xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003420
The distribution and population dynamics of the golden mouse (_ochrotomys nuttalli_) at its southern range periphery
h [electronic resource] /
by Sarah Smiley.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (M.S.)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: This research assesses the status of the golden mouse (Ochrotomys nuttalli) in Florida by taking a multi-pronged approach. Geographic Information Systems (GIS) was used to understand the distribution of habitats and occurrence records for this species within the state. Presence-absence trapping occurred at 13 study sites to determine if historic southern periphery populations were still occupied, gauge if more central populations were being maintained, and document golden mice in previously unrecorded areas. In addition, surveys for O. nuttalli took place at regular intervals at the USF Ecological Research Area to understand how populations of this species fluctuate over time and ensure that individuals could be caught during the months when statewide trapping was occurring. Trapping data from all 14 sites were combined to determine a level of confidence for absences at each site which did not yield a golden mouse capture. Finally, I determined the relative abundance of golden mice relative to other small mammal species caught. Locality records for this species align closely with the distribution of hardwood-containing habitats in Florida. The distribution of O. nuttalli is not continuous across Florida and becomes increasingly patchy near the southern range periphery of this species. In south-central Florida, populations are restricted to regions where hardwoods extend south along one of three upland ridges. Golden mice were determined to be present in the vicinity of the southernmost historic sites on each of these ridges. Ochrotomys nuttalli were captured at six of the 13 sites surveyed. At the USF Ecological Research Area, O. nuttalli were captured in all months surveyed although abundances remained relatively low from October through January and then increased from February through May. At study sites which did not catch a golden mouse, 78.6 to 100% of the trapping periods which successfully caught a golden mouse had done so by the effort levels invested at these absent sites. Ochrotomys nuttalli was the fourth most abundant of 12 species captured, but several of the species caught less frequently than golden mice are non-native or too large to have their true abundance reflected by these trapping methods.
Advisor: Earl D. McCoy, Ph.D.
x Biology Integrative
t USF Electronic Theses and Dissertations.
The Distribution and Population Dynamics of the Gol den Mouse ( Ochrotomys nuttalli ) at Its Southern Range Periphery by Sarah A. Smiley A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Integrative Biology College of Arts and Sciences University of South Florida Co-Major Professor: Earl D. McCoy, Ph.D. Co-Major Professor: Henry R. Mushinsky, Ph.D. Gordon A. Fox, Ph.D. Date of Approval: February 22, 2010 Keywords: biogeography, conservation, scrub, small mammal, landscape ecology Copyright 2010, Sarah A. Smiley
To my parents, who have given their love and suppor t. This is possible because of you.
Acknowledgements This research would not have taken place without th e cooperation of many state biologists and agencies. The following state biolo gists were helpful in providing capture locales of golden mice: A. Bard, R. Bolt, X. Desegu in, J. Duby, T. Hingtgen, H. Hoffman, B. Kaiser, E. McCoy, J. Stout, and D. Sump ter. Site access was granted by the following organizations: Archbold Biological Statio n, The Nature Conservancy, The University of South Florida, the Southwest Florida Water Management District, Florida State Parks (Division III), Tall Timbers Research S tation, and the governments of Lee, Charlotte, Hillsborough, and Brevard counties. Aly son Dagly, Diana Daniels, Leo Gude, Mike Grispo, Damitha Gunasekera, and Bill Ryerson a ssisted with field work at the University of South FloridaÂ’s Ecological Research A rea. Dr. Gordon Fox contributed rodent bait to the project. Above all, I would lik e to acknowledge my advisors, Dr. Henry Mushinsky and Dr. Earl McCoy, for their suppo rt and patience (on a mammal project nonetheless).
i Table of Contents List of Tables iii List of Figures iv Abstract v Introduction 1 Methods 6 Determining the Geographic Range of Golden Mice an d Distribution of Habitats Using Geographic Information Systems (GIS) 6 Temporal Changes in Abundance 9 Assessment of Current Extent of Occurrence 10 Relative Abundance of the Golden Mouse 14 Results 15 Determining the Geographic Range of Golden Mice an d Distribution of Habitats Using Geographic Information Systems (GIS) 15 Temporal Changes in Abundance 17 Assessment of Current Extent of Occurrence 18 Relative Abundance of the Golden Mouse 21 Discussion 22 Literature Cited 40
ii List of Tables Table 1 Capture success of Ochrotomys nuttalli at the eleven sites within the documented range of this species durin g 2007 and 2008 35 Table 2 Effort on Â‘absentÂ’ study sites and the prop ortion of Â‘presentÂ’ sites with a first Ochrotomys nuttalli capture before the amount of effort at each Â‘absentÂ’ site 37
iii List of Figures Figure 1 Probable historic distribution of Ochrotomys nuttalli floridanus 31 Figure 2 A probable current distribution of Ochrotomys nuttalli with focus on south-central Florida where histori cally populations were the most naturally fragmented 32 Figure 3 Layout of the sixty trap stations operated at the USF Ecological Research Area 33 Figure 4 The locations of the fourteen study sites on which trapping occurred in search of Ochrotomys nuttalli 34 Figure 5 The percentage of Â‘presentÂ’ trapping perio ds which resulted in a capture of Ochrotomys nuttalli by a given level of uncorrected (A) and corrected (B) effort 36 Figure 6 Abundance of Ochrotomys nuttalli over time based on the number of individuals captured each trap period a t the USF Eco Area 38 Figure 7 Relative abundance of small mammal species captured in Sherman live traps at the eleven locales within t he documented geographic range of Ochrotomys nuttalli 39
iv The Distribution and Population Dynamics of the Gol den Mouse ( Ochrotomys nuttalli ) at its Southern Range Periphery Sarah A. Smiley ABSTRACT This research assesses the status of the golden mou se ( Ochrotomys nuttalli ) in Florida by taking a multi-pronged approach. Geographic Inform ation Systems (GIS) was used to understand the distribution of habitats and occurre nce records for this species within the state. Presence-absence trapping occurred at 13 st udy sites to determine if historic southern periphery populations were still occupied, gauge if more central populations were being maintained, and document golden mice in previously unrecorded areas. In addition, surveys for O. nuttalli took place at regular intervals at the USF Ecologi cal Research Area to understand how populations of this species fluctuate over time and ensure that individuals could be caught during the months when statewide trapping was occurring. Trapping data from all 14 sites were co mbined to determine a level of confidence for absences at each site which did not yield a golden mouse capture. Finally, I determined the relative abundance of golden mice relative to other small mammal species caught. Locality records for this species align closely with the distribution of hardwood-containing habitats in Florida. The distr ibution of O. nuttalli is not continuous across Florida and becomes increasingly patchy near the southern range periphery of this species. In south-central Florida, populations are restricted to regions where hardwoods extend south along one of three upland ridges. Gol den mice were determined to be
v present in the vicinity of the southernmost histori c sites on each of these ridges. Ochrotomys nuttalli were captured at six of the 13 sites surveyed. At the USF Ecological Research Area, O. nuttalli were captured in all months surveyed although abun dances remained relatively low from October through Januar y and then increased from February through May. At study sites which did not catch a golden mouse, 78.6 to 100% of the trapping periods which successfully caught a golden mouse had done so by the effort levels invested at these absent sites. Ochrotomys nuttalli was the fourth most abundant of 12 species captured, but several of the species cau ght less frequently than golden mice are non-native or too large to have their true abun dance reflected by these trapping methods.
1 Introduction A geographic range basically is Â“a description of the regions in which a species has been recordedÂ” (Gaston 1991). Geographic range s can be measured in terms of extent of occurrence and area of occupancy. Extent of occurrence is the smallest area contained within an imaginary polygon drawn to incl ude all locality records of an organism; it may include areas that are not current ly occupied by the species or habitat regions that are unsuitable for the organism of int erest (Gaston 1991). Area of occupancy accounts for the pockets of non-occurrence that fal l within the extent of occurrence (Gaston 1991). No matter how a geographic range is measured, changes in range size can occur as a result of habitat alteration caused by humans (Lomolino and Smith 2001; Ceballos and Ehrlich 2002) and climate change (Parm esan et al. 1999; Moritz et al. 2008). In landscapes that are being modified at a fast rate, populations of species may become highly fragmented, subject to the effects of low population size, and eventually extirpated. At the periphery of a speciesÂ’ geograp hic range, individuals may already be stressed by living near their physiological limits and it may not take much additional stress for them to succumb to external pressures. The end result could be a contraction in the speciesÂ’ geographic range. The conservation status of a species is determined by the distribution and abundance of the species and the change in these me asures of commonness over time (McGowan, Gillman, and Dodd 1998). Thus, to determ ine the conservation status of
2 species, distributional data over time are of parti cular importance. Unfortunately, systematically-collected distributional data are be coming increasingly difficult to find at a time when assessing conservation status and maintai ning biodiversity are of great concern (Schipper et al. 2008). Natural historians are less prevalent than they were 60 years ago (Schmidly 2005; Hafner 2007) and biologic al monitoring conducted by government agencies has been reduced because of ins ufficient funding to maintain such programs (Smiley 2008). This study aims to assess the status of the golden mouse ( Ochrotomys nuttalli Harlan, 1832) in Florida using distribution and abu ndance data. The golden mouse is a mid-size cricetine rodent whose geographic range la rgely coincides with the extent of the deciduous (oak-hickory) hardwood and pine forest of the southeastern United States (Packard 1969). South-central Florida is the south ern range periphery of this species. Five subspecies of O. nuttalli are recognized based on the habitat region where individuals are collected (i.e. coastal plain, pied mont, mountains) as well as pelage and morphological variation of museum specimens (Packar d 1969). The subspecies O. nuttalli floridanus is found in the northern two-thirds of Florida, ex cept for the region of the Florida panhandle west of the Apalachicola Rive r where O. nuttalli nuttalli occurs. Like most small mammals, Ochrotomys nuttalli is mainly crepuscular and nocturnal (Linzey and Packard 1977). Golden mice a re semi-arboreal, using their semiprehensile tails, stomach musculature, and well dev eloped plantar tubercles to move around above ground level in thick vegetation. Ochrotomys nuttalli build arboreal nests in some habitats (Linzey and Packard 1977). They h ave a preference for thick shrubby habitats (Stelljes 1982; Wagner, Feldhamer, and New man 2000). The structure of
3 preferred Ochrotomys habitats is sufficiently overgrown so as to challe nge human movement through them. Although this species occur s over a relatively large geographic area (Feldhamer and Morzillo 2008) and occupies sev eral habitat types (Linzey and Packard, 1977), it is usually highly selective at t he microhabitat scale (Wagner, Feldhamer, and Newman 2000). Ochrotomys nuttalli prefer some stages of succession over others. In the central and northern portion o f their range, golden mice tend to be more common in early to mid-successional forests wh ere saplings, shrubs, and vines are present (Seagle 2008), while in southern population s they are more abundant in the later stages of succession when leaf litter, vines, and S panish moss ( Tillandsia usneoides ) accumulate in the absence of fire (Landers and Craw ford 1995). Overall, golden mice tend to be the most abundant in habitats which supp ort the thickest underbrush vegetation for the region. Ochrotomys nuttalli are typically found in densities much lower than s ympatric Peromyscus species (Feldhamer and Linzey 2008; Rose 2008). D ensities of golden mice, as summarized by Linzey and Packard (1977) as well as Rose (2008), range widely from 0 to 74.1 individuals per hectare. In a review of 26 density estimates found in the literature, Rose (2008) reported a median of 4.3 in dividuals per hectare. Much of the variation in density estimates may be attributed to whether or not researchers established their study grids within habitat optimal for golden mice (Rose 2008). Ochrotomys nuttalli often have patchy distributions (Rose 2008) and li ve in highly localized populations (Feldhamer and Linzey 2008). Thus, den sity estimates that include or exclude these concentrations of individuals would l ead to very different results. Density estimates also will vary with season. In the south O. nuttalli attains its highest densities
4 in the winter (McCarley 1958) and data generally su pport a winter breeding season in southern populations (see Rose 2008). Ochrotomys nuttalli is not a species of conservation concern over the majority of its range with the exception of the peripheral stat es of West Virginia (rank of S2: Six to 20 documented occurrences or few remaining individu als within the state; very rare and imperiled) and Illinois (threatened). In Oklahoma, golden mice are a species of greatest conservation need. In sum, the golden mouse is sel dom targeted for monitoring. Because it is nocturnally active above ground, high ly selective of thick and unpleasant to sample habitats, patchy in distribution, and low in relative abundance, the golden mouse is an organism that is rarely encountered while con ducting surveys for other species; it requires a species-specific sampling regime. In Florida, reports of O. nuttalli captures are infrequent and incidental. Locality data for the state have not been centralized beyond the extent of museum specimen databases. This is a species whose changes in dist ribution and abundance over time have gone largely unmonitored, especially at the souther n extent of its range. The conservation status of the golden mouse in Florida is described as Â“not ranked/under reviewÂ” by Feldhamer and Morzillo (2008), signaling the general lack of data that have been collected for this species in the state. Desp ite these challenges, assessing the distribution and conservation status of O. nuttalli in Florida is of importance because widespread habitat loss has been occurring in the s tate. Landscape level changes of natural areas to urban centers, suburban sprawl, ag ricultural fields, and large scale mining operations has been brought about by the pressures of a burgeoning human population. To exemplify the magnitude of the changes seen over the past quarter century, the human
5 population of the state was estimated by the U.S. C ensus Bureau at 9,746,342 individuals in the 1980 census, but has grown to 18,251,243 per sons in 2007, a population increase of 87% in 27 years (United States Census Bureau). The objective of this study was to assess the stat us of the golden mouse in Florida. I compared the current geographic range of Ochrotomys nuttalli floridanus to that of historic records focusing on the extent of occurren ce and examined the level of fragmentation between existing populations by looki ng at the distribution of suitable habitats in the state. In addition, I documented t he abundance of O. nuttalli relative to other small mammals captured in this study as a mea ns of describing rarity at the population levelone of three types of rarity used to classify organisms (Rabinowitz et al. 1986; McCoy and Mushinsky 1992). I utilized Geogra phical Information Systems (GIS) as well as small mammal live trapping techniques to address these objectives.
6 Methods Determining the Geographic Range of Golden Mice and Distribution of Habitats Using Geographic Information Systems (GIS) To determine if the distribution of Ochrotomys nuttalli floridanus has shifted from historic data, it was first necessary to estab lish a baseline using historic occurrence records of Ochrotomys nuttalli floridanus and its potential habitats. ArcGIS 9.2 (ESRI, Redlands, CA) was used to construct a map (Figure 1 ) showing the historic distribution of golden mice and the natural fragmentation of sui table habitats within the state. This map also served to visualize how closely locality r ecords were aligned with specific habitat types and identify areas with appropriate h abitat that lacked O. nuttalli records for trapping in the field. Next, changes to this histo ric distribution were examined using current land use data. A second map (Figure 2) was developed to estimate changes in the historic occurrence of habitat types in south-centr al Florida and how these landscapelevel changes further fragment O. nuttalli populations already existing in a patchy environment. To create these maps, records of O. nuttalli floridanus (n=195) were accumulated from various sources including museum specimens, pe rsonal communications, literature survey (Packard 1969, Pinkham 1971), and live trapp ing in the field (see subsequent sections for trapping details). Museum specimens u sed were those from the Florida Museum of Natural History (FLMNH), the Florida Stat e University (FSU), Tall Timbers
7 Research Station (TTRS), and the Smithsonian Nation al Museum of Natural History (NMNH). It was assumed that the current range is p robably a subset of the historic range and so all new localities identified from field wor k under this project were included in the historic range estimate. Land use data layers, including a 1967 distributio n of the vegetative communities of Florida, locations of shopping centers, and deve lopments of regional impact (DRIs), were downloaded from the Florida Geographic Data Li brary ( http://www.fgdl.org/ accessed March 2007). These layers, along with lat itudes and longitudes of the golden mouse locations, were imported into ArcGIS 9.2. Ea ch feature class of the 1967 vegetative communities of Florida was exported into a separate data layer. Habitats were then placed into one of three categories (inappropr iate, marginal, and appropriate) based on the suitability of that habitat type for golden mice. Categorization was done using an educated opinion built upon experience trapping in the field and the known preference of the golden mouse for habitats with a thick shrubby component (Stelljes 1982; Wagner, Feldhamer, and Newman 2000). Data within each cate gory were merged with the resulting layers: 1) inappropriate habitat: include s urban, mangrove coastal marsh, cypress swamp, southern slash pine, scrub cypress, rockland marshes, Everglades wet prairie, Everglades sawgrass, freshwater marsh, and prairie grasslands, 2) marginal habitat: pine flatwoods and cabbage palm forest, 3) appropriate hardwood-containing habitats: hardwood swamps, longleaf pine-xeric oak, sand pine scrub, mixed hardwoods and pine, hardwood forest, and coastal straind, and 4) bodies of water. A probable historic range map of O. nuttalli floridanus was constructed by placing an 80 kilometer (km) buffer around each loc ale record and dissolving these
8 buffers into one occurrence layer. As museum recor ds were concentrated around the institution of origin, the addition of a buffer aro und locality records helped compensate for under-sampling in some regions. The buffer was set at 80 kilometers because this was the minimum distance at which under-sampled are as not questioned to be within the range of Ochrotomys -in the center of the state and along the Georgi a border --were included in the analysis. Next, this occurrence la yer was intersected with two habitat layers, the first containing only hardwood habitats and the second including marginal habitats, as previously defined. This intersection of occurrence data and habitat layers generated the final distribution map (Figure 1). A more updated (current) range map was developed by taking the historic habitat data layers and erasing natural habitat lost to urb anization and mining uses. Developments of regional impact (DRIs) were used to approximate land lost to mining, while a layer of dissolved 1.6 kilometer buffers ar ound shopping centers represented urban centers. A buffer around shopping centers wa s utilized to reflect urban areas in preference of commercially available urban area lay ers because it permitted patches of suitable habitat to remain within urban boundaries; this coincides with known O. nuttalli populations persisting in natural reserves surround ed by urban development such as the University of South FloridaÂ’s Ecological Research A rea (USF Eco Area). Erasing these anthropogenic land uses from the historic map creat ed a probable current distribution of O. nuttalli in Florida. As this distribution ignores conversi on of habitat to land uses other than urban and mining (such as agricultural), it is intended to be a conservative estimate of habitat loss. Shown in Figure 2 is a subset of the range map that was developed as land use changes were most pronounced in the southcentral region of Florida.
9 Temporal Changes in Abundance To assure that Ochrotomys nuttalli were able to be captured in all months during which state-wide surveying was occurring (discussed below), the abundance of golden mice was monitored over a 7.5 month period (October 2007 Â– May 2008) at a location where O. nuttalli was known to occur. Monitoring took place at the University of South FloridaÂ’s Ecological Research Area (USF Eco Area) ( latitude = 28.07, longitude = -82.38), a preserve of sandhill, overgrown scrub, f latwoods, and cypress habitats that is surrounded on at least three sides by suburban deve lopment in north Tampa. Sherman live traps (dimensions: 22.9cm x 8.9cm x 7.6cm, H.B Sherman company, Tallahassee, FL) were placed in pairs (hereafter referred to as a Â“trap stationÂ”) in habitat that had been previously identified as containing O. nuttalli by trapping conducted in January and March 2007. A total of 120 traps were placed at 60 trap stations (Figure 3). Trap stations were minimally spaced 10m apart and placed to maxim ize captures within the heterogeneous habitat. Traps were baited with a mi xture of sunflower seeds and rolled oats, set each evening, and checked and closed just after sunrise. Trapping occurred over a period of four consecutive nights (hereafter, a Â“ trap periodÂ”) and except for October to November was conducted at monthly intervals. Dates of trapping were October 20-24, November 3-7, December 1-5, January 6-10, February 3-7, March 2-6, April 6-10, and May 4-8. Each trapping period consisted of 480 tra p nights (a measure of effort equal to the number of traps set out times the number of nig hts they were open) for a total of 3,840 trap nights over the 7.5 month period.
10 All captured mice, rats, and squirrels were identi fied to species, sexed, ear tagged (Monel, size 1005-1, National Band and Tag, Newport KY), and released at the point of capture. Captured shrews were weighed, measured, a nd released at the place of capture. Captured Ochrotomys nuttalli also were weighed and subject to genetic tissue sa mpling (see Smiley et al., in review) prior to release. Assessment of Current Extent of Occurrence Presence/absence surveying was conducted around th e state of Florida between November 2007 and May 2008, with a focus on the sou thern periphery of Ochrotomys nuttalli Â’s range. The primary goal of the survey was to de termine if the southern range periphery of O. nuttalli had shifted from the historic locale data. The mo st likely direction of a shift would be the contraction of O. nuttalli Â’s extent of occurrence northward resulting from increased fragmentation of the landscape and the resulting extinction of isolated populations. Ancillary goal s were to determine if documented populations more central to the range were being ma intained (reflecting a stable area of occupancy), to document previously undescribed popu lations in areas determined to have suitable habitat by the GIS model, and to determine the abundance of O. nuttalli relative to that of other small mammals captured. The prese nce or absence of O. nuttalli was determined on each property visited using direct an d indirect measures. Direct evidence of golden mice on site was obtained from live trapp ing and indirect evidence came from searching for arboreal nests. To assess if the southern range periphery of the g olden mouse was stable, the southernmost documented locales along the east coas t, central ridge, and west coast were
11 identified from accumulated location data. These s outhernmost locales are Little Manatee River Â– South Fork State Preserve (also kno wn as the Beker Parcel) in north Manatee County, Archbold Biological Station (ABS) i n Highlands County, and 5 miles south of Melbourne Beach in Brevard County. O. nuttalli were last documented at South Fork State Preserve in 1990 by state biologist Terr y Hingtgen (Florida State Parks, District 8 annual report), at ABS in 1998 by Richar d Lavoy (personal communication), and at south Melbourne Beach in 1945 from a specime n at the Florida Museum of Natural History (catalog number 2723). Accessible lands that were in close proximity to t hese three southern sites were surveyed in 2008 to determine the current extent of Ochrotomys nuttalli Â’s range (Figure 4; Table 1). On the west coast, trapping was perfo rmed at Southwest Florida Water Management DistrictÂ’s (SWFWMD) Little Manatee River Southfork Tract (LMRSF), a property adjacent to the south boundary of South Fo rk State Preserve. On the central ridge, trapping occurred directly on ABS, including some of Jim LayneÂ’s former grids (for a description see Packer and Layne 1991). On the east coast, the habitat in the vicinity of Melbourne Beach has become quite fragme nted in recent years. Trapping in this area occurred on small, scattered undeveloped parcels that are a part of Brevard CountyÂ’s Environmentally Endangered Lands Program ( EELP) including Coconut Point Preserve (25.1 hectares), Maritime Hammock Preserve (56.7 hectares), and the Washburn Property (15.8 hectares). In addition to surveying for O. nuttalli on the barrier island near Melbourne Beach, trapping also occurred at approxim ately the same latitude on the mainland at Brevard CountyÂ’s EELP Malabar Scrub Pro perty in an area of the county where the landscape was less developed.
12 Nine additional sites across peninsular Florida we re surveyed to determine if the area of occupancy of golden mice was stable, docume nt previously unknown locality records, and assess the relative abundance of O. nuttalli (Figure 4; Table 1). Six of these sites were within the historic extent of occurrence (with county of Florida indicated where appropriate): Hillsborough CountyÂ’s Balm Boye tte Scrub, Brevard CountyÂ’s Indian Mounds, Faver Dykes State Park (St. Johns), The Nature ConservancyÂ’s Tiger Creek (Polk), SWFWMDÂ’s Jack Creek (Highlands), and Wekiwa Springs State Park (Orange). Additionally, three sites were trapped i n the southwest portion of the state outside the documented range of this mouse, but wit hin habitat identified as marginal by the GIS model. Trapping was performed in this port ion of the state to ensure that O. nuttalli Â’s apparent absence here was not caused by lack of monitoring. The properties surveyed were Lee CountyÂ’s HickeyÂ’s Creek Mitigatio n Park, Charlotte CountyÂ’s Shell Creek, and SWFWMDÂ’s Deep Creek (Desoto County). It would have been ideal to survey for golden mice in the region between north Manatee and south Desoto Counties (between LMRSF and Deep Creek) particularly in Hard ee and Desoto Counties where hardwood habitats extend south along the Peace Rive r corridor, but the lack of public lands and a growing mining pressure in this region limited where trapping could occur. At all sites, traps were placed in habitats that w ere visually deemed the most suitable for golden mice compared to what was avail able at each location. Traps were baited with a mixture of sunflower seeds and rolled oats in the same manner as at the USF Eco Area. Captured small mammals were identifi ed to species, sexed, given a short term mark by fur clipping, and released at the poin t of capture. Ochrotomys nuttalli that were caught were also weighed and subject to geneti c tissue sampling prior to release.
13 The number of trap nights accumulated per site dep ended on a number of factors including travel distance to site, weather, study s ite policies (state parks were not accessible on weekends), amount and layout of favor able habitat, and density of small mammals (in warm weather and in areas of high trap success it becomes difficult to process all animals before heat mortalities occur). Trapping dates and the number of trap nights for the ten sites within the previously docu mented range are detailed in Table 1. Dates of trapping and amount of effort for those in the southwest portion of the state were as follows: HickeyÂ’s Creek Mitigation Park (Decembe r 19-24, 2007) 573 trap nights, Shell Creek (December 26-31, 2007) 536 trap nights, and Deep Creek (January 19-23, 2008) 411 trap nights. In summary, a total of 5,72 3 trap nights were conducted in search of O. nuttalli at thirteen sites across peninsular Florida. The following procedures were performed to assign a level of confidence to the determination of absence at study sites on which li ve trapping did not result in the capture of Ochrotomys nuttalli First, the effort until first capture (expressed in trap nights) was determined for all fourteen trapping periods in whi ch golden mice were caught. This effort was then plotted as the percentage of Â‘prese ntÂ’ trapping periods which resulted in a capture by a given quantity of effort (Figure 5). Next, the total effort (in trap nights) was calculated on absent sites. Finally, each absent s ite was assigned a percentage of the present trapping periods which yielded a golden mou se capture at or below the effort expended on the absent site (Table 2). These steps were repeated for both uncorrected and corrected trapping efforts. Corrected efforts were calculated by taking the uncorrected effort and subtracting one-half times t he number of traps that were closed in the morning which did not contain O. nuttalli This correction was intended to adjust the
14 effort expended on study sites where non-target spe cies were ubiquitous or traps were tripped by raccoons ( Procyon lotor ) or other animals (for more information on correct ion factors see Nelson and Clark 1973 and Beauvais and Buskirk 1999). These events would lower the likelihood of catching O. nuttalli In addition to live trapping for Ochrotomys nuttalli indirect evidence of their presence was obtained from arboreal nests found on study sites. Although no formal surveys for the nests were conducted because it was not known how common or identifiable these structures would be at the onset of the study, places where nests were found indicated that O. nuttalli was likely present there at least in the recent pa st. Relative Abundance of the Golden Mouse Relative abundance data was examined in order to c ompare population numbers of Ochrotomys nuttalli to sympatric species, in essence trying to assess if the golden mouse is relatively rare at a local level. The rel ative abundance of individuals of each small mammal species captured was examined for the eleven study sites which fall within the documented range of Ochrotomys nuttalli These relative abundance data are based on a combined effort of 8,043 trap nights. Bar gra phs in which species were ranked from most to least common were constructed twiceonce f or all study sites and again using only those sites where O. nuttalli were captured.
15 Results Determining the Geographic Range of Golden Mice and Distribution of Habitats Using Geographic Information Systems (GIS) Recorded occurrences of Ochrotomys nuttalli align closely with the distribution of hardwood-containing habitats in Florida (Figure 1). GIS places most locales (146 of 195) directly within hardwood-containing habitats and 48 of the remaining 49 locales within 4 km of this vegetation class (mean distance = 1.82 k m, median distance = 1.13 km). Chuluota Wilderness Area in Seminole County is the location furthest from the nearest appropriate hardwood-containing habitat at a linear distance of 10.3 km (Figure 1). As depicted in the 1967 vegetation data layer, har dwood-containing habitats have a non-continuous distribution in the northern two-t hirds of Florida (Figure 1). At the golden mouseÂ’s southern range periphery in south-ce ntral Florida, these vegetation types occur in patches usually surrounded by a matrix of pine-palmetto flatwoods. The east coast of Florida contains a long, narrow stretch of hardwood habitat (on the Atlantic Coastal Ridge) that is isolated from other such veg etation types. As the distribution of hardwood-containing habitats in Florida is naturall y fragmented and Ochrotomys nuttalli locality records are closely aligned with these veg etation types, the distribution of the golden mouse in Florida should also be non-continuo us. This distribution is reflected in the occurrence data. O. nuttalli are found approximately two-thirds of the way down the peninsula and have an interrupted distribution from east to west within this region that
16 coincides with the prevalence of appropriate habita t patches. The most isolated occupied hardwood patch is LMRSF, a site on the west side of the peninsula at the range periphery of O. nuttalli (Figure 4). The golden mouse population at LMRSF i s 12.8 km from another hardwood habitat patch and 21.4 km from a h ardwood patch that is known to contain O. nuttalli (Balm-Boyette Scrub). Golden mice found on Atlant ic Coastal Ridge at Malabar Scrub (Figure 4) are historically connec ted to other populations north (and possibly south) along the coast, but are approximat ely 74 km linear distance from the large patch of hardwood-containing habitat found at the same latitude on the central ridge. The addition of urban and mining land use categori es to the map of Ochrotomys nuttalli occurrence records has the effect of making natura lly fragmented southern periphery populations even more isolated (Figure 2) Although changes in land use have led to all populations becoming increasingly isolat ed from one another, the effects are most dramatic in two regions of the state: the Atla ntic Coastal Ridge on the east coast and Bone Valley Â– an area of extensive phosphate mining in Polk, Hardee, Hillsborough, and Manatee Counties to the west of the central Lake Wa les Ridge. Urban development and sprawl on the Atlantic Coastal Ridge has broken the narrow north-south linkage that historically existed between coastal populations. Southern populations such as those at Malabar Scrub are no longer connected to northern p opulations by either hardwoodcontaining or marginal habitats. Meanwhile in Bone Valley, mining developments of regional impact have been the greatest perpetrator in reducing hardwood habitats in the region. Large scale mining not only reduces availa ble habitat for local populations (one locality record lies in the midst of a development of regional impact (DRI)), but has the
17 effect of making west coast Ochrotomys nuttalli more isolated from those inhabiting the central ridge. Temporal changes in abundance Ochrotomys nuttalli were captured during all trapping periods at the U SF Eco Area. The abundance of individuals varied over tim e (Figure 6). The number of individuals captured during each trap period remain ed relatively low from the beginning of the study in October 2007 through January of 200 8, with only two to three individuals caught during these months. Similarly, the capture rate (including recaptures) was minimal for this period ranging from 0.0042 to 0.01 04 golden mice per trap night. Beginning in February 2008 through the end of the s tudy in May 2008, the number of golden mice increased with each successive trap per iod. Twelve individuals were caught in February, 16 in March, 24 in April, and 26 in Ma y. Capture rates (including recaptures) increased accordingly, going from 0.035 4 golden mice per trap night in February to 0.1021 golden mice per trap night in Ma y, about an order of magnitude larger than in the fall and early winter. New individuals were basically non-existent in the traps prior to February 2008 (captured marked individuals were tagged in January or March 2007), after which time between 7 and 11 new indivi duals were caught each month. A total of 39 individuals were captured 162 times ove r the course of study at the USF Eco Area for an across-month average capture rate of 0. 0422 golden mice per trap night (Table 1). These results confirm that O. nuttalli were able to be captured during all months when state-wide surveys were taking place.
18 Assessment of current extent of occurrence Ochrotomys nuttalli were captured at six of the thirteen statewide loc ations that were surveyed (Figure 2; Table 1). In focusing on the southernmost sites, the presence of O. nuttalli was first confirmed at LMRSF on the west coast. A t this site, five individuals were captured seven times in 464 trap nights for a capture rate of 0.0151 mice per trap night. On the central ridge, no golden mice were c aught at ABS despite considerable effort here (976 trap nights), but arboreal nests w ere observed on station. Further support for possible low densities of O. nuttalli present on ABS is that golden mice were documented at the Lake Placid Scrub, a preserve tha t lies just north of ABS, in 2008 (H. Hoffman, personal communication). On the east coas t, O. nuttalli was not captured on any of the three small properties surveyed to the s outh of Melbourne Beach (482 trap nights). The habitat present on this barrier islan d at the time of field work consisted mostly of maritime hammock with sparse understory, exotic plants or areas where exotics had been cleared, and a patch of recently burned co astal scrub. On the mainland, trapping was successful at Brevard CountyÂ’s Malabar Scrub where six individuals were captured six times in 456 trap nights for a capture rate of 0.0132 mice per trap night. In summary, Ochrotomys nuttalli were found to be present in the vicinity of each o f the southernmost documented localities in 2008. Ochrotomys nuttalli were also captured at Hillsborough CountyÂ’s Balm-B oyette Scrub (one individual), The Nature ConservancyÂ’s Ti ger Creek (one individual), SWFWMDÂ’s Jack Creek (four individuals), and Brevard CountyÂ’s Indian Mounds (four individuals). Balm-Boyette Scrub was a confirmatio n of a previous record, while golden mice had not been documented before at the remainin g three sites. Ochrotomys nuttalli
19 were not found at Faver-Dykes State Park and Wekiwa Spring State Park, however arboreal nests were identified at the former. Thes e state parks both had previous records of golden mice and the amount of effort invested at each was low (266 and 233 trap nights, respectively). Faver-Dykes and Wekiwa Spri ng are fairly large parks, preserving 2,446 and 3,157 hectares, respectively. The result s of presence-absence trapping for the eleven study sites that lie within the documented r ange (including the USF Eco Area) are summarized in Table 1 and Figure 2. Ochrotomys nuttalli were not found at any of the locations surveyed in the southwest portion of the state that lie outside of the documented range (Deep Creek, Shell Creek, and HickeyÂ’s Creek), despite on the ground v erification of oak species ( Quercus sp. ) occurring along the stream corridor of each site. Hickory species ( Carya sp. ) were also present at Shell Creek. On study sites where Ochrotomys nuttalli was found to be present using live traps, the first capture of a golden mouse in a trapping p eriod occurred with an uncorrected effort ranging from 50 to 442 trap nights. Correct ed effort at first capture ranged from 35 to 420 trap nights. The percentage of trapping per iods with a first capture occurring by an invested amount of effort is shown in Figure 5. A large percentage of trapping periods had a first capture with a relatively low amount of effort. For example, 79% of trapping periods in which a golden mouse would be caught had already documented the species as present at an effort of 146 uncorrected trap nights At 278, 318, and 442 uncorrected trap nights, 86%, 93%, and 100% of trapping periods resp ectively had caught O. nuttalli Correcting effort for closed traps leads to a faste r accumulation of the percentage of trapping periods (Figure 5). At sites where O. nuttalli was absent, all had an invested
20 effort greater than 150 trap nights and five of the sites had an uncorrected effort greater than 400 trap nights. Thus, a great amount of conf idence can be placed in the fact that either O. nuttalli is indeed absent from the sites where they were no t caught or that they occur there in exceedingly low densities. Table 2 details the percentage of trapping periods which yielded a golden mouse capture at or below the effort expended on each absent site. Arboreal nests were seen at all the sites where go lden mice were caught, with the exception of Hillsborough CountyÂ’s Balm-Boyette Scr ub. In addition to nests seen at study sites where Ochrotomys nuttalli were present, arboreal nests were identified at tw o locales where O. nuttalli were not captured Â– Archbold Biological Station an d FaverDykes State Park. Nests identified as likely built by golden mice were constructed primarily of saw palmetto ( Serenoa repens ) fibers. At Brevard CountyÂ’s Indian Mounds, several nests with a similarity in appearance to th ose seen at other locations were found to be active, with golden mice flushing from them w hen disturbed. The culmination of evidence from GIS mapping, pres ence-absence surveying, and nest surveying, indicates that the range of Ochrotomys nuttalli has not experienced any substantial northward latitudinal shift over the pa st 70 years. However, hardwood habitat patches are becoming increasingly isolated from one another in south-central Florida and some local populations of golden mice, such as thos e at Melbourne Beach and in Bone Valley, have likely been extirpated as a result of modifications to the landscape.
21 Relative abundance of the golden mouse Three small mammal species dominated the 1,217 ind ividuals that were captured across the eleven study sites within the documented range of the golden mouse, including the USF Eco Area (Figure 7). The Florida mouse ( Podomys floridanus ), the cotton mouse ( Peromyscus gossypinus ), and the hispid cotton rat ( Sigmodon hispidus ) accounted for combined 93.2% of all individuals. Podomys floridanus were far more abundant than any other species where they were present with 512 individuals caught at six different locales. Peromyscus gossypinus were the second most frequently caught species, to taling 484 individuals. There were 138 Sigmodon hispidus individuals found. P. gossypinus and S. hispidus were cosmopolitan in distribution, occurring on al l eleven and ten of eleven sites, respectively. Ochrotomys nuttalli was the fourth most common species with 60 individuals or 4.9% of total captures. Most gol den mice (n=39) were caught at the USF Eco Area. There were nine shrew captures durin g the course of this study (shrews were not individually marked or identified to speci es). Other species documented, each with 3 or fewer individuals, were Peromyscus polionotus Didelphis virginia Rattus rattus (not native), Neotoma floridana Mus musculus (not native), Sciurus carolinensis and Glaucomys volans The distribution of species differed between sites where Ochrotomys nuttalli was present and that of all sites combined (Figure 7). Although Podomys floridanus was slightly more common overall than Peromyscus gossypinus when only sites where O. nuttalli were captured are considered, the relative abundan ce of P. gossypinus (n = 324) is greater than that of P. floridanus (n = 191).
22 Discussion Ochrotomys nuttalli displays a geographic range that is closely tied to the distribution of hardwood-containing habitats in Flo rida. These habitats are naturally fragmented and are becoming increasingly so under r ecent land use changes. This study suggests that the extent of occurrence of O. nuttalli remains relatively unchanged from historic records, but isolated populations of golde n mice are vulnerable to extirpation as a result of extensive habitat fragmentation. A few h istoric populations are likely already extinct. Data from temporal trapping reveals that O. nuttalli were able to be captured during all months in which statewide presence/absen ce surveys took place, but peak abundances occur February through May. When focusi ng trapping on the thickest habitats available on each study site, Ochrotomys nuttalli were the fourth most abundant of eleven small mammal species captured. This project describes the close alignment of Ochrotomys nuttalli locality data with the distribution of hardwood habitats in Flori da. Hardwood-containing habitats are naturally fragmented, especially near the periphery of O. nuttalli Â’s range. Some suitable habitats, such as those on the Atlantic Coastal Rid ge, are separated from other hardwood areas by a considerable distance. This situation s uggests that there is little movement of golden mice between non-continuous hardwood areas. Mitochondrial sequence data from O. nuttalli support this claim, showing a high degree of struc ture between disjunct populations (Smiley et al., in review). In fact, p opulations on the Atlantic Coastal Ridge
23 appear to be on an independent evolutionary traject ory from other populations in Florida (Smiley et al., in review). Smiley et al. also fou nd that southern periphery populations contain unique haplotypes not found elsewhere. Thi s information is somewhat distressing given the highly isolated nature of som e O. nuttalli populations and the pace of development in Florida. For example, as the nor th-south dispersal corridor is broken for populations inhabiting the Atlantic Coastal Rid ge, individuals will not be able to colonize formerly occupied patches if extirpated by fire or some other disturbance. The historic range map developed using GIS shows t hat Ochrotomys nuttalli have not been documented in all regions of Florida which it suggests they should occur. Additional trapping is needed, especially in centra l and north Florida to determine if O. nuttalli is present here. Also, while O. nuttalli occur near the limit of hardwood habitats in the western and central parts of the Florida pen insula, hardwood habitat on the east coast Atlantic Coastal Ridge continues well south o f documented populations. Despite high levels of development on some stretches of the east coast of Florida, it is possible that O. nuttalli populations do occur further south. For example, arboreal nests likely made by golden mice were seen at Brevard CountyÂ’s V alkaria-Grant Scrub Sanctuary, a few kilometers south of Malabar Scrub. However, th e presence of arsonist-ignited wildfires in the area prevented live trapping at th e time when nests were identified. If Brevard County is in fact as far south as O. nuttali extend on the east coast, one explanation for a range limit in this area is that forests begin to shift from temperate to mixed temperate and tropical species along this str etch of coastline. As arboreal nests were present on all but one stud y site where golden mice were trapped, they seem to be a good indicator of O. nuttalli Â’s presence. Nest searching could
24 be utilized if circumstances do not permit the use of live traps. Nest searching is not a new technique as Ivey (1949) collected golden mice only from arboreal nests in his natural history study. If using nests as an indica tor of presence, it is important to note that arboreal nests may remain intact for greater t han a year after being deserted by O. nuttalli (Ivey 1949). The utility of nest searching could be limited if structures of similar appearance are in fact made by other species. Although occurrence records aligned well with hard wood-containing habitats, some fell outside this habitat category. This cond ition can be explained in several ways: 1) error in the geographical data, 2) fluctuation i n distribution of hardwoods based on land management practices, and 3) Ochrotomys nuttalli Â’s use of marginal habitats such as pine flatwoods. Error could occur in the locality data from an incorrect assignment of latitude and longitude coordinates to a specimen co llection point. The Florida Museum of Natural History uses the georeferencing software GEOlocate which was developed at Tulane University ( http://www.museum.tulane.edu/geolocate/ ) to assign latitude and longitude coordinates to museum specimens. If a mu seum label does not provide a thorough description of the point of collection, th e software could assign an incorrect location to this specimen. Also, some degree of er ror is inevitably involved in categorizing regions of a state into one of several habitat types. In the fire-maintained habitats of Florida the veg etation species composition of a particular area is likely to change with land manag ement practices. Suppression of fire in pine flatwoods often leads to hardwood encroachment (Abrahamson and Hartnett 1990). Thus, the current distribution of hardwoods at a sm all scale may differ in some areas of Florida compared to their 1967 distribution if fire has been excluded. Fire suppression
25 occurs most often in small parcels of wildlands in close proximity to developed areas. This situation is often the case with county-owned properties. Lastly, not enough is known concerning the biology of Ochrotomys nuttalli to justify that they do not occasionally make use of p ine flatwoods habitat. To the contrary, when population abundances were high, individuals o f O. nuttalli were captured at the USF Eco Area in areas of pine flatwoods (Smiley, un published data), albeit not far from more characteristic microhabitat. In addition, in the absence of fire, saw palmetto can become thick and grow to mid-story heights, possibl y providing refuge for golden mice in the absence of shrubby hardwoods. More research is needed to determine what, besides the structure of the vegetation, may limit O. nuttalli to hardwood habitats. The latitudinal extent of Ochrotomys nuttalli Â’s occurrence has not contracted in Florida, however two peripheral study sites where g olden mice where not captured indicate the susceptibility of this species to land use change and land management practices. These locales are Melbourne Beach and A rchbold Biological Station. Melbourne Beach has undergone a great deal of land use change since the collection of an O. nuttalli specimen here in 1945, although the Melbourne Beac h region has faired better than regions to the south in terms of amount of development. The importance of this stretch of coastline to marine sea turtle nest ing has driven conservation efforts, including the establishment of the Archie Carr Nati onal Wildlife Refuge. Despite this, remaining natural areas, particularly of scrub and hardwood hammock, are scattered in distribution and small in total size. Trapping dat a provided by this study and the minimal amount of remaining suitable habitat are strong ind icators that O. nuttalli is absence in this area. However, further trapping is needed to confirm this assertion.
26 At Archbold Biological Station the potential absen ce of Ochrotomys nuttalli is mainly driven by land management practices. The co mbined effect of an increased fire regime to benefit Florida scrub jays ( Aphelocoma coerulescens ) on one side of the property and two seasons of intensive hurricanes (2 004-2005) greatly altering habitat by knocking down canopy sand pines on the long-unburne d Red Hill side of the station (F. Lohrer, personal communication) has resulted in the potential absence of O. nuttalli at this study site. No golden mice were captured at A rchbold despite the greatest amount of effort invested here and continual preservation of the land. Only the presence of unoccupied arboreal nests and a reported capture by H. Hoffman at Lake Placid Scrub, just north of Archbold Biological Station, suggest that O. nuttalli are still present in the area. An alternative explanation for the absence o f O. nuttalli on this study site centers on the high abundance of Florida mice ( Podomys floridanus ) on station. At Archbold, 301 Podomys floridanus individuals were caught 449 times with an effort o f 976 trap nights. Such high numbers of Podomys floridanus could possibly drive down abundances of other competing small mammals or simp ly occupy traps. However, correcting effort for trap closures at this study s till resulted in an effort that greatly surpasses first captures on sites where O. nuttalli were observed. Although high abundances of Florida mice themselves might help ex plain the absence of O. nuttalli at ABS, this explanation is not independent of land ma nagement practices as numbers of Podomys floridanus decline with fire suppression (Jones and Layne 199 3). Temporal changes in abundance and the timing of un marked individuals entering the population at the USF Ecological Research Area point towards a mid-winter to spring breeding season of Ochrotomys nuttalli in south-central Florida. Two additional
27 observations lend support to this conclusion. Firs t, a trap mortality caught on 29 January at Little Manatee River Southfork Tract in north Ma natee County had 2 embryos present when prepped at the museum (Florida Museum of Natur al History specimen number UF31664). Second, a female golden mouse that was f lushed from an arboreal nest on 27 Febuary at Indian Mounds in north Brevard County ha d two suckling young attached to her (Smiley, person observation). When taken toget her, these data point toward a late January to May breeding season for O. nuttalli in south-central Florida. Data from this study both support and contradict o ther information available in the literature concerning the breeding season of so uthern Ochrotomys nuttalli populations. Pearson (1953) reported higher abunda nces of golden mice from January through May in comparison to the remainder of the y ear from live trapping in Gulf Hammock, Florida. In Texas, mature golden mice hav e been reported in breeding condition in January and February (McCarley 1958). In contrast, Ivey (1949) reported a female with suckling young and four embryos on 3 No vember in eastern Florida and young about one week of age on 21 December, suggest ing an October to December breeding season. Also in partial disagreement with the present study, Layne (1960) reports O. nuttalli litters born in June and July, pregnant and lactat ing females in July, September, and November, as well as a female with n ewborn young on the 2 March. LayneÂ’s (1960) observations in central Florida poin t toward a longer eight to nine month breeding season than the five to six month season s uggested by the present study. The discordance among breeding season data suggest that a factor besides season may regulate breeding times in southern populations of O. nuttalli
28 When trapping in the thickest habitats available o n each study site, Ochrotomys nuttalli was the fourth most abundant of twelve species cap tured (with all shrews potentially erroneously lumped into one species). However, some qualification is needed for a few of the more uncommon species. Two specie s, the Virginia opossum ( Didelphis virginia ) and the grey squirrel ( Sciurus carolinensis ), are typically too large to physically fit in Sherman live traps. Thus, these captures sh ould not be taken as a true indicator of abundance for these species. Grey squirrels were i ndeed seen more than they were captured and the trap that captured a grey squirrel had to be physically dismantled to release the large animal (Smiley, personal observat ion). Two other species, the black rat ( Rattus rattus ) and the house mouse ( Mus musculus ), are not native to Florida and seldom reach high densities outside of urban areas. This result leaves O. nuttalli as the fourth most abundant of eight species. In habitats with a thick understory to mid-story, one would not expect to find high abundances of oldfiel d mice ( Peromyscus polionotus ) which prefer more open areas or southern flying squ irrels ( Glaucomys volans ) which would encounter trouble gliding through thick scrub Thus, of species that typically occur in thick habitats, only shrews and the wood r at ( Neotoma floridana ) are less abundant than O. nuttalli The only captured species that is categorized in the state of Florida as imperiled under the designation of Speci es of Special Concern because of significant vulnerability to habitat modification i s the Florida mouse ( Podomys floridanaus ). The Florida mouse was also the species of great est abundance in this study, but was not found on all study sites. It is not po ssible to give density estimates for O. nuttalli from this research with which to compare to the li terature as trapping was not carried out on established grids.
29 The distribution of species differed between site s where Ochrotomys nuttalli was present and that of all sites combined. This diffe rence is caused by the fact that Podomys floridanus can reach very high abundances on study sites that are managed to be relatively open, a situation which typically leads to lower O. nuttalli abundances. The removal of just one study site, Archbold Biological Station where 301 Podomys floridanus individuals were caught, changes the species abund ance rankings and makes Â‘presentÂ’ sites closer in species distribution to t hose places where O. nuttalli was not captured. My data should not be taken as evidence that preserve managers must make a choice between maintaining Florida mice versus gold en mice on their properties. To the contrary, several scrub sites including SWFWMDÂ’s Ja ck Creek and Little Manatee River Southfork Tract contain heterogenous scrub habitat Â– managed to be open and of low height in some portions and with taller thickets of scrub oaks in other areas of the site Â– a management strategy that seems to be able to suppor t both species given a minimal preserve size. Ochrotomys nuttalli and Podomys floridanus represent only two of the numerous species that have adapted to living in dif ferent subsets of the various successional stages of scrub, a habitat that histor ically burned every 10 to 100 years (Myers, 1990). The best option for maintaining mul tiple species would be to protect large areas of land from development. As large tra cts of scrub have already been converted to other land uses, the reality of this o ption is limited. The second best option for maintaining multiple species with different hab itat requirements on preserves may be to manage these properties in a way so that they ma intain the heterogenous nature of the historic landscape. If hardwoods habitats in Flori da are managed exclusively for those
30 species adapted to frequent fires, then the future persistence of O. nuttalli in the remaining hardwood habitat patches of south-central Florida may be bleak. Several of the parcels of land where golden mice w ere easily captured in this study contain patches of habitat that have not burn ed for decades. As the attitudes of land managers shift from one of fire exclusion to that o f active management, many of these properties are undergoing tremendous alterations. After so many years of fire exclusion, fire-dependent species may no longer be present on these properties, especially if preserves are small in size. Land managers should set clear goals for their management techniques. In the case of scrub, managers might a im for manipulating the land until it has the visual characteristics of ideal scrub habit at, but in doing so they should realize that the biodiversity of their property may decline as species that are adapted to overgrown conditions go locally extinct and those a dapted to open conditions are unable to colonize the property because of habitat fragmen tation. These consequences should be considered when making management decisions.
31 Figure 1. Probable historic distribution of Ochrotomys nuttalli floridanus Locality records align with the distribution of hardwood-containing habitats. See text for detailed description of habitat classi fication.
32 Figure 2. A probable current distribution of Ochrotomys nuttalli with focus on south-central Florida where historically populations were the mos t naturally fragmented. Removal of select anthropogenic land uses creates a landscape where hardwood-containing habitats are even more isolated from one another. Populations evalua ted during 2007-2008 are identified based on the category they best fall into based on direct tr apping evidence: Confirmed Present: O. nuttalli caught on site, Determined Absent: O. nuttalli not captured on site, and Not Assessed: site statu s not evaluated by the present study. Marginal habit at is not shown here for sake of clarity.
33 Figure 3. Layout of the sixty trap stations operat ed at the USF Ecological Research Area. Stations were spaced at least 10 meters apart in ar eas where O. nuttalli had been previously captured. Two traps were set at each station.
34 Figure 4. The locations of the fourteen study site s on which trapping occurred in search of Ochrotomys nuttalli Locales sampled included: USF Ecological Research Area (ECO), BalmBoyette Scrub (BBS), Little Manatee River Southfork (LMRSF), Wekiwa Springs State Park (WSSP), Tiger Creek (TC), Jack Creek (JC), Archbold Biological Station (ABS), Faver-Dykes State Park (FDSP), Indian Mounds (IM), Malabar Scru b (MS), south Melbourne Beach (MB), Deep Creek (DC), Shell Creek (SC), and HickeyÂ’s Cre ek Mitigation Park (HC).
35 Site County # Trap Nights Dates of Trapping O. nuttalli caught? O. nuttalli capture rate O. nuttalli corrected capture rate O. nuttalli nests present? USF Eco Area Hillsborough 3,840 Oct 2007May 2008 Yes 0.0422 0.0455 Yes Balm-Boyette Scrub Hillsborough 542 Nov 16-21, 2007 Yes 0.0037 0.0039 No SWFWMD LMRSF Manatee 464 Jan 12-16 and 26-30, 2008 Yes 0.0151 0.0245 Yes Wekiwa Springs State Park Orange 233 Feb 11-13, 2008 No 0.0000 0.0000 No TNC Tiger Creek Polk 316 Apr 19-23, 2008 Yes 0.00 32 0.0037 Yes SWFWMD Jack Creek Highlands 190 Mar 9-12, 2008 Ye s 0.0263 0.0369 Yes Archbold Biological Station Highlands 976 Mar 13-19, 24-26, and Apr 4-6, 2008 No 0.0000 0.0000 Yes Faver-Dykes State Park St. Johns 266 Apr 28May 1, 2008 No 0.0000 0.0000 Yes Indian Mounds EELP Brevard 278 Feb 23-26, 2008 Ye s 0.0144 0.0212 Yes Melbourne Beach EELP Brevard 482 Apr 12-17, 2008 No 0.0000 0.0000 No Malabar Scrub EELP Brevard 456 May 10-15, 2008 Yes 0.0132 0.0143 Yes Table 1. Capture success of Ochrotomys nuttalli at the eleven sites within the documented range of this species during 2007 and 2008. Capture rates reflect the number of captures (including recapture s) divided by number of trap nights. Corrected cap ture rates was computed similarly (number of captures/corrected trap nights), but with corrected trap nights calculated by the total number of trap nights minus the product of the number of traps cl osed in the morning that did not contain O. nuttalli times one half. Shading indicates sites where O. nuttalli were not caught, but nests were seen (light grey) or those sites where nests were not observed and golden mice were not caught (dark grey). Asterisks (*) indicate pre viously undescribed locations were O. nuttalli were captured.
36 A) 0 5 0 1 00 1 5 0 2 00 2 5 0 3 00 3 50 4 00 4 50 5 00 Effort (Uncorrected Trap Nights) 0 20 40 60 80 100 Percentage of Trap Periods B) 0 50 10 0 150 200 250 300 3 5 0 400 450 500 Effort (Corrected Trap Nights) 0 20 40 60 80 100 Percentage of Trap Periods Figure 5. The percentage of Â‘presentÂ’ trapping per iods which resulted in a capture of Ochrotomys nuttalli by a given level of uncorrected (A) and corrected (B) effort.
37 Site Uncorrected Effort % Trapping periods below Corrected Effort % Trapping periods below Wekiwa Springs State Park 233 78.6 201 85.7 Faver-Dykes State Park 266 78.6 234 92.9 SWFWMDÂ’s Deep Creek 411 92.9 335 92.9 Brevard CountyÂ’s Melbourne Beach 482 100 363 92.9 Charlotte CountyÂ’s Shell Creek 536 100 428.5 100 Lee CountyÂ’s HickeyÂ’s Creek 573 100 553 100 Archbold Biological Station 976 100 673 100 Table 2. Effort on Â‘absentÂ’ study sites and the pr oportion of Â‘presentÂ’ sites with a first Ochrotomys nuttalli capture before the amount of effort at each Â‘absen tÂ’ site.
38 0 5 10 15 20 25 30Oct-07Nov-07Dec-07Jan-08Feb-08Mar-08Apr-08May-08Time# O. nuttalli Total Indiv New Indiv Figure 6. Abundance of Ochrotomys nuttalli over time based on the number of individuals captured each trap period at the USF Eco Area.
39 Relative Abundance of Small Mammals0 100 200 300 400 500 PFLPGOSHIONUSHRPPODVIRRANFLMMUSCAGVOSpeciesNumber of Individuals All Sites within Range 'Present' Sites Figure 7. Relative abundance of small mammal speci es captured in Sherman live traps at the eleven locales within the documented geographic range of Ochrotomys nuttalli Each category of sites graphed is a subset of the previous bar graph group. Â‘Pres entÂ’ sites are those on which O. nuttalli were captured. Species codes are as follows: PFL = Podomys floridanus PGO = Peromyscus gossypinus SHI = Sigmodon hispidus ONU = Ochrotomys nuttalli SHR = Shrew sp ., PPO = Peromyscus polionotus DVI = Didelphis virginia RRA = Rattus rattus NFL = Neotoma floridana MMU = Mus musculus SCA = Sciurus carolinensis and GVO = Glaucomys volans
40 Literature Cited Abrahamson, W. G. and D. C. Hartnett. 1990. Pine Flatwoods and Dry Prairies. Pp. 103-149 in Ecosystems of Florida (R. L. Myers and J J. Ewel, eds.). University of Central Florida, Orlando, FL. Beauvais, G. P. and S. W. Buskirk. 1999. Modifyin g estimates of sampling effort to account for sprung traps. Wildlife Society Bulleti n 27:39-43. Ceballos, G. and P. R. Ehrlich. 2002. Mammal popu lation losses and the extinction crisis. Science 296:904-907. Feldhamer, G. A. and D. W. Linzey. 2008. The gold en mouse: taxonomy and natural history. Pp. 21Â–35 in The golden mouse: ecology an d conservation (G. W. Barrett and G. A. Feldhamer, eds.). Springer, New York. Feldhamer, G. A. and A. T. Morzillo. 2008. Relati ve abundance and conservation: Is the golden mouse a rare species? Pp. 117Â–133 in The go lden mouse: ecology and conservation (G. W. Barrett and G. A. Feldhamer, ed s.). Springer, New York. Gaston, K. J. 1991. How large is a speciesÂ’ geogr aphic range? Oikos 61:434-438. Hafner, M. S. 2007. Field research in mammalogy: an enterprise in peril. Journal of Mammalogy 88:1119-1128. Hingtgen, T. 1990. Florida State Parks, District 8: South Fork State Preserve Annual Report.
41 Ivey, R. D. 1949. Life History Notes on Three Mic e from the Florida East Coast. Journal of Mammalogy 30:157-162. Jones, C. A. and J. N. Lane. 1993. Podomys floridanus Mammalian Species 427:1-5. Landers, L. J. and R. L. Crawford. 1995. NB66: A study of habitat relations of nongame birds and mammals in series of oldfield pinewoods f ollowing fire exclusion. Florida game and fresh water fish commission. Nong ame wildlife program project report GFC-84-004. 39 pp. Layne, J. N. 1960. The growth and development of young golden mice, Ochrotomys nuttalli Quarterly Journal of the Florida Academy of Scie nces 23:36-58. Linzey, D. W., and R. L. Packard. 1977. Ochrotomys nuttalli Mammalian Species 75:1-6. Lomolino, M. V. and G. A. Smith. 2001. Dynamic bi ogeography of prairie dog ( Cynomys ludovicianus ) towns near the edge of their range. Journal of Mammalogy 82:937-945. McCarley, W. H. 1958. Ecology, behavior and popul ation dynamics of Peromyscus nuttalli in eastern Texas. Texas Journal of Science 10:147 -171. McCoy, E. D. and H. R. Mushinsky. 1992. Rarity of organisms in the sand pine scrub habitat of Florida. Conservation Biology 6:537-548 McGowan, P., M. Gillman, and M. Dodd. 1998. Asses sing the status of poorly known species: lessons from partridges and pheasants of s outheast Asia. Biological Conservation 83:1-7.
42 Moritz, C., J. L. Patton, C. J. Conroy, J. L. Parra G. C. White, and S. R. Beissinger. 2008. Impact of a century of climate change on sma ll-mammal communities in Yosemite National Park, USA. Science 322:261-264. Myers, R. L. 1990. Scrub and High Pine. Pp. 150193 in Ecosystems of Florida (R. L. Myers and J. J. Ewel, eds.). University of Central Florida, Orlando, FL. Nelson, L. Jr. and F. W. Clark. 1973. Correction for sprung traps in catch/effort calculations of trapping results. Journal of Mamma logy 54:295-298. Packard, R. L. 1969. Taxonomic review of the gold en mouse, Ochrotomys nuttalli. Misc. Publications of the Museum of Natural History University of Kansas, 51:373-406. Packer, W. C. and J. N. Layne. 1991. Foraging sit e preference and relative arboreality of small rodents in Florida. American Midland Natural ist, 125:187-194. Pearson, P. G. 1953. A Field Study of Peromyscus Populations in Gulf Hammock, Florida. Ecology 34:199-207. Parmesan, C., et al. 1999. Poleward shifts in geog raphical ranges of butterfly species associated with regional warming. Nature 399:579-5 83. Pinkham, C. F. A. 1971. Peromyscus and Ochrotomys from the Pleistocene of Reddick, Florida. Journal of Mammalogy, 52:28-40. Rabinowitz, D., S. Cairns, and T. Dillon. 1986. S even Forms of rarity and their frequency in the flora of the British Isles. Pp. 1 82-204 in Conservation biology: the science of scarcity and diversity (M.E. Soule, ed.). Sinauer Associates, Sunderland, Massachusetts.
43 Rose, R. K. 2008. Population ecology of the golde n mouse. Pp. 39Â–58 in The golden mouse: ecology and conservation (G. W. Barrett and G. A. Feldhamer, eds.). Springer, New York. Schipper, J., et al. 2008. The status of the worl dÂ’s land and marine mammals: diversity, threat, and knowledge. Science 322:225-230. Schmidly, D. J. 2005. What it means to be a natur alist and the future of natural history at American universities. Journal of Mammalogy 86: 449-456. Seagle, S. W. 2008. Ecosystem ecology of the gold en mouse. Pp. 81Â–97 in The golden mouse: ecology and conservation (G. W. Barrett and G. A. Feldhamer, eds.). Springer, New York. Smiley, S. A. 2008. The cost of conservation: The National Wildlife Refuge System. Bioscience 58:1014. Smiley, S. A., A. W. Schrey, E. D. McCoy, and H. R. Mushinsky. In review. Mitochondrial DNA variation and phylogeography of g olden mice ( Ochrotomys nuttalli ) from fragmented and continuous landscapes. Stelljes, M. E. 1982. Microhabitats of four speci es of cricetine mice in a Florida scrub community. MasterÂ’s thesis, University of Missouri -Columbia. United States Census Bureau. Annual Estimates of t he Population for the United States, Regions, States, and Puerto Rico: April 1, 2000 to July 1, 2007 (NST-EST200701) [Data file]. Retrieved from http://www.census.gov/popest/states/NST-annest2007.html
44 United States Census Bureau. Historical Annual Tim e Series of State Population Estimates and Demographic Components of Change 1900 to 1990 Total Population Estimates [Data file]. Retrieved from http://www.census.gov/popest/archives/1980s/80s_st_ totals.html Wagner, D. M., G. A. Feldhamer, and J. A. Newman. 2000. Microhabitat selection by golden mice ( Ochrotomys nuttalli ) at arboreal nest sites. The American Midland Naturalist 144:220-225.