Temporal and spatial dynamics of competitive parapatry in chewing lice

Abstract We synthesize observations from 1979 to 2016 of a contact zone involving two subspecies of pocket gophers (Thomomys bottae connectens and T. b. opulentus) and their respective chewing lice (Geomydoecus aurei and G. centralis) along the Río Grande Valley in New Mexico, U.S.A., to test predictions about the dynamics of the zone. Historically, the natural flood cycle of the Rio Grande prevented contact between the two subspecies of pocket gophers. Flood control measures completed in the 1930s permitted contact, thus establishing the hybrid zone between the pocket gophers and the contact zone between their lice (without hybridization). Since that time, the pocket gopher hybrid zone has stabilized, whereas the northern chewing louse species has replaced the southern louse species at a consistent rate of ~150 m/year. The 0.2–0.8 width of the replacement zone has remained constant, reflecting the constant rate of chewing louse species turnover on a single gopher and within a local pocket gopher population. In contrast, the full width of the replacement zone (northernmost G. centralis to southernmost G. aurei) has increased annually. By employing a variety of metrics of the species replacement zone, we are better able to understand the dynamics of interactions between and among the chewing lice and their pocket gopher hosts. This research provides an opportunity to observe active species replacement and resulting distributional shifts in a parasitic organism in its natural setting.

distributional changes resulting from climate change can occur over far shorter periods (decades, even years;Frey, 1992), and rapid distributional expansions or reductions (including extirpation or extinction) can result from environmental perturbations that create or eliminate dispersal corridors and barriers (e.g., bridges, dams, river diversion, deforestation, or interference with natural fire cycles).
Because such changes occur rapidly, opportunities to study them while they are actively occurring are rare, yet have the potential to provide insight into the dynamics of species distributions. Study of species at the margins of their geographic distributions may reveal factors critical to their ecological limitations or other life-history parameters (Hall, 1946;Hargreaves & Eckert, 2014). In a similar sense, study of the interactions of species in zones of contact has the potential to reveal emergent features of their natural histories that could not be discovered by separate studies of each species in isolation.
Contact zones involving host species and their obligate parasites are a special case of multiple-species contact zones, in that the dynamics of the parasite species' contact zone and transmission of parasites to alternate hosts are determined by specializations and behaviors of both the parasites and their hosts. The contact between the host species is generally viewed as reflecting the current distribution of their respective habitats, whether along a wide front or in patchy islands of habitat. These habitats shift over ecological and geological time. In contrast, from the perspective of the parasite species, the hosts are patches of habitat that move constantly both in daily activity and in annual dispersal. Thus to understand the dynamics of the parasite contact zone, it is important to consider the basic life history and dynamics of both the parasite and the host.

| Pocket gophers, chewing lice, and history of the San Acacia contact zone
Pocket gophers of the genus Thomomys (Rodentia: Geomyidae) are fossorial, solitary, and aggressively territorial, and encounters among individuals, except when mating, are generally avoided. Individuals probably live from 1 to 3 years, with high juvenile mortality during above-ground dispersal following birth in the spring .
Chewing lice of the genus Geomydoecus (Phthiraptera: Trichodectidae; Figure 1) are found only on pocket gophers. They are wingless insects that feed on skin detritus, and they spend their entire lives on their host and are highly host-specific Marshall, 1981;Murray, 1957).

T. bottae come into contact in the Río Grande Valley of central New
Mexico (Belfiore, Liu, & Moritz, 2008;Hall, 1981;Patton & Smith, 1990;Smith, 1998). Pocket gophers are largely restricted to the friable soils of the valley floor, and the broad 2-4-km wide valley is constricted at the contact zone by an elevated ridge that forces the Río Grande through a narrow (300-m wide), steep-walled canyon known as the San Acacia Constriction. Thomomys b. connectens (representing the northern, Great Basin genetic group; Patton & Smith, 1990) occurs in the Albuquerque Basin south to La Joya (just north of the constriction), whereas T. b. opulentus (the southern, Basin and Range genetic group; op. cit.) occurs south of the constriction (Figure 2a). The two subspecies are more differentiated genetically than most congeneric species of other mammals (Harper et al., 2015;Patton & Yang, 1977), yet they exhibit limited hybridization with introgression at this zone (Smith, Patton, Hafner, & Hafner, 1983; sampling localities shown in Figure 2b). Subsequent studies Harper et al., 2015) demonstrated that the pocket gopher hybrid zone has not changed location since its initial discovery in .
At the contact zone, T. b. connectens and T. b. opulentus host different species of chewing lice (Geomydoecus aurei and G. centralis, respectively) belonging to different species groups within the G. californicus species complex (Price & Hellenthal, 1981a). Unlike the two subspecies of pocket gophers, the two species of chewing lice that meet at this zone show fixed allelic differences, and there is no evidence of interbreeding (Demastes, 1990 While the pocket gopher hybrid zone has stabilized at the San Acacia Constriction, the chewing louse contact zone has moved continuously southward at a steady rate since , with the northern species of chewing louse (G. aurei) replacing the southern species (G. centralis; initial louse sampling shown in Figure 2c). Concentrated sampling in 1991 fixed the midpoint of the chewing louse replacement zone at ~6 km south of the midpoint of the pocket gopher hybrid zone (Figure 3a). Subsequent sampling 5 years later (1996) revealed that the chewing louse replacement zone had moved another 700-900 m to the south (Figure 3b). Based on the estimated annual rate of movement of the chewing louse replacement zone (140-190 m/year between 1991(140-190 m/year between and 1996(140-190 m/year between ), Hafner et al. (1998 concluded that initial contact between the two species of chewing louse, and possibly the most recent contact between the two subspecies of pocket gophers, had occurred after the catastrophic San Marcial floods of 1929 rather than 10,000 years ago, as suggested by Smith et al. (1983). Hafner et al. (1998) reasoned that any pocket gophers (and their chewing lice) inhabiting the narrow canyon of the San Acacia Constriction would have been extirpated by the catastrophic floods, thus obliterating any genetic signature of the pocket gopher hybrid zone remaining from previous contact. Hafner et al. (1998) concluded that the hybrid zone between the two subspecies of pocket gophers had stabilized at the San Acacia Constriction; theoretical cline models (Endler, 1977;Kohlmann & Shaw, 1991) show that partial barriers, such as the San Acacia Constriction, can attract and anchor the geographic position of clines and contact zones.
Herein, we analyze in detail data collected over nearly four decades at the San Acacia contact zone to test predictions about the dynamics F I G U R E 2 (a) Distribution of Botta's pocket gophers, Thomomys bottae, in New Mexico. Dark shading indicates Great Basin genetic group; light shading indicates Basin and Range genetic group (Patton & Smith, 1990). Distributions of the subspecies T. b. connectens and T. b. opulentus are indicated by dotted lines. (b) La Joya-Escondida study area showing 1979-1980 pocket gopher samples . (c) Detail of the San Acacia-Lemitar study area, showing 1989  of the zone. Our data, from five time periods (1979-1980, 1989-1991, 1996, 2001, and 2016), have enabled a more complete description of the rate and dynamics of the moving chewing louse replacement zone, while leading to a more precise explanation for the recent contacts between the two pocket gopher subspecies and their chewing lice. We evaluate the interactions among both pairs of taxa, employing additional metrics of the replacement zone and extrapolating back almost a century to the time of establishment of the zone. By employing these additional zone metrics, we are able to better understand the dynamics of interactions between and among the chewing lice and their pocket gopher hosts.

| Specimens examined
Between 1979 and 2016, 589 pocket gophers were collected for analyses of the contact zone along the Río Grande between La Joya and Escondida, Socorro Co., New Mexico; 425 of these were brushed to collect chewing lice Smith et al., 1983; this study; see Appendix 1; Figures  There is no evidence of interbreeding between the two louse species at this replacement zone (Demastes, 1990); thus, we were

| Genetic analyses
The pocket gopher hybrid zone initially was described based on chromosomes, allozyme electrophoresis, and morphometric analysis of cranial and pelage characters . Subsequent studies analyzed diagnostic allozymes (glucose-6 phosphate dehydrogenase and mannose phosphate isomerase), mitochondrial DNA haplotypes , and mitochondrial and nuclear DNA (Harper et al., 2015).
More than 6,300 individual Geomydoecus lice were identified to species. Specimens collected in 1991-2001 were identified to species using allozyme electrophoresis methods of Nadler and Hafner (1989) and Demastes (1990). Chewing lice from 2016 and selected individuals from 1991-2001 were identified to species using one of three molecular methods, each of which began with isolation of DNA from individual lice as described in Harper et al. (2015). Some individuals were identified to species using the Group 1 microsatellite primers and conditions published by Light, Harper, Johnson, Demastes, and Spradling (2018), which consistently amplify three loci for G. aurei and only two for G. centralis (Loci Ga3702 and Ga6020, but not Ga4103). Others were identified using a 379-bp (base pair) region of the COI gene amplified using conditions described by Hafner et al. (1994) and either sequenced or cut with the restriction enzyme, Sau3AI (Optizyme; Thermo Fisher Scientific). Digested COI products for G. aurei yielded two fragments, 96 and 283 bp plus primer length, while G. centralis samples were not cut by Sau3AI.

| Zone characteristics
Five-parameter (5-p) logistic regressions were conducted using the nplr package (version 0.1-7) of R (version 3.3.3, R Core Team, 2013) to model the sigmoidal nature of the variables over geography. Unlike the typically employed tanh curve (Barton, 1979;Barton & Hewitt, 1981,1985Bull & Burzacott, 2001;Szymura & Barton, 1986), the 5-p logistic regression allows for asymmetry in the resultant sigmoidal curve. To assess position and width of the louse replacement zone for each of four time periods (1989-1991; 1996; 2001; and 2016), nine variables describing the resulting logistic curves were recorded as follows: (a) "nGc," defined as the northernmost location of the southern louse, G. centralis; (b) "0.8," the point at which G. aurei represents 80% of the lice on gophers (the 0.8 frequency point of a conventional zone width; May, Endler, & McMurtrie, 1975)

| Movement of the chewing louse replacement zone
We identified samples of chewing lice from 377 pocket gophers from the study area (San Acacia to south of Lemitar, Figures 2c and   3) as pure G. aurei, mixed-species samples, or pure G. centralis. These

| D I S C U S S I O N
In this study, the dispersal rate of chewing lice across the landscape is dependent on a complicated set of nonindependent factors including pocket gopher density, rate of pocket gopher dispersal, frequency of physical contact among pocket gophers, rate of successful colonization of new host individuals by chewing lice, and rate of species replacement on a newly colonized pocket gopher. Further study is needed to reveal the nature of competitive interaction between these two species of chewing lice, be it competition for specific microhabitats on an individual pocket gopher, different reproductive rates, breeding interference, or some other interaction. Each of the metrics that we have employed to describe the replacement zone, and the rate of change in each (Tables 1 and 2 Figure 3d). However, it is more likely that these departures are artifacts of the relatively narrow width of sampling along the north-south transect in 1996, and that additional mixed-species samples, northern G. centralis, and southern G. aurei were outside of our sampling zone ( Figure 5).

| Interpretations of zone metrics
Studies of contact zones usually describe only the line of contact or, at most, the width of the contact zone (Buggs, 2007). Our use of additional metrics allows more detailed description of the replacement zone, its movement, and the relative roles of both pocket gopher and chewing louse dispersal in effecting zone width and other param-  Figure 4). First, the southernmost G. aurei (sGa), which marks the leading edge of southward dispersal of the species, moved 5,800 m from 1991-2016 (Table 1), or 232 m/year ( Table 2). The location of sGa in 2016 was at the 9,700-m point on the north-south transect (Table 1) (Scurlock, 1998).
TA B L E 1 Replacement zone parameters (position south of the San Acacia diversion dam, in meters) and goodness-of-fit values for 5-parameter logistic curves for each time period  year. This estimate, based on 25 years of monitoring over a distance of 10 km, is higher than those reported in the literature based on direct observation, usually trapping data (e.g., 62 m/year, Howard & Childs, 1959;78 m/year, Vaughan, 1963;117 m/year, Daly & Patton, 1990), but it is widely acknowledged that annual dispersal distances are probably site specific and depend on resource availability, population density, and potentially many other ecological factors that influence animal movement.

| Geographic and temporal context of the contact zone
Contact between these two subspecies of pocket gophers occurred at a biogeographically unique location along the Río Grande at a singular time in the history of the river's flood regime. In a larger geographic context, the San Acacia Constriction marks the meeting place between the Great Basin and Chihuahuan Desert biomes (Bailey, 1913;Küchler, 1985),  (Scurlock, 1998), and the bosque ecosystem evolved in this annual flood regime (Cartron, Lightfoot, Mygatt, Brantley, & Lowrey, 2008). These floods must have regularly scoured pocket gophers from the San Acacia Constriction and temporarily reduced population density elsewhere along the river. The catastrophic San Marcial floods of 1929 occurred prior to completion of flood control measures (Harden, 2006;Lee, 1907;Patterson, 1965;Scurlock, 1998;Sodrensen & Linford, 1967). The southern louse species in this study, G. centralis, represents an isolated and geographically restricted population of a species that is widespread across the Great Basin (Price & Hellenthal, 1981a). Prior to ~1971, this isolated population of G. centralis hosted by T. b. opulentus ranged from San Acacia southward 70 km to San Marcial, but has surrendered nearly 15% of its estimated 17,500 ha distribution to G. aurei during the course of our study. At the current rate of replacement and zone movement, G. aurei will reach San Marcial in ~250 years, and, unless conditions change, will completely extirpate this population of G. centralis in ~650 years.
TA B L E 3 Geographic range overlap between species and genetic groups of Thomomys (subgenus Megascapheus) pocket gophers and species of Geomydoecus chewing lice: █ indicates the major host of each chewing louse species; "s" indicates alternate louse species associations suggested by range overlap; "X" indicates alternate louse species associations that are confirmed by one or more known localities (Price & Hellenthal, 1979, 1981a, 1981b Host Geomydoecus sp.
This contact zone, a zone of competitive parapatry (sensu Haffer, 1986;Bull, 1991) from the louse's perspective, is likely to be only one of many such zones among the mosaic of 29 species of chewing lice hosted by the seven species of Thomomys in the subgenus Megascapheus that cover most of the western United States and northern Mexico. Comparison of the distributions of Geomydoecus species  and Thomomys (subgenus Megascapheus) species and genetic groups (Mathis et al., 2014;Patton & Smith, 1981Smith, 1998) indicate multiple instances of apparent range overlap that may have resulted from past host switching or may represent current zones of competitive parapatry between species of chewing lice (summarized in Table 3).
Published records of louse localities Price & Hellenthal, 1979, 1981a, 1981b already reveal 35 specific sites where multiple species of lice have been collected from a single population of pocket gophers (Table 3). These localities of sympatry not only confirm many of the apparent instances of range overlap, but may represent geographically stable zones of competitive parapatry or moving zones of species replacement. Just as "movement of hybrid zones in the present and recent past could be a widespread phenomenon" (Buggs, 2007:307), untested assumptions of stasis of species contact (without hybridization) may mask a more widespread existence of ongoing species replacement following recent environmental perturbations or ongoing climatic change. Thus, the localities of sympatry may provide additional opportunities to evaluate the spatio-temporal dynamics of both the louse species and their pocket gopher hosts over long periods of time, as well as the nature of species interactions in the zone of competitive parapatry. Such zones may signal past or ongoing human disturbances (as in this study) or other ecological change not previously detected. Our demonstration that zone metrics can be used to estimate natural history parameters of the species involved is a relatively new finding that portends well for future contact zone studies in which preserved specimens are not available to verify dating estimates based exclusively on zone metrics and empirical studies of dispersal distance are not available to corroborate distance estimates, again based solely on zone metrics.
Comparison of multiple zones of parapatry within Geomydoecus may reveal common patterns among zones of parapatry in Geomydoecus that may better inform future studies of zones of competitive parapatry in other taxa.

ACK N OWLED G EM ENTS
Fieldwork in 2016 and portions of the laboratory work represented here were funded by National Science Foundation Grant DEB 1445708 to JWD, JEL, and TAS. We are grateful for the gracious cooperation of landowners at our study sites. For assistance in the field, we thank Lucas Pietan, Sarah Huebner, and Aleyda Galán. For assistance in the laboratory, we thank Sheree Harper, Wyatt Andersen, Lucas Pietan, Natalie Espinosa, Sarah Huebner, and Brian Ross. Roger D. Price kindly confirmed louse species identity in the initial phase of this study in 1989. This is publication 1600 of the Biodiversity Research and Teaching Collections at Texas A&M University.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N
DJH and MSH conceived the idea for this synthesis and wrote the article; JWD and TAS prepared and analyzed the data; all authors participated in revising the manuscript; JWD, TAS, and JEL secured funding for the project.