and number of chicks assumed to have fledged in 2019 as response variables, total Hg concentration in blood and δ 15 N as explanatory variables, and nest identity as a random effect.
We used the likelihood-ratio test to assess the significance of each explanatory variable in all models using the Anova function in the car package (Fox and Weisberg 2011). The results were considered significant when P < 0.05.

RESULTS
Total Hg concentrations in females were 2.48 μg/g dry weight ± 0.55 (1.28-3.28) in 2018 and 2.42 μg/g dry weight ± 0.55 (1.87-3.29) in 2019. The difference between years was not statistically significant; however, the δ 15 N of female parents was significantly higher in 2019 than in 2018 (Table 1). None of the reproductive parameters (i.e., clutch size, number of hatchlings per nest, and fledging success) of female Black-tailed Gulls differed significantly between 2018 and 2019 except for fledgling success in nests examined in two consecutive years ( Table 1).

DISCUSSION
In our two-year field study, we found that the δ 15 Ns of female parents was significantly higher in 2019 than in 2018. The difference in δ 15 Ns, suggesting a dietary difference of female parents between the years noted in previous studies (Deguchi et al. 2004, Kwon et al. 2013, could be related to differences in fledging success between the two years. However, despite the fact that diets of the female parents differed between the two years, Hg exposure was not significantly related to any of the reproductive parameters of clutch size, hatching success, or nest success. This result was different from one of our initial predictions, that females with higher Hg levels would have lower reproductive success than females with lower levels. Adverse effects of Hg on behavior, reproduction, physiology, and health in birds generally tend to occur at 1.0 μg/g wet weight of bloodequivalent total Hg concentration, and more substantial impairments to health and reproduction occur at approximately 2.0 μg/g wet weight (Ackerman et al. 2016). Converting published Hg toxicity benchmarks among birds into blood-equivalent total Hg concentrations based on Ackerman et al. (2016), the lowest observed effect level of blood-equivalent total Hg concentration was 0.2 μg/g wet weight. If avian blood is assumed to contain 80% water, as indicated by Bell (1957), our Black-tailed Gulls would accumulate 0.256-0.656 μg/g wet weight total Hg in their blood, which is below the threshold for direct effects on breeding success, indicating that they should show normal breeding behavior to their chicks under the current level of Hg exposure. However, Hg exposure had a significant negative relationship with chick growth rate. This may be because the sensitivity to lower levels of Hg contamination and their impacts on reproductive success may differ among species and sexes of seabirds. For example, male Grey-headed Albatrosses (Thalassarche chrysostome) that breed successfully have lower levels of Hg in their contour feathers than those that failed. Females of the same species that breed successfully and failed have similar levels of total Hg to successful males, thus Hg contamination does not relate to the breeding success of females (Mills et al. 2020). Wandering Albatrosses (Diomedea exulans) that had higher concentrations of Hg than Grey-headed Albatrosses, showed no significant effect of Hg on hatching success or fledging success . Hg concentrations in both sexes of the Antarctic Petrel (Thalassoica antarctica) (0.42-2.71 μg/g dry weight in blood) were not found to be related to hatching success (Carravieri et al. 2018). Black-legged Kittiwakes (Rissa tridactyla) breeding in Svalbard, Norway, had similar Hg levels in their blood to our gulls (Tartu et al. 2013(Tartu et al. , 2016. Male and female kittiwakes that skipped breeding had higher total Hg blood concentrations than those that bred (Tartu et al. 2013), and male kittiwakes that raised only one chick successfully had higher total Hg concentrations in their blood than those that raised two chicks successfully, but this difference did not apply to females (Tartu et al. 2016). Although we did not measure Hg concentration in male gulls, the lack of a relation to Hg on the reproductive success of female Black-tailed Gulls was similar to that observed in two species of albatrosses and in Black-legged Kittiwakes. It is not clear why the effects of Hg exposure on reproduction differ between sexes or species; however, the differences may be related to differences between sexes or among species in their foraging areas, trip lengths, and trip frequencies. Chick growth was not related to the δ 15 N of female gulls in this study. This result suggests that the provisioning behavior of female parents was not related to their own diet, which is a different result from that found by González-Medina et al. (2018). Among seabirds, variation in parental provisioning rates influences the growth rate of chicks (Morbey and Ydenberg 1997), so the negative relationship of the Hg exposure of females to chick growth rate suggests that parents with higher Hg concentrations would provide less food to their chicks than parents with lower concentrations. The absence of a consistent relationship between Fig. 2. Relationships between total mercury (Hg) concentrations in blood, nitrogen stable isotope (δ15N) ratios, and reproductive performance of female Black-tailed Gulls (Larus crassirostris) in 2018 and 2019 (combined). Error bars represent standard deviation. Fig. 3. Relationships between chick growth rate and (a) total mercury (Hg) concentrations in blood and (b) nitrogen stable isotope (δ15N) ratios. Circles represent raw counts and dashed lines and shaded areas represent the GLMM-estimated results ± 95% confidence intervals.
heavy metal contamination and breeding performance may arise from the disparate life history cycles and foraging habitats of the affected species. Seabirds that inhabit extensive oceanic domains may exhibit disparities in their intake of contaminants, owing to their proximity to human settlements and variations in foraging behavior that are contingent on sex and age (Hindell et al. 1999).
We cannot exclude the possibility that low-level mercury contamination may have affected the hormonal balance of the female Black-tailed Gull parents. All of the female parents in this study had higher levels than Ackerman et al.'s (2016) lowest observed effect level of blood-equivalent total Hg. The values that we observed may cause alterations to parental behavior through disruption of hormone secretions (Burger and Gochfeld 2002), because Hg compounds absorbed into the body are distributed and accumulated in endocrine tissues under lower-level Hg exposure during the early process of accumulation in body tissues (Niizuma et al. 2021). The reproductive performance of Blacklegged Kittiwakes may be impaired by Hg contamination disrupting gonadotropin-releasing hormone input to the pituitary or prolactin hormone (Tartu et al. 2013(Tartu et al. , 2016. Among incubating male Snow Petrels (Pagodroma nivea) with concentrations of Hg in their blood of 0.9-4.0 µg/g dry weight, increasing Hg concentrations were related to decreasing stressinduced prolactin concentrations (Tartu et al. 2015). It is known that prolactin mediates avian parental behavior such as incubation and provisioning (Schoech et al. 1996, Angelier and Chastel 2009, Tartu et al. 2015. Therefore, it is possible that the lower growth rate of Black-tailed Gull chicks observed in this study is a result of females with higher total Hg concentrations taking less care of their offspring because of abnormal prolactin secretion.
In conclusion, while Black-tailed Gulls may not currently experience remarkable fitness consequences at their current level of Hg exposure, our detailed observations have shown for the first time that the chicks of females with higher Hg concentrations grew more slowly than those of females with lower concentrations. Slow growth of young seabird chicks is considered to influence their quality or survival after fledging (Kitaysky et al. 2006). In addition to adverse effect of Hg through altered behavior of their female parents, chicks are also affected in several ways by Hg contamination through intake of foods, such as altered behavior, reduced growth rates, impaired immune function, and decreased survival (Spalding et al. 2000, Kenow et al. 2007, Ackerman et al. 2008, Heddle et al. 2020. Over many generations, any negative effects of Hg contamination on their breeding performance, through their behavioral changes, could be more considerable than at the current level. If anthropogenic Hg emissions into the environment continue at the current level, therefore, Hg contamination may threaten gull numbers in the future, even though the breeding numbers of Black-tailed Gulls are not thought to be declining at colonies around Japan at the moment.