Personality, space use, network, and tick infestation data from a field study of sleepy lizards, 2010 and 2014 - 2017
Data files
Dec 12, 2024 version files 58.82 KB
-
lizard_data.csv
53.13 KB
-
README.md
5.69 KB
Abstract
Host personality can markedly affect parasite transmission. Especially for parasites with indirect transmission through the environment, the effects of consistent among-individual differences in behavior may have both direct and indirect components. For example, personality may mediate both how hosts respond to infected individuals and the likelihood that hosts indirectly interact with infected conspecifics (e.g., by visiting patches infected hosts have previously contaminated). Integrating parasites, personality, and these different kinds of interaction networks constitutes a key step towards understanding transmission in natural systems. In the original manuscript for this dataset (Payne et al. Ecological Monographs 2024), we evaluated these elements using a five-year field study of a wild population of sleepy lizards, Tiliqua rugosa, and their tick parasites, which transmit among lizards through lizards’ shared use of refuges. Using Bayesian models, we evaluated (1) predictors of lizard infestation probability and intensity (i.e., average tick count when infested) and (2) relationships among the predictors. We used the latter set of models to assess indirect relationships between the predictors and the infestation metrics. As predictors, we used lizards’ infestation ‘risk’ (derived from a time-lagged refuge-sharing transmission network), traits (sex, mass, and the personality axes aggression and boldness), space use (number of unique refuges used and home range overlap with other lizards), and measures of synchronous social interactions (i.e., edge weight and degree). We found both indirect and direct connections between our predictors and tick infestation. For example, boldness was positively directly associated with infection intensity and indirectly positively associated with both infestation probability and intensity via intermediary connections with social network interaction and risk. Using more unique refuges, on the other hand, was indirectly negatively associated with infestation probability (via reduced risk), but directly positively associated with infestation probability, indicating a potential trade-off in the anti-parasite benefits of using more refuges. Our results emphasize that (1) multiple aspects of host behavior may be associated with parasite infection, (2) these components may proceed through both direct and indirect pathways, and (3) multiple pathways should be considered together because the pathways may have compounding or counteracting effects.
README: Personality, space use, network, and tick infestation data from a field study of sleepy lizards, 2010 and 2014 - 2017
Dataset DOI: https://doi.org/10.5061/dryad.n8pk0p344
Provided data file
- lizard_data.csv
- Includes the tick count, infestation risk, and other variables used in the tick transmission network analysis of sleepy lizards (Tiliqua rugosa) reported in Payne et al. 2024 (Ecological Monographs).
- Columns in the dataset are described below.
Columns
- lizard: the unique individual to which the current row refers (the focal lizard).
- year: the year of the observation.
- sex: the sex of the lizard. Female = 0. Male = 1.
- mass.avg: the lizard's average mass within a given year (in grams). Lizard mass was recorded at each tick count event.
- obsthreat.avg: the lizard's average aggression score within a given year. Aggression scores come from an observer threat assay in which an observer attempted to grab the lizard. Higher ranks correspond to more aggressive responses. More details on this assay can be found in Godfrey et al. 2012 (Animal Behaviour), Payne et al. 2021 (Animal Behaviour), and Payne et al. 2023 (Behavioral Ecology).
- bananaboldness.avg: the lizard's average boldness score within a given year. Boldness scores come from a banana boldness assay in which a lizard was offered a piece of banana, but choosing to obtain the banana meant coming closer to the looming observer (a presumed threat). Higher ranks correspond to bolder responses. More details on the boldness assay can be found in Payne et al. 2021 (Animal Behaviour) and Payne et al. 2023 (Behavioral Ecology). Note that boldness averages should be and were standardized within year prior to use in models because boldness ranks differed among years.
- center.site.dist.95: the distance (in meters) between the center of the lizard's 95% home range polygon (HR) and the center of the field site. A lizard's HR polygon was derived from its utilization distribution (UD), which was calculated using the kernalUD function from the adehabitatHR package, using a bivariate normal kernel, an ad hoc smoothing method, and a grid set to 500.
- max.degree: the number of other lizards with which the focal lizard could have potentially interacted (i.e., the number of other lizards that were ever observed at the same time as the focal lizard within the year of the row).
- degree: the number of other lizards with which the focal lizard interacted in the given year, where an interaction was defined as the focal being within 14 m of another lizard at the same time as the other lizard.
- degree.std: degree divided by max.degree (degree / max.degree).
- sn.edge.sri.sum: the focal lizard's social network edge weight sum (i.e., the sum of a focal lizard's edge weights with all other lizards in a given year), where edge weight was calculated as the simple ratio index. See Farine and Whitehead 2015 (Journal of Animal Ecology). An edge between two lizards occurred when they were within 14 m of each other at the same time.
- phr.overlap.95.sum: PHR refers to the proportion of a focal lizard's home range (HR) polygon that is overlapped by the utilization distribution (UD) of another animal, which can be thought of as the probability of finding the other animal within the HR polygon of the focal animal. PHR was calculated as the proportion of the focal lizard's 95% HR polygon that was overlapped by the 95% UD of another animal. Then, we summed a given focal lizard's PHR overlaps with other lizards, producing the total extent to which the focal lizard's 95% HR was overlapped by other lizards. PHR values were calculated with the kerneloverlap function in the adehabitatHR package, using a bivariate normal kernel, an ad hoc smoothing method, a grid set to 500, and the same grid for all animals within a given year, which is necessary for deriving overlap.
- nights.observed: the number of nights for which a refuge location was able to be identified. More details on the refuge derivation process can be found in the manuscript associated with this dataset.
- num.unique.refuges.4m: the number of unique refuges that a lizard used within a given year.
- num.unique.refuges.4m.std: the number of unique refuges that a lizard used divided by the number of nights for which a refuge could be determined (num.unique.refuges.4m / nights.observed).
- cross.risk: an estimate of a lizard's risk of acquiring infestation from other lizards, approximated as how much a lizard interacted with other lizards via time-lagged shared refuge use (i.e., using the same refuge at time point X as another lizard did at an earlier time).
- self.risk: an estimate of a lizard's risk of self-infestation, approximated as how much a lizard interacted with itself via time-lagged shared refuge use (i.e., using the same refuge at time point X as it did at an earlier time).
- trials: the number of times that a lizard was assessed for ticks in a given year.
- success: the number of times that a lizard was observed with tick larvae in a given year. For example, if a lizard's ticks were counted 4 times in a year, and the lizard had tick larvae at 3 of those counting events, then trials would be 4 and success would be 3.
- tick.avg: the average of a lizard's non-zero larval counts. For instance, if a lizard's ticks were counted 4 times, and the lizard had 0, 15, 0, and 5 larvae at each counting, respectively, then the non-zero average would be 20/2 = 10.
- tick.total: the total number of larvae that a lizard had across all of its tick counting events within a given year. In the previous example, the lizard's larval total would be 20.
Methods
The following description comes the Methods section of Payne et al. (Ecological Monographs, 2024), modified slightly to focus on data acquisition and derivation. See the paper for a more detailed description of the statistical analyses that we performed with our metrics.
Study System
We studied sleepy lizards, a skink common in South Australia (Cogger 2014), within an approximately 1.2km2 field site roughly centered on 33.887997° S, 139.311814° E. Sleepy lizards are large (400-950 g as adults) and long-lived (up to 50 years; Bull 1995, Reinke et al. 2022). At our field site, which has a Mediterranean climate and is dominated by chenopod shrubs, sleepy lizards are primarily active during the Austral spring, September through December (Bull 1987), which constitutes our field season. During their activity season, lizards overnight in refugia, such as dense bushes (Kerr et al. 2003).
Sleepy lizards at the site harbor two reptile-specific tick species, Amblyomma limbatum and Bothriocroton hydrosauri (Bull 1991). Due to sleepy lizards’ relatively large size and abundance at the site, they are the principal hosts for the two tick species (Bull et al. 1981). Both ticks have a three-host life cycle, infesting a host at the larval, nymph, and adult stages (Smyth 1973). Each tick stage typically detaches from the lizard host while the host is in a refuge (e.g., into the litter underneath a bush), as ticks have low survival outside of refuges (Bull and Smyth 1973, Chilton and Bull 1993a). While detached from a host, neither tick species quests (Petney et al. 1983). Thus, tick transmission occurs primarily in lizard refuges (Petney et al. 1983, Chilton and Bull 1993b).
Data collection
During the activity seasons of 2010 and 2014 through 2017, we assayed lizard aggression and boldness, tracked lizard movement via GPS, and counted lizard ticks at regular intervals. We monitored 61 lizards in 2010, 49 in 2014, 76 in 2015, 76 in 2016, and 73 in 2017. In 2014, we had fewer lizards because monitoring effort was split between the main and an accessory site; in this work, we use data from the main site. To identify lizards, we permanently marked them with toe clips (described in Leu et al. 2015). Toe clipping causes relatively low stress in these lizards and is less stressful than other techniques, such as microchipping, for similar species (Langkilde and Shine 2006, Perry et al. 2011). Lizard sex was identified once per year based on lizard morphology (Bull and Pamula 1996).
We assayed lizard aggression and boldness approximately three times per lizard per year. The lizard boldness and behavioral assays have been described elsewhere (Godfrey et al. 2012, Spiegel et al. 2015, Payne et al. 2021, 2023). Briefly, in the aggression assay, we ranked a lizard’s response to a simulated grab. More aggressive responses (e.g., attempting to bite the observer) received higher ranks. In the boldness assay, we ranked a lizard’s response to a preferred food item (a banana) in a perceived situation of risk (a looming observer). Bolder behavior (e.g., immediately approaching and eating the banana, which required approaching the observer) received higher ranks. Since these metrics have been shown to be repeatable (Payne et al. 2021), and we needed a single value per year to align with our other metrics, we averaged boldness and aggression scores within year.
Our GPS monitoring of lizards has been previously described (Leu et al. 2010a, e.g., Godfrey et al. 2012, Spiegel et al. 2015, 2018). At the beginning of each activity season (year), we attached GPS units and radio-transmitters with surgical tape to lizards’ tails. Tracking units were removed at the end of each year. We have never observed adverse effects of this protocol, such as skin irritation or changes in lizard behavior (Leu et al. 2010a, e.g., Godfrey et al. 2012). GPS fix frequency varied among years, but was always at least one fix per every 10 minutes. GPS data corresponding to lizard handling was removed, as were errors identified by manual inspection and an automated filtering algorithm (Bjorneraas et al. 2010).
Each year, we relocated lizards via radio-tracking approximately every two weeks to download GPS data, replace GPS batteries, count ticks, and weigh lizards. These assessments were usually done in the morning to avoid interfering with daily lizard activity (Kerr et al. 2004). We visually inspected lizards to count the number of larvae, nymphs, and adults of each tick species. For this study, we used only larval counts, combined across species. We used larvae because in 2010, 2016, and 2017, we experimentally added larvae to several lizards in the site. After molting, these added ticks would bias our counts of the nymph and adult stages on all lizards, not just the treated ones. However, removing lizards that received the added larvae removes the bias from the larval counts. We combined tick species because larvae are difficult to reliably identify to species in the field and the two species both transmit indirectly through refuges (Leu et al. 2010b). Since larvae may stay on a lizard for two to four weeks, we filtered non-zero larval tick count observations to those separated by at least 28 days (Payne et al. 2020); zero tick counts were treated as independent. At each tick count, we recorded lizard mass with a digital scale. We averaged mass within year due to defecation, seasonal fat accumulation, and measurement error.
Derivation of social network and space use metrics
We calculated two social network metrics for each lizard within each year: degree and edge weight sum. Following our previous work in this system, we defined an edge to occur between two lizards whenever they were within 14 m of each other at the same time (Leu et al. 2010a), which is based on GPS error but well within lizard to lizard detection ability (Auburn et al. 2009). Degree was calculated as a proportion: the number of different individuals (partners) with which a focal individual ever formed an edge divided by the total number of possible partners, where possible partners was the number of lizards that were ever observed at the same time as the focal lizard. We calculated degree as a proportion (P|Degree in figures) because the number of possible partners varied among focal lizards. Edge weight was calculated as the simple ratio index (Farine and Whitehead 2015), and we summed all edge weights within a focal lizard.
For space use, we calculated each lizard’s distance from the center of the site, home range overlap with other lizards, and the number of unique refuges that it used within a year. For the former two metrics, we first estimated each lizard’s home range utilization distribution (Van Winkle 1975) for each year, calculated using a bivariate kernel with the ad hoc smoothing method in the R adehabitatHR package (Calenge 2006). Distance from the center of the site was then calculated as the distance in meters between the center of the field site (33.8879971726° S, 139.311814433° E) and the center of a lizard’s 95% home range polygon (i.e., the smallest polygon encompassing 95% of the lizard home range utilization distribution probability). Home range overlap (Fieberg and Kochanny 2005) was calculated as the total extent to which a focal lizard’s 95% home range polygon was overlapped by the utilization distributions other lizards (i.e., the probability of finding another lizard in the focal’s home range, hence PHR, and then summed across all nonfocals). Number of unique refuges was the number of different overnight refuges that a lizard used within a year divided by the number of a nights a lizard’s overnight refuge was determined (e.g., if a lizard was observed for 20 nights and used 15 different refuges, then the value would be 15 / 20). We calculated number of unique refuges as a proportion (P|Refuges in figures) because the number of nights a refuge could be determined varied among lizards. We discuss the full details of our nightly and unique refuge derivation process in Appendix S1. Essentially, we identified a refuge as a coherent cluster of GPS points using density-based clustering (Ester et al. 1996, Campello et al. 2013). Unique refuges were then derived by clustering refuges; all refuges belonging to a single cluster were assumed to represent a single unique refuge.
Derivation time-lagged network metrics
After ticks are deposited in a refuge, they have an infectious interval controlled by their developmental period and survival times off host. Based on previous work in this system, we assume an infectious interval of 9 through 39 days post-deposition in the refuge (Leu et al. 2010b, Wohlfiel et al. 2013, Wohlfeil et al. 2020). The total risk that a recipient lizard acquires ticks should stem from (1) how often the recipient uses refuges 9 – 39 days after those refuges were used by a donor lizard, and (2) how often donor lizards used the recipient’s refuge, in that greater use by donors should translate to a greater probability of depositing ticks at a refuge (Leu et al. 2010b).
Given these components of risk and following Leu, Kappeler, & Bull (2010b), we calculated two metrics: cross risk and self risk. Cross risk describes a focal individual’s time-lagged refuge sharing with other lizards, whereas self risk refers to a focal lizard’s lagged interaction with itself. Briefly, for a given donor-recipient lizard dyad, we counted the number of times that the donor used each of the recipient’s nightly refuges 9 – 39 days prior to the recipient’s use of each refuge. We then divided this count by a standardization factor (the sum of the number of nights where only the recipient was observed, only the donor was observed, and both were observed) to account for different observation extents among donor-recipient dyads (see the Appendix to the main manuscript, Section S3.2), analogously to the simple ratio index (Farine and Whitehead 2015). Cross risk was the sum of these weighted values from each of a recipient lizard’s non-self donors, whereas self risk was the recipient’s single weight into itself. For more details, see the Appendix to the main manuscript, Section S3.
Overview of provided dataset
Based on the aforementioned data collection and metric derivation, the dataset provided in this repository contains one row for each lizard-year observation (e.g., if a lizard were observed in 2010, 2014, and 2015, then that lizard would have three rows in the dataset). Columns identify the year and the lizard for each row. The dataset also contains columns for lizard traits (sex, mass, behavior), space use metrics (center site distance, home range overlap, number unique refuges used), social network metrics (degree, edge weight), transmission network metrics (cross risk, self risk), and tick infestation (number of tick counting events, number of tick counting events where a lizard was observed with larvae, average larval counts when a lizard was infested). The ReadMe more fully describes each column in the dataset.
References
Auburn, Z. M., C. M. Bull, and G. D. Kerr. 2009. The visual perceptual range of a lizard, Tiliqua rugosa. Journal of Ethology 27:75–81.
Bjorneraas, K., B. Van Moorter, C. M. Rolandsen, and I. Herfindal. 2010. Screening Global Positioning System Location Data for Errors Using Animal Movement Characteristics. Journal of Wildlife Management 74:1361–1366.
Bull, C. M. 1987. A Population Study of the Viviparous Australian Lizard, Trachydosaurus rugosus (Scincidae). Copeia 1987:749–757.
Bull, C. M. 1991. Ecology of Parapatric Distributions. Annual Review of Ecology and Systematics 22:19–36.
Bull, C. M. 1995. Population Ecology of the Sleepy Lizard, Tiliqua rugosa, at Mt-Mary, South-Australia. Australian Journal of Ecology 20:393–402.
Bull, C. M., and Y. Pamula. 1996. Sexually dimorphic head sizes and reproductive success in the sleepy lizard Tiliqua rugosa. Journal of Zoology 240:511–521.
Bull, C. M., R. D. Sharrad, and T. N. Petney. 1981. Parapatric boundaries between Australian reptile ticks. Pages 95–107.
Bull, C. M., and M. Smyth. 1973. The distribution of three species of reptile ticks, Aponomma hydrosauri (Denny), Amblyomma albolimbatum Neumann, and Amb. limbatum Neumann II. Water balance of nymphs and adults in relation to distribution. Australian Journal of Zoology 21:103–110.
Calenge, C. 2006. The package “adehabitat” for the R software: A tool for the analysis of space and habitat use by animals. Ecological Modelling 197:516–519.
Campello, R. J. G. B., D. Moulavi, and J. Sander. 2013. Density-Based Clustering Based on Hierarchical Density Estimates. Pages 160–172 in J. Pei, V. S. Tseng, L. Cao, H. Motoda, and G. Xu, editors. Advances in Knowledge Discovery and Data Mining. Springer Berlin Heidelberg.
Chilton, N. B., and C. M. Bull. 1993a. A Comparison of the Off-Host Survival Times of Larvae and Nymphs of 2 Species of Reptile Ticks. International Journal for Parasitology 23:693–696.
Chilton, N. B., and C. M. Bull. 1993b. Interspecific Differences in Microhabitat Choice by 2 Species of Australian Reptile Tick. International Journal for Parasitology 23:1045–1051.
Ester, M., H.-P. Kriegel, and X. Xu. 1996. A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. Pages 226–231 Knowledge Discovery and Data Mining. Institute for Computer Science, University of Munich.
Farine, D. R., and H. Whitehead. 2015. Constructing, conducting and interpreting animal social network analysis. J Anim Ecol 84:1144–1163.
Fieberg, J., and C. O. Kochanny. 2005. Quantifying home-range overlap: The importance of the utilization distribution. Journal of Wildlife Management 69:1346–1359.
Godfrey, S. S., J. K. Bradley, A. Sih, and C. M. Bull. 2012. Lovers and fighters in sleepy lizard land: where do aggressive males fit in a social network? Animal Behaviour 83:209–215.
Kerr, G. D., C. M. Bull, and D. Burzacott. 2003. Refuge sites used by the scincid lizard Tiliqua rugosa. Austral Ecology 28:152–160.
Kerr, G. D., C. M. Bull, and D. Mackay. 2004. Human Disturbance and Stride Frequency in the Sleepy Lizard (Tiliqua rugosa): Implications for Behavioral Studies. Journal of Herpetology 38:519–526.
Langkilde, T., and R. Shine. 2006. How much stress do researchers inflict on their study animals? A case study using a scincid lizard, Eulamprus heatwolei. Journal of experimental biology 209:1035–1043.
Leu, S. T., J. Bashford, P. M. Kappeler, and C. M. Bull. 2010a. Association networks reveal social organization in the sleepy lizard. Animal Behaviour 79:217–225.
Leu, S. T., D. Burzacott, M. J. Whiting, and C. M. Bull. 2015. Mate Familiarity Affects Pairing Behaviour in a Long-Term Monogamous Lizard: Evidence from Detailed Bio-Logging and a 31-Year Field Study. Ethology 121:760–768.
Leu, S. T., P. M. Kappeler, and C. M. Bull. 2010b. Refuge sharing network predicts ectoparasite load in a lizard. Behavioral Ecology and Sociobiology 64:1495–1503.
Payne, E., D. L. Sinn, O. Spiegel, M. G. Gardner, and A. Sih. 2023. A field experiment reveals reciprocal effects of host personality and parasitism in wild lizards. Behavioral Ecology 35.
Payne, E., D. L. Sinn, O. Spiegel, S. T. Leu, M. G. Gardner, S. S. Godfrey, C. Wohlfeil, and A. Sih. 2021. Consistent after all: behavioural repeatability in a long-lived lizard across a 6-year field study. Animal Behaviour 174:263–277.
Payne, E., D. L. Sinn, O. Spiegel, S. T. Leu, C. Wohlfeil, S. S. Godfrey, M. G. Gardner, and A. Sih. 2020. Consistent individual differences in ecto-parasitism of a long-lived lizard host. Oikos 129:1061–1071.
Perry, G., M. C. Wallace, D. Perry, H. Curzer, and P. Muhlberger. 2011. Toe Clipping of Amphibians and Reptiles: Science, Ethics, and the Law. Journal of Herpetology 45:547–555.
Petney, T. N., R. H. Andrews, and C. M. Bull. 1983. Movement and Host Finding by Unfed Nymphs of 2 Australian Reptile Ticks. Australian Journal of Zoology 31:717–721.
Reinke, B. A., H. Cayuela, F. J. Janzen, J.-F. Lemaître, J.-M. Gaillard, A. M. Lawing, J. B. Iverson, D. G. Christiansen, I. Martínez-Solano, G. Sánchez-Montes, J. Gutiérrez-Rodríguez, F. L. Rose, N. Nelson, S. Keall, A. J. Crivelli, T. Nazirides, A. Grimm-Seyfarth, K. Henle, E. Mori, G. Guiller, R. Homan, A. Olivier, E. Muths, B. R. Hossack, X. Bonnet, D. S. Pilliod, M. Lettink, T. Whitaker, B. R. Schmidt, M. G. Gardner, M. Cheylan, F. Poitevin, A. Golubović, L. Tomović, D. Arsovski, R. A. Griffiths, J. W. Arntzen, J.-P. Baron, J.-F. Le Galliard, T. Tully, L. Luiselli, M. Capula, L. Rugiero, R. McCaffery, L. A. Eby, V. Briggs-Gonzalez, F. Mazzotti, D. Pearson, B. A. Lambert, D. M. Green, N. Jreidini, C. Angelini, G. Pyke, J.-M. Thirion, P. Joly, J.-P. Léna, A. D. Tucker, C. Limpus, P. Priol, A. Besnard, P. Bernard, K. Stanford, R. King, J. Garwood, J. Bosch, F. L. Souza, J. Bertoluci, S. Famelli, K. Grossenbacher, O. Lenzi, K. Matthews, S. Boitaud, D. H. Olson, T. S. Jessop, G. R. Gillespie, J. Clobert, M. Richard, A. Valenzuela-Sánchez, G. M. Fellers, P. M. Kleeman, B. J. Halstead, E. H. C. Grant, P. G. Byrne, T. Frétey, B. Le Garff, P. Levionnois, J. C. Maerz, J. Pichenot, K. Olgun, N. Üzüm, A. Avcı, C. Miaud, J. Elmberg, G. P. Brown, R. Shine, N. F. Bendik, L. O’Donnell, C. L. Davis, M. J. Lannoo, R. M. Stiles, R. M. Cox, A. M. Reedy, D. A. Warner, E. Bonnaire, K. Grayson, R. Ramos-Targarona, E. Baskale, D. Muñoz, J. Measey, F. A. de Villiers, W. Selman, V. Ronget, A. M. Bronikowski, and D. A. W. Miller. 2022. Diverse aging rates in ectothermic tetrapods provide insights for the evolution of aging and longevity. Science 376:1459–1466.
Smyth, M. 1973. The distribution of three species of reptile ticks, Aponomma hydrosauri (Denny), Amblyomma albolimbatum Neumann, and Amb. limbatum Neumann I. Distribution and hosts. Australian Journal of Zoology 21:91–101.
Spiegel, O., S. T. Leu, A. Sih, S. S. Godfrey, and C. M. Bull. 2015. When the going gets tough: behavioural type-dependent space use in the sleepy lizard changes as the season dries. Proceedings of the Royal Society B: Biological Sciences 282:20151768.
Spiegel, O., A. Sih, S. T. Leu, and C. M. Bull. 2018. Where should we meet? Mapping social network interactions of sleepy lizards shows sex-dependent social network structure. Animal Behaviour 136:207–215.
Van Winkle, W. 1975. Comparison of Several Probabilistic Home-Range Models. The Journal of Wildlife Management 39.
Wohlfeil, C. K., S. S. Godfrey, S. T. Leu, J. Clayton, and M. G. Gardner. 2020. Spatial proximity and asynchronous refuge sharing networks both explain patterns of tick genetic relatedness among lizards, but in different years. Austral Ecology 45:493–501.
Wohlfiel, C. K., S. T. Leu, S. S. Godfrey, and C. M. Bull. 2013. Testing the robustness of transmission network models to predict ectoparasite loads. One lizard, two ticks and four years. International Journal for Parasitology: Parasites and Wildlife 2:271–7.