Although Borrelia burgdorferi sensu lato (s.l.) are found in a great diversity of vertebrates, most studies in North America have focused on the role of mammals as spirochete reservoir hosts. We investigated the roles of birds as hosts for subadult Ixodes pacificus ticks and potential reservoirs of the Lyme disease spirochete B. burgdorferi sensu stricto (s.s.) in northwestern California. Overall, 623 birds representing 53 species yielded 284 I. pacificus larvae and nymphs. We used generalized linear models and zero-inflated negative binomial models to determine associations of bird behaviors, taxonomic relationships and infestation by I. pacificus with borrelial infection in the birds. Infection status in birds was best explained by taxonomic order, number of infesting nymphs, sampling year, and log-transformed average body weight. Presence and counts of larvae and nymphs could be predicted by ground- or bark-foraging behavior and contact with dense oak woodland. Molecular analysis yielded the first reported detection of Borrelia bissettii in birds. Moreover, our data suggest that the Golden-crowned Sparrow (Zonotrichia atricapilla), a non-resident species, could be an important reservoir for B. burgdorferi s.s. Of 12 individual birds (9 species) that carried B. burgdorferi s.l.-infected larvae, no birds carried the same genospecies of B. burgdorferi s.l. in their blood as were present in the infected larvae removed from them. Possible reasons for this discrepancy are discussed. Our study is the first to explicitly incorporate both taxonomic relationships and behaviors as predictor variables to identify putative avian reservoirs of B. burgdorferi s.l. Our findings underscore the importance of bird behavior to explain local tick infestation and Borrelia infection in these animals, and suggest the potential for bird-mediated geographic spread of vector ticks and spirochetes in the far-western United States.
Citation: Newman EA, Eisen L, Eisen RJ, Fedorova N, Hasty JM, Vaughn C, et al. (2015) Borrelia burgdorferi Sensu Lato Spirochetes in Wild Birds in Northwestern California: Associations with Ecological Factors, Bird Behavior and Tick Infestation. PLoS ONE 10(2): e0118146. doi:10.1371/journal.pone.0118146
Academic Editor: Brian Stevenson, University of Kentucky College of Medicine, UNITED STATES
Received: September 18, 2014; Accepted: January 5, 2015; Published: February 25, 2015
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication
Data Availability: All files are available from the Dryad database (doi:10.5061/dryad.p3b61).
Funding: This research was funded in part by a grant from the National Institutes of Health (AI22501) to RSL, and by a Graduate Research Fellowship from the National Science Foundation to EAN. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Lyme disease is the most prevalent vector-borne disease in the United States, with 20,000 to 30,000 confirmed cases reported annually from 2003–2011 . Worldwide, clinical manifestations are caused by a subset of genospecies within the expanding Borrelia burgdorferi sensu lato (s.l.) spirochete complex [2,3]. These bacteria are transmitted by ticks in the genus Ixodes. Larval or nymphal (subadult) ticks ingest spirochetes while feeding upon a bacteremic host, and may then infect other vertebrate hosts while feeding during the subsequent nymphal or adult stages. In North America, Lyme disease is caused by B. burgdorferi sensu stricto (s.s.). The primary bridging vectors to humans are the nymphal stages of the black-legged tick, Ixodes scapularis, in the east, and the western black-legged tick, Ixodes pacificus, in the far west [4–6].
In California, I. pacificus is known to infest >100 species of lizards, birds or mammals . The dusky-footed woodrat (Neotoma fuscipes) and the western gray squirrel (Sciurus griseus) serve as primary reservoirs for B. burgdorferi s.l. [8–12], but little is known about the importance of non-mammalian tick hosts, particularly birds, as spirochete reservoirs. Studies attempting to identify potential avian reservoirs of B. burgdorferi s.l., based on tick loads or infection in birds or their attached ticks, have been conducted in the far-western United States [13–21], as well as in the eastern United States, Europe and Asia (e.g., [22–26]). Birds are potentially important spirochete reservoirs because they are species-rich and ecologically diverse, abundant, occur in the same habitats utilized by the primary mammalian reservoirs, are often infested by vector ticks, and are widely found to carry B. burgdorferi s.l. (e.g., [27–30]). Birds therefore may be important local reservoirs for various B. burgdorferi s.l. spirochetes [29,31–34]. However, the ability of spirochete-infected birds to infect feeding ticks is not well understood and varies among bird species [29,35].
Moreover, birds may play an important role in the geographic spread of ticks and borreliae. Dispersal distances for birds are generally much larger than for small mammalian hosts of Ixodes ticks, and because of their annual migrations, birds have been implicated as long-distance dispersal agents of vector ticks and pathogenic borreliae (e.g. [22,36]). Bird communities also are changing radically due to human impacts, with large shifts in species composition, total abundance, and relative abundances among species. Geographic distributions of individual species are changing with climate [37,38], land use  and introduction of exotic species . Although Lyme disease cases in the far western United States are rare compared to other regions of North America, birds, because of their high mobility, are likely to be the most important vertebrate hosts of B. burgdorferi s.l. in determining the northward spread of Lyme disease in western North America due to climate and related anthropogenic changes.
Although the primary reservoirs of B. burgdorferi s.l. in the United States are mammals, there is growing evidence that different vertebrate species contribute to the enzootic maintenance of Lyme disease spirochetes in the far west. As birds are species-rich (433 regularly-occurring species in California ) and their community composition varies greatly among California’s diverse vegetation communities, clarifying the role of birds as reservoirs for B. burgdorferi s.l. is important both for more accurate modeling of the local ecology of Lyme disease spirochetes and to better understand the risk of human exposure to infected ticks.
The goals of this study are to identify bird species that are potentially important to the enzootic maintenance of Lyme disease spirochetes in northwestern California, to evaluate the individual nesting and feeding behaviors of birds that contribute to the transmission of spirochetes, and to test for the first time whether avian taxonomic relationships are predictive of I. pacificus larval and nymphal loads, and of B. burgdorferi s.l. infection.
The study was conducted at 14 sites within the 2,168-ha (5,358-ac) University of California, Hopland Research and Extension Center (HREC) (39°00’03 N, 123°04’59 W), located in the foothills of the Mayacmas Mountains (often alternately spelled as “Mayacamas”) in Mendocino County in northwestern California. The HREC ranges in elevation from 152–914 m (500–3000 ft). The region’s climate can be characterized as Mediterranean, with hot, dry summers and cool, wet winters. Average yearly temperature is 14°C (57°F), with maximum monthly mean temperatures of 13°C (56°F) in the winter and 33°C (91°F) in the summer. The average annual rainfall is 940 mm (37 in) (HREC, published online at http://hrec.ucanr.edu/Weather,_Physical,_and_Biological_Data/). The four dominant plant communities are dense oak woodland including Pacific madrone (Arbutus menziesii) with an understory mainly composed of leaf litter; oak-woodland communities with sparse oaks surrounded by grass and herbaceous forbs; open grasslands; and a diverse chamise-type chaparral (Adenostoma fasciculatum) community containing significant amounts of California lilac (Ceanothus spp.) and manzanita (Arctostaphylos spp.). Oak species at HREC include interior live oak (Quercus wislizeni), coast live oak (Quercus agrifolia), valley oak (Quercus lobata), California black oak (Quercus kelloggii), and blue oak (Quercus douglasii). Over two hundred species of birds and 50 species of mammals have been reported from HREC (http://ucanr.edu/sites/hopland/Natural_Resources/). Primary vertebrate reservoirs of B. burgdorferi s.l. at HREC include the western gray squirrel and the dusky-footed woodrat, and zooprophylactic (refractory) hosts include the southern alligator lizard (Elgaria multicarinata) and the western fence lizard (Sceloporus occidentalis) [8–12,42,43].
Bird capture methods were described by Eisen et al. . In brief, birds were captured by mist netting during the breeding seasons of 2003 and 2004. Mist netting occurred over four-day sampling periods at 14 sites during >20 sessions per year from March through July (nets measuring 9–12 m, 30 mm mesh size). Although nymphs of I. pacificus have been found in low numbers in all seasons, the main period of nymphal activity in Mendocino County lasts from March through June or July, with peak numbers from late April to late May [44,45]. In 2003, five nets were placed at each site to maximize capture success, with two nets sited in dense oak woodland, one net placed either in chaparral or grassland, and two nets sited along the ecotone between habitats (sampling the oak woodland community). In 2004, mist netting occurred at 3 of the 13 original sites, and one new site. Potter ground traps baited with birdseed were also used (4-cell Potter traps; Boreal King Supply Co., Bigfork, Minnesota). Nets and ground traps were opened within 15 min of sunrise and kept open for 4–8 h, depending on daily environmental conditions and capture success.
Bird and tick processing
Birds were identified to species and examined for presence of ticks, particularly around the eyes and beak where most ixodid ticks attach. All ticks found were removed with fine-tipped forceps. Search time was not standardized because of variation in the birds’ sizes, behavior of different individuals when handled (calm or not), and because of the time involved in removing attached ticks. Blood was drawn from the bird’s ulnar vein, and stored in EDTA-coated blood storage tubes at-80 C°. A tail feather was clipped to indicate when a bird was recaptured within a season. All handling and sampling of birds were conducted following protocols approved by the Institutional Animal Care and Use Committee at the University of California at Berkeley and the California Department of Fish and Game.
Ticks were preserved in 95% ethanol. A stereomicroscope at 30–40x magnification was used to distinguish I. pacificus larvae from those of Ixodes woodi and Ixodes spinipalpis. Larvae not easily identified as either I. pacificus or I. spinipalpis were treated for 1 h in 10% potassium hydroxide, mounted in Hoyer’s medium on a microscope slide, and examined by light microscopy at magnifications up to 400x.
Taxonomy and nomenclature of Borrelia
The B. burgdorferi s.l. (BBSL) complex consists of 20 confirmed or proposed genospecies [2,3], at least six of which (B. americana, B. bissettii (BBIS), B. burgdorferi s.s. (BBSS), B. californiensis, B. carolinensis, and B. genomospecies 2 [sensu Postic et al. ]) occur in California. BBSS is of special interest because it causes Lyme disease, and BBIS has been repeatedly detected in or isolated from California mammals , and occasionally, humans . We use the code BBSL(un) to refer to undifferentiated genospecies in BBSL that are neither BBSS nor BBIS, and use the term BBSLmix to refer to mixed infections detected in a single sample.
Tissue DNA extraction, PCR and sequencing
Genomic DNA from ticks and for bird blood samples was extracted using DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol as described previously [48,49]. All DNA extracts were tested with a nested polymerase chain reaction (PCR) targeting the 5S-23S rRNA intergenic spacer (IGS) of BBSL. Primers and the PCR protocol used were described previously by Lane et al. . Due to its nucleated erythrocytes, bird blood contains large amounts of DNA and presents challenges for the detection of Borrelia DNA, which has far fewer copies per sample than does bird DNA. Blood DNA extracts were first screened using the Lane et al.  PCR protocol, but because of inconsistent results, the protocol was optimized and all samples were retested. The PCR protocol was modified by (a) adjusting the concentration of the Amplitaq DNA polymerase buffer (Life Technologies, Grand Island, New York) from 10x to 12x; (b) decreasing the denaturation temperature to 88°C; and (c) increasing the annealing time to 2 min and the annealing temperature to 55°C in the first round of amplification. Each PCR run contained multiple wells of B. burgdorferi strain CA4 and UV-treated nuclease free water as positive and negative controls, respectively. PCR products were electrophoresed with 1.5% agarose gels, stained with ethidium bromide, and visualized by UV transillumination. All bird blood DNA extracts were tested at least twice in separate runs, and if results were in disagreement, a third test was performed. Samples that tested positive twice were considered to be positive. Products from positive samples were purified using QIAquick PCR Purification Kit (Qiagen, Valencia, CA). Sequencing was performed at the University of California, Berkeley DNA Sequencing Facility using inner PCR primers . Sequences were assembled and manually edited using Sequencher 4.6 (Gene Codes Corp., Ann Arbor, MI). Contigs with more than one nucleotide at the same position in forward and reverse directions were considered mixed infections (BBSLmix). Each sequence was directly compared with sequences of the same loci from various spirochete genospecies available in the GenBank database using the Basic Local Alignment Search Tool (http://www.ncbi.nlm.nih.gov/BLAST/). The criterion for inclusion within a Borrelia species was ≥99% similarity. Data for the 5S-23S rRNA IGS DNA sequences have been deposited in Dryad, in FASTA format.
Modeling infection and infestation: modeling goals
Explanatory variables, including ecological and taxonomic variables, were evaluated in models of BBSL infection status in birds or attached I. pacificus larvae, and in modeling presence of larvae and of nymphs on birds using bionomial models in a generalized linear modeling (GLM) framework. The same variables were used to predict I. pacificus larval and nymphal loads on birds using zero-inflated models for count data. Life-history traits included as explanatory variables were log-transformed average body weight for each bird species [50–52], resident status, breeding status, feeding substrate, main food item, main habitat type, and nest placement . For modeling infection, we also used the number of nymphs on the individual bird and year as explanatory variables. We attempted to control for differences in immune system with the taxonomic variables order, family, and genus . “Species” was also investigated as an explanatory variable with 52 levels, but could not be included in any model given the generally low sample sizes.
Bird taxonomic relationships are presented in Table 1. Bird species’ preferences for a main habitat were derived from the Cornell Lab of Ornithology online, and then matched to the four main habitat categories reported from the HREC: chaparral, grass, oak woodland-grass, and dense oak woodland. Body weights for birds by species were also taken from the Cornell Lab of Ornithology online, and where ranges of body weights were provided, an arithmetic mean was calculated and used in each analysis (see Table A in S1 File). The resident status of birds was determined by long-term HREC observations . Explanatory variables and their levels are summarized in Table 2, and bird species categorized by ecological guilds are listed in Table A in S1 File. Ecological variables included foraging and other behaviors based on classifications presented in De Graaf , and modified by the Cornell Lab of Ornithology .
We chose specific modeling frameworks (including selection of link functions) after performing extensive preliminary analyses. After zero-truncating the dataset such that only positive results (non-zero instances of bird infection, or larval or nymphal presence, depending on analysis) were examined, we found that in each case, data were not normally distributed, and also violated assumptions of homoscedasticity of errors. Additionally, bird individuals per species varied, leading to “unbalanced” data structure for each analysis. These issues are inherent to multiple-species studies that examine parasite loads and require appropriate models that can model both zero-inflated data and account for unbalanced data.
For these reasons, a bionomial model in a GLM framework was appropriate for explaining (1) presence of infection in birds and larvae removed from individual birds, (2) presence of larval ticks on individual birds, and (3) presence of nymphs on individual birds. We used zero-inflated models for count data for (4) number of larvae and (5) number of nymphs removed from individual birds. Zero-inflated negative binomial models used here have two parts: a negative binomial count model with a log link, and a zero-inflation binomial model with a logit link that models the presence of extra zeros. Here, true zeroes represent ecologically meaningful information in that, for example, the bird was inspected, but no larvae or nymphs were found for infestation models. Likewise for infection models, zeroes might represent larvae removed from birds that are tested for but lack infection either due to lack of exposure to an infected bird, or lack of infection due to low level of engorgement and short feeding time prior to testing. Both modeling approaches, that is, binomial models in a GLM framework and zero-inflated negative binomial models, generated model outcomes that were then subjected to model selection based on Akaike Information Criterion (AIC) values . All modeling goals are summarized in Table 3.
Modeling infection and infestation: explanatory variables
Criteria used to establish that a particular bird species may serve as a Borrelia reservoir host were as follows: (a) a bird can obtain spirochetes only from an infected I. pacificus nymph; (b) the spirochetes must survive in the tissues or blood of a reservoir bird for some period of time; and (c) the reservoir bird must be able to transmit spirochetes to feeding ticks. Criteria (a) and (c) were established based on the prior knowledge that unfed I. pacificus larvae are BBSL-free (transovarial transmission occurs rarely, if ever); therefore, detection of BBSL in a larva removed from a bird can be assumed to very likely demonstrate infection in that bird as long as no infected nymphs were feeding on the bird at the same time (potentially resulting in co-feeding infection). For tick infestation, modeling was done separately for nymphs and for larvae.
Bird contact with larvae or nymphs may lead to an increase in overall tick abundance in an ecosystem, and contact with nymphs may increase the prevalence of BBSL in bird populations. Using species-level body weight and behavioral data, we hypothesize that:
- (1). The number of nymphs on a bird is predictive of BBSL infection status. Because only ~5–15% of host-seeking nymphs in northwestern California carry BBSL [5,16,57], a higher nymphal load should increase the likelihood that a bird is bitten by a BBSL-infected nymph;
- (2). Average body weight of birds by species correlate to tick loads and therefore higher likelihood of exposure to BBSL. Body weight may therefore serve as an adequate null model of infestation and infection prevalence, either because of larger surface area of skin, or correlation with life history traits such as home range size , ground nesting and foraging behaviors;
- (3). Contact with leaf litter through either ground foraging or ground nesting is predictive of tick load for both larvae and nymphs [16,59,60] for individual birds;
- (4). Contact with bark from tree bases through feeding or use as nesting substrate is predictive of tick load [16,22,60] for individual birds;
- (5). Breeding birds have higher tick loads than non-breeding birds that migrate and/or are winter residents because of increased foraging activity associated with nest-building and feeding their young, and therefore breeding birds will sample more of their surroundings than non-breeding birds;
- (6). The main food item of each species of bird (categories include insects, seeds, mammals and “omnivorous”), while defining much about foraging behaviors, is not predictive of tick load in individual birds, because it may be too coarse a category in explaining bird behavior given generalizations about feeding behavior, and due to dietary switching ;
- (7). Year-round residents do not have higher tick loads than migratory birds, because subadult ticks are attached to the birds for <1 week. In our data, it is not the case that all migratory birds arrived one week or less before they were captured. We test this variable because migratory birds live in tropical regions outside the geographic distribution of I. pacificus ticks for much of the year and there may be a difference between these and resident birds;
- (8). Sampling year, as a proxy for interannual variation in environmental variables such as temperature and annual rainfall (both correlated to density of questing ticks), is predictive of the number of larvae and nymphs found on individual birds;
- (9). The main habitat of birds by species is predictive of the presence of larvae and nymphs, and of BBSL infection, for individual birds. In the ecosystems studied, dense oak woodland was found to have BBSS-infected nymphs, although the drag-sampling method used to show this is ineffective in grass, chaparral, and grass-understory oak-woodland, which may also harbor BBSS-infected nymphs .
- (10). Certain taxonomic groups (within family, order or genus) of birds will have more infected members than other taxonomic groups.
None of the birds evaluated exhibit social grooming, so a variable related to this behavior of interest was not included in models. All birds were sampled in or close to their breeding seasons, so their immunological status was assumed to be similar. Because some species could not be accurately sexed in the field, we did not distinguish males from females in our analyses. One accidental bird species (Red-eyed Vireo) is excluded from most analyses (except body weight null models) because its interactions with California ecosystems are unknown. Two birds that were the sole representatives of their taxonomic orders (the Western Screech-Owl and California Quail) were excluded from model analyses. One Brown-headed Cowbird was excluded from the analysis of nesting substrate because of that species’ opportunistic and brood-parasitic behavior, and four Brown Creepers were reclassified as “bark” nesters rather than placed in the suggested “tree” nesting category because of the species’ unique nesting strategy .
Modeling infection and infestation: model specification
We define “infection” as any genospecies of BBSL detected from a bird and/or in a larva found attached to a bird. Bird infection was modeled as a binomial distribution (with dispersion parameter = 1), with the individual bird as a unit of analysis. The dependent variable was constructed as a column vector of “successes” (infected) and “failures” (not infected): (1) A full, or maximal model was specified: (2) (with variables and their abbreviations above defined in Table 2) and was reduced to form a candidate set, which were then tested through extensive model comparison with Akaike Information Criterion (AIC) values and second order AIC (AICc) “corrected” values, which assign heavier penalties for additional parameters. Similar analyses were performed for presence of larvae on individual birds, and presence of nymphs on individual birds. Count models for number of larvae and number of nymphs removed from individual birds reduced the same full model above for both the negative binomial count model and the zero-inflation binomial model parts of the zero-inflated negative binomial model.
Capture effort and success
Sampling efforts for birds included 3,117 mist net hours at 14 sites and 910 Potter trap hours (at 8 sites only in 2003) for birds. In total, 623 individual birds comprising 4 orders, 24 families, 44 genera and 53 species were captured, all by mist netting. Ixodes pacificus made up > 99% (284 of 286) of ticks removed from birds. The remaining two ticks were the rabbit tick, Haemaphysalis leporispalustrus.
Summary statistics of infestation and infection
The majority of birds were not infested with ticks and the majority of examined larval ticks were not infected with BBSL. Data for tick infestation, and tick-infection prevalence were highly zero-inflated. Only 100 (16.1%) of 623 examined birds carried I. pacificus ticks, with 62 (10.0%) birds infested with a total of 192 larvae and 55 birds (8.8%) infested with 92 nymphs, for a total of 284 ticks. BBSL was detected only from the bird itself for 57 (9.1%) individuals, while 66 birds had BBSL present either in blood and attached larvae or only in attached larvae. Among the four orders of birds represented, BBSL was detected from 23 species, and infected larvae were recovered from 9 species, all in the order Passeriformes. The 23 bird species containing individuals with Borrelia present, and the remaining uninfected species, are organized by larval and nymphal tick loads (Table 4). Of 192 larvae tested, 25 (13.0%) were infected. Of 85 nymphs tested, 21 (24.7%) were infected. Seven ticks were not tested because they were either damaged or lost.
BBSL genospecies detected
The prevalence of BBSL infection differed among samples from birds (57 positive findings/623 specimens; 9.2%), larvae removed from birds (25/192; 13.0%), and nymphs removed from birds (21/92; 22.8%). All four BBSL subcategories—BBSS, BBIS, BBSL(un) and BBSLmix—were detected in birds and nymphs, whereas larvae were infected with only BBSS and BBIS. The representations of subcategories among, respectively, birds, larvae, and nymphs (with the most common genospecies for each in bold) were BBSS (29.8, 96.0, 57.1), BBIS (40.4, 4.0, 23.8), BBSL(un) (21.1, 0.0, 9.5), and BBSLmix (8.8, 0.0, 9.5). See Fig. 1. Data on BBSL diversity in birds, larvae or nymphs are presented in Tables 5–7.
The Lesser Goldfinch (Spinus psaltria), Oak Titmouse (Baeolophus inornatus) and Dark-eyed Junco (Junco hyemalis) were the only species harboring all four subcategories of BBSL infection, whereas the Bewick’s Wren (Thryomanes bewickii) yielded 3 subcategories (BBSS, BBIS, BBSL(un)). The American Robin and Dark-eyed Junco were the only species with infections in both individual birds and the larvae removed from them, although the genospecies differed in the birds and larvae.
Compared to the sample population, the Golden-crowned Sparrow (Zonotrichia atricapilla) was infected with BBSS more frequently than other species (4 of 14 infected; two-tailed P < 0.0001, Pearson’s Chi-squared test). Lark Sparrows (Chondestes grammacus) yielded a much higher percentage of BBSS-infected larvae removed from birds than did other bird species (14/79, 17.7%). However, these larvae were removed from two heavily infested individual birds, and therefore the results must be interpreted cautiously.
Null model for infection and infestation: average body weight
Table 3 summarizes the main results of all modeling goals. To test the hypotheses that larval and nymphal loads are proportional to the (logarithmically transformed) average body weight of a bird species, and that infection depends similarly on body weight, species’ average body weights were log-transformed. Single-variable linear regression models showed that average body weight of birds by species (n = 53) did not predict the number of larvae (F1,50 = 0.07912; P = 0.7797) or the number of nymphs (F1,51 = 0.0124; P = 0.9118) carried per bird, or BBSL infection prevalence in a given species (F1,51 = 3.124x10–5; P = 0.9956). Models constructed from positive results only, or a “zero-truncated” dataset, showed no significant correlations between body weight by species and number of larvae carried per bird (F1,22 = 0.0237; P = 0.8791) or number of nymphs (F1,19 = 0.1954; P = 0.6635) carried per bird, or BBSL infection prevalence (F1,25 = 2.133; P = 0.1566). See Fig. 2.
Graphs depict: (a) larvae per bird by log(average body weight), (b) nymphs per bird by log(average body weight), and (c) bird infection by log(average body weight). Each data point represents one bird species. One outlier point in plot (a) represents Lark Sparrows (Chondestes grammacus). The solid line is a regression on all species, while the dotted-line is a regression on zero-truncated data (species with positive results only).
Model exploration and selection, BBSL infection
In binomial models for BBSL infection in birds that included the orders Piciformes (woodpeckers) and Passeriformes, the presence of BBSL infection was best explained by number of nymphs removed from a bird, year, taxonomic order, and breeding status (AIC evidence ratio 2.41 between best and next best model; AIC weight = 0.55). BBSL was detected in 23 bird species in the order Passeriformes and in larvae removed from 9 species, but no members of the order Piciformes (31 individuals in 4 genera and 4 species) were infected or harbored infected larvae. In terms of I. pacificus infestation, only two nymphs were found on two separate Acorn Woodpeckers (Melanerpes formicivorus), and only one larva was found on one of four Northern Flickers (Colaptes auratus), the only ground-feeding woodpecker captured. Restricting further BBSL-infection analyses to the order Passeriformes, we find that presence of BBSL infection was best explained by number of nymphs removed from a bird, year, and breeding status (AIC evidence ratio 1.2 between best and next best model; AIC weight = 0.31).
In both the combined Passeriformes and Piciformes model, and the Passeriformes-only model, the non-breeding status of birds was highly significant. However, only 20 of the 623 sampled birds were non-breeders; 14 of these were Golden-crowned Sparrows (Zonotrichia atricapilla). This species contained all four infected, non-breeding birds. This indicates that the significance of non-breeding status in these models was attributable to the prevalence of infection in this one species. Re-evaluating both models without breeding status as a variable, we found that bird infection was best predicted by number of nymphs removed from a bird, year, and taxonomic order for models including both Passeriformes and Piciformes; and number of nymphs removed from a bird, year, and log (average body weight) for the Passeriformes-only model. Details of model selection, parameter estimates and significance for these and all modeling goals are available in Tables B-G in S2 File, and Tables H-I in S3 File. Raw data were graphed by variable of interest for visual comparisons and is available in Figures A-F in S1 File.
Model selection, presence of larvae on birds
Only 62 of 623 individual birds were found to carry one or more larvae. The binomial model that was best supported by AICc values for the orders Passeriformes and Piciformes was presence of larvae on individual birds ~ year + taxonomic order + feeding substrate. The significant variables and their levels were the intercept and year (highly significant), the order Piciformes (which was negatively correlated with the presence of larvae), and in the category of feeding substrate, the levels “bark” and “ground” both showed a significant positive correlation with larval presence.
When limited to the order Passeriformes only, the best-supported model was presence of larvae on individual birds ~ year + nesting substrate + feeding substrate, where the intercept and year again showed high significance. The nesting substrates “shrub” and “tree” both showed significantly lower risk of larval infestation than other nest placement substrates (ground, bark). Bark and foliage feeding substrates were both significant and positively correlated with larval presence, and ground as a feeding substrate showed high significance for predicting larval presence. See Tables D-E in S2 File.
Model selection, presence of nymphs on birds
We considered the presence of nymphs on the orders Passeriformes and Piciformes, and noted that only 54 of 620 birds (8.7%) had nymphs. The best-supported binomial model for nymphal presence was presence of nymphs on individual birds ~ year + taxonomic family, where the intercept was not significant, and year was highly significant. While no level of family was significant by itself, the inclusion of this variable was responsible for best model fits. Most instances of nymphs removed from birds came from the taxonomic families Certhiidae, Corvidae, and Troglodytidae, however, the observations are too limited to yield significant results. Because the number of genera was greater than the number of family included in this study, it appears that model selection was not biased in favor of the category with the greatest number of levels.
When the Passeriformes-only models were compared, the best-supported model was y ~ year + taxonomic family + main habitat, where the intercept was again not significant, and year was highly significant. Additionally, the main habitat category “dense oak woodland” showed a highly significant and positive correlation with nymphal presence. Once again, inclusion of the variable “family” yields the best-model fits, though none of the levels of family was by itself significant. See Tables F-G in S2 File.
Model selection, counts of larvae on birds
Both the orders Passeriformes and Piciformes were considered in these analyses. Models consisted of a negative binomial count model with a log link, and a zero-inflation binomial model with a logit link, in the form y ~ (count model | zero-inflated model). For the top supported model counts of larvae on individual birds ~ (main habitat + feeding substrate | year + feeding substrate), the count model showed the following levels of main habitat to be highly significant: grass (P < 0.0001); oak woodland (P = 0.0248); dense oak woodland (P = 0.0059). The level “grass” has the largest predicted positive effect, followed by dense oak woodland, oak woodland, and finally chaparral, which was negatively correlated with larval counts. Although feeding substrate was significant, none of the levels were by themselves significant. For the zero-inflation part of the model, the intercept is highly significant (P < 0.0001). Complete model results are available in Table H in S3 File.
Model selection, counts of nymphs on birds
These analyses also include the orders Passeriformes and Piciformes. Here, y = number of nymphs removed from an individual bird. For the top supported model, y ~ (main habitat | log (average body weight) + year), the count model variable levels predicting higher numbers of nymphs on birds were main habitat = grass (P = 0.0004), oak woodland (P = 0.0505), and dense oak woodland (P = 0.0082), than the reference level “chaparral,” again with “grass” having the largest effect, followed by dense oak woodland, and oak woodland. Intercept was also significant in the count model (P = 0.0002). Intercept and log(average body weight) were significant in predicting presence of nymphs on birds. Complete model results are available in Table I in S3 File.
Our findings demonstrate that in northwestern California, birds contribute to the enzootic maintenance of BBSL spirochetes, given the number of bird species (32/53) and individual birds (100/623) hosting I. pacificus, and the prevalence of BBSL both in birds and in larvae removed from them. Our results also demonstrate that in addition to taxonomic relatedness, certain bird life history traits predict BBSL infection. In contrast, null models based on bird body weight by species as predictive models of BBSL infection, larval loads, and nymphal loads on birds did not explain any of these phenomena. Models that performed best were those that incorporated bird behaviors, taxonomic relationships, and sampling year. With regard to our original hypotheses, we found:
- (1). As expected, the number of infesting nymphs was highly significant in predicting bird infection with BBSL;
- (2). Counter to expectation, average body weight by species did not serve as an adequate null model of infestation and infection prevalence;
- (3). Although ground-feeding behavior was a highly significant predictor of larval presence (P = 0.0009 among passerines) in agreement with Weisbrod and Johnson  and Wright et al. , , contact with the ground through nest placement was not a significant factor in any of the models we tested. Our results support previously published findings indicating a much higher prevalence of infestation among ground-foraging birds [16,22,26,59], but did not support the hypothesis that ground-nesting per se is an important factor in predicting infestation, as implied in other studies [17,19,22];
- (4). As expected, contact with bark while feeding significantly explained larval presence on individual birds in binomial models (in contrast to the findings of ), but use of bark as a nesting substrate did not significantly predict either tick-infestation prevalence or bird infection prevalence;
- (5). Counter to our hypothesis, non-breeding birds manifested a higher level of BBSL infection (25%) compared to breeding birds (11.5%) (Passeriformes only). This result was entirely due to high levels of infection among Golden-crowned Sparrows, a non-resident species, which led us to abandon breeding status as a testable variable. Although it may be that infection is more prevalent among breeding or non-breeding birds, we were not able to test that hypothesis with our limited sample size;
- (6). As hypothesized, the main food item of birds did not explain larval or nymphal infestation;
- (7). As hypothesized, resident status was not significant in infection or infestation models;
- (8). As hypothesized, sampling year was a highly significant variable for bird infection, presence of larvae and nymphs, and the number of ticks (both larvae and nymphs) a bird carried;
- (9). Although main habitat was not by itself a significant predictor of infection in birds, the sub-category (or level) “dense oak woodland” explained presence of nymphs in binomial (presence-absence) models, and was highly significant in explaining counts of larvae removed from birds in zero-inflated negative binomial models, as expected. “Grass” as habitat type was also highly significant as a predictor of the number of larvae removed from a bird (at least in part due to the presence of one heavily infested Lark Sparrow outlier in the dataset), and “oak woodland” was also significant. In contrast, the reference level “chaparral” correlated with the lowest counts of larvae and nymphs on birds of all habitat types considered;
- (10). As hypothesized, best-performing models for infection and infestation contained variables representing taxonomic structure at the level of order and below. Our results suggest that the order Piciformes is unlikely to contain major reservoirs of B. burgdorferi.
Although birds contribute to the ecology of BBSL in our study system, in general, their role is likely to be ecosystem-specific, and therefore more complex than that of the primary rodent-reservoir hosts. The large number of bird species in one location, each with generally unknown reservoir competency, greatly increases the difficulty of community-level modeling . Other inherent difficulties in evaluating the importance of birds in BBSL ecology is that the species make-up of the community  and their true abundances are notoriously difficult to estimate [70–72].
Prevalence of BBSL genospecies in birds
Borrelia bissettii (BBIS) had not been reported in birds before. This spirochete was, in fact, the most prevalent genospecies detected in birds in this study (40.4%), followed by BBSS (29.8%). In contrast, most infected larvae removed from birds instead carried BBSS (96.0%, versus 4.0% for BBIS). This discrepancy may have occurred because of the uneven abundance distribution of larvae across individual birds (leading to biased sampling), but also suggests the possibilities of a differential rate of transmission of BBSL genospecies from birds to I. pacificus larvae, or more transient infection of birds with BBIS. As we only sampled bird blood, we cannot exclude the possibility that BBSS is present in surface tissues (e.g., skin) more often than BBIS, and is therefore more easily transmitted to feeding ticks. Similarly, infected I. pacificus nymphs predominantly contained BBSS (57.1%, versus 23.8% for BBIS). Co-feeding infection, in which larvae become infected via feeding in close proximity to infected nymphs on a non-infected bird, might account for some infected larvae. Mixed and uncategorized Borrelia infections were present at much lower levels in birds or ticks removed from them.
Previously, BBIS had only been found in various species of mammals in North America (reviewed in ). Although nearly all cases of Lyme disease in North America are attributed to B. burgdorferi, BBIS occasionally infects people in central and southern Europe , and a genospecies closely related to, if not identical with, BBIS was detected in serum specimens from three persons residing in northwestern California .
Suggestions for future studies
Four aspects of our data suggest important considerations in the design of future studies. First, no birds carried the same genospecies of BBSL in both their blood and in larvae feeding on them. Of 12 individual birds (in 9 species) that carried infected larvae, birds either had a different genospecies of BBSL in their blood (3 individuals in 2 species) or tested negative for blood infection (9 individuals in 8 species; Tables 4–7). Investigations of skin, feather attachments, and other surface tissues of birds may be useful in discovering the tissue tropisms of BBSL in birds . Considering the skin-tropisms of certain Lyme disease spirochetes in mammals, and that cultures are often obtained from ear-punch biopsies from rodents, examination of more types of bird tissues may result in (non-lethal) field-sampling techniques that generate a higher rate of positive test results among infected animals [19,74].
Second, certain bird species’ immune systems may clear BBSL infections from their blood. If so, and the feeding attraction of infection-cleared hosts is of sufficient magnitude, these species would constitute “dilution hosts” of BBSL by reducing the overall prevalence of spirochetes in the vector-tick population. Clearance, or marked suppression, of infection could also lead to the observation cited in the previous paragraph: infected, blood-feeding larvae on birds lacking blood infection may signify transient bacteremia in those birds. Investigations of host competency of birds by species, and mechanisms by which birds clear infection from their blood are needed. For example, blood pH-raising effects of high-altitude migrations  may affect the abundance of Borrelia spirochetes in bird blood, and blood-temperature increases induced by migration and periods of high activity  may exceed the tolerance range of some Borrelia genospecies . Additionally, the age class of birds, unmeasured in this study, may correlate with naturally acquired immunity.
Third, attempts to model the dynamics of BBSL infections in wild bird populations will require assessments of the true species diversity and estimations of true abundances by species within each habitat type, including birds that are potential hosts but are not adequately sampled with mist-netting techniques. Mist netting, compared to other bird survey techniques, can give highly biased estimates of bird abundances and species composition by excluding birds that do not fly at the level of the mist net or are otherwise excluded from this survey technique, such as large birds, nocturnal birds and ground-dwelling birds [78,79]. With true bird abundance information obtained through a variety of methods such as point counts and trapping, total tick loads and relative importance of each host species could be estimated from tick data obtained from caught birds.
Of the 219 bird species observed in the Hopland study area, many species known to be residents or breeders in the ecosystems studied were either underrepresented or not represented in our dataset . These include the American Robin, which is a competent reservoir host for B. burgdorferi in the northeastern United States . Although American Robins are common to abundant during all seasons at the HREC, only five birds were collected during this study. Similarly, two abundant species of quail at the HREC yielded only 1 individual bird to our mist netting and ground-trapping efforts, and the fairly common Wild Turkey (Meleagris gallopavo) was not captured at all . Among raptors (hawks, falcons, eagles, vultures, and owls), 1 of 22 species occurring there was captured during the present study. Such predatory and scavenging birds, especially those that feed primarily on mammals, may be attractive to partially fed nymphs that detach from their dead hosts . Capture methods appropriate to the life histories of each species of birds may increase the diversity of birds that can be sampled . A more complete understanding of the entire bird community and their relative abundances may reveal additional important reservoirs of BBSL.
Lastly, landscape dynamics are particularly critical to understand in California, where losses of chaparral and oak woodland ecosystems due to fragmentation, removal, and other anthropogenically-induced changes are causing biodiversity loss and extreme changes to the existing bird communities [82–87]. Chaparral yielded the lowest prevalence of larval and nymphs ticks on birds at the HREC, although recent studies targeting rodents and lizards at the there [62,88] demonstrated abundant subadult I. pacificus larval and nymphal populations in this habitat type. Rapid shifts in California bird community structures towards locally increased abundances of some known carriers of Borrelia spirochetes (e.g. American Robins and Dark-eyed Juncos) as a result of urbanization  and chaparral removal in northwest California (Newman et al., in preparation) represents changes in local disease ecology dynamics  and the potential for increased transmission of Lyme disease spirochetes to humans.
S1 File. (Table A, Figure A, Figure B, Figure C, Figure D, Figure E, Figure F).
Table A. Information on body weight and ecological guilds for all species of birds included in this study. Categories of analysis, their levels, and level abbreviations are shown in the main text in Table 2. Figures A-F. Boxplots of raw data larval infestation, nymphal infestation, and bird infection status by variable of interest.
S2 File. (Tables B, Tables C, Tables D, Tables E, Tables F, Tables G).
Presence models, model selection AIC values, and parameter estimates for binomial models of bird infection, larval infestation, and nymphal infestation of birds. Table B. Model selection second-order Akaike Information Criterion (AICc) values for binomial models of BBSL blood infection in birds (and in larvae removed from those birds). Models include factors specified in the main article and bird traits including the orders Passeriformes and Piciformes. Table C. Model selection second-order Akaike Information Criterion (AICc) values and approximate P-values for variables in the top selected model. The models below are binomial models of BBSL blood infection in birds (and in larvae removed from those birds). Models include factors specified in the main article and bird traits limited to the order Passeriformes. Table D. Model selection second-order Akaike Information Criterion (AICc) values for bird life history traits that lead to Ixodes pacificus larvae presence on a bird, including the orders Passeriformes and Piciformes. Table E. Model selection on second-order Akaike Information Criterion (AICc) values for binomial models of bird life history traits and sampling conditions that lead to I. pacificus larvae presence on a bird, limited to the order Passeriformes. Table F. Model selection for bird traits and sampling conditions that lead to Ixodes pacificus nymph presence on a bird, including the orders Passeriformes and Piciformes. Table G. Model selection for bird traits and sampling conditions that result in Ixodes pacificus nymph presence on a bird, limited to the order Passeriformes.
S3 File. (Tables H, Tables I).
Count models, model selection AIC values, and parameter estimates for zero-inflated negative binomial models of larval and nymphal infestations of birds Tables H. Model selection AIC values for bird life history traits that predict number of Ixodes pacificus larvae removed from a bird, including the orders Passeriformes and Piciformes Tables I. Model selection AIC values for bird life history traits that lead to number of Ixodes pacificus nymphs removed from a bird, including the orders Passeriformes and Piciformes
We thank the University of California Hopland Research and Extension Center for providing logistical support and use of its facilities. We also thank J. Patrick Kelley for guidance on statistical methods, Sarah Leonhard for preliminary molecular studies, and David Hembry, Nyeema Harris and the Steve Beissinger Lab at the University of California, Berkeley for constructive edits. Michael McCloy provided the photograph of the Golden-crowned Sparrow in this article.
Conceived and designed the experiments: RSL RJE LE CV. Performed the experiments: NF JMH RSL RJE LE CV. Analyzed the data: EAN NF JMH. Contributed reagents/materials/analysis tools: EAN RSL. Wrote the paper: EAN NF RJE LE JMH RSL. Designed and executed models: EAN.
- 1. CDC. Reported Cases of Lyme Disease by Year, 2003–2012 [website]. Atlanta, GA: Division of Vector-Borne Diseases, National Center for Emerging and Zoonotic Infectious Diseases, U.S. Centers for Disease Control and Prevention. Available: http://www.cdc.gov/lyme/stats/chartstables/casesbyyear.html. Accessed 28 November 2012.
- 2. Margos G, Vollmer SA, Ogden NH, Fish D (2011) Population genetics, taxonomy, phylogeny and evolution of Borrelia burgdorferi sensu lato. Infect Genet Evol 11: 1545–1563. doi: 10.1016/j.meegid.2011.07.022. pmid:21843658
- 3. Ivanova LB, Tomova A, González-Acuña D, Murúa R, Moreno CX, et al. (2013) Borrelia chilensis, a new member of the Borrelia burgdorferi sensu lato complex that extends the range of this genospecies in the Southern Hemisphere. Environ Microbiol 16: 1069–1080. doi: 10.1111/1462-2920.12310. pmid:24148079
- 4. Burgdorfer W, Lane RS, Barbour AG, Gresbrink RA, and Anderson JR (1985) The western black-legged tick, Ixodes pacificus: a vector of Borrelia burgdorferi. Am J Trop Med Hyg 34: 925–930. pmid:3898886
- 5. Clover JR, Lane RS (1995) Evidence implicating nymphal Ixodes pacificus (Acari: Ixodidae) in the epidemiology of Lyme disease in California. Am J Trop Med Hyg 53: 237–240. pmid:7573703
- 6. Dennis DT, Nekomoto TS, Victor JC, Paul WS, Piesman J (1998) Reported distribution of Ixodes scapularis and Ixodes pacificus (Acari: Ixodidae) in the United States. J Med Entomol 35: 629–638. pmid:9775584 doi: 10.1093/jmedent/35.5.629
- 7. Castro MB, Wright SA (2007) Vertebrate hosts of Ixodes pacificus (Acari: Ixodidae) in California. J Vector Ecol 32: 140–149. pmid:17633435 doi: 10.3376/1081-1710(2007)32[140:vhoipa]2.0.co;2
- 8. Lane RS, Brown RN (1991) Wood rats and kangaroo rats: potential reservoirs of the Lyme disease spirochete in California. J Med Entomol 28:299–302. pmid:1875357 doi: 10.1093/jmedent/28.3.299
- 9. Brown RN, Lane RS (1992) Lyme disease in California: a novel enzootic transmission cycle of Borrelia burgdorferi. Science 256: 1439–1442. pmid:1604318 doi: 10.1126/science.1604318
- 10. Brown RN, Lane RS (1996) Reservoir competence of four chaparral-dwelling rodents for Borrelia burgdorferi in California. Am J Trop Med Hyg 54: 84–91. pmid:8651378
- 11. Lane RS, Mun J, Eisen RJ, Eisen L (2005) Western grey squirrel (Rodentia: Sciuridae): a primary reservoir host of Borrelia burgdorferi in Californian oak woodlands? J Med Entomol 36: 329–340. doi: 10.1093/jmedent/42.3.388
- 12. Salkeld DJ, Leonhard S, Girard YA, Hahn N, Mun J, et al. (2008) Identifying the reservoir hosts of the Lyme disease spirochete Borrelia burgdorferi in California: the role of the western gray squirrel (Sciurus griseus). Am J Trop Med Hyg 79: 535–540. pmid:18840740
- 13. Lane RS, Piesman J, Burgdorfer W (1991) Lyme borreliosis: Relation of its causative agents to its vectors and hosts in North America and Europe. Annu Rev Entomol 36: 587–609. pmid:2006870 doi: 10.1146/annurev.en.36.010191.003103
- 14. Manweiler SA, Lane RS, Block WM, Morrison ML (1990) Survey of birds and lizards for ixodid ticks (Acari) and spirochetal infection in northern California. J Med Entomol 27: 1011–1015. pmid:2280384 doi: 10.1093/jmedent/27.6.1011
- 15. Slowik TJ, Lane RS (2001) Birds and their ticks in northwestern California: minimal contribution to Borrelia Burgdorferi enzootiology. J Parasitol 87: 755–761. pmid:11534638 doi: 10.1645/0022-3395(2001)087[0755:battin]2.0.co;2
- 16. Eisen L, Eisen RJ, Lane RS (2004) The roles of birds, lizards, and rodents as hosts for the western black-legged tick Ixodes pacificus. J Vector Ecol 29: 295–308. pmid:15709249
- 17. Wright SA, Thompson MA, Miller MJ, Knerl KM, Elms SL, et al. (2000) Ecology of Borrelia burgdorferi in ticks (Acari: Ixodidae), rodents, and birds in the Sierra Nevada foothills, Placer County, California. J Med Entomol 37: 909–918. pmid:11126549 doi: 10.1603/0022-2585-37.6.909
- 18. Wright SA, Lemenager DA, Tucker JR, Armijos MV, Yamamoto SA (2006) An avian contribution to the presence of Ixodes pacificus (Acari: Ixodidae) and Borrelia burgdorferi on the Sutter Buttes of California. J Med Entomol 43: 368–374. pmid:16619623 doi: 10.1603/0022-2585(2006)043[0368:aacttp]2.0.co;2
- 19. Wright SA, Tucker JR, Donohue AM, Castro MB, Kelley KL (2011) Avian hosts of Ixodes pacificus (Acari: Ixodidae) and the detection of Borrelia burgdorferi in larvae feeding on the Oregon junco. J Med Entomol 48: 852–859. pmid:21845945 doi: 10.1603/me11001
- 20. Lane RS, Kucera TF, Barrett RH, Mun J, Wu C, et al. (2006) Wild turkey (Meleagris gallopavo) as a host of ixodid ticks, lice, and Lyme disease spirochetes (Borrelia burgdorferi sensu lato) in California state parks. J Wildlife Dis 42: 759–771. pmid:17255442 doi: 10.7589/0090-3558-42.4.759
- 21. Dingler RJ, Wright SA, Donohue AM, Macedo PA, Foley JE (2014) Surveillance for Ixodes pacificus and the tick-borne pathogens Anaplasma phagocytophilum and Borrelia burgdorferi in birds from California’s Inner Coast Range. Ticks Tick-borne Dis 5: 436–445. doi: 10.1016/j.ttbdis.2014.02.002. pmid:24690191
- 22. Weisbrod AR, Johnson RC (1989) Lyme disease and migrating birds in the Saint Croix River Valley. Appl Environ Microbiol 55: 1921–1924. pmid:2782872
- 23. Ishiguro F, Takada NT, Masuzawa FT (2000) Prevalence of Lyme disease Borrelia spp. In ticks from migratory birds on the Japanese mainland. Appl Environ Microbiol 66: 982–986. pmid:10698761 doi: 10.1128/aem.66.3.982-986.2000
- 24. Comstedt P, Bergström S, Olsen B, Garpmo U, Marjavaara L, et al. (2006) Migratory passerine birds as reservoirs of Lyme borreliosis in Europe. Emerg Infect Dis 12: 1087–1095. pmid:16836825 doi: 10.3201/eid1207.060127
- 25. Poupon MA, Lommano E, Humair P-F, Douet V, Rais O, et al. (2006) Prevalence of Borrelia burgdorferi sensu lato in ticks collected from migratory birds in Switzerland. Appl Environ Microbiol 72: 976–979. pmid:16391149 doi: 10.1128/aem.72.1.976-979.2006
- 26. Marsot M, Henry P- Y, Vourc’h G, Gasqui P, Ferquel E, et al. (2012) Which forest bird species are the main hosts of the tick, Ixodes ricinus, the vector of Borrelia burgdorferi sensu lato, during the breeding season? Int J Med Microbiol 42: 781–788. doi: 10.1016/j.ijpara.2012.05.010
- 27. Craine NG, Nuttall PA, Marriott AC, Randolph SE (1996) Role of grey squirrels and pheasants in the transmission of Borrelia burgdorferi sensu lato, the Lyme disease spirochaete, in the UK. Folia Parasitol 44: 155–160.
- 28. Miyamoto K, Sato Y, Okada K, Fukunaga M, Sato F (1997) Competence of a migratory bird, red-bellied thrush (Turdus chrysolaus), as an avian reservoir for the Lyme disease spirochetes in Japan. Acta Trop 65: 43–51. pmid:9140513 doi: 10.1016/s0001-706x(97)00651-7
- 29. Ginsberg HS, Buckley PA, Balmforth MG, Zhioua E, Mitra S, et al. (2005) Reservoir competence of native North American birds for the Lyme disease spirochete, Borrelia burgdorferi. J Med Entomol 42: 445–449. pmid:15962798 doi: 10.1603/0022-2585(2005)042[0445:rconna]2.0.co;2
- 30. Scott JD, Anderson JF, Durden LA (2012) Widespread dispersal of Borrelia burgdorferi-infected ticks collected from songbirds across Canada. J Parasitol 98: 49–59. doi: 10.1645/GE-2874.1. pmid:21864130
- 31. Olsen B, Jaenson TGT, Noppa L, Bunikis J, Bergström S (1993) A Lyme borreliosis cycle in seabirds and Ixodes uriae ticks. Nature 362: 340–342. pmid:8455718 doi: 10.1038/362340a0
- 32. Richter D, Spielman A, Komar N, Matuschka FR (2000) Competence of American robins as reservoir hosts for Lyme disease spirochetes. Emerg Infect Dis 6: 133–138. pmid:10756146 doi: 10.3201/eid0602.000205
- 33. Humair PF (2002) Birds and Borrelia. Int J Med Microbiol 291: 70–74. pmid:12141764 doi: 10.1016/s1438-4221(02)80015-7
- 34. Piesman J, Gern L (2004) Lyme borreliosis in Europe and North America. Parasitol 129: S191. doi: 10.1017/s0031182003004694
- 35. Mather TN, Telford SR III, MacLachlan AB, Spielman A (1989) Incompetence of catbirds as reservoirs for the Lyme disease spirochete (Borrelia burgdorferi). J Parasitol 7: 66–69. doi: 10.2307/3282938
- 36. Brinkerhoff RJ, Folsom-O’Keefe CM, Tsao K, Diuk-Wasser MA (2011) Do birds affect Lyme disease risk? Range expansion of the vector-borne pathogen Borrelia burgdorferi. Front Ecol Environ 9: 103–110. doi: 10.1890/090062
- 37. Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, et al. (2003) Fingerprints of global warming on wild animals and plants. Nature 421: 57–60. pmid:12511952 doi: 10.1038/nature01333
- 38. Møller AP, Fiedler W, Berthold P (Eds) (2010) Effects of climate change on birds. Oxford, UK: Oxford University Press.
- 39. Gaston KJ, Blackburn TM, Goldewijk KK (2003) Habitat conversion and global avian biodiversity loss. Proc R Soc B, 270: 1293–1300. pmid:12816643 doi: 10.1098/rspb.2002.2303
- 40. Butchart SHM, Walpole M, Collen B, van Strien A, Scharlemann JPW, et al. (2010) Global Biodiversity: Indicators of Recent Declines. Science 328: 1164–1168. doi: 10.1126/science.1187512. pmid:20430971
- 41. Atlas of the Biodiversity of California, California Department of Fish and Game (2003) Third printing: (June 2011).
- 42. Lane RS, Quistad GB (1998) Borreliacidal factor in the blood of the western fence lizard (Sceloporus occidentalis). J Parasitol 84: 29–34. pmid:9488334 doi: 10.2307/3284524
- 43. Wright SA, Lane RS, Clover JR (1998) Infestation of the southern alligator lizard (Squamata: Anguidae) by Ixodes pacificus (Acari: Ixodidae) and its susceptibility to Borrelia burgdorferi. J Med Entomol 6: 1044–1049. pmid:9835700 doi: 10.1093/jmedent/35.6.1044
- 44. Eisen L, Eisen RJ, and Lane RS (2002) Seasonal activity patterns of Ixodes pacificus nymphs in relation to climatic conditions. Med Vet Entomol 16: 235–244. pmid:12243224 doi: 10.1046/j.1365-2915.2002.00372.x
- 45. Eisen L, Eisen RJ, Chang C- C, Mun J, Lane RS (2004) Acarological risk of exposure to Borrelia burgdorferi spirochaetes: long-term evaluations in northwestern California, with implications for Lyme borreliosis risk-assessment models. Med Vet Entomol 18: 38–49. pmid:15009444 doi: 10.1111/j.1365-2915.2004.0476.x
- 46. Postic D, Garnier M, Baranton G (2007) Multilocus sequence analysis of atypical Borrelia burgdorferi sensu lato isolates-Description of Borrelia californiensis sp. nov., and genomospecies 1 and 2. Int J Med Microbiol 297: 263–271. pmid:17374507 doi: 10.1016/j.ijmm.2007.01.006
- 47. Girard YA, Fedorova N, Lane RS (2011) Genetic diversity of Borrelia burgdorferi and detection of B. bissettii-like DNA in serum of north-coastal California residents. J Clin Microbiol 49: 945–954. doi: 10.1128/JCM.01689-10. pmid:21177909
- 48. Lane RS, Steinlein DB, Mun J (2004) Human behaviors elevating exposure to Ixodes pacificus nymphs and their associated bacterial zoonotic agents in a hardwood forest. J Med Entomol 41: 239–248. pmid:15061284 doi: 10.1603/0022-2585-41.2.239
- 49. Lane RS, Mun J, Eisen L, Eisen RJ (2006) Refractoriness of the western fence lizard (Sceloporus occidentalis) to the Lyme disease group spirochete Borrelia bissettii. Parasitol 92: 691–696. doi: 10.1645/ge-738r1.1
- 50. Collins PW (1999) Rufous-crowned Sparrow (Aimophila ruficeps), The Birds of North America Online (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology. Available: http://bna.birds.cornell.edu/bna/species/472; doi:10.2173/bna.472. Accessed 15 February 2012.
- 51. Cornell Lab of Ornithology (2011) Life histories of birds. Available: http://www.allaboutbirds.org/. Accessed 15 February 2012.
- 52. Birds of North America online (2011) Rufous-crowned Sparrow Measurements. Available: http://bna.birds.cornell.edu/bna/species/472/articles/measurements. Accessed 5 January 2013.
- 53. Chesser RT, Banks RC, Keith Barker F., Carla Cicero, Dunn Jon L., Kratter Andrew W., Lovette Irby J. et al. (2013) Fifty-fourth supplement to the American Ornithologists' Union check-list of North American birds. Auk 130: 558–572. doi: 10.1525/auk.2013.130.3.558
- 54. Vaughn C, Keiffer R (2007) Seasonal checklist of the Hopland Research & Extension Center. Available: http://ucanr.org/sites/hopland/files/109161.pdf. Accessed 15 February 2012.
- 55. De Graaf RM, Tighman NG, Anderson SH (1985) Foraging guilds of North American birds. Environ Manage 9: 493–536. doi: 10.1007/bf01867324
- 56. Burnham KP, Anderson DR (1998) Model Selection and Inference: A practical information-theoretic approach. Springer-Verlag. New York.
- 57. Eisen RJ, Eisen L, Castro MB, Lane RS (2003) Environmentally related variability in risk of exposure to Lyme disease spirochetes in northern California: effect of climatic conditions and habitat type. Environ Entomol 32: 1010–1018. doi: 10.1603/0046-225x-32.5.1010
- 58. Mace GM, Harvey PH (1983) Energetic constraints on home-range size. Am Nat 120–132. doi: 10.1086/284043
- 59. Reed KD, Meece JK, Henkel JS, Shukla SK (2003) Birds, migration and emerging zoonoses: West Nile Virus, Lyme disease, influenza A and enteropathogens. Clin Med Res 1: 5–12. pmid:15931279 doi: 10.3121/cmr.1.1.5
- 60. Lane RS, Mun J, Peribanez MA, Stubbs HA (2007) Host-seeking behavior of Ixodes pacificus (Acari: Ixodidae) nymphs in relation to environmental parameters in dense-woodland and woodland-grass habitats. J Vector Ecol 32: 342–357. pmid:18260527 doi: 10.3376/1081-1710(2007)32[342:hboipa]2.0.co;2
- 61. Bairlein F (1990) Nutrition and food selection in migratory birds. In Gwinner E., editor. Bird migration: physiology and ecophysiology. Springer-Verlag, Berlin. pp. 198–213.
- 62. Eisen L, Eisen RJ, Mun J, Salkeld DJ, Lane RS (2009) Transmission cycles of Borrelia burgdorferi and B. bissettii in relation to habitat type in northwestern California. J Vector Ecol 34: 81–91. doi: 10.3376/038.034.0110. pmid:20514140
- 63. Biaich PJ, Harrison CJO (1997) A Guide to the Nests, Eggs, and Nestlings of North American Birds (second edition). San Diego, CA: Academic Press.
- 64. R Core Team. (2012) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3–900051–07–0. Available: http://www.R-project.org/. Accessed 10 August 2013.
- 65. Jackman S (2012) pscl: Classes and Methods for R Developed in the Political Science Computational Laboratory, Stanford University. Department of Political Science, Stanford University. Stanford, California. R package version 1.04.4. doi: 10.1007/s12070-012-0514-9. pmid:25621271
- 66. Zeileis A, Kleiber C, Jackman S (2008) Regression Models for Count Data in R. J Stat Softw 27 (8). Available: http://www.jstatsoft.org/v27/i08/paper. Accessed 10 February 2012.
- 67. Mazerolle MJ (2012) AICcmodavg: Model selection and multimodel inference based on (Q)AICI. R package version 1.26. Available: http://CRAN.R-project.org/package=AICcmodavg. Accessed 28 November 2012. doi: 10.1007/s12070-012-0514-9. pmid:25621271
- 68. Ostfeld RS, Keesing F (2000) Biodiversity series: the function of biodiversity in the ecology of vector-borne zoonotic diseases. Can J Zoolog 78: 2061–2078. doi: 10.1139/cjz-78-12-2061
- 69. Iknayan KJ, Tingley MW, Furnas BJ, Beissinger SR (2014) Detecting diversity: emerging methods to estimate species diversity. Trends Ecol Evol 29: 97–106. doi: 10.1016/j.tree.2013.10.012. pmid:24315534
- 70. MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm CA (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83: 2248–2255. doi: 10.2307/3072056
- 71. Royle JA, Nichols JD (2003) Estimating abundance from repeated presence-absence data or point counts. Ecology 84: 777–790. doi: 10.1890/0012-9658(2003)084[0777:eafrpa]2.0.co;2
- 72. Amundson CL, Royle JA, Handel CM (2014) A hierarchical model combining distance sampling and time removal to estimate detection probability during avian point counts. Auk 131: 476–494. doi: 10.1642/auk-14-11.1
- 73. Barbour AG, Bunikis J, Travinsky B, Hoen AG, Diuk-Wasser MA, et al. (2009) Niche partitioning of Borrelia burgdorferi and Borrelia miyamotoi in the same tick vector and mammalian reservoir species. Am J Trop Med Hyg 81: 1120–1131. doi: 10.4269/ajtmh.2009.09-0208. pmid:19996447
- 74. Sinsky RJ, Piesman J (1989) Ear punch biopsy method for detection and isolation of Borrelis burgdorferi from rodents. J Clin Microbiol 27: 1723–1727. pmid:2768461
- 75. Gill FB (2006) Ornithology, 3rd Edition. New York, NY: W. H. Freeman.
- 76. Prinzinger R, Preßmar A, Schleucher E (1991) Body temperature in birds. Comp Biochem Phys A 99: 499–506. doi: 10.1016/0300-9629(91)90122-s
- 77. Hubálek Z, Halouzka J, Heroldova M (1998) Growth temperature ranges of Borrelia burgdorferi sensu lato strains. J Med Microbiol 47: 929–932. pmid:9788818 doi: 10.1099/00222615-47-10-929
- 78. Remsen JV Jr., Good DA (1996) Misuse of data from mist-net captures to assess relative abundance in bird populations. Auk 113: 381–398. doi: 10.2307/4088905
- 79. Wang Y, Finch DM (2002) Consistency of mist netting and point counts in assessing landbird species richness and relative abundance during migration. Condor 104: 59–72. doi: 10.1650/0010-5422(2002)104[0059:comnap]2.0.co;2
- 80. Millán J, Gortazar C, Martín-Mateo MP, Villafuerte R (2004) Comparative survey of the ectoparasite fauna of wild and farm-reared red-legged partridges (Alectoris rufa), with an ecological study in wild populations. Parasitol Res 93: 79–85. pmid:15103557 doi: 10.1007/s00436-004-1113-9
- 81. Joint Food and Agriculture Organization (FAO). FAO Animal Production and Health Manual, No. 5. (2007) Wild Birds and Avian Influenza: an introduction to applied field research and disease sampling techniques. Chapter 3: Wild bird capture techniques. pp. 33–50.
- 82. Soulé ME, Bolger DT, Alberts AC, Wrights J, Sorice M, et al. (1988) Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conserv Biol 2: 75–92. doi: 10.1111/j.1523-1739.1988.tb00337.x
- 83. Soulé ME, Alberts AC, and Bolger DT. (1992) The effects of habitat fragmentation on chaparral plants and vertebrates. Oikos 39–47. doi: 10.2307/3545514
- 84. Bolger DT, Scott TA, Rotenberry JT (1997) Breeding bird abundance in an urbanizing landscape in coastal southern California. Conserv Biol 11: 406–421. doi: 10.1046/j.1523-1739.1997.96307.x
- 85. Stralberg D, Williams B (2002) Effects of residential development and landscape composition on the breeding birds of Placer County’s foothill oak woodlands. In: Richard B. Standiford, Douglas McCreary, Kathryn L. Purcell, technical coordinators. Proceedings of the fifth symposium on oak woodlands: Oaks in California’s changing landscape; 2001 October 22–25; San Diego, CA. Gen. Tech. Rep. PSW-GTR-184. Albany, CA: Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture; 341–366.
- 86. Blair RB (1996) Land use and avian species diversity along an urban gradient. Ecol Appl 6: 506–519. doi: 10.2307/2269387
- 87. Whipple AA, Grossinger RM, Davis FW (2011) Shifting baselines in a California oak savanna: nineteenth century data to inform restoration scenarios. Restor Ecol 19: 88–101. doi: 10.1111/j.1526-100x.2009.00633.x
- 88. Casher L, Lane RS, Barrett R, Eisen L (2002) Relative importance of lizards and mammals as hosts for ixodid ticks in northern California. Exp Appl Acarol 26: 127–143. pmid:12475082
- 89. Shultz AJ, Tingley MW, Bowie RC (2012) A century of avian community turnover in an urban green space in northern California. Condor 114: 258–267. doi: 10.1525/cond.2012.110029
- 90. LoGiudice K, Ostfeld RS, Schmidt KA, Keesing F (2003) The ecology of infectious disease: Effects of host diversity and community composition on Lyme disease risk. Proc Natl Acad Sci USA 100: 567–571. pmid:12525705 doi: 10.1073/pnas.0233733100