Body Mass, Wing Length, and Condition of Wintering Ducks Relative to Hematozoa Infection

Protozoan blood parasites (i.e., hematozoa) are commonly found in birds and can be related to reduced wing length (Dunn et al. 2013), body mass (Pierce 1984; Shutler et al. 1999a; Schrader et al. 2003; Dyrcz et al. 2005; Garvin et al. 2006; Rojo et al. 2013), body condition (Dawson and Bortolotti 2000; Merino et al. 2000; Schrader et al. 2003; DeGroote 2006; Garvin et al. 2006; Ishak et al. 2010; Meixell et al. 2016), productivity (Merino et al. 2000; Dyrcz et al. 2005; Asghar et al. 2015), and survival (Fallis and Bennett 1966; Herman et al. 1975; Dawson and Bortolotti 2000; Remple 2004; Lachish et al. 2011). However, in some cases, relationships between hematozoa infection and measures of body mass, body condition, and survival varied among hematozoa genera infecting birds and demographic characteristics of host populations including species, age, and sex (Schrader et al. 2003; Ishak et al. 2010; Lachish et al. 2011; Rojo et al. 2013; Meixell et al. 2016). Other studies have found no relationship between hematozoa infection and bird growth (Shutler et al. 1999b), wing length (Shurulinkov et al. 2012), body mass (Bennett et al. 1988; Shurulinkov et al. 2012; Shutler et al. 1999b), condition (Ashford 1971; Shurulinkov et al. 2012; Sorensen et al. 2016), or other health metrics (Shutler et al. 1996, 1999b; Stjerman et al. 2004). Thus, targeted investigations are necessary to assess the effects of hematozoa infection on specific avian populations of interest.

Waterfowl (order Anseriformes) comprise a diverse and geographically widespread group of birds inhabiting North America. Body condition of waterfowl during winter may greatly affect their survival (Conroy et. al. 1989; Hohman et al. 1995; Fleskes et al. 2002), productivity (Arnold et al. 2010), and ultimately population size (Raveling and Heitmeyer 1989). Thus, the Central Valley Joint Venture (CVJV; a partnership of nine conservation organizations, 11 state and federal agencies, and one corporation established under the North American Waterfowl Management Plan to help conserve the continent's waterfowl populations and habitats), similar to other large-scale habitat conservation programs focused on waterfowl wintering areas, uses body condition of wintering waterfowl as one of several biological metrics of its success in the Central Valley of California (hereafter the Central Valley; CVJV 2006). Body mass and body mass corrected for structural size with some morphometric measure such as wing length are commonly used indices of body condition (Miller 1989; Thomas 2009; Labocha and Hayes 2012). Therefore, understanding factors affecting body mass (or indices thereof) of waterfowl wintering in the Central Valley, one of the most important waterfowl areas in the world (Gilmer et al. 1982; Fleskes 2012), is critical for assessing the effectiveness of the CVJV and similar conservation programs.

Waterfowl are commonly infected by at least three genera of hematozoa, Leucocytozoon, Haemoproteus, and Plasmodium, transmitted by black flies, biting midges, and mosquitoes, respectively (Valkiūnas 2005). Hematozoa transmission in North America presumably occurs mostly in summer at breeding areas, but infections, which may be long-lasting, have been detected in waterfowl throughout North America and at other times of year (Greiner et al. 1975; Valkiunas 2005; Ramey et al. 2016). Reeves et al. (2015) reported that prevalence of Leucocytozoon, Haemoproteus, and Plasmodium in ducks wintering in the Central Valley was 56%, 10%, and 4%, respectively. Thus, given the high rates at which waterfowl in the Central Valley were infected with hematozoa, assessing whether infection is related to body condition may be useful for better understanding factors affecting the response of wintering waterfowl to CVJV habitat improvements and the impact of conservation programs (Fleskes et al. 2016). Such an assessment may also be useful for determining the need for management actions to reduce potential negative influences of parasite infection. Therefore, to understand whether body condition of wintering ducks in the Central Valley is related to hematozoa infection, we studied body mass, wing length, and a commonly calculated body condition index (BCI; i.e., body mass divided by wing length) relative to hematozoa infection of the five most common duck species wintering in this region (northern pintail [Anas acuta], northern shoveler [A. clypeata], green-winged teal [A. crecca], American wigeon [A. americana], and mallard [A. platyrhynchos]; U.S. Fish and Wildlife Service 2007).

Totaling about 52,000 km², the Central Valley spans 640 km from Red Bluff in the north to the Tehachapi Mountains in the south, and 48–112 km wide between the foothills of the Sierra Nevada and Pacific Coastal Ranges (Gilmer et al. 1982; Fleskes 2012). The Central Valley is composed of three major regions: the Sacramento Valley in the north, the San Joaquin Valley in the south, and the Suisun Marsh-Delta region in between (Figure 1). Central Valley wetlands, once estimated at 1.6–2 million ha, were reduced by > 90% by the early 1900s (U.S. Fish and Wildlife Service 1978). The magnitude of wetland loss and types of waterfowl habitats remaining in each region differ. During our study, about 162–256 km² of managed wetlands on private lands, National Wildlife Refuges (NWRs; e.g., Sacramento, Colusa, Delevan, Sutter, and Llano Seco NWRs), and state Wildlife Areas (WAs; e.g., Gray Lodge and Upper Butte Basin WAs), and 537–847 km² of agricultural fields (mostly rice) flooded after harvest provided a relatively contiguous block of habitat in the Sacramento Valley (Fleskes et al. 2005a, 2005b). In the San Joaquin Valley, habitat consisted of 257–315 km² of seasonally flooded managed wetlands in three blocks (about 80% in the Grasslands Ecological Area and 10% each in Mendota WA and the Tulare Lake Basin–Kern NWR vicinity) and 31–47 km² of postharvest flooded fields, mostly in the Tulare Lake Basin (Fleskes 1999). The Suisun Marsh-Delta region provided about 70 km² of unmanaged and 179 km² of managed (mostly brackish) wetlands and 128 km² of postharvest flooded corn, wheat, and other cropland. Climate in the Central Valley is Mediterranean, with dry, warm summers and wet, mild winters. September–January rainfall at Sacramento, California averaged 23.8 cm during 1941–2007 but was only 46% of average during 2006–2007 (Western Regional Climate Center 2013).

We weighed (±1 g) ducks shot by hunters in the Sacramento Valley region at public hunting area check stations on Sacramento and Delevan NWRs and in the San Joaquin Valley region at Mendota WA and public hunting area check stations and private duck clubs in the Grasslands Ecological Area (Figure 1) during October–January 2006–2007. After we weighed each duck, we removed one wing and placed it in a labeled clear plastic bag. We refrigerated the wings and within 3 d measured (±1 mm) flat wing (Dzubin and Cooch 1992; i.e., wing length) and identified species, sex, and age class (hatch year [immature] or after hatch year [adult]; Larson and Taber 1980; Duncan 1985; Carney 1992) for each bird. We then froze all the wings in their individual bags, stratified samples by age–sex cohort (adult female [F], adult male [M], immature F, immature M), month period (October–November, December–January; i.e., month), and Central Valley region (Sacramento Valley, San Joaquin Valley) and randomly selected 206 northern pintail, 200 mallard, 200 American wigeon, 200 green-winged teal, and 200 northern shoveler wings for hematozoa screening. Age–sex cohorts were equally represented (i.e., 50 each) for each species, except 42 adult F, 53 adult M, 55 immature F, and 56 immature M northern pintails were selected. Samples were also nearly evenly distributed by month (i.e., 502 in October–November, 504 in December–January), but more samples from the Sacramento Valley (n = 601) than the San Joaquin Valley (n = 405) were selected to better represent regional distribution of wintering waterfowl in the Central Valley (Fleskes et al. 2005a; U.S. Fish and Wildlife Service 2007).

The frozen wings were shipped to U.S. Geological Survey–Alaska Science Center's Molecular Ecology Laboratory where they were thawed and tested for hematozoa via molecular methods. Analyzing wing muscle tissue has previously been shown to be effective for detection of blood parasites (Ramey et al. 2013). Host and parasite deoxyribonucleic acid (DNA) were extracted from all wing muscle samples using the DNeasy blood and tissue kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. Amplification and visualization of a fragment of host cytochrome oxidase I gene, following protocols described by Kerr et al. (2007), was completed to verify the competency of the DNA extraction before molecular screening for hematozoa DNA. Extracted DNA was screened for presence of hematozoa using a nested polymerase chain reaction as described by Hellgren et al. (2004) and following verification methods described by Ramey et al. (2013). All samples were tested twice to increase probability that infection was detected (Ramey et al. 2013; Meixell et al. 2016). Samples were considered positive for hematozoa infection only if the target product was visualized on a 0.8% agarose gel and confirmed by bidirectional Sanger sequencing of target DNA, including genus identification using maximum identity scores > 90% for Leucocytozoon, Haemoproteus, or Plasmodium parasites, respectively, using the nucleotide basic local alignment search tool function (U.S. National Library of Medicine 2016). Infections were further classified as coinfected if the cleaned and sequenced double-stranded target product contained three or more ambiguous nucleotides (Szymanski and Lovette 2005). Coinfections were classified as infection by parasites of greater than one lineage of 1) Leucocytozoon, 2) Plasmodium, 3) Haemoproteus, 4) Leucocytozoon and Plasmodium, 5) Leucocytozoon and Haemoproteus, or 6) Plasmodium and Haemoproteus on the basis of top basic local alignment search tool results.

Apparent prevalence of three species of hematozoa for 60 of our 206 northern pintails and all 200 mallards, 200 northern shovelers, 200 green-winged teal, and 200 American wigeon included in our study have previously been reported (Reeves et al. 2015), although not for coinfections, by month or Central Valley region, or in the context of body condition of birds.

We used analysis of variance models and Akaike's information criterion corrected for small sample sizes (AIC_c) to examine relationships (Burnham and Anderson 2002) among duck wing length, body mass, and body mass divided by wing length (i.e., BCI; all natural log transformed for normality) and 12 measures of hematozoa infection: 1) infection by “any hematozoa”, 2) infection by Leucocytozoon, 3) infection by Plasmodium, 4) infection by Haemoproteus, 5) infection by both Leucocytozoon and Plasmodium, 6) infection by both Leucocytozoon and Haemoproteus, 7) infection by both Plasmodium and Haemoproteus, 8) infection by any two genera of hematozoa, 9) infection by any three genera of hematozoa, 10) coinfection by more than one haplotype of Leucocytozoon, 11) coinfection by more than one haplotype of Plasmodium, and 12) coinfection by more than one haplotype of Haemoproteus.

We conducted model selection for each response variable (i.e., ln wing length, ln body mass, ln BCI) in three stages. In the first stage, we established a null model without any measure of hematozoa infection by fitting and comparing models (n = 166) with all different combinations of species (northern pintail, mallard, American wigeon, green-winged teal, northern shoveler), age–sex cohort (adult F, adult M, immature F, immature M), sampling region (Sacramento Valley, San Joaquin Valley), sample month (October–November, December–January) and two-way and three-way interaction effects. During the second stage, we added each of the 12 measures of hematozoa infection, one at a time, to the best null model (i.e., model with smallest AIC_c from the first stage) and evaluated whether any of the 12 new models ranked higher than the null model. For the third stage, we chose the highest-ranking model with a hematozoa effect, as well as any other model that ranked higher than the null model, and added all combinations of two-way interactions between the hematozoa and other main effects in the model (e.g., species, age–sex cohort, region, month). We evaluated how these models with hematozoa interaction effects ranked in relation to all of our fitted models. For any of the 12 measures of hematozoa infection that improved the model compared with the null model, we computed the effect of each, if any, in terms of predicted responses (i.e., wing length, body mass, BCI) separated between infected and noninfected individuals as well as their percent difference. We back-transformed means for every subgroup, i.e., Y = exp(α + βX + β*W), where Y is the response variable, X represents species, month, age–sex, and region variables, W represents infection status, and α, β, and β* are model intercept and coefficients. We computed 85% confidence limits in accordance with the common use of ΔAIC_c = 2 as a threshold for importance (Arnold 2010), and similarly back-transformed them. The proportion difference for an infected bird, where W = 1, when compared with an uninfected but otherwise similar bird, where W = 0, is (exp[α + βX + β*] – exp[α +βX])/exp(α + βX), which simplifies to exp(β*) − 1 after canceling the exp(α + βX) factor. We converted each important hematozoa coefficient to an estimated percent change by calculating (exp[β*] – 1) × 100%, and similarly for the 85% confidence limits. We estimated the standard error of the percent change using the delta method (Williams et al. 2002). If other factors were found to interact with the infection effect, then this indicated that the infection effect varied among the different levels of the other factor, and we also computed the effect of infection separately for the different levels of those factors by taking the relevant linear combinations of hematozoa and interaction coefficients from the highest-ranking model, which contains that interaction, averaging as needed, and repeating the conversion to percent change. We conducted all statistical analyses using R statistical software (R Core Team 2016), including customized functions to generate all models based on a priori variables (M. P. Herzog, personal communication).

Prevalence of infection with Leucocytozoon, Haemoproteus, and Plasmodium parasites varied by genus among duck species, as did prevalence of coinfections with more than one parasite genus or hematozoa haplotype (Table 1, Table S1). Consistent with hematozoa prevalence in these California-wintering ducks reported by Reeves et al. (2015), infection by Leucocytozoon (55%) was far more common than infection by Haemoproteus (9%) or Plasmodium (4%). Also, Leucocytozoon infection was less common in mallards (13%) than in the other four species (50–76%), Haemoproteus (9%) infection was more common in green-winged teal (20%) than in the other four species (3–11%), and prevalence of Plasmodium was similarly low (2–6%) in all species (Table 1). Infection prevalence did not vary greatly among months or regions (e.g., overall infection by any hematozoa = 55% in October–November vs. 58% in December–January and 61% in Sacramento Valley vs. 51% in San Joaquin Valley; Table 1). Prevalence of infection by multiple hematozoa genera was generally low, ranging from < 1% for dual Plasmodium–Haemoproteus infection and infection by all three genera to 11% for infection by any two hematozoa genera (Table 1). Prevalence of coinfection by multiple haplotypes of parasites within each genus ranged from < 1% for Plasmodium and Haemoproteus to 25% for Leucocytozoon (Table 1).

Model selection of factors related to duck body mass indicated that body mass of ducks infected by any hematozoa or Leucocytozoon was less than for ducks apparently lacking hematozoa infections (Table 2). Some evidence indicated that the infection effect also varied by month and region. Models including the other 10 measures of hematozoa infection ranked lower than a model without any measure of hematozoa infection (Table 2). The top models differed only in which hematozoa infection measures (i.e., any hematozoa or Leucocytozoon) were included, suggesting that the any-hematozoa effect was mostly due to Leucocytozoon infection status. On the basis of the best model, body mass of ducks infected by any hematozoa was −1.63% (−2.79% to −0.47%) compared with body mass of uninfected ducks, across all other factors (see Table S2 for body mass values for all combinations of factors). On the basis of the second best model, which included an interaction effect between infection status and month, the effect of hematozoa infection on body mass was −2.79% (−4.40% to −1.15%) during October–November and −0.49% (−2.13% to 1.17%) during December–January.

Model selection of factors related to duck wing length provided marginal evidence that wing length of ducks infected by Leucocytozoon differed from wing length of uninfected ducks and infection effect varied by species and age–sex cohort (Table 2). On the basis of the best model, wing length of ducks infected by Leucocytozoon was −0.12% (−0.41% to 0.17%) compared with wing length of uninfected ducks, averaged across all other factors. The effect varied from −1.49% (−2.16% to −0.81%) for immature female northern pintails to 1.51% (0.69% to 2.34%) for immature male green-winged teals (Table S2).

Model selection of factors related to BCI indicated that the BCI of ducks infected by any hematozoa or by Leucocytozoon was less than the BCI of uninfected ducks (Table 2). Some evidence indicated that the infection effect also varied by month and region. Models including the other 10 measures of hematozoa infection ranked lower than a model without any measure of hematozoa infection (Table 2). The top models differed only in which hematozoa infection measures (i.e., any hematozoa or Leucocytozoon) were included, suggesting that the any-hematozoa effect was mostly due to Leucocytozoon infection status. On the basis of the best model, which included an interaction effect between infection status and month, the BCI of ducks infected by any hematozoa compared with BCI of uninfected ducks was −1.38% (−2.49% to −0.26%) overall, with the effect stronger in October–November (−2.55%, −4.10% to −0.98%) than December–January (−0.20%, −1.77% to 1.39%) and in the Sacramento Valley (−2.24%, −3.68% to −0.77%) than San Joaquin Valley (−0.21%, −1.93% to 1.53%; Table S2).

Our results indicate that body mass, wing length, and BCI were negatively related to Leucocytozoon infection for five dabbling duck species (northern pintail, mallard, American wigeon, green-winged teal, northern shoveler) wintering in the Sacramento and San Joaquin valleys during 2006–2007, with stronger support of a negative relation for mass and BCI than for wing length. However, like other observational studies that reported similar relationships between infection and body mass (Pierce 1984; Shutler et al. 1999a; Schrader et al. 2003; Dyrcz et al. 2005; Garvin et al. 2006; Rojo et al. 2013), wing length (Dunn et al. 2013), or body condition (Dawson and Bortolotti 2000; Schrader et al. 2003; DeGroote 2006; Garvin et al. 2006; Ishak et al. 2010; Meixell et al. 2016), we did not determine whether hematozoa infection directly reduced these measures or whether birds in poor condition were simply more vulnerable to infection (Flint and Franson 2009). Although experimental studies to ascertain effects of hematozoa infection on postfledgling waterfowl have not been conducted to date to our knowledge, Merino et al. (2000) demonstrated detrimental effects of hematozoa infection on postbreeding body condition of blue tits (Cyanistes caeruleus) in a previous experimental field study. Thus, field experiments would be useful to better understand what, if any, direct effects hematozoa have on wintering ducks.

Because any significant effect of hematozoa infection on duck body mass, wing length, or BCI complicates evaluation of how duck body condition responds to habitat changes, we considered all models in our exploratory study that ranked better than the null model (i.e., the best model lacking a hematozoa infection factor). Although this approach could result in retention of some uninformative factors in the best models (Arnold 2010), we erred on the side of inclusion in this exploratory study to help guide future studies and suggest that an experimental approach in the future would be most definitive (e.g., Merino et al. 2000). Furthermore, we suggest that experimental approaches would be more definitive for either supporting or refuting hematozoa infection as causative for declines in waterfowl body condition. Additionally, it should be noted that many of the 12 measures of hematozoa infection we included in our analysis were represented by very few individuals (e.g., few Plasmodium infections were detected in our study) and therefore, other classifications of infections could also be associated with body mass, wing length, and BCI in wintering Central Valley ducks, given sufficient sample size.

Our results provide support for only relatively small differences in body mass, wing length, and BCI between ducks infected and noninfected with Leucocytozoon parasites; however, even these small differences might be biologically significant. For instance, the mean difference in body mass between infected and uninfected ducks that we measured (i.e., −1.63%) represents 38% (−1.63% divided by 4.30% = 38%) of the change in northern pintail body mass observed between 1982 and 1984 and 2006 and 2008 in the Central Valley (i.e., 4.30%; Fleskes et al. 2016). Also, even though evidence for a relationship between wing length and infection was very weak, wing length is commonly used to account for body size when deriving body condition indices (Miller 1989, Thomas 2009, Labocha and Hayes 2012) and changes in wing length related to hematozoa infection could bias BCI estimates. For instance, if reduction in wing length related to hematozoa infection increases, whereas the infection effect on body mass does not increase, BCI estimates would be biased high. Thus, assessments of the success of habitat management programs and practices aimed at increasing body condition of wintering waterfowl might be compromised if hematozoa infection or its effect on wing length or body mass increases (Zamora-Vilchis et al. 2012). Therefore, fully understanding how hematozoa infection affects wintering ducks and exploring ways to mitigate negative consequences might be important for maintaining the effectiveness of wintering waterfowl habitat conservation programs.

Because infection prevalence might change over time because of climate or other factors (Zamora-Vilchis et al. 2012), learning more about factors related to parasite infection rates and tracking hematozoa infection prevalence might be important to inform and accurately judge the effect of conservation programs designed to improve waterfowl body condition. For instance, the effect of hematozoa infection on BCI was slightly greater in the Sacramento Valley than in the San Joaquin Valley. This was surprising and warrants further investigation since infection is thought to mostly occur on the breeding grounds and genetic evidence indicates that at least one duck species (i.e., northern pintail) that winters in the Sacramento and San Joaquin valleys likely originates from the same breeding areas (Fleskes et al. 2010). The mechanism by which hematozoa infection affects wing length in wintering ducks is not well known. Parasites can influence avian feather length through competition for host resources in adults during molt (see Dunn et al. 2013) and presumably in immatures during development, but how this effect changes over time is not well understood. Averaged across other factors, the effect of hematozoa infection on body mass and BCI was slightly greater during October–November than during December–January. We suspect that since infection primarily occurs during spring–summer, effects declined over time.

Hunter check stations provided us opportunity to collect a large sample of body mass data from several duck species throughout California. Thus, tracking of hematozoa infection prevalence and relationships to body condition could be accomplished by similar periodic collection of body mass, wing length, and wing tissue samples at hunter check stations. Results suggest that in addition to measuring body mass and wing length, future monitoring would benefit from inclusion of other structural measurements (e.g., culmen, tarsus) that might improve the body condition correlation and better enable detection of any changes in body size unrelated to feather growth that could confound interpretation of body mass changes (Guillemain et al. 2010). Although hunter-shot birds might be in poorer condition and have different body condition dynamics than the general population (Greenwood et al. 1986; Heitmeyer et al. 1993), results using hunter-shot ducks (Fleskes et al. 2016) agreed with Thomas (2009), who also reported increased body mass and condition of northern pintails, American wigeons, and (to a lesser degree) northern shovelers that he collected in the Sacramento Valley. In addition, similar to other studies (Miller 1986; Hine et al. 1996; Haukos et al. 2001; Moon et al. 2007), the study of Fleskes et al. (2016) using hunter-shot ducks also found that body condition of wintering waterfowl varied by month, age, sex, and rainfall. However, any interaction between hunter bias and the effects of infections could complicate interpretation of results.

Body mass, wing length, and BCI were found to be negatively related to infection by Leucocytozoon and by any hematozoa but not related to infection by only Plasmodium or Haemoproteus, or coinfections of more than one genus or haplotype of these hematozoa. However, prevalence of Plasmodium and Haemoproteus infections and coinfections of more than one genus or haplotype of these parasites was low in the sample of ducks we analyzed, resulting in few positive samples to test any relationship with body mass, wing length, and BCI. Thus, our evidence for a lack of correlation between body mass, wing length, and BCI and infection with Plasmodium or Haemoproteus or coinfections with multiple parasites should be considered weak and interpreted cautiously. Meixell et al. (2016) found that Haemoproteus infection was negatively related to body condition of ducks and relationships between Leucocytozoon infection and body condition varied among duck species.

Our study demonstrates that variation in duck body condition might be related to many factors. Similar to other investigations, we found that body condition of wintering waterfowl varied among species, age–sex cohorts, regions, and month (Miller 1986; Hine et al. 1996; Haukos et al. 2001; Moon et al. 2007; Fleskes et al. 2016); however, we also found some evidence that mass and body condition also varied with Leucocytozoon infection. Therefore, accounting for all these factors might be necessary to accurately assess the effectiveness of habitat programs for wintering waterfowl.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Microsoft Excel file containing data from 1,006 ducks shot by hunters in the Sacramento Valley region of California at public hunting area check stations on Sacramento and Delevan NWRs, and in the San Joaquin Valley region at Mendota WA, public hunting area check stations, and private duck clubs in the Grasslands Ecological Area during October–January 2006–2007 and used in the analysis relating duck flat wing length, body mass, and body mass divided by flat wing length (i.e., body condition index [BCI]; all natural log transformed for normality) and 12 measures of hematozoa infection. Columns are a) species (AGWT = green-winged teal [Anas crecca], AMWI = American wigeon [A. americana], MALL = mallard [A. platyrhynchos], NOPI = northern pintail [A. acuta], NSHO = northern shoveler), b) date (month/day/year), c) month (Oct–Nov, Dec–Jan), d) age (A = adult, I = immature), e) sex (F = female, M = male), f) body mass (g), g) flat wing (mm), h) region (SACV= Sacramento Valley, SJV = San Joaquin Valley), and infection (1 = positive, 0 = negative) by i) any hematozoa, j) Leucocytzoon (Leuc), k) Plasmodium (Plas), l) Haemoproteus (Haem), m) both Leucocytzoon and Plasmodium (LeucPlas), n) both Leucocytzoon and Haemoproteus (LeucHaem), o) both Plasmodium and Haemoproteus (PlasHaem), p) any two genera of hematozoa, q) any three genera of hematozoa, r) coinfection by greater than one haplotype of Leucocytzoon (CoinfLeuc), s) coinfection by greater than one haplotype of Plasmodium (CoinfPlas), and t) coinfection by greater than one haplotype of Haemoproteus (CoinfHaem).

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S1 (90 KB XLSX).

Table S2. Microsoft Excel file containing back-transformed mean estimates and 85% confidence level relating three morphometric dependent variables (body mass [g], flat wing length [mm], BCI [body mass divided by flat wing length]) and hematozoa infection for 1,006 ducks shot by hunters in the Sacramento Valley (SACV) and San Joaquin Valley (SJV) during October–January 2006–2007. Estimates were calculated by taking the anti-log of mean estimates from the top-ranked analysis of variance model (i.e., model having smallest Akaike information criterion corrected for small sample sizes [AIC_c]) with interaction effects for each log-transformed dependent variable, and varied by species, month period, sex, region, and infection status (any-hematozoa infection for body mass and BCI models; Leucocytzoon infection for flat wing length model). Results are presented for all 160 combinations of duck species (5), month periods (2), sex (2), region (2), and infection status (2). Columns are a) species (mallard [Anas platyrhynchos], northern pintail [A. acuta], northern shoveler [A. clypeata]), green-winged teal [A. crecca], American wigeon [A. americana]), b) month (October–November, December–January), c) age–sex (AF = adult female, AM = adult male, IF = immature female, IM = immature male), d) SACV infected, e) SACV noninfected, f) SJV infected, g) SJV noninfected.

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S2 (19 KB XLSX).

Reference S1. [CVJV] Central Valley Joint Venture. 2006. Central Valley Joint Venture implementation plan—conserving bird habitat. Sacramento, California: U.S. Fish and Wildlife Service. Found at http://dx.doi.org/10.3996/082016-JFWM-063.S3 (16,441 KB PDF).

Reference S2. Dzubin A, Cooch EG. 1992. Measurements of geese: general field methods. Sacramento: California Waterfowl Association.

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S4 (6,742 KB PDF).

Reference S3. Fleskes JP, Yee JL, Casazza ML, Miller MR, Takekawa JY, Orthmeyer DL. 2005a. Waterfowl distribution, movements and habitat use relative to recent habitat changes in the Central Valley of California: a cooperative project to investigate impacts of the Central Valley Habitat Joint Venture and changing agricultural practices on the ecology of wintering waterfowl. Dixon, California: U.S. Geological Survey–Western Ecological Research Center, Dixon Field Station,Published Final Report.

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S5 (10,130 KB PDF).

Reference S4. Haukos DA, Neaville JE, Meyers JE. 2001. Body condition of waterfowl harvested on the Upper Gulf Coast of Texas, 1986–2000. Lubbock: U.S. Fish and Wildlife Service and Texas Tech University.

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S6 (9,832 KB PDF).

Reference S5. U.S. Fish and Wildlife Service. 1978. Concept plan for waterfowl wintering habitat preservation, Central Valley, California. Portland, Oregon: U.S. Fish and Wildlife Service. Found at http://dx.doi.org/10.3996/082016-JFWM-063.S7 (48,727 KB PDF).

Reference S6. U.S. Fish and Wildlife Service. 2007. Winter waterfowl survey, Pacific Flyway-2007. Portland, Oregon: U.S. Fish and Wildlife Service.

Found at http://dx.doi.org/10.3996/082016-JFWM-063.S8 (2,577 KB PDF).

We thank hunters, check station personnel, California Department of Fish and Wildlife, California Waterfowl Association, Ducks Unlimited, Inc., U.S. Fish and Wildlife Service, and U.S. Geological Survey for cooperation, funding, or personnel to collect body mass and wing data at check stations. Funding to process and analyze wing tissue for hematozoa infection was provided by the U.S. Fish and Wildlife Service Avian Health and Disease Program to U.S. Geological Survey via an interagency agreement. Statistical code written by M. P. Herzog greatly facilitated our analyses. We appreciate reviews provided by M. A. Ricca, G. S. Yarris, and an anonymous reviewer. None of the authors has any financial interests or conflicts of interest with this article. Procedures were part of a study plan approved by U.S. Geological Survey. An Animal Care and Use Committee reviewed and approved our methods to ensure that they were in compliance with the Animal Welfare Act and U.S. Government Principals for the Utilization and Care of Vertebrate Animals Used in Testing, Research, and Training policies. Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.