Negligible evidence for detrimental effects of Leucocytozoon infections among Emperor Geese (Anser canagicus) breeding on the Yukon-Kuskokwim Delta, Alaska

Emperor Geese (Anser canagicus) are iconic waterfowl endemic to Alaska and adjacent areas of northeastern Russia that are considered to be near threatened by the International Union for Conservation. This species has been identified as harboring diverse viruses and parasites which have, at times, been associated with disease in other avian taxa. To better assess if disease represents a vulnerability for Emperor Geese breeding on the Yukon-Kuskokwim Delta, Alaska, we evaluated if haemosporidian parasites were associated with decreased mass or survival among adult female nesting birds captured during 2006–2016. Through molecular analyses, we detected genetically diverse Leucocytozoon, Haemoproteus, and Plasmodium parasites in 28%, 1%, and 1% of 607 blood samples screened in triplicate, respectively. Using regression analysis, we found evidence for a small effect of Leucocytozoon infection on the mass of incubating adult female Emperor Geese. The estimated mass of infected individuals was approximately 43 g (95% CI: 20–67 g), or approximately 2%, less than uninfected birds when captured during the second half of incubation (days 11–25). We did not, however, find support for an effect of Leucocytozoon infection on survival of adult female nesting Emperor Geese using a multi-state hidden Markov framework to analyze mark-resight and recapture data. Using parasite mitochondrial DNA cytochrome b sequences, we identified 23 haplotypes among infected Emperor Geese. Leucocytozoon haplotypes clustered into three phylogenetically supported clades designated as ‘L. simondi clade A’, ‘L. simondi clade B’, and ‘other Leucocytozoon’. We did not find evidence that parasites assigned to any of these clades were associated with differential mass measures among nesting adult female Emperor Geese. Collectively, our results provide negligible evidence for Leucocytozoon parasites as causing detrimental effects to adult female Emperor Geese breeding on the Yukon-Kuskokwim Delta.


Introduction
Emperor Geese (Anser canagicus) have a geographically limited distribution within coastal western Alaska and northeastern Russia, and adjacent waters (Schmutz et al., 2020). This species is less abundant than other North American geese, with an estimated population during 2007-2014 of approximately 130,000 birds (Pacific Flyway Council 2016). Emperor Geese are a culturally important subsistence resource in western and southwestern Alaska and are subject to limited sport harvest (Pacific Flyway Council 2016). The conservation status of Emperor Geese is considered 'near threatened' by the International Union for Conservation given the species' limited geographic range, relatively small population size, and harvest pressure (BirdLife International 2016).
Though not well understood, disease may also represent a potential vulnerability for Emperor Geese. For example, diverse coronaviruses (Muradrasoli et al., 2010), influenza A viruses (Reeves et al. , 2018Ramey et al., 2019), paramyxoviruses , and haemosporidian parasites (Hollmén et al., 1998;Ramey et al., 2014;Buchheit et al., 2021a) have previously been identified in Emperor Geese inhabiting Alaska and Russia. Various strains and lineages of these viral and parasitic agents have previously been associated with disease in wild and domestic birds (Valkiunas 2004;Swayne 2009;Dimitrov et al., 2016;Lee et al., 2017;Miłek and Blicharz-Domańska, 2018). Though none of these infectious agents has been associated with clinical disease in Emperor Geese, potential detrimental subclinical effects of infection have not yet been evaluated in these birds.
In this study, we sought to assess if haemosporidian parasites may be associated with decreased mass or survival among breeding adult female Emperor Geese, suggestive of detrimental effects of infection on host fitness and sub-clinical disease. Specifically, we aimed to: (1) estimate the detection probability and prevalence of haemosporidian parasites among Emperor Geese captured on the Yukon-Kuskokwim Delta during 2006-2016; (2) assess potential impacts of haemosporidian infection on the mass of adult female Emperor Geese during incubation as well as the inter-annual survival of infected birds of this same age/sex cohort; and (3) explore the genetic diversity of haemosporidians infecting Emperor Geese relative to potential associations with fitness costs of infection. We conducted our investigation based upon previous reports for associations between haemosporidian infection in waterfowl and possible detrimental fitness effects including anemia (Desser and Ryckman 1976), reduced body condition (Shutler et al., 1999;Meixell et al., 2016;Fleskes et al., 2017, Merrill et al., 2018, and decreased survival (Herman et al., 1975;Sorci and Møller 1997).

Study area and sampling
Emperor Geese were captured at breeding sites adjacent to the Manokinak River (approx. 61.2 • N, 165.1 • W) on the Yukon-Kuskokwim Delta of Alaska during nest incubation in May and June 2006-2016 and during brood-rearing in August 2014. Incubating females were caught on nests utilizing bow or mist nets, and molting flightless adults and unflighted goslings were herded into corrals and sampled during brood drives. Capture and sampling procedures were conducted under U.S. Geological Survey (USGS) federal bird banding permit #20022 and reviewed, as applicable, by the Animal Care and Use Committee at the USGS Alaska Science Center . Whole blood samples were collected from captured individuals using jugular venipuncture and stored in Longmire buffer (Longmire et al., 1988). Captured geese were weighed using hanging scales or electronic balances and differentiated by sex through cloacal examination. The nest initiation date for birds captured at nest sites was estimated via egg floating and/or candling (Fischer et al., 2017).

DNA extraction and molecular identification of haemosporidian parasites
Methodology for molecular assessment of haemosporidian infection generally followed Ramey et al. (2012) with few exceptions as indicated in the brief description that follows. DNA was extracted from blood samples using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, USA) per the manufacturer's protocol. Extracted DNA was screened in triplicate to assess the presence of haemosporidians using a nested PCR as described by Hellgren et al. (2004). A minimum of one negative control was incorporated every 24 wells in each set of PCR reactions. Amplified products were visualized on 0.8% agarose gels stained with Gel Red Nucleic Acid Gel Stain (Biotium, Hayward, USA). Products appearing to correspond to our 479 base pair (bp) mitochondrial DNA (mtDNA) cytochrome b (cyt b) target were treated with ExoSap-IT (USB Inc., Cleveland, USA) and sequenced with identical primers used for PCR with BigDye Terminator version 3.1 mix (Applied Biosystems, Foster City, USA) on an Applied Biosystems 3730xl automated DNA sequencer (Applied Biosystems, Foster City, USA). Sequence data were cleaned and edited using Sequencher version 5.1 (Gene Codes Corp., Ann Arbor, USA).
Parasite infections were characterized to genus (Plasmodium, Haemoproteus, or Leucocytozoon) using the nucleotide BLAST function (Altschul et al., 1997) available through the National Center for Biotechnology Information (NCBI). Assignment of sequences to parasite genus was based on the top NCBI BLAST result with a minimum max identity score of 90%. A sample was considered positive for a parasite infection if any one of the triplicate PCR reactions resulted in a bi-directional double-stranded mtDNA cyt b target product that was verified through genetic sequencing. Samples from which only single stranded or otherwise ambiguous molecular products were obtained were considered haemosporidian negative. Samples were classified as co-infected with parasites of more than one genetic lineage if the sequenced and cleaned target product for one or more nested PCR runs yielded > one ambiguous nucleotide or if > one mtDNA cyt b haplotype was inferred through resolution with SeaView v4.3.5 (Gouy et al., 2010). Only sequences with ≤ one ambiguous nucleotide within a mtDNA cyt b haplotype were included in downstream genetic analyses.
To confirm the presence of quality DNA within extracted samples, a 481 bp fragment of the mtDNA cytochrome oxidase I (COI) gene of the Emperor Goose host was amplified using 2 μl of template DNA, primers specifically designed for geese (Ramey et al., 2014), and the same proportion of reagents as previously reported for the nested PCR (Hellgren et al., 2004). Thermalcycling conditions followed Kerr et al. (2007).

Occupancy modeling and assessment of detrimental effects of parasite infection
To estimate detection probability (ρ 1 ) and prevalence (ψ 1 ), we analyzed results from triplicate screening of each DNA extraction for haemosporidians in an occupancy modeling framework. More specifically, we assessed variation in ρ 1 and ψ 1 with a parameter for sample year using occupancy modeling procedures in program MARK (version 9.0; White and Burnham 1999). For this analysis, we included information derived from all blood samples (n = 607), including those from hatch-year (n = 52) and male (n = 37; 28 of which were also hatch-year) birds sampled as part of brood drives in 2014. We limited our estimates of detection probability and prevalence to parasites of the genus Leucocytozoon given few detections for Plasmodium and Haemoproteus parasites in our data set (see Results). We used Fletcher's ĉ to check for overdispersion in the data and assess the fit of our most highly parameterized model (ρ 1 (Year) ψ 1 (Year)) to the data. Our estimate of Fletcher's ĉ was 2.22, and we adjusted ĉ in MARK accordingly. We assessed model performance using quasi-Akaike Information Criterion adjusted for small sample sizes (QAICc).
We next performed a regression analysis to assess the potential impacts of haemosporidian infection on the mass of female Emperor Geese during incubation. Specifically, we assessed whether Leucocytozoon parasite infections (i.e., those comprising >90% of our dataset; see Results) were associated with lower mass during incubation for female Emperor Geese. For this analysis, we constrained our data to mass measures after the 10th day of incubation (when field crews were present), and observations for birds with mass measures ≥1200 g but ≤2100 g (to remove potentially spurious measures). Constraining data in this way resulted in the omission of six mass measures, which we considered outliers relative to our data set and prior reports (Thompson and Raveling 1987;Schmutz et al., 2020), and which may have been influenced by observer error in the field (e.g., failing to tare electronic balances prior to measuring mass). For the regression analysis, we included data on the infection status of 337 individual female geese using triplicate parasite screening results for 493 blood samples. Numerous (n = 113) individuals were captured and bled in multiple years (≤6), and infection status was assessed from each sample. We included mass measures for the same individual from different years in our final data set, relaxing the assumption of independence of measures in modeling efforts, given inter-annual variability in mass. We contrasted six models in our regression analysis to explain variability in the mass of incubating female Emperor Geese incorporating parameters for day of incubation (Inc), Leucocytozoon infection status (Leu), and year. We assessed model performance by applying an information-theoretic approach (Burnham and Anderson 2002) and using the Akaike Information Criterion adjusted for small sample size (AICc).
To evaluate the impact of haemosporidian infection on adult female Emperor Goose survival and resighting probabilities, we used markresight and recapture data for 424 tagged adult females from the same 11 year interval assessed in regression analysis (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016). Differences in sample size between this survival analysis and the previously described regression evaluating possible effects of infection on mass were on account of the inclusion of numerous unsampled birds of unknown infection status in the survival analysis and the exclusion from the regression analysis of some tagged individuals from which mass data was not obtained. For the survival analysis, we used a multi-state hidden Markov framework (Conn and Cooch 2009) in Program MARK to account for state uncertainty where individually marked Emperor Geese were resighted but no infection status data were available (i.e., resighted alive but not recaptured). Parameters in these models included survival probability (ϕ), resight probability (ρ 2 ), state transition probability (ψ 2 ), probability of an event at first capture (π), and the probability of accurate state assessment (δ). We used two states in our analyses to account for infection status of individuals (Leucocytozoon infected and uninfected), and one event to account for detection of a marked individual where the infection status was uncertain. We evaluated a candidate set of eight models without covariates. We evaluated potential change in survival over time, but small sample sizes prevented us from evaluating temporal variation between parameters and infection states. We used a combination of logit and mlogit link functions to bound parameter estimates. As with the regression analysis, we constrained our data to parasites of the genus Leucocytozoon and assessed model performance using AICc.

Genetic and phylogenetic analyses
To identify the genetic diversity of haemosporidians infecting Emperor Geese, and to explore potential associations between parasite genotype and fitness metrics for incubating female birds, we used mtDNA cyt b sequences, a marker widely used to investigate haemosporidians in birds (Pacheco et al., 2018). Specifically, we estimated the frequency of distinct haplotypes, compared them to those available on public databases, and created a median-joining minimum spanning network for the haplotypes of the most common parasite genus detected (i.e., Leucocytozoon; see Results). To first identify the number of parasite mtDNA cyt b haplotypes among our sample, we constructed three alignments, one for each Plasmodium, Haemoproteus, and Leucocytozoon sequence generated from our sample collection. We performed alignments using ClustalX v2.0.12 and Muscle as implemented in SeaView v4.3.5 (Gouy et al., 2010). Alignments for Plasmodium, Haemoproteus, and Leucocytozoon sequences were 393 base pairs (bp), 320 bp, and 290 bp, respectively. Identical sequences within each alignment were considered to represent the same parasite haplotypes. Next, each parasite haplotype was compared to those reported on the MalAvi (Bensch et al., 2009) and GenBank (Benson et al., 2012) public databases (accessed 25-31 March 2021) to identify identical or similar sequences previously reported for haemosporidians infecting wild birds. Subsequently, to depict the relative frequency of the most commonly detected parasites in sample collection and their comparative genetic distances to one another, we estimated a median-joining minimum spanning network for Leucocytozoon haplotypes using the software Network, version 10.2.0.0 (Bandelt et al., 1999) and calculated genetic distance using the p-distance method as implemented in MEGA version 7.0.14 (Kumar et al., 2016).
To infer the genetic relationships among the observed haplotypes in our sample and those closely related as identified through our query of MalAvi and Genbank databases, we estimated phylogenetic trees using Bayesian methods as implemented in MrBayes v3.2.6 (Ronquist and Huelsenbeck 2003). For Plasmodium, we used genetic information for one representative sequence for each parasite haplotype we detected in Emperor Geese, sequences from well-known Plasmodium morphospecies, and information for L. caulleryi and Haemoproteus spp. (as outgroups) in our 393 bp alignment. For Haemoproteus, we analyzed each haplotype detected in Emperor Geese with four previously reported lineages of unknown morphospecies and 16 sequences from previously identified morphospecies that emerged from our query of public databases. Seven sequences of Leucocytozoon or Haemoproteus spp. were used as an outgroup in our 320 bp alignment. For Leucocytozoon, our 320 bp alignment was comprised of one representative sequence for each distinct parasite haplotype we detected in this study, sequences from 15 previously reported lineages of unidentified Leucocytozoon species, and information for 13 previously identified Leucocytozoon morphospecies. The sequences of Leucocytozoon caulleryi and Haemoproteus spp. were used as outgroups. In all cases, the phylogenetic trees were estimated using the Bayesian method with the default priors and the best model that fit the data (general time-reversible model with gamma-distributed substitution rates and a proportion of invariant sites, GTR +Γ + I). This model was selected as it had the lowest Bayesian Information Criterion (BIC) score, as estimated by MEGA v7.0.14 (Kumar et al., 2016). The posterior probability was inferred for the nodes by sampling every 1000 generations from two independent chains lasting 3 × 10 6 Markov Chain Monte Carlo (MCMC) steps. Chains were assumed to have converged once the average S.D. of the posterior probability was <0.01 and the value of the potential scale reduction factor (PSRF) was between 1.00 and 1.02 (Ronquist and Huelsenbeck 2003). Once convergence was reached, 25% of the samples were discarded as burn-in.

Assessment of effects of parasite infection on mass incorporating genetic information
To explore potential associations between genetic lineages of parasites and possible detrimental fitness effects, we related the genetic haplotypes for Leucocytozoon infected incubating female Emperor Geese to measures of mass used in our regression analysis post hoc. In particular, we sought to identify if groups of phylogenetically related parasite haplotypes were consistently associated with lower mass measures. We characterized Leucocytozoon infections among nesting female birds into three groups based upon the haplotypes detected and the results of our phylogenetic analysis: (1) L. simondi clade A (including co-infections with more than one L. simondi clade A haplotype), (2) L. simondi clade B (including co-infections with more than one L. simondi clade B haplotype), or (3) other/mixed Leucocytozoon (including co-infections with haplotypes not exclusively nested with L. simondi clade A or L. simondi clade B). We then conducted a linear regression using the program R (version 4.0.5; R Core Team 2021) to compare mass as a function of incubation day among the geese infected with Leucocytozoon parasites assigned to three phylogenetic clades as previously described. Given the lack of support for an association between survival and Leucocytozoon infection (see Results), we did not explore associations between survival and specific Leucocytozoon clades. Furthermore, we did not assess possible associations between Haemoproteus or Plasmodium parasite haplotypes and mass or survival given the few detections of parasites of these two genera (see Results).
All data supporting conclusions in this product have been made publicly available in an associated data release (Buchheit et al., 2021b) and via GenBank (accession numbers MW885252-MW885274, MW960777-MW960958, and MW990224).

Results
Through the screening of 607 blood samples collected from Emperor Geese inhabiting the Yukon-Kuskokwim Delta, Alaska during the 11year interval from 2006 to 2016, we found consistent evidence for infection with parasites of the genus Leucocytozoon among 11-45% of samples collected in each year (Table 1). In contrast, the molecular detection of Plasmodium and Haemoproteus parasites was more sporadic with the detection of parasites of these genera in samples collected from three and five years, respectively, including no more than 5% of samples collected in any given year (Table 1). The molecular methods used did not provide information regarding the intensity of these haemosporidan infections. Most geese that were sampled in more than one year (82/113 or 73%) remained either negative (n = 62) or positive (n = 20) for haemosporidian parasite infection(s) across years (Buchheit et al., 2021b). However, a smaller proportion of birds (31/113 or 27%) changed status from haemosporidian negative to haemosporidian positive (n = 17), haemosporidian positive to haemosporidian negative (n = 11), or some combination thereof (i.e., changed infection status more than once; n = 3) (Buchheit et al., 2021b).
Through the application of an occupancy modeling approach, we estimated our molecular detection probability for Leucocytozoon parasites to be relatively high at approximately 0.72 for any given PCR run or 0.98 per sample when run in triplicate [1 − (1-0.72) * (1-0.72) * (1-0.72)]. We also found estimated prevalence to be fairly consistent between years (approximately 28%; 95% confidence interval or CI: 25-32%) with limited support for sample year as an informative parameter (Table 2). Given the low prevalence of Plasmodium and Haemoproteus infections, we did not model the detection probability or prevalence of parasites of these genera among our sample of Emperor Geese nor did we assess potential detrimental fitness effects thereof.
Using regression analysis, we found evidence for a small effect of Leucocytozoon infection on the mass of incubating female geese (Table 3). More specifically, our results provided the strongest support for models explaining the mass of incubating female Emperor Geese as a function of incubation day and Leucocytozoon infection status, with weaker evidence for a year effect (Table 3). The most parsimonious model in our candidate set estimated the mass of incubating females infected with Leucocytozoon parasites to be approximately 43 g (95% CI: 20-67 g) less than uninfected females during the second half of incubation (days 11-25; Fig. 1). The rate of loss for both infected and uninfected females was estimated to be 20 g/day (95% CI: 17-23 g/day; Fig. 1).
Using a multi-state hidden Markov framework, we did not find evidence for an effect of Leucocytozoon infection on survival of adult female    Emperor Geese breeding on the Yukon-Kuskokwim Delta (Table 4). Specifically, our top-ranked model (AICc wi = 0.37) did not include parameters for survival differences between infected and uninfected geese (Table 4). Instead, our top model included parameters for resight probability based on infection status and state transition probabilities (e. g., switching from uninfected to infected and vice versa). Furthermore, survival rates from the best-supported model that incorporated From the amplification and genetic sequencing of parasite mtDNA cyt b sequences, we identified 23 haplotypes among Emperor Geese infected with haemosporidian parasites assigned to the genera Plasmodium (n = 2), Haemoproteus (n = 5), and Leucocytozoon (n = 16; Table 5). Thirteen of these haplotypes were identical to those previously identified on public databases, including eight Leucocytozoon haplotypes ( Table 5). All three parasite mtDNA cyt b haplotypes detected among at least ten individuals were of the genus Leucocytozoon, two of which (EMGO11, and EMGO12) were differentiated by a single nucleotide polymorphism (Fig. 2).
In our first phylogenetic analysis, both Plasmodium mtDNA cyt b haplotypes found in Emperor Goose blood samples formed a monophyletic clade with sequences previously reported for Plasmodium circumflexum (posterior probability 1.0), though the topology with a sister clade including sequences for Plasmodium relictum and Plasmodium megaglobularis was not conclusively resolved (posterior probability 0.57; Fig. 3). In our second phylogenetic analysis involving Haemoproteus mtDNA cyt b haplotypes, four parasite sequences from Emperor Goose formed a moderately supported clade (posterior probability 0.62) with Table 4 Model selection results for influences on survival (ϕ), detection probabilities (ρ 2 ), state transition probabilities (ψ 2 ), probability of event at first capture (π), and probability of determining states (δ)     haplotypes previously reported for Haemoproteus sacharovi (from doves) and unidentified Haemoproteus morphospecies detected in waterfowl (Fig. 4). A fifth Haemoproteus mtDNA cyt b haplotype from a single goose sample was nested within a strongly supported clade (posterior probability 1.0) with haplotypes previously reported for Haemoproteus belopolski, Haemoproteus parabelopolskyi, Haemoproteus lanii, and an unidentified Haemoproteus morphospecies (Fig. 4). In our final phylogenetic analysis, we found 16 representative Leucocytozoon mtDNA cyt b haplotypes sequences to cluster into three strongly supported major clades (posterior probabilities >0.94) with those previously reported on public databases (Fig. 5). Nine Leucocytozoon parasite mtDNA sequences from Emperor Goose blood samples formed a clade with a sequence previously identified for L. simondi as well as numerous other sequences previously detected in waterfowl samples from unidentified Leucocytozoon morphospecies. We designated this group of sequences as 'L. simondi clade A' (Fig. 5). Four additional Leucocytozoon parasite mtDNA sequences identified in goose blood samples were part of a sister clade with another sequence previously identified for L. simondi and several other sequences previously detected in waterfowl samples from unidentified Leucocytozoon morphospecies. We designated this group as 'L. simondi clade B' (Fig. 5). The four remaining Leucocytozoon parasite mtDNA sequences from Emperor Goose blood samples were part of a third clade with sequences previously reported for L. fringillinarum and an unidentified Leucocytozoon morphospecies (Fig. 5).
In a post hoc analysis to explore potential associations between genetic lineages of parasites and possible detrimental fitness effects, we did not find evidence that specific groups of phylogenetically related Leucocytozoon parasite haplotypes were consistently associated with differential mass measures among adult female nesting Emperor Geese (Fig. 6). That is, exploratory analyses did not support a difference in mass as a function of incubation day among adult females infected with Leucocytozoon parasites genetically characterized in this study as

Discussion
In this study, we found Emperor Geese inhabiting the Yukon-Kuskokwim Delta of Alaska to be infected with Leucocytozoon parasites in each of 11 sample years with an annual prevalence of approximately 28%. These results are consistent with the microscopic identification of Leucocytozoon simondi in an adult Emperor Goose sampled at this same location in 1996 (Hollmén et al., 1998) as well as two prior molecular assessments estimating the prevalence of Leucocytozoon parasites among adult Emperor Geese at this location to be approximately 20-27% (Ramey et al., 2014;Buchheit et al., 2021a). Furthermore, geese harbored phylogenetically diverse Leucocytozoon parasites, consistent with previously reports for parasites infecting other North American waterfowl species and suggestive of cryptic parasite speciation (Reeves et  Additionally, we detected few infections with Haemoproteus parasites and even fewer infections with Plasmodium parasites. These findings are consistent with the infrequent molecular detection of Haemoproteus and Plasmodium parasites in Emperor Geese from the Yukon-Kuskokwim Delta in prior studies (Ramey et al., 2014;Buchheit et al., 2021a). Similar relative prevalence for parasites of these genera, including a higher prevalence of Leucocytozoon as compared to Haemoproteus and Plasmodium parasites, has been reported in Black Brant (Branta bernicla nigricans), Northern Pintails (Anas acuta), and Spectacled Eiders (Somateria fischeri) sampled on the Yukon-Kuskokwim Delta (Ramey et al. 2014Reed et al., 2018), which might be related to differences among these parasite genera with regard to minimum temperatures required for parasite development (Valkiunas 2004). We incorporated an occupancy modeling approach in the current study to account for imperfect sensitivity of molecular methods to detect haemosporidian parasites and better estimate prevalence, which achieved near-perfect detection probability (0.98) by incorporating triplicate PCR runs. Given the consistency of our findings with previously published work, including our relatively high estimated detection probability, we believe our results provide robust estimates of haemosporidian prevalence for Emperor Geese inhabiting the Yukon-Kuskokwim Delta.
Using a combination of regression analysis and multi-state hidden Markov modeling, we found negligible support for potential fitness consequences of Leucocytozoon infection among adult breeding Emperor Geese. Though our results provided evidence for an association between reduced mass and Leucocytozoon infection among incubating females during the second half of incubation, the effect was relatively small (approximately 2% of total mass) and appeared unlikely to lead to detrimental fitness consequences as inferred from the lack of support for an effect of Leucocytozoon infection on annual survival. Furthermore, our genetic analyses indicated that the most common parasite lineages detected in adult female nesting Emperor Geese may also be common in waterfowl and other birds sampled elsewhere. To our knowledge, these parasite lineages have not previously been associated with widespread clinical disease in birds. Finally, our post-hoc analyses did not provide evidence for specific clades of Leucocytozoon parasites as being associated with comparatively lower mass among infected nesting adult female Emperor Geese. Therefore, our collective results provide negligible evidence for Leucocytozoon parasites as causing detrimental effects to adult breeding female Emperor Geese on the Yukon-Kuskokwim Delta, and indicate that these parasites are unlikely to represent a contemporary disease threat to the population health of this species.
Though our results did not provide clear evidence for fitness consequences of haemosporidian infections among nesting adult female Emperor Geese, we acknowledge limitations of our study design that precluded rigorous inference pertaining to potential effects of parasites at the individual level, to other age/sex cohorts, and birds infected under adverse ecological conditions (e.g., during extended periods of nutritional stress or prolonged exposure to extreme weather events). For example, as our molecular approach did not evaluate the parasitemia of haemosporidian infections or distinguish between pre-erythrocytic or patent infections, it is plausible that we failed to recognize important fitness consequences of birds suffering from high intensity, acute parasite infections. Similarly, our investigation did not assess the potential fitness consequences of relatively rare parasite genera and haplotypes. Our sample was almost entirely of birds of the same breeding status and age/sex cohort and we did not stratify our analyses based upon ecological conditions encountered by birds during the breeding season. Therefore, the current evaluation should not be considered exhaustive and future assessments of potential effects of haemosporidian parasites may be warranted, particularly if mortality events occur that are suspected of being the result of avian malaria or Leucocytozoonosis, or if observations are made indicating increases in occurrence of Haemoproteus or Plasmodium parasites among birds inhabiting the Yukon-Kuskokwim Delta.

Conclusions
Using molecular analysis, we found evidence that Emperor Geese inhabiting the Yukon-Kuskokwim Delta, Alaska are consistently infected with genetically diverse Leucocytozoon parasites at a prevalence of approximately 28%. Through the application of modeling approaches, we found evidence for a small effect of Leucocytozoon infection on the mass of incubating adult female Emperor Geese; however, we did not find support for an effect of Leucocytozoon infection on survival. Furthermore, we did not find evidence that parasites assigned to two genetically diverse clades were associated with differential mass measures among nesting adult female Emperor Geese. Collectively, our results provide negligible evidence for Leucocytozoon parasites as causing detrimental effects to adult female Emperor Geese breeding on the Yukon-Kuskokwim Delta.

Declaration of competing interest
The authors declare that they have no conflicts of interest.