LONG-TERM ASSOCIATION BETWEEN THE COMMENSAL FLORIDA MOUSE (PODOMYS FLORIDANUS) AND THE GOPHER TORTOISE (GOPHERUS POLYPHEMUS) IN THE FOSSIL RECORD OF FLORIDA NATASHA S. VITEK1,2, JAMES D. AUSTIN3, AND JONATHAN I. BLOCH1 1Division of Vertebrate Paleontology, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611-7800 USA 2Department of Biology, University of Florida, Gainesville, Florida 32611 USA 3 Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, Florida 32611 USA Vitek, N. S., J. D. Austin, and J. I. Bloch. 2017. Long-term association between the commensal Florida Mouse (Podomys floridanus) and the Gopher Tortoise (Gopherus polyphemus) in the fossil record of Florida. Bulletin of the Florida Museum of Natural History 55(5):105–116. ABSTRACT The Florida Mouse Podomys floridanus and the Gopher Tortoise Gopherus polyphemus are linked in modern ecosystems by a commensal relationship in which Podomys uses the burrows of Gopherus. How- ever, previous paleoecological research demonstrated that species interactions, including commensalisms, are not necessarily stable through geologic time. Given that the longevity of species interactions cannot be assumed, we asked: is the current association between Podomys and Gopherus a coincidental feature of the modern biota or is it a long-term phenomenon detectable in the fossil record? We explored that ques- tion using literature and collection records of Podomys and Gopherus from throughout the published geo- graphic and temporal extent of the fossil record of Podomys (Florida, ~1.35 Ma – 0.01 Ma). We expected that a long-term commensal relationship would result in the presence of Gopherus in sites preserving Podomys significantly more often than expected by chance. Given the asymmetric nature of the relation- ship, a complementary expectation is that a search for sites containing Gopherus within the same spatio- temporal extent should not result more sites containing both species than would be expected by chance. After accounting for potential collection biases, fourteen Irvingtonian and Rancholabrean sites preserving Podomys were searched for occurrences of Gopherus as well as other vertebrates whose burrows Podo- mys is reported to use. Twelve of those fourteen sites contained fossils of Gopherus. The two remaining sites contained fossils of the other candidate taxa, consistent with the hypothesis that if Podomys flori- danus did not use the burrows of Gopherus, it could have used the burrows of other species, as it does in modern ecosystems. Among the sites that are from the same spatiotemporal extent and contain Gopherus, fossils of Podomys were no more likely to be present at a site than would be expected by chance. Overall, we find evidence that Gopherus and Podomys have a long-term association consistent with their current relationship that extends at least to early Irvingtonian faunas. Key words: ecology, Gopher Tortoise, Florida Mouse, Pleistocene, Rancholabrean, Irvingtonian. Published on-line: May 8, 2017 Open access download at https://www.flmnh.ufl.edu/bulletin/publications/ ISSN 2373-9991 Copyright © 2017 by the Florida Museum of Natural History, University of Florida. All rights reserved. Text, images and other media are for nonprofit, educational, or personal use of students, scholars, and the public. Any commercial use or republication by printed or electronic media is strictly prohibited without written permission of the museum. 106 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) INTRODUCTION The fossil record contains data that extend our understanding of the stability of ecological associa- tions. A key insight from paleoecological research is that, despite the faunal stasis that is prevalent in the fossil record (Morris et al., 1995; McGill et al., 2005; Blois and Hadly, 2009; Gill et al., 2009), the responses of individual species to climate change leads to no-analogue communities through time (e.g., Hoffman, 1979; Stewart, 2008). Those same no-analog assemblages demonstrate that species distributions can change through time, and that two species currently in syntopy may not represent a long-term sympatric association (Semken Jr, 1983; Graham, 1986; Davis et al., 2014). Species pairs linked by ecological inter- actions such as mutualism, parasitism or com- mensalism might be predicted to be more closely associated through geologic time than ecologi- cally unassociated pairs from a community (Wilf et al., 2000). However, even ecological associations between species pairs are subject to change over time. For example, the ‘obligate’ predator-prey relationship between black-footed ferrets and prai- rie dogs may have emerged within the last 800,000 years, as black-footed ferrets colonized areas of North America containing prairie dogs (Owen et al., 2000). Therefore, study of the fossil record of species pairs is a potentially powerful way of understanding the evolutionary stability of such relationships. An example of such a species pair is the Florida Mouse (Podomys floridanus, hereafter Podomys) and the Gopher Tortoise (Gopherus polyphemus, hereafter Gopherus) in peninsular Florida. Podomys has an asymmetric, commensal relationship with Gopherus in which Podomys ben- efits from the presence of Gopherus but Gopherus is unaffected by the presence of Podomys. Podo- mys has been characterized as exclusively bur- row-dwelling (Layne and Jackson, 1994). Unlike the sympatric Peromyscus gossypinus, which also utilizes large burrows dug by other animals, Podo- mys is a poor burrower itself and is considered to be specialized for burrow-dwelling, as opposed to burrow-excavation, due to its limited burrowing ability, weaker nest-building behavior (Layne and Jackson, 1994), and physiological traits that make burrow-dwelling a key for its specialized xeric hab- itat (Fertig and Layne, 1963). Gopherus was previ- ously described as an ecosystem engineer due to its excavation of prominent burrows that can be re- engineered by Podomys, which uses the main bur- row or digs rudimentary small pockets and escape chimneys off of the main burrow (Jones and Franz, 1990; Kinlaw and Grasmueck, 2012). Podomys prefers the burrows of Gopherus to those of other species (Blair and Kilby, 1936; Jones and Franz, 1990; Layne, 1990; Layne and Jackson, 1994) but has also been observed entering burrows of other animals such as Peromyscus polionotus, Sigmodon hispidus, Geomys pinetis, and Dasypus novemcinc- tus (Layne, 1990). It is unclear when burrow-associated behav- ior evolved in the ancestors of Podomys and Gopherus. It may be related to the extension of a xeric province along the northern Gulf of Mexico, which would have been followed by an eastward range extension of the ancestors of Podomys, Gopherus, and other scrub associated biota from the American southwest (5–2 mya; Axelrod, 1948; Myers, 1990; Haywood et al., 2001; Reynoso and Montellano-Ballesteros, 2004). Alternatively, Podomys may have evolved as a xeric specialist dur- ing the Pleistocene when major ecological changes occurred in peninsular Florida, including the for- mation of much of the current xeric scrub ridges (Myers, 1990). The applicability of a phylogenetic bracket to infer the evolution of burrow-association is limited because the evolutionary relationships of Podomys to other neotomine species remain unre- solved (Bradley et al., 2007; Platt et al., 2015). It is also possible that a xeric habitat association alone may explain an association between the two taxa in the fossil record without the need to infer burrow- association. The longevity of the association between Gopherus and Podomys can be studied in the fos- sil record. Here, we use the known fossil record of Podomys, which is limited to the Pleistocene of Florida, to test for a long-term relationship between Podomys and Gopherus (Jones and Layne, 1993). We hypothesize that in the fossil record, Podomys VITEK ET AL.: Long-term Commensalism Between Podomys and Gopherus in Florida 107 and Gopherus do not have a significant, long-term association, consistent with other records of spe- cies disassociations over geologic time (Owen et al., 2000). Given the asymmetric nature of the rela- tionship between Podomys and Gopherus in the modern biota, the expectations for the presence of one species at a fossil site given the presence of the other are also asymmetric. Podomys benefits from the microhabitat provided by burrows of Gopherus, and it is most appropriate to test the longevity of the association through the presence or absence of Gopherus at fossils sites where Podomys is found. If the two species have a long-term association con- sistent with the preference of Podomys to occupy the burrows of Gopherus, then Gopherus should be present at sites where Podomys is found more often than expected by chance. If the current commensal relationship is a coincidental result of overlapping geographic ranges of the two species in the present day, then the 1.35 million year-long fossil record of sites containing Podomys should have no more occurrences of Gopherus than expected by chance. Examining the reverse condition, or the presence or absence of Podomys at sites where Gopherus is found, provides additional evidence about the potential for association. In contrast to the benefit that Gopherus provides Podomys, Podomys provides no cost or benefit to Gopherus. A long- term relationship consistent with the current one should result in sites containing Gopherus having no more occurrences of Podomys than expected by chance. MATERIALS AND METHODS We developed a list of sites where Podomys flori- danus is documented by searching the literature for records of the species (Bader, 1957; Pinkham, 1971; Martin, 1974; Martin and Webb, 1974; Webb, 1974; Ober, 1978; Morgan, 1991; Morgan and White, 1995; Franz and Quitmyer, 2005). In addi- tion, we queried the Florida Museum of Natural History’s Vertebrate Paleontology (FLMNH VP) online database (http://www.flmnh.ufl.edu/vertpa- leo-search/) for the genus and species ‘Podomys floridanus’ (accessed 14 April 2016). From each site where fossils of Podomys were reported, the Figure 1. Map of fossil sites containing Podomys floridanus and Gopherus polyphemus among Rancholabrean and Irvingtonian faunas of Florida. FLMNH VP database was also queried for records of Gopherus (Fig. 1). The list of sites was also checked against the most recent review of the fossil record of Gopherus (Franz and Quitmyer, 2005). Database identifications were assumed to be cor- rect. In the literature, teeth or mandibles of Podo- mys are diagnosed based primarily on their large size and the relative rarity or absence of accessory upper and lower molar cusps such as the mesostyle/ mesostylid and ectostyle/ectostylid (Martin, 1967; Pinkham, 1971). The identification of Gopherus is based primarily on shell features, as reviewed in Franz and Quitmyer (Franz and Quitmyer, 2005). A single specimen of Podomys was reported from the Fort Meade Mine #7 Dragline, but we excluded this site from further analysis. The mine generally preserves fossils from the Pliocene Pal- metto Fauna, but that specimen likely comes from overlying Pleistocene sands because all other fos- sils of Podomys are from Irvingtonian, Rancho- labrean, or Holocene North American Land Mam- mal Age (NALMA) sites and other collections http://www.flmnh.ufl.edu/vertpaleo-search/ http://www.flmnh.ufl.edu/vertpaleo-search/ 108 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) from the mine draglines are known to be tempo- rally mixed (Webb et al., 2008). We also excluded he site Inglis 1A from analyses, despite reports of the presence of Podomys by Martin and Webb (1974), because of contrary reports in Webb (1974) and Ruez (2001). In order to account for poorly sampled sites, we recorded the total number of specimens cata- logued at each site to evaluate the possibility that fossils of Gopherus were present at the site but missed by chance. We excluded sites contain- ing fewer than 50 fossils from further analyses (Forcino, 2012). In addition, we assembled faunal lists and site descriptions for localities at which Gopherus was not present using the FLMNH VP database to explore what might explain the absence of Gopherus. The FLMNH VP database contains over 37,000 records of Pleistocene small mam- mals and more than 560,000 records in total, and is therefore appropriate for such searches. If a significant number of sites containing Podomys did not also contain Gopherus, we were interested to know if other burrowing animals were associated with Podomys. For each site, we also searched the FLMNH VP database for other taxa whose burrows Podomys floridanus is associated with in modern ecosystems: Dasypus spp., Geomys spp., Peromyscus polionotus, and Sigmodon spp. (Layne, 1990). Queries to the FLMNH VP data- base used genus names in the absence of species names to account for chronospecies as well as a range of identification precision. The exception to the practice of query-by-genus was for Peromyscus polionotus, because other extant species of Pero- myscus are present in the Pleistocene fossil record of Florida. Although many species of Peromyscus are difficult to discriminate from each other based on isolated teeth in the fossil record, P. polionotus can be discriminated from other species of Pero- myscus in the fossil record of Florida based on size, similar to Podomys (Pinkham, 1971). It is not clear whether or not Dasypus bel- lus constructed burrows that could have been used by Podomys. However, the current hypotheses of close relationships to either (a) Dasypus kappleri or (b) Dasypus novemcincus correspond to the recon- struction of D. bellus as a burrower commensurate to Dasypus novemcinctus (Vizcaíno and Milne, 2002; Rincón et al., 2008). We therefore included records of D. bellus in our analyses of potential burrow associates of Podomys. We constructed a second, complementary dataset by searching the FLMNH VP database for all sites containing Gopherus within the same spatiotemporal extent as the fossil record of Podo- mys (i.e., Irvingtonian to Rancholabrean faunas of Florida; Fig. 1). The site list was checked against the review of Franz and Quitmyer (2005) and all missing sites were added to the dataset. We also collected the total number of fossils catalogued per site and records for the same suite of taxa exam- ined in the first analysis. In addition, it is possible that Podomys was present, but not collected, at a site that wasn’t screenwashed. Furthermore, it is possible that specimens of Podomys, as a member of Peromyscus sensu lato (Platt et al., 2015), and P. polionotus were more coarsely identified as Pero- myscus sp. at certain sites where they are present. To account for these possibilities, records for all rodents from each candidate site were collected as a proxy for screenwash effort. In those cases where small rodent teeth were not collected, it is highly unlikely that taxa such as Podomys would be iden- tified even if they were present in the fauna. Those sites that did not contain any rodents, or which did not contain teeth of small rodents such as Sigmo- don, Geomys, Neotoma, Peromyscus, or even inde- terminate Muridae were excluded from analyses. We further searched for any records of Peromyscus sp. at remaining sites to examine whether identifi- cation practices could have affected our results in regards to Podomys and Peromyscus polionotus. Chi-square tests are sometimes used to com- pare species occurrences to test for association with the null hypothesis that the two species are both independent of each other in locality occu- pation (McCulloch, 1985; Calede et al., 2011). However, in this case it is not informative to test for symmetrical independence of the two species because the alternative hypothesis under investiga- tion addresses an asymmetrical presence-absence expectation. In the particular case of Podomys and VITEK ET AL.: Long-term Commensalism Between Podomys and Gopherus in Florida 109 Gopherus, the independence of the presence of Gopherus with respect to Podomys might result in a failure to reject the null hypothesis using a chi- square test, but that result would be uninformative about the potential non-independence of Podomys with respect to Gopherus. Instead, exact, one-tailed binomial tests were used to evaluate the probability of significant species-pair associations at a greater frequency than the null hypothesis. The null hypothesis of no significant association was mod- elled with a probability of success of 0.5. For each dataset, the p-values of the five tests were classi- cally Bonferroni corrected by dividing original p-values by the number of tests and accepting as significant those transformed values that remained less than an alpha of 0.05 (Holm, 1979). Data were organized and analyzed in R 3.2.4 (R Core Team, 2015). RESULTS Podomys floridanus is known from a total of 17 sites, containing Irvingtonian to Rancholabrean faunas. Based on the estimated age of those fau- nas, the fossil record of P. floridanus spans ~1.35 Ma – 0.01 Ma (Bell et al., 2004). Of those sites, three (Arredondo 2C, Haile 13B, Haile 14B) were excluded because of low total sample sizes. Of the remaining 14 sites, Gopherus was present at 12 (86%, Fig. 2A, Table 1), signifi- cantly more than would be expected by chance (p = 0.032). The 14 sites span the Irvingtonian and Rancholabrean NALMA. The two remaining sites where Podomys was present but Gopherus was absent were both Rancholabrean and contained other burrowing vertebrates. At Warm Mineral Springs, Peromyscus polionotus and Sigmodon hispidus were recovered. At Haile 11B, Dasypus bellus, Geomys pinetis, P. polionotus, and S. hispi- dus were all present. Of the other burrowing taxa with which Podomys is known to associate, only S. hispidus had a significantly strong association with Podomys (p = 0.005, Fig. 2A). Based on a second database and literature search, Gopherus was present at 67 sites in the same spatiotemporal extent as the fossil record of Podomys. Of those 67 sites, 23 were removed for Figure 2. Counts of burrowing taxon occurrenc- es at A. fourteen fossil sites in Florida containing Podomys floridanus or B. thirty-three fossil sites from the same spatial and biostratigraphic extent that contain Gopherus polyphemus. Sites are first discriminated by North American Land Mammal Age (left), then combined to show total site counts (right). Thin, black lines on bars indicate confi- dence intervals for one-tailed binomial tests. Thick, dashed black lines indicate the number of sites ex- pected according to the null hypothesis (species present at 50% of sites that preserve a given spe- cies of interest). 110 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) TA B L E 1 . F lo rid a fo ss il si te s co nt ai ni ng P od om ys fl or id an us th at m ee t m in im um c rit er ia o f f os si l a bu nd an ce a nd p re se nc e of tu rtl es (s ee M at er ia ls & M et ho ds f or ju st ifi ca tio n) , a s w el l a s re co rd s of r ep or te d bu rr ow a ss oc ia te s. A bb re vi at io ns : n = to ta l n um be r of s pe ci m en s ca ta lo gu ed fr om th e si te . Si te n Po do m ys G op he ru s D as yp us G eo m ys P. p ol io no tu s Si gm od on La nd M am m al A ge A rr ed on do 1 A 37 2 + + + + + R an ch ol ab re an A rr ed on do 2 A 69 2 + + + + + + R an ch ol ab re an C ol em an 2 A 94 9 + + + + + Ir vi ng to ni an D ev il' s D en 34 10 + + + + + + R an ch ol ab re an H ai le 1 1A 73 + + + + + R an ch ol ab re an H ai le 1 1B 13 35 + + + + + R an ch ol ab re an H ai le 2 1A 20 19 + + + Ir vi ng to ni an L ec an to 2 A 21 31 + + + + + R an ch ol ab re an M on ke y Ju ng le H am m oc k 10 21 + + + + R an ch ol ab re an R ed di ck 1 A 10 44 9 + + + + + + R an ch ol ab re an S ab er to ot h C av e 11 69 + + + + + R an ch ol ab re an S ur pr is e C av e 22 81 + + + + + R an ch ol ab re an V er o 16 94 + + + + + + R an ch ol ab re an W ar m M in er al S pr in gs 26 3 + + + R an ch ol ab re an VITEK ET AL.: Long-term Commensalism Between Podomys and Gopherus in Florida 111 having low sample sizes. An additional 11 sites were removed based on an absence of fossilized teeth of small rodents, which would preclude the identification of taxa such as Podomys. Of the remaining 33 sites containing fossils of Gopherus, Podomys was only recovered in the same 12 sites discussed above, or 36% of the sites (Table 1). Of the five other mammalian taxa studied, three (Dasypus, Geomys, and Sigmodon) had a signifi- cant association with Gopherus (p = 0.003, 0.003, <0.000001, respectively, Fig. 2B). Peromyscus sp., but not Podomys floridanus or Peromyscus polionotus, was found at four of the 33 candidate sites (Haile 8A, Haile 13A, Haile 14A, and Reddick 1C). It is possible that fossils of P. polionotus and Podomys, as members of Peromys- cus sensu lato, may not have been identified to the species level in these four sites. However, even if those collections of Peromyscus sp. were to contain fossils of both Podomys and P. polionotus, then the two taxa would still be present in fewer than 50% of the candidate sites containing Gopherus (16 and 11 sites, 48% and 33%, respectively). DISCUSSION Throughout the known fossil record of Podo- mys, the species has a significant association with Gopherus. Podomys also has a close association with Sigmodon spp. throughout the same inter- val. Both Podomys and Sigmodon are known to inhabit burrows of Gopherus (Lips, 1991; Witz et al., 1991). Although co-occurrence at the same site does not demonstrate a commensal relationship between any pair of these species, it is consistent with such a hypothesis. A reciprocal, close associa- tion of Podomys in the fossil record of Gopherus was not found in our study (Fig. 2B). That lack of association is also consistent with a commen- sal relationship in which individuals of Podomys benefits from the presence of Gopherus burrows, but individuals of Gopherus are unaffected by the presence or absence of Podomys. Division of the faunas into Rancholabrean (160-10 ka; Bell et al., 2004) and Irvingtonian (1,350-160 ka; Bell et al., 2004) NALMAs support the hypothesis that the association between the two taxa is not limited to a single portion of the fossil record of Podomys, but is present throughout. In addition, the presence of other burrowing species, specifically Geomys pinetis, Peromyscus poliono- tus, and Sigmodon hispidus, and likely the extinct Dasypus bellus (Vizcaíno and Milne, 2002; Rincón et al., 2008), at the same fossil sites as Podomys supports the hypothesis that individuals of Podo- mys need not have dug their own burrows since the Irvingtonian. The two sites containing Podomys but lack- ing Gopherus are not taphonomically remarkable in comparison to other fossil sites from the fissure fills and sinkholes of other springs and the Haile quarry (Clausen et al., 1975; Morgan and Emslie, 2010). Based on the faunal lists assembled, both sites contain fossils of the terrestrial, giant tortoise Hesperotestudo as well as other terrestrial mam- mals. It is not known if Hesperotestudo burrowed like Gopherus. There is no a priori reason to expect Gopherus to be absent from either site. Alternative hypotheses could be proposed to explain the co-occurrence patterns found in this study. The co-occurrence of Podomys, Peromyscus polionotus, and Geomys together was previously proposed to be a proxy for a xeric environment (Franz and Quitmyer, 2005). In the modern biota, those three taxa are restricted to well-drained soils (Gentry and Smith, 1968; Wilkins, 1987; Franz and Quitmyer, 2005). It is possible that Geomys, P. polionotus, and Gopherus are frequently found at sites with Podomys because they all inhabited the same xeric habitat and not because any taxon was inhabiting the burrows of any other taxon. The two hypotheses, species interaction and habitat similar- ity, are not mutually exclusive and mirror present- day conditions for those species (Jones and Layne, 1993). The two hypotheses would only be sepa- rable in the fossil record if both Podomys and one of the other candidate species were found together at a site that was demonstrably not derived from a xeric habitat. Such a situation would indicate that the habitat tolerances had changed over time (Findley, 1964; Davis et al. 2014). If all four taxa had habitat restrictions in the past similar to their current habitat restrictions, then we expect to see a 112 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) large number of sites that preserve Gopherus, Geo- mys, and P. polionotus given the presence Podo- mys. The lack of significant association of Geomys and P. polionotus with Podomys in the fossil record after p-values are corrected for multiple tests (Fig. 2A) is evidence against the habitat similarity hypothesis. Although the fossil record of Gopherus in Florida extends as far back as the hypothesized origin of the local, xeric scrub habitat in the late Pliocene or early Pleistocene (Myers, 1990; Franz and Quitmyer, 2005), the published fossil record of Podomys, including Podomys nov. sp. or Podomys sp., is notably shorter (Morgan and White, 1995; Ruez, 2001). The expectation for the reverse condition, the presence or absence of other xeric taxa given the presence of Gopherus, may not be as informa- tive for predictions about Podomys. In the modern biota, Geomys and Gopherus are found in a wider range of habitats than Podomys (Wilkins, 1987; Endries et al., 2009), and they might therefore be more likely to be preserved at a wider range of sites. The result of a significant association of Geomys, but not Podomys, with the presence of Gopherus may reflect those differences in habitats. An additional hypothesis is that Podomys floridanus and Peromyscus polionotus are found at relatively fewer sites than other taxa because of differences in identification. That is, they were searched for in the database at the species level while other taxa were searched for at the generic level. When additional records of Peromyscus spp. are provisionally added to the counts of either spe- cies, in order to account for this potential bias, the two species are still not significantly associated with occurrences of Gopherus. In short, the evidence supports the hypoth- esis that (1) Podomys has been associated with bur- rows of Gopherus since the early Pleistocene, or (2) Podomys occupied a xeric habitat even more restricted than that occupied by other taxa such as Geomys and Gopherus, or (3) some combination of both of the previous two hypotheses. All three scenarios support the hypothesis that Podomys evolved from an ancestor that migrated from arid regions of Mexico and the southwestern United States during the late Miocene or Pliocene (Wil- liams et al., 1985; Platt et al., 2015). An improved understanding of the paleon- tological history of species associations can help inform conservation management strategies (Willis and Birks, 2006; Hadly and Barnosky, 2009; Sed- don et al., 2014). On one hand, uncovering broader association patterns that are less strict than they appear in the modern biota can open up manage- ment strategies that are not apparent from studying living animals alone (Owen et al., 2000; Hadly and Barnosky, 2009). On the other hand, documenting long-term species associations that extend through periods of abiotic change can highlight the evolu- tionary and ecological importance of those specific relationships and lend weight to certain conser- vation priorities (Willis et al., 2010). In the case documented here, evidence from the fossil record supports the hypothesis that a close association between Podomys floridanus and Gopherus poly- phemus is part of the baseline ecological state of Floridian scrub and sandhill communities since the early Pleistocene. ACKNOWLEDGMENTS Thanks to C.J. Bell for critical discussions of paleo- ecology that inspired this research, and to R. Hul- bert and the course ZOO 6927, Florida Vertebrate Paleontology, for providing the opportunity for this research, and to two anonymous reviewers whose comments improved the manuscript. This material is based upon work supported by NSF Graduate Research Fellowship DGE-1315138 to N.S.V. and NSF CSBR 1203222 to J.I.B. This is the University of Florida Contribution to Paleobiology 828. LITERATURE CITED Axelrod, D. I. 1948. Climate and evolution in western North America during mid- dle Pliocene time. Evolution 2:127–144: DOI:10.2307/2405373. Bader, R. S. 1957. Two Pleistocene mammalian faunas from Alachua County, Florida. Bulle- tin of the Florida State Museum, Biological Sciences 2:53–75. Bell, C. J., E. L. Lundelius Jr, A. D. Barnosky, R. VITEK ET AL.: Long-term Commensalism Between Podomys and Gopherus in Florida 113 W. Graham, E. H. Lindsay, D. R. Ruez Jr, H. A. Semken Jr, S. D. Webb, and R. J. Zakrze- wski. 2004. The Blancan, Irvingtonian, and Rancholabrean mammal ages. Pp. 232–314 in M. O. Woodburne (ed.), Late Cretaceous and Cenozoic Mammals of North America. Columbia University Press, New York. Blair, W. F., and J. D. Kilby. 1936. The Gopher Mouse —Peromyscus floridanus. Journal of Mam- malogy 17:421–422: DOI:10.2307/1374409. Blois, J. L., and E. A. Hadly. 2009. Mamma- lian response to Cenozoic climatic change. Annual Review of Earth and Planetary Sci- ences 37:181–208: DOI:10.1146/annurev. earth.031208.100055. Bradley, R. D., N. D. Durish, D. S. Rogers, J. R. Miller, M. D. Engstrom, and C. W. Kilpatrick. 2007. Toward a molecular phylogeny for Pero- myscus: evidence from mitochondrial cyto- chrome-b sequences. Journal of Mammalogy 88:1146–1159. DOI: 10.1644/06-MAMM-A- 342R.1. Calede, J. J. M., S. S. B. Hopkins, and E. B. Davis. 2011. Turnover in burrowing rodents: The roles of competition and habitat change. Pal- aeogeography, Palaeoclimatology, Palaeo- ecology 311:242–255: DOI:10.1016/j.pal- aeo.2011.09.002. Clausen, C. J., H. K. Brooks, and A. B. Wesolowsky. 1975. Florida spring confirmed as 10,000 year old early man site. The Florida Anthropolo- gist 28:1–38. Davis, E. B., J. L. McGuire, and J. D. Orcutt. 2014. Ecological niche models of mammalian gla- cial refugia show consistent bias. Ecography 37:1133–1138: DOI:10.1111/ecog.01294. Endries, M., B. Stys, G. Mohr, G. Kratimenos, S. Langley, K. Root, and R. Kautz. 2009. Wildlife habitat conservation needs in Florida: updated recommendations for strategic habitat con- servation areas. Fish and Wildlife Research Institute Technical Reports TR-15:1–178. Fertig, D. S., and J. N. Layne. 1963. Water relation- ships in the Florida mouse. Journal of Mam- malogy 44:322: DOI:10.2307/1377202. Findley, J. S. 1964. Paleoecologic reconstruction: vertebrate limitations. Pp. 23–25 in J. J. Hes- ter and J. Schoenwetter (eds.), The recon- struction of past environments. Publication 3. Fort Burgwin Research Center, Rancho de Taos, New Mexico. Forcino, F. L. 2012. Multivariate assessment of the required sample size for community paleoeco- logical research. Palaeogeography, Palaeocli- matology, Palaeoecology 315–316:134–141: DOI:10.1016/j.palaeo.2011.11.019. Franz, R., and I. R. Quitmyer. 2005. A fossil and zooarchaeological history of the gopher tor- toise (Gopherus polyphemus) in the south- eastern United States. Bulletin of the Florida Museum of Natural History 45:179–199. Gentry, J. B., and M. H. Smith. 1968. Food habits and burrow associates of Peromyscus polion- otus. Journal of Mammalogy 49:562–565: DOI: 10.2307/1378235. Gill, J. L., J. W. Williams, S. T. Jackson, K. B. Lininger, and G. S. Robinson. 2009. Pleis- tocene megafaunal collapse, novel plant communities, and enhanced fire regimes in North America. Science 326:1100–1103: DOI:10.1126/science.1179504. Graham, R. W. 1986. Responses of mammalian communities to environmental changes dur- ing the late Quaternary. Pp. 300–313 in J. Diamond and T. J. Case (eds.), Community Ecology. Harper and Row, New York. Hadly, E. A., and A. D. Barnosky. 2009. Vertebrate fossils and the future of conservation biol- ogy. Pp. 39–59 in G. P. Dietl and K. W. Flessa (eds.), Conservation Paleobiology: Using the Past to Manage for the Future, Paleontologi- cal Society Short Course, October 17th, 2009. The Paleontological Society Papers vol. 15. Haywood, A. M., P. J. Valdes, B. W. Sellwood, J. O. Kaplan, and H. J. Dowsett. 2001. Model- ling Middle Pliocene warm climates of the USA. Palaeontologia Electronica 4:1–21. Hoffman, A. 1979. Community paleoecology as an epiphenomenal science. Paleobiology 5:357– 379: DOI: 10.1017/S0094837300016894. Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian Journal 114 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) of Statistics 6:65-70. Jones, C. A., and R. Franz. 1990. Use of gopher tortoise burrows by Florida mice (Podomys floridanus) in Putnam County, Florida. Flor- ida Field Naturalist 18:45–51. Jones, C. A., and J. N. Layne. 1993. Podomys flori- danus. Mammalian Species 427:1–5. Kinlaw, A., and M. Grasmueck. 2012. Evidence for and geomorphologic consequences of a reptilian ecosystem engineer: the burrow- ing cascade initiated by the Gopher Tor- toise. Geomorphology 157–158:108–121: DOI:10.1016/j.geomorph.2011.06.030. Layne, J. N. 1990. The Florida mouse. Pp. 1–21 in C. K. Dodd Jr, R. E. Ashton, R. Franz, and E. Wester (eds.), Burrow Associates of the Gopher Tortoise. Proceedings of the 8th Annual Meeting of the Gopher Tortoise Council, Auburn University, Auburn, Ala- bama, 13-15 November 1987. Gopher Tor- toise Council, Auburn, Alabama. Layne, J. N., and R. J. Jackson. 1994. Burrow use by the Florida Mouse (Podomys floridanus) in South-Central Florida. American Midland Naturalist 131:17–23: DOI:10.2307/2426604. Lips, K. R. 1991. Vertebrates associated with tor- toise (Gopherus polyphemus) burrows in four habitats in south-central Florida. Journal of Herpetology 25:477: DOI:10.2307/1564772. Martin, R. A. 1967. A comparison of two mandibu- lar dimensions in Peromyscus with regard to identification of Pleistocene Peromyscus from Florida. Tulane Studies in Zoology 14:75–79. Martin, R. A. 1974. Fossil mammals from the Cole- man IIA Fauna, Sumter County. Pp. 35–99 in S. D. Webb (ed.), Pleistocene Mammals of Florida. The University Press of Florida, Gainesville. Martin, R. A., and S. D. Webb. 1974. Late Pleisto- cene mammals from the Devil’s Den Fauna, Levy County. Pp. 114–145 in S. D. Webb (ed.), Pleistocene Mammals of Florida. The University Presses of Florida, Gainesville. McCulloch, C. E. 1985. Variance tests for spe- cies association. Ecology 66:1676–1681: DOI:10.2307/1938030. McGill, B. J., E. A. Hadly, and B. A. Maurer. 2005. Community inertia of Quaternary small mam- mal assemblages in North America. Proceed- ings of the National Academy of Sciences of the United States of America 102:16701– 16706. Morgan, G. S. 1991. Neotropical Chiroptera from the Pliocene and Pleistocene of Florida. Pp. 176–213 in T. A. Griffiths, D. Klingener, and K. F. Koopman (eds.), Contributions to Mam- malogy in honor of Karl F. Koopman. Bulle- tin of the American Museum of Natural His- tory vol. 206. American Museum of Natural History, New York. Morgan, G. S., and J. A. White. 1995. Small mam- mals (Insectivora, Lagomorpha, and Roden- tia) from the early Pleistocene (early Irving- tonian) Leisey Shell Pit Local Fauna, Hills- borough Florida. Pp. 397–462 in R. C. Hul- bert, Jr., G. S. Morgan, and S. D. Webb (eds.), Paleontology and Geology of the Leisey Shell Pits, Early Pleistocene of Florida, Part II. Bul- letin of the Florida Museum of Natural His- tory vol. 37. Morgan, G. S., and S. D. Emslie. 2010. Tropical and western influences in vertebrate faunas from the Pliocene and Pleistocene of Flor- ida. Quaternary International 217:143–158: DOI:10.1016/j.quaint.2009.11.030. Morris, P. J., L. C. Ivany, K. M. Schopf, and C. E. Brett. 1995. The challenge of paleoecologi- cal stasis: reassessing sources of evolutionary stability. Proceedings of the National Acad- emy of Sciences 92:11269–11273. Myers, R. L. 1990. Scrub and high pine. Pp. 150– 193 in Ecosystems of Florida. University of Central Florida Press, Orlando. Ober, L. D. 1978. The Monkey Jungle, a Late Pleistocene fossil site in southern Florida. The Plaster Jacket 23:1–13. Owen, P. R., C. J. Bell, and E. M. Mead. 2000. Fossils, diet, and conservation of black- footed ferrets (Mustela nigripes). Jour- nal of Mammalogy 81:422–433: DOI: 10.1644/1545-1542(2000)081<0422:FDAC OB>2.0.CO;2. VITEK ET AL.: Long-term Commensalism Between Podomys and Gopherus in Florida 115 Pinkham, C. F. A. 1971. Peromyscus and Ochrot- omys from the Pleistocene of Reddick, Florida. Journal of Mammalogy 52:28–40: DOI:10.2307/1378429. Platt, R. N., B. R. Amman, M. S. Keith, C. W. Thompson, and R. D. Bradley. 2015. What is Peromyscus? Evidence from nuclear and mitochondrial DNA sequences suggests the need for a new classification. Journal of Mam- malogy 96:708–719: DOI:10.1093/jmammal/ gyv067. R Core Team. 2015. R: A Language and Environ- ment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, pp. Reynoso, V.-H., and M. Montellano-Balleste- ros. 2004. A new giant turtle of the genus Gopherus (Chelonia: Testudinidae) from the Pleistocene of Tamaulipas, México, and a review of the phylogeny and biogeography of gopher tortoises. Journal of Vertebrate Pale- ontology 24:822–837: DOI:10.1671/0272- 4634(2004)024[0822:ANGTOT]2.0.CO;2. Rincón, A. D., R. S. White, and H. G. Mcdonald. 2008. Late Pleistocene cingulates (Mamma- lia: Xenarthra) from Mene de Inciarte Tar Pits, Sierra de Perijá, western Venezuela. Jour- nal of Vertebrate Paleontology 28:197–207: DOI:10.1671/0272-4634(2008)28[197:LPC MXF]2.0.CO;2. Ruez, D. R. 2001. Early Irvingtonian (Latest Pliocene) rodents from Inglis 1C, Citrus County, Florida. Journal of Vertebrate Pale- ontology 21:153–171: DOI:10.1671/0272- 4634(2001)021[0153:EILPRF]2.0.CO;2. Seddon, A. W. R., A. W. Mackay, A. G. Baker, H. J. B. Birks, E. Breman, C. E. Buck, E. C. Ellis, C. A. Froyd, J. L. Gill, L. Gillson, E. A. John- son, V. J. Jones, S. Juggins, M. Macias-Fauria, K. Mills, J. L. Morris, D. Nogués-Bravo, S. W. Punyasena, T. P. Roland, A. J. Tanentzap, K. J. Willis, M. Aberhan, E. N. van Asperen, W. E. N. Austin, R. W. Battarbee, S. Bhagwat, C. L. Belanger, K. D. Bennett, H. H. Birks, C. Bronk Ramsey, S. J. Brooks, M. de Bruyn, P. G. Butler, F. M. Chambers, S. J. Clarke, A. L. Davies, J. A. Dearing, T. H. G. Ezard, A. Feurdean, R. J. Flower, P. Gell, S. Hausmann, E. J. Hogan, M. J. Hopkins, E. S. Jeffers, A. A. Korhola, R. Marchant, T. Kiefer, M. Lamen- towicz, I. Larocque-Tobler, L. López-Merino, L. H. Liow, S. McGowan, J. H. Miller, E. Montoya, O. Morton, S. Nogué, C. Onou- friou, L. P. Boush, F. Rodriguez-Sanchez, N. L. Rose, C. D. Sayer, H. E. Shaw, R. Payne, G. Simpson, K. Sohar, N. J. Whitehouse, J. W. Williams, and A. Witkowski. 2014. Look- ing forward through the past: identification of 50 priority research questions in palaeo- ecology. Journal of Ecology 102:256–267: DOI:10.1111/1365-2745.12195. Semken Jr, H. A. 1983. Holocene mammalian bio- geography and climatic change in the eastern and central United States. Pp. 182–207 in H. E. Wright Jr (ed.), Late-Quaternary Environ- ments of the United States, Volume 2: The Holocene. University of Minnesota Press. Stewart, J. 2008. The progressive effect of the individualistic response of species to Qua- ternary climate change: an analysis of Brit- ish mammalian faunas. Quaternary Science Reviews 27:2499–2508: DOI:10.1016/j.qua- scirev.2008.08.003. Vizcaíno, S. F., and N. Milne. 2002. Structure and function in armadillo limbs (Mam- malia: Xenarthra: Dasypodidae). Journal of Zoology 257:117–127: DOI:10.1017/ S0952836902000717. Webb, S. D. 1974. Chronology of Florida Pleisto- cene mammals. Pp. 5–31 in S. D. Webb (ed.), Pleistocene Mammals of Florida. The Univer- sity Presses of Florida, Gainesville. Webb, S. D., R. C. Hulbert, Jr., G. S. Morgan, and H. F. Evans. 2008. Terrestrial mammals of the Palmetto Fauna (early Pliocene, latest Hemphillian) from the Central Florida Phos- phate District. Pp. 293–312 in X. Wang and L. G. Barnes (eds.), Geology and Vertebrate Paleontology of Western and Southern North America, Contributions in Honor of David P. Whistler. Science Series vol. 41. Natural His- tory Museum of Los Angeles County, Los 116 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 55(5) Angeles. Wilf, P., C. C. Labandeira, W. J. Kress, C. L. Staines, D. M. Windsor, A. L. Allen, and K. R. Johnson. 2000. Timing the radiations of leaf beetles: hispines on gingers from latest Cretaceous to Recent. Science 289:291–294: DOI:0.1126/science.289.5477.291. Wilkins, K. T. 1987. Zoogeographic analysis of variation in recent Geomys pinetis (Geomy- idae) in Florida. Bulletin of the Florida State Museum, Biological Sciences 30:1–28. Williams, S. L., J. Ramirez-Pulido, and R. J. Baker. 1985. Peromyscus alstoni. Mammalian Spe- cies 242:1–4. Willis, K. J., and H. J. B. Birks. 2006. What is natu- ral? The need for a long-term perspective in biodiversity conservation. Science 314:1261– 1265: DOI:10.1126/science.1122667. Willis, K. J., R. M. Bailey, S. A. Bhagwat, and H. J. B. Birks. 2010. Biodiversity baselines, thresholds and resilience: testing predictions and assumptions using palaeoecological data. Trends in Ecology & Evolution 25:583–591: DOI:10.1016/j.tree.2010.07.006. Witz, B. W., D. S. Wilson, and M. D. Palmer. 1991. Distribution of Gopherus polyphemus and its vertebrate symbionts in three burrow catego- ries. American Midland Naturalist 126:152– 158.