Microhabitat use by Podomys floridanus in the high pine lands of Putnam County, Florida

MISSING IMAGE

Material Information

Title:
Microhabitat use by Podomys floridanus in the high pine lands of Putnam County, Florida
Physical Description:
xii, 160 leaves : ill. ; 29 cm.
Language:
English
Creator:
Jones, Cheri A., 1957-
Publication Date:

Subjects

Subjects / Keywords:
Mice -- Florida   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 146-158).
Statement of Responsibility:
by Cheri A. Jones.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001558330
notis - AHH1974
oclc - 79166853
System ID:
AA00003749:00001

Full Text









MICROHABITAT USE BY PODOMYS FLORIDANUS IN THE HIGH
PINE LANDS OF PUTNAM COUNTY, FLORIDA










By

CHERI A. JONES


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1990






















Copyright 1990

by

Cheri Ann Jones













ACKNOWLEDGMENTS

I thank my chairman, J. F. Eisenberg, for providing

encouragement, advice, and equipment during this project.

Members of my committee (R. Franz, J. H. Kaufmann, R. A.

Kiltie, F. E. Putz, and J. G. Robinson) made useful comments

and suggestions regarding this paper. John Eisenberg and

Dick Franz bear responsibility for introducing me to the

Ordway Preserve, where I formed a strong attachment to

Florida mice, gophers, and sandhills. I also benefited from

discussions with C. A. Gates, J. I. Glendinning, J. N.

Layne, and members of the Eco-lunch seminar group in the

Department of Zoology.

I am grateful to the Florida Game and Fresh Water Fish

Commission for issuing permits allowing me to trap Florida

mice and to establish a captive colony. I borrowed aquaria,

traps, and other equipment from J. F. Anderson, L. S. Fink,

W. H. Kern, R. F. Labisky, C. A. Langtimm, F. Nordlie, P.

Ryschkewitsch, and C. A. Woods. The Florida Foundation

provided funds for additional traps, flags, and other field

supplies. I thank T. Perry, the preserve caretaker, for

providing information about the history of Ordway and for

pulling my truck out of the sand more frequently than I care

to remember. Help in caring for captive animals, counting


iii








plants, and other forms of assistance were provided by R. A.

Drury, J. Dudley, J. F. Eisenberg, K. Eisenberg, J. Ellis,

N. Etheridge, R. Franz, S. E. Franz, B. Furlow, A. Gannon,

M. K. Lacy, C. A. Langtimm, A. Mahler, M. Marquez, R. Mor-

rison, C. Partin, L. Partin, J. R. Thomson, and W. W. Tim-

merman. I thank R. Franz and D. W. Hall, who patiently

identified plant specimens and taught me much about the high

pine flora. J. Boucher, J. I. Glendinning and K. M. Portier

helped solve statistical and computer problems.

Burns on the Ordway-Swisher Preserve were conducted by

J. F. Eisenberg, R. Franz, T. Perry, and M. E. Sunquist.

These burns would be impossible without additional help; I

thank all those volunteers (including R. Burke, P. Carnell,

B. Charest, C. K. Dodd, Jr., R. A. Drury, S. E. Franz, C. A.

Gates, H. Gold, L. Hay-Smith, C. A. Langtimm, M. E. Ludlow,

R. Morrison, D. Pearson, F. E. Putz, G. W. Tanner, K. K.

Timmerman, D. D. Wright, and students from the departments

of Botany, Wildlife, and Zoology) who cheerfully incinerated

my study sites and other sandhills.

Finally, special thanks are due to my parents, Clyde

and Charlene Jones, and my grandparents, Ellis and Maye

Owen, who have provided constant encouragement and moral

support throughout my academic career. I also am grateful

to R. A. Drury, P. Foster-Turley, R. and S. E. Franz, and C.

A. Langtimm, who offered friendship and support during my

years in Gainesville.

















TABLE OF CONTENTS


ACKNOWLEDGMENTS . .

LIST OF TABLES . .

LIST OF FIGURES . .

ABSTRACT . .

INTRODUCTION . .

TAXONOMY AND SYSTEMATICS .

DISTRIBUTION . .

MATERIALS AND METHODS .

USE OF TORTOISE BURROWS .

Dependence on Burrows .
Description of Burrows .
Podomys and the Burrow Community
Life History . .

HOME RANGE . .

Fluorescent Powder .
Trapping on Grids .
Trapping at Burrows .
Conclusions . .

DIET . .

Field Observations .
Laboratory Observations .
Acorn Preference Tests .. ..
Supplemental Feeding Experiment

ANALYSES OF VEGETATION .

Sampling Technique .
Results and Discussion .


......... iii
. . vii
. . Vii

. . ix

. . xi

. . 1

. . 4

. . 7

. . 10

. . 18

. . 18
. . 26
. . 31
. 34

. 40

. 41
. 44
. . 51
. . 65

. . 68

. 68
. 70
. . 72
. . 78

. 84

. . 84
. . 89







PRESCRIBED BURNS . . 100

Materials and Methods . 100
Description of Procedure for Prescribed Burns 102
Results and Discussion . 103
Conclusions . . 129

SUMMARY AND DISCUSSION . . 132

APPENDIX LIST OF PLANT SPECIES IDENTIFIED IN
VEGETATION SAMPLES . ... 140

LITERATURE CITED . . .. 146

BIOGRAPHICAL SKETCH . ... .159













LIST OF TABLES


Table Page

1 Numbers of captures of non-target species in
traps at burrows (22,085 trapnights) and on
grids (11,300 trapnights) . 24

2 Data describing distribution of burrows on SL
(top) and AC (bottom) with the probabilities
expected from the Poisson distribution (with
means of 1.5 and 1.7 burrows per square) .. 30

3 Longevity records exceeding 360 days for
Podomys floridanus at SL . ... .39

4 Activity data from ten Podomys dusted with
fluorescent powder at SL . ... 43

5 Home range data for 20 Podomys trapped on two
grids .... . . 48

6 Capture data for 36 Podomys at Anderson-Cue 49

7 Capture data for the 62 adults whose home
ranges are shown on Fig. 9 .. 61

8 Persistence data for animals on grids 83

9 Habitat variables measured in each quadrat 86

10 Frequency (% = percent of quadrats in which
a species occurred) and density (# = average
number of stems per 314 m2 quadrat) of woody
species identified at SL . ... .90

11 Frequency (% = percent of quadrats in which
a species occurred) and density (# = average
number of stems per 314 m2 quadrat) of woody
plant species identified at BP in 1988 .. .92

12 Habitat profile of Podomys at all BP and SL
burrows where vegetation was sampled .. 94


vii








13 Number (%) of vegetation samples reassigned
according to discriminant function analysis
to two categories of total number of captures
(ALL) of Podomys . .. 96

14 Burrows on burned sandhills ..... 111

15 Burrows on unburned sandhills ......... 112

16 Comparison of capture indices for Podomys
at burrows on burned and unburned sandhills 113

17 Comparison of capture indices for Podomys at
burrows on 3 burned sandhills . 115

18 Survival of Podomys following three prescribed
burns on the Ordway Preserve . .120

19 Comparison of trapping effort on unburned and
burned sandhills, where TN is the total number
of trapnights and trapping effort is expressed
as trapnight/capture/ha (TN/CAP/HA) ... 122

20 Residence of young Podomys on the SL sandhill,
where N is the total number of young marked
and R is the percentage of marked young that
became residents on that sandhill ....... .123

21 Responses of Peromyscus species to fire .... 125


viii














LIST OF FIGURES

Figure Page

1 Distribution of Podomys floridanus . 8

2 Distribution of Podomys on the Ordway/Swisher
Preserve . . ... .. .14

3 Trapping success (%) and minimum numbers of
individuals (MNI) (adults and subadults) of
Podomys at burrows on the Smith Lake Sandhill 22

4 Trapping success (expressed as the total num-
ber of captures/adjusted trapnights) for Podo-
mvs on the AC (top) and SL (bottom) grids .. 27

5 Numbers of juveniles (dots) and subadults
(circles) marked on the SL sandhill, expressed
as minimum numbers of individuals/number of
trapnights (top) and as numbers of individuals
(bottom) . . ... .. 36

6 Reproductive data for Podomys trapped on the
SL sandhill, 1983-1988: the proportion of
adult females who were pregnant (top) and the
proportion of males who were scrotal (bottom)
are shown . .... .37

7 Trapping data for Podomys on the ACI, ACII,
and SL grids . . ... 45

8 Home ranges of Podomys trapped on the ACI grid
in May 1987 . . 52

9 Annual home ranges of 62 Podomys on the SL sand-
hill, 1983-1988. ... . .55

10 Minimum numbers of individuals (MNI) on the ACI,
ACII, and SL grids . .. 81

11 Apparent densities (expressed as minimum numbers
of individuals/ha) of Podomys on the burned and
unburned areas of the Blue Pond sandhill ..... 105








12 Apparent densities (expressed as minimum numbers
of individuals/ha) of Podomys on the burned and
unburned areas of the Longleaf Pine sandhill 107

13 Apparent densities (expressed as minimum numbers
of individual/ha) of Podomys on the burned and
unburned areas of the Smith Lake Sandhill 109

14 Home ranges of eight Podomys on the SL sandhill
before (stippled) and after (hatched) the 1985
fire . . 117













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

MICROHABITAT USE BY PODOMYS FLORIDANUS IN THE HIGH
PINE LANDS OF PUTNAM COUNTY, FLORIDA

By

Cheri A. Jones

May 1990

Chairman: John F. Eisenberg
Major Department: Zoology

The purpose of this project was to examine the use of

tortoise burrows (Gopherus polyphemus) by the Florida mouse

(Podomys floridanus) in high pine (longleaf pine sandhills)

on the Katharine Ordway Preserve/Swisher Memorial Sanctuary

in Putnam County, Florida. Mark-and-recapture techniques

were used at tortoise burrows and on grids (10000 m2) in

order to describe burrow use by mice and to measure home

ranges. I also performed a food supplementation experiment

to test the hypothesis that population sizes were food-

limited. Finally, I examined how local distributions of

Florida mice were correlated with vegetation structure and

affected by prescribed burns.

Podomys floridanus are commensals of Gopherus in high

pine habitats at Ordway, as indicated by trapping success,

escape responses, burrow excavations, and observations of








mice in the field. In five years of trapping at burrows on

a single sandhill, 255 mice were marked. Of this total, 32%

were captured only once and 8.6% survived for 360-920 days.

Reproductive activity was concentrated in August, September,

and October. Florida mice appeared to be opportunistic

feeders, accepting a wide variety of fruits, seeds, and

insects. Populations appeared to be limited by factors

other than food. Mean annual home ranges were 2601 m2 for

adult females, whose ranges were mutually exclusive. Home

ranges of adult males were significantly larger (4042 m2)

and overlapped with those of other adults. Numbers of

captures at burrows were best correlated with large numbers

of herbaceous species and high stem counts of Pinus palus-

tris, Quercus laevis, and Diospyros virginiana; there was

little correlation with ground cover characteristics.

Immediate responses of Podomys to prescribed burns were not

spectacular, but in the long term, populations appeared to

be more stable on sites that were burned periodically than

on areas not burned. Implications for conservation of

Podomys on the Ordway Preserve also were discussed.


xii













INTRODUCTION


An understanding of home range, diet, burrow use, and

other interactions between an animal and its physical en-

vironment is crucial for the management and conservation of

threatened and endangered species. However, little is known

about the ecology of Florida's only endemic mammal, the

Florida or gopher mouse. This mouse, Podomys floridanus

(Rodentia: Muridae), is listed as Threatened by the Florida

Committee on Rare and Endangered Plants and Animals (Layne,

1978) and as a Species of Special Concern by the Florida

Game and Fresh Water Fish Commission (1987). Podomys flori-

danus lives in a narrow range of habitats; my study, and

those by Layne at the Archbold Biological Station in High-

lands County, suggest that P. floridanus populations are

higher in the southern sand pine scrub than in the high pine

lands (also known as turkey oak-longleaf pine sandhills or

high pine hills). The special listing of P. floridanus is

due to destruction of high pine and scrub habitats and to

the decline of gopher tortoise (Gopherus polyphemus) popula-

tions. Nash (1895) stated that high pine was the most

extensive of the five habitats he recognized in Florida; by










1982 approximately 57% of this habitat had been lost (Auf-

fenberg and Franz, 1982).

In the high pine lands of northern Florida, the Florida

mouse is associated closely with burrows of the gopher

tortoise (Blair and Kilby, 1936; Humphrey et al., 1985;

Brand, 1987). My preliminary field studies showed that

populations of Florida mice were relatively high and stable

on certain sandhills, whereas populations fluctuated to a

greater degree on other sites. I was intrigued by the

differences I observed in these populations. How did these

sandhills differ from each other? Why were certain tortoise

burrows favored nesting sites, year after year, whereas

other burrows never were used? In an attempt to determine

microhabitat preferences in the high pine habitats, I tried

five approaches:


1) Do Podomys in the high pine depend exclusively on
tortoise burrows? Mouse-use of tortoise burrows
was monitored over a five-year period so that
long-term "burrow histories" could be constructed.


2) How large are home ranges in the high pine? Home
range areas of Podomys in high pine were measured.

3) Do dietary data explain why some tortoise burrows
are preferred over others? Diets of these mice
were observed in the field and laboratory. A
supplemental feeding experiment was performed to
test the hypothesis that population size is food-
limited (i.e., does increased food supply produce
an increase in numbers and body size of mice and
an increase in the length of the breeding sea-
son?).








3

4) Is the local distribution of these mice predic-
table? Physical parameters of burrows, and the
vegetation surrounding them, were quantified.

5) Do populations respond positively to burns?
Effects of prescribed burns on populations of
Podomys were determined.













TAXONOMY AND SYSTEMATICS


Podomys floridanus is the only mammalian species en-

demic to Florida. The species was described as Hesperomys

floridanus by Frank Chapman in 1889, with the type locality

in Gainesville, Alachua County.

Podomys floridanus has long been recognized as distinct

from North American species of Peromyscus. Osgood (1909)

considered Podomys a distinct subgenus within the genus

Peromyscus. Apparently Hooper (1958) was the first to

suggest that the closest relatives of Podomys were not the

cotton mouse and other local Peromyscus, but peromyscines in

Central America. In Carleton's (1980) major revision of the

genus Peromyscus, Podomys was raised from subgenus to genus.

Carleton suggested that Podomys shared a common ancestor

with Habromys and Neotomodon from Mexico and Guatemala;

these three genera grouped together consistently in his

cladistic analysis of 79 morphological characters.

The proposed affinity between Podomys and mice from

Central America is supported by three lines of evidence:

1) Reproductive tract. The baculum of Podomys is
relatively small and slender, compared to that of
northern Peromyscus. Linzey and Layne (1969)
examined additional features of the male reproduc-
tive tract and noted reduced vesicular and










prostate glands, enlarged ampullary glands, and
other features that suggested affinity with Habro-
mys.

2) Ectoparasites. The "typical and only common flea"
of Podomys, described by Johnson and Layne (1961),
is Polygenis floridanus, which most closely resem-
bles a species found in South America. According
to Wenzel and Tipton (1966: 718), "Polygenis is an
"expanding" South American genus which has, quite
clearly, dispersed into Middle America along with
complex-penis-type Cricetinae and caviomorph
rodents from South America".

3) Electrophoresis. In their analysis of genic
variation among Peromyscus, Smith et al. (1973)
considered Podomys one of the least variable
species of the group. One possible explanation
for this low variability is the maintenance of low
numbers in a relictual population.

The fossil record of the Florida mouse is limited to

six Rancholabrean assemblages in this state, so there is not

much light shed on the distribution of Podomys prior to the

Illinoisan. However, a western affinity might be indicated

by a Pleistocene species (Peromyscus oklahomensis) from the

Illinoisan and Sangamonian of Oklahoma and Texas, that might

be related to Podomys (Stephens, 1960). Unfortunately, only

a few isolated teeth of this extinct species have been

found.

Podomys is not unique among the flora and fauna of

Florida in its apparent affinity to Central America. Several

plants (such as Opuntia, the prickly pear) and several

reptiles, birds, and mammals are considered species of

western affinity. Some, such as Gopherus polyphemus, show a

disjunct distribution (Blair, 1958). Among the mammals, Webb

and Wilkins (1984) found that 15-28% of the mammals from








6

five Pleistocene sites in Florida were neotropical species.

They suggested that the arrival of most of these species

could be explained by a Gulf Coastal Corridor, which served

as a pathway for southern and western immigrants in the

early Pleistocene.













DISTRIBUTION


In recent times P. floridanus occurred from St. Johns,

Clay, Alachua, Putnam, Suwannee, and Taylor counties south

to Sarasota, Highlands, and Dade counties, with isolated

populations in Franklin County and on Merritt Island (Os-

good, 1909; Layne, 1978; Stout et al., 1979). A survey of

the present distribution of Podomys is urgently needed.

Podomys is the characteristic mouse of some of the most

xeric habitats in Florida. It is most common in sand pine

scrub and high pine, although it also is found in slash

pine-turkey oak and coastal strand associations (Bangs,

1898; Layne, 1978; U.S. Dept. Agric., 1980). On Fig. 1, I

have mapped the distribution of scrub and high pine, as

indicated by the Soil Conservation Service (U.S. Dept.

Agric., 1980), and localities from which Podomvs have been

reported (Layne, 1978). Clearly there is a close associa-

tion between Podomys and these two habitats. Layne (in

press) has suggested that the scrub was the ancestral home

of this mouse, with the high pine a secondary habitat.

In some areas, P. floridanus and P. gossypinus appear

to coexist (Layne, 1970). In Putnam County, however, trapp-

ing by Suzanne Brand (1987), John Eisenberg (1983; 1988),






































80 km


as 0


Fig. 1. Distribution of Podomys floridanus (triangle
designates type locality). Stippling represents
distribution of high pine and scrub habitats.








9

Eric Milstrey (1987), and me has indicated a clear separa-

tion between these two species, with P. aossypinus occurring

only in lower, wetter microhabitats, and Podomys only in

sandhills, or along edges between sandhills and old pastures

or sandhills and xeric hammocks.













MATERIALS AND METHODS


I conducted this research on the Katharine Ordway

Preserve-Carl S. Swisher Memorial Sanctuary, located about

4 km E of Melrose in Putnam County. The preserve, 37 km2 in

size, is administered jointly by the Florida Museum of Na-

tural History, the University of Florida's School of Forest

Resources and Conservation, and the Nature Conservancy.

Approximately one-third of the Ordway Preserve is

covered by high pine land (sensu Nash, 1895). This habitat

is characterized by an open canopy of longleaf pine (Pinus

palustris) and turkey oak (Quercus laevis), although other

trees, including sand pine (P. clausa), Chapman's oak (Q.

chapmanii), and bluejack (Q. incana), also are present. The

understory consists of scattered shrubs, and herbs and

grasses. On Ordway the characteristic grass of the high

pine land is wiregrass (Aristida stricta); other common

grasses include pineywoods dropseed (Sporobolus junceus),

six species of bluestem (Andropogon spp.), and three species

of threeawn (Aristida spp.). Additional grasses identified

on the preserve are included in the list compiled by David

W. Hall and Richard Franz (unpubl. data). Reindeer moss









(Cladonia and Cladina spp.) is common. The herbs I

identified on my study sites are listed in the Appendix.

The soils of high pine are sandy and well-drained.

There are many patches of bare ground, which become covered

with oak leaves, pine needles, tree branches, and other

debris in the absence of fire. Under the pines, the needle

litter is several inches deep; in areas where the turkey

oaks predominate, leaf litter is distributed unevenly by the

wind. Burrows of the gopher tortoise (Gopherus polyphemus)

are prominent features on this landscape, although popula-

tion densities probably were higher in coastal areas before

human development (Carr, 1940; Auffenberg and Franz, 1982;

Cox et al., 1987).

Other mammals found on the high pine land are the

opossum (Didelphis virainianus), pocket gopher (Geomys

pinetis), fox squirrel (Sciurus niger), flying squirrel

(Glaucomys volans), armadillo (Dasypus novemcinctus), rac-

coon (Procyon lotor), and white-tailed deer (Odocoileus

virainianus). Eisenberg (1988) and Franz (1989) compiled

lists of mammals known from the various habitats of Ordway.

Carr (1940) listed the amphibians and reptiles of high pine

habitats. On my study sites, I saw gopher frogs (Rana

areolata), oak toads (Bufo quercicus), fence lizards (Scelo-

porus undulatus), six-lined racerunners (Cnemidophorus

sexlineatus), pine snakes (Pituophis melanoleucus), crowned

snakes (Tantilla spp.), southern black racers (Coluber









constrictor), coachwhips (Masticophis flagellum), coral

snakes (Micrurus fulvius), pigmy rattlesnakes (Sistrurus

miliarius), eastern diamondbacks (Crotalus adamanteus), and

tortoises (G. polyphemus). Birds common to high pine were

listed by Nelson (1952), who thought the avifauna of the

high pine did not differ substantially from those of oak

hammocks and other nearby communities. However, Maehr et

al. (1982) suggested that the high pine might support a

greater density of birds than flatwoods and xeric hammocks.

I did not attempt to identify birds on my study sites in

detail.

On the Ordway Preserve, all locations at which P.

floridanus have been found are in high pine habitats, or at

the ecotone between high pine and hammock or high pine and

old pasture (Brand, 1987; Eisenberg, 1988) (Fig. 2). I

trapped mice on four independent high-pine sites, designated

Anderson-Cue (AC), Blue Pond (BP), Longleaf Pine Pasture

(LP), and Smith Lake (SL). The history of fire on these

sites is described below in the Prescribed Burns section.

Smith Lake is my longterm study site; Eisenberg and Franz

started research projects there in 1983, marking tortoise

burrows and trapping Florida mice, gopher frogs, and other

burrow inhabitants. I trapped mice at SL from 1984 through

1988. SL is approximately 16.8 ha in size. Vegetation

consists of P. palustris, Q. laevis, Q. chapmanii, Q. qemin-

ata, and grasses and forbs characteristic of the high pine.













4-4 wQ)


>

.4J ..-I
mC
d 0p 0


0 *
HUE




U-S
0 0

44 cP4I





4 0 ( 0


.044





OPm
*Q) i0




















OW
: 0 .0

*tMO

























Q -,C
go (
a )*
VC 0)







05
0P4
4)


0
'Q 0)


m 3






>Mj r n c a
c4 Q) 0 U

a -1






rcz
0e

















El
-r


r I
I


I
I
I
),.





0
-J


0
Qe

c I
m
0 c


o o

)
CL







15

The Anderson-Cue sandhill is the highest in elevation of the

four sites; typical vegetation there includes P. palustris

and Q. laevis, with fewer Q. geminata than on SL. I worked

there from 1986 to 1988. AC was burned in the winters of

1983, 1986, and 1989. It was not included in my study of

prescribed burns because I did no pre-burn trapping there.

The Longleaf Pine Pasture Sandhill is south of the pine

pasture surrounding the old Swisher Estate (Fig. 2). Part

of this sandhill, the most disturbed of my study areas, has

been planted in slash pine. Dominant plants here include

maidencane (Panicum hemitomon) and centipede grass (Eremoch-

loa ophiuroides), which are not typical of natural high pine

vegetation. There are many turkey oaks and several large

patches of gopher apple, but fewer longleaf pines, pawpaws,

and other high pine species. I worked on LP for only one

trapping season for the purpose of gathering additional

information regarding prescribed burns. The fourth study

site, Blue Pond, was logged intensively. There are few

longleaf pines, and turkey oaks are by far the most numerous

trees. Ground cover is sparse in some areas but rather

dense in others. Species composition of the ground cover

appeared similar to that of SL.

As described in the following chapters, I trapped E.

floridanus at tortoise burrows and on grids using standard

mark-and-recapture techniques (Davis, 1956). While trapping

at burrows, I numbered, marked, and mapped each burrow, and








16
then placed a pair of Sherman traps (baited with oatmeal) at

the entrance, or just inside the mouth, of the burrow.

Animals were marked by toe-clipping and then weighed. I

also noted sex, reproductive condition, and presence of

ectoparasites. Animals were released near the burrow en-

trance, and in most cases I recorded whether they entered

the burrow or ran elsewhere. I collected the same data on

grids. I placed one grid (designated SL) on my study site

on the south side of SL; two more grids (ACI and ACII) were

set north and east of the Anderson-Cue ponds. Grids were

more than 100 m apart to reduce movement of animals between

grids. Each grid consisted of 10 columns and 10 rows 10 m

apart (area = 10000 m2), with a single Sherman at each

intersection. Each tortoise burrow and grid location was

given a unique number and marked with forestry flags. On

the burrow flags I also attached an aluminum tag (Al Tag,

Forestry Suppliers, 205 W. Rankin Street, Jackson, Missis-

sippi 39204) bearing the burrow number. These tags survived

all but the hottest burns, thus serving as more permanent

markers of burrow locations.

Traps were opened near sundown and checked at or short-

ly following sunrise. Because of temperature extremes,

morning condensation, and (in some months) heavy rain, I

found it necessary to check traps in the morning as quickly

as possible. Animals developed hypothermia when ambient

temperatures reached 10 C; I extended the trapping season










into November by insulating traps with a handful of excel-

sior, wood shavings used in packing glassware. Mice com-

press the excelsior into a pad on top of the treadle.

Spanish moss also can be used, but I found the wood to be

superior because it does not absorb moisture. I did not

trap in December or January; fieldwork resumed when warm

weather resumed (February-April).

I measured trapping effort as trapnights (one trapnight

being one trap/night). Results of the vegetation analyses

were run on SYSTAT at the IFAS Consulting Center at the

University of Florida; other statistics were performed on a

hand-held calculator. Statistical methods are described in

each chapter.













USE OF TORTOISE BURROWS


Dependence on Burrows

Previous students of peromyscine behavior and morphol-

ogy concluded that Podomys was primarily a terrestrial

species; although capable of climbing and digging, Podomys

climbed and dug less than several species of Peromyscus to

which it was compared (King, 1968; Layne 1970; Layne and

Ehrhart, 1970; Siegel and Van Meter, 1973). In some areas,

these mice seemed capable of excavating their own burrows;

Lee (1968) stated that P. floridanus built burrows greater

than 120 cm in depth in scrub habitats, and Starner (1956)

trapped one animal in a small burrow that she thought might

have been dug by the mouse (in the ecotone between the mesic

hammock and longleaf pine flatwoods at San Felasco Hammock,

Alachua County). Layne and Ehrhart (1970) suggested that

Podomys rarely dug its own burrows, but Layne (in press)

noted high populations of Florida mice in scrubs where

tortoises were absent or occurred in low numbers. However,

observations by Eisenberg (1983), Brand (1987), and myself

suggested a much closer relationship between Podomys and

Gopherus in the high pine in Putnam County.










Blair and Kilby (1936) were the first to note in the

literature that Podomys floridanus used burrows of Gopherus

polyphemus in northern and central Florida. They saw mice

using active burrows (i.e., burrows that had tortoises in

them) in an old field in Alachua County. Three of the 15

mice captured by Starner (1956) were taken from traps placed

inside entrances of tortoise burrows. Milstrey (1987:115)

reported that "Florida mice were continuously present" in

many tortoise burrows at his study site in Putnam County.

Johnson and Layne (1961) also noted that Podomys inhabited

tortoise burrows, and possibly burrows of pocket gophers

(Geomys pinetis). At other sites Podomys used burrows

constructed by Peromyscus polionotus, Sigmodon hispidus, G.

pinetis, and Dasypus novemcinctus (Gentry and Smith, 1968;

Layne, in press). However, in my 33,385 trapnights on

burrows and grids at Ordway I have observed Florida mice

entering small holes of unknown origin only three times

(twice on SL and once on AC), and I have no evidence that

they inhabited logs or other shelters.

I found that mice almost always ran down tortoise

burrows when released. For example, I tallied escape re-

sponses of 18 Podomys caught on the AC grid (and therefore

released at grid coordinates, not at burrows). Out of 35

captures, mice ran down tortoise burrows 25 times; they hid

in the base of turkey oaks or in fallen logs and other

debris 10 times. In my regular trapping sessions, I never








20
saw mice climb vegetation as an escape response; Eisenberg

(pers. comm.) noted one such response in 1989. However, my

field observations have confirmed climbing ability by these

mice (Jones, 1989, and Home Range below). In T-maze tests

of arboreality performed by C. A. Langtimm at SL (in which a

T-maze was propped against an oak), most mice jumped off the

T, but the animals who chose to climb oaks reached heights

of 8 m (C. A. Langtimm, unpubl. data). Clearly these ani-

mals are capable of climbing, even though the time spent on

this activity in any one night remains unknown.

In preliminary studies on the Ordway Preserve, Eisen-

berg (1983) reported that trapping success was much higher

if traps were placed at the mouths of tortoise burrows (33%)

than on transects (0.4%). Humphrey et al. (1985) caught 113

mice in 1,144 trapnights at 51 burrows, also at Ordway. In

146 trapnights on transects--in which traps were not placed

at burrows--I caught no mice. At the SL burrows in August,

September, October, and November 1983, Eisenberg (unpubl.

data) had trapping successes of 35, 43, 62, and 44%, respec-

tively. I calculated trapping success using adjusted trap-

nights, which indicated how many traps were available after

subtracting half the number of traps sprung and containing

non-target species from the total number of traps set. My

trap success for Podomys ranged from 0 to 16%, with SL being

the most successful (2-16% with a mean of 7%) (Fig. 3).

Variations in total trapnights (within and among sites)















W)

'-I








4J

00





t4-4 a :
0d 4JC
~4J t

p .0
4.1













a0
to

















0 0 t

-AI


E I4J
li
















dP N4.4
a)a



















a) -r4
r.-I 0
u to 04
















U 4J
0. r-1
to 0 $-4
V4 41i
EE03












r:4




















0
* |

-3

.4

0


o




.4.


0




*-3

O)
--
z
0

o s

<)




z

o)
0"1
-


en
Wco
oa
T"


INYN


aN ( C

GNV (%) SSmoonS ONIddV.l









reflect my attempt to maximize sample size by trapping at

all tortoise burrows on each sandhill until burrows disap-

peared.

On all four sandhills, traps at burrows occasionally

were sprung by falling branches or by insects removing bait.

A large number of sprung traps was an occasional problem at

AC, but a more major one at BP, particularly on the north

side. There I identified tracks of armadillos, raccoons,

and dogs, but I suspected dogs were the major culprits. One

morning I found two traps covered with urine and several

times I saw dogs on the sandhill, which was the closest of

my sites to the periphery of the preserve. It is also worth

noting that I found more carcasses of turtles and tortoises

here than on any other study site. Following a suggestion

of M. E. Sunquist, I fashioned "croquet hoops" out of alu-

minum welding rods and placed them around the traps. They

helped decrease the number of traps moved, but did not

always prevent closing of trap doors.

In traps at burrows, the most frequently-captured non-

target species was Rana areolata (Table 1). At SL, from

September 1984 to the end of the study, gopher frogs were

captured 105 times. Other occasional captures were camel

crickets (Ceuthophilus sp.), centipedes, southern toads

(Bufo terrestris, once on AC, LP, and SL), six-lined race-

runners (Cnemidophorus sexlineatus), and a young rabbit

(presumably Svlvilaqus floridanus, on SL). All of these








24

TABLE 1. Numbers of captures of non-target species in traps
at burrows (22,085 trapnights) and on grids (11,300 trap-
nights). Captures of insects and centipedes were not
counted. Scientific names are listed in text.


Burrows

camel crickets

1 spider

3 southern toad

129 gopher frog



3 six-lined racerunner

1 cotton mouse

38 flying squirrel

1 rabbit


Grids

grasshoppers

cockroaches

centipedes

3 six-lined

racerunner

2 fence lizard

1 black racer

2 rice rat

3 cotton rat

148 flying squirrel










species are known to use Gopherus burrows (Jackson and

Milstrey, 1989). In 1988 I began catching flying squirrels

(Glaucomys volans) at the AC and SL burrows; at SL I cap-

tured squirrels 27 times from May to October, 15 times in

May alone. These results must indicate some change in the

Glaucomys population that resulted in additional terrestrial

foraging, or perhaps they reflect learned behavior of ani-

mals who found oats at traps on grids. At any rate, I

observed only one flying squirrel enter a burrow upon re-

lease, and I do not believe they reside in tortoise burrows

on these sandhills.

At study sites other than Ordway, additional murids,

including P. gossypinus and P. polionotus, visit tortoise

burrows (Jackson and Milstrey, 1989). In 22,085 trapnights

at burrows on the AC,BP, LP, and SL sandhills I caught only

one cotton mouse (P. gossypinus, on AC). In other Ordway

studies, Eisenberg (1983) and Brand (1987) found Podomys in

the high pine and Peromyscus in the hammocks and flatwoods.

Clearly, Podomys is the only rodent that regularly inhabits

tortoise burrows in these sandhills. At burrows in other

Ordway habitats, the absence or lower numbers of Podomys

might represent a case of competitive exclusion; after the

1987 drought Eisenberg (pers. comm.) reported trapping

Florida mice in hammocks closer to lakes than in previous

years, and Peromyscus were absent.







26

For comparative purposes, trapping results on two grids

on AC and one grid on SL also are shown (Fig. 4). Trapping

success on grids ranged from 0 to 7% (AC) and from 0 to 6%

(SL); mean trapping successes were 0.47%, 0.70%, and 3.6%

for ACI, ACII, and SL. Traps on grids were more likely to

be sprung by rain, falling branches, pine cones, and rac-

coons than were traps at burrows. Only one Florida mouse

(female 18, a juvenile known to be alive 36 days at SL) was

found on a grid and not during burrow trapping. Flying

squirrels were caught more frequently (Table 1), usually at

traps near the bases of pines, oaks, or dead trees. One

cotton rat (Siqmodon hispidus) was captured once at SL; this

was a juvenile male (which I marked) presumably dispersing

across the high pine. The other cotton rats and the rice

rats (Orvzomys palustris) were caught on the AC grids on 28

and 29 June 1988; their appearance on the grids might re-

flect drought-related dispersal out of the vegetation sur-

rounding the AC ponds. Centipedes occasionally were caught

on all grids. On ACI I also captured grasshoppers, cock-

roaches, fence lizards (Sceloporus undulatus), six-lined

racerunners (C. sexlineatus), and a black racer (Coluber

constrictor).



Description of Burrows

The size and configuration of tortoise burrows have

been described by Hallinan (1923) and Hansen (1964). Blair












---0-- ACI
-- ACII


I
%
0, ,S

M J J AS O M A M J J AS O N


1987


1988


MONTH


A M J J ASOM A M J J A S O N
1987 MONTH 1988
MONTH


Fig. 4. Trapping success (expressed as the total number of
captures/adjusted trapnights) for Podomys on the
AC (top) and SL (bottom) grids.







28

and Kilby (1936) noted the use of a small hole approximately

182 cm from the entrance of a burrow. However, the only

description of structures used by Podomys was that of Layne

(in press), who described nests, nest chambers, and small

tunnels found in two tortoise burrows in Alachua County.

Richard Franz and I excavated five burrows on Ordway in

order to further describe structures utilized by mice.

Typically, mice used small U-shaped tunnels constructed off

one side of the tortoise burrow, with both entrances to the

U opening into the main burrow. We also noted use of tun-

nels we dubbed "chimneys", which ran from the ceiling or

side of the burrow to an outside opening a meter or two past

the main entrance. In all five burrows, mouse activity

appeared concentrated in the upper 2 m of the burrow.

Additional details of the actual excavations were discussed

by Jones and Franz (in prep.). Chimneys allowed occupation

of burrows after the main entrance has collapsed, and might

serve as escape routes from predators. Although not all

burrows have chimneys (at my four study sites from 6.8% (AC)

to 26.9% (BP) of the burrows had chimneys), the presence of

chimneys is an indicator of past or present occupation of

tortoise burrows by Podomys. Exploratory trapping for mice

in the sandhills should not omit observation of these small

holes.

To determine the pattern of dispersion of tortoise

burrows on the SL sandhill, I used methods described by







29

Brower and Zar (1984). In this procedure, a grid is placed

over a map of the study area and numbers of squares contain-

ing no burrows, squares containing one burrow, and etc., are

counted. I used a grid with squares 35 mm by 35 mm (which

corresponds to 1225 m2 on the map). My SL study site con-

sisted of 93 squares with a mean of 1.46 burrows per square

(or 0.11 burrow/m2). I included all 136 burrows monitored

from 1983 through 1988, without consideration of burrow

classification or tortoise demography; that is, I considered

only burrow locations without burrow size, duration of

residence, or other characteristics. The data fit the

Poisson distribution, leading to the conclusion that the 136

burrows on SL are randomly distributed (Table 2). Another

method to determine whether distribution is random is the X2

test, for which I obtained a value of 0.547 (not signifi-

cantly different from random at p = 0.05). As a third me-

thod, I tried the Morisita's index of dispersion, which is

calculated as I = n [(EX2-N)/(N(N-1))] where n is the number

of squares, N is the total number of burrows counted on n

squares, and ZX2 is the squares of the numbers of burrows

per plots for all plots. A value of 1.0 for I indicates

random dispersion; value of 0 indicates uniform dispersion.

I obtained I = 0.95, which indicates random dispersion.

Finally, I calculated the ratio of observed to expected

density; I obtained a value of 0.94, which is close to the

value of 1.00 which again indicates random distribution









TABLE 2. Data describing distribution of burrows on SL
(top) and AC (bottom) with the probabilities expected from
the Poisson distribution (with means of 1.5 and 1.7 burrows
per square). X indicates number of burrows per square; f(X)
indicates the observed frequency; p(X) is the observed
probability; and P(X) the expected Poisson probability.


X f(XIl pX_ P(X)

0 20 0.215 0.223

1 33 0.354 0.335

2 24 0.258 0.251

3 10 0.108 0.126

4 5 0.054 0.047

5 1 0.011 0.014


0 15 0.259 0.183

1 13 0.224 0.311

2 13 0.224 0.264

3 12 0.207 0.150

4 3 0.052 0.064

5 0 0 0.022

6 0 0 0.006


2 0.035


0.001









(Brower and Zar, 1984). The 101 AC burrows also were dis-

persed randomly (X2 = 5, significant at p = 0.10; I = 0.98;

ratio of observed to expected = 1.29).

Observations of dispersion of other gopher tortoise

burrows vary. McRae et al. (1981) and Auffenberg and Franz

(1982) recognized burrows clumped as colonies, whereas

others reported random distributions (e.g., Kushlan, 1982;

Kushlan and Mazzotti, 1984). Whether these discrepancies

reflect variations in tortoise behavior or in data collec-

tion remains unknown.



Podomys and the Burrow Community

In the high pine, Gopherus burrows provide shelter from

the fires and extreme fluctuations in temperature and rain-

fall that characterize this environment. Burrows of the

southeastern pocket gopher (Geomys pinetis) also are common

in the high pine, but the two types of burrows are used by

different species (Funderburg and Lee, 1968). Eisenberg

(1983) suggested that G. polyphemus could be considered a

keystone species in the high pine due to the numerous and

diverse species that use its burrows. Franz (1986) listed

19 species of amphibians and reptiles that used the SL bur-

rows. The most recent state-wide list of species associated

with Gopherus burrows consists of 60 vertebrates (Amphibia,

Reptilia, Aves, and Mammalia) and 302 invertebrate species









(Gastropoda, Malacostraca, Arachnida, Chilopoda, Diplopoda,

and Insecta) (Jackson and Milstrey, 1989).

Besides using burrows as shelter, I suspect Podomys

takes advantage of the invertebrate commensals and visitors.

My captive mice readily accepted live crickets and meal-

worms. Milstrey (1987) described Podomys eating blood and

body fluids of engorged ticks (Ornithodoros turicata) that

parasitize Gopherus and Rana areolata.

Florida mice are presumably preyed upon by birds,

foxes, bobcats, and raccoons (Layne, 1978; Maehr and Brady,

1986; Wassmer et al., 1988). I suspect that in the high

pine snakes are major predators, although predation by

snakes has not been documented. I occasionally saw coral

snakes (M. fulvius) near burrows, but omitted the species

from consideration because coral snakes do not prey upon

mammals (Jackson and Franz, 1981). Mushinsky (1987) sum-

marized predation records for snakes; some of the species

known to take mammals include Coluber constrictor, Drymar-

chon corais, Elaphe obsoleta, and Pituophis melanoleucus.

Several species of snakes utilize tortoise burrows at Ordway

and must be considered potential predators of Podomys,

including C. constrictor, E. quttata, P. melanoleucus,

Masticophis flagellum, D. corais, Sistrurus miliarius, and

Crotalus adamanteus (Franz, 1986; Timmerman, 1989). In

particular, eastern diamondbacks are known to take Peromys-

cus, and Timmerman (1989) considered Podomys a potential









prey item. During my time at Ordway, I saw only one dia-

mondback in the high pine (an adult female near BP burrow

#319, 6 November 1988). Timmerman (1989) examined habitats

used by eight eastern diamondbacks at Ordway from 1985 to

1989. His snakes used high pine and old field habitats less

than expected (approximately 13% of all observations) al-

though these two habitats comprised more than 60% of the

area available. The diamondbacks spent little time under-

ground, and only 16% of the burrow locations were in Gopher-

us burrows. However, as Timmerman pointed out, 75% of the

locations were in dens of the armadillo, which frequently

takes over old Gopherus burrows. Tortoise burrows are used

as winter refugia and, in the fall, as parturition sites,

leading to the suggestion that Podomys might be important in

the diet of young diamondbacks (Timmerman, 1989).

Mortality rates of Podomys due to disease and parasites

are unknown. Thirty-two species of endo- and exoparasites

of Podomys have been reported (Layne, 1963, 1968a; Lichten-

fels, 1970; Whitaker, 1968). Milstrey (1984, 1987) sum-

marized recent studies of the soft tick Ornithodorus, a

known carrier of relapsing fever that feeds on residents of

gopher burrows, but, as he points out, the relationships

between the vertebrates and invertebrates of these burrow

systems are still poorly understood.









Life History

I examined demographic data for my largest sample, the

255 P. floridanus marked at the SL burrows, August 1983 to

November 1988. Of this total, 52 females and 44 males were

marked in 1983. Eighty-one (38 females, 43 males)of the 255

(32%) were captured only once. I calculated minimum trap-

pability ((number of captures 2)/(possible captures -2))

for animals captured three or more times (Hilborn et al.,

1976).

Individual trappability ranged from 14 to 100%. Aver-

age trappability was 58% for 27 mice in 1983, 55% for n = 17

in 1984, 58% for n = 16 in 1985, 62% for n = 16 in 1986, 48%

for n = 13 in 1987, and 60% for n = 24 in 1988. The average

for all five years was 57%. According to Hilborn et al.

(1976), when trappability = < 50%, estimates of minimum num-

bers of individuals (MNI) by direct enumeration will under-

estimate actual population size; estimates become more

reliable as trappability increases. Trappability on grids

was lower than on burrows. For example, individual trap-

pability on the SL grid ranges from 38 to 50%.

The reproduction, growth, and development of Podomys

were described by Layne (1966). He demonstrated that sea-

sonal distribution of pregnancies in a wild population in

Alachua County was bimodal, with the major peak occurring

between August and November. At SL, I classified young

animals as juveniles if the gray and white juvenile pelage







35

was still present, and as subadults if the dorsum was gray

with gold and agouti hairs appearing along sides and flanks

(minimum weights of subadults were approximately 25 g).

Numbers of young marked (Fig. 5) and numbers of reproductive

adults (Fig. 6) at SL indicate that seasonality of reproduc-

tion at Ordway also peaked in the late summer and fall,

although juveniles and reproductive adults might appear in

small numbers at other times of the year (Layne, 1966).

Juveniles usually were first captured at the mouths of

burrows at approximately three weeks of age. At this point

the young were still completely gray and white and weighed

approximately 9 to 14 g (Layne, 1966). In some cases young

appeared at burrows with females (and sometimes males) who

were long-term residents, but I had no way of determining

parentage of some juveniles and almost all subadults. This,

and the fact that many juveniles disappeared before comple-

tion of adult molt, suggests that both pre- and post-

reproductive dispersal occur.

Captive animals breed year-round (Dice, 1954; Layne,

1966; Drickamer and Vestal, 1973). In captivity mean litter

size (1.7-2.6) is slightly smaller than that reported (3.1)

for wild-conceived litters (Dice, 1954; Layne, 1966; Rood,

1966; Drickamer and Vestal, 1973; Glazier, 1985). Maternal

care in captivity consisted of creating nests, plugging

entrances to nest boxes, nursing and grooming young, and

herding juveniles into the nest, as described previously by










an
I-
z
0






I-


z
2-


M A M J J A S O N


MONTH


-0-"


M A M J J A S O N


MONTH


Fig. 5. Numbers of juveniles (dots) and subadults
(circles) marked on the SL sandhill, expressed as
minimum numbers of individuals/number of
trapnights (top) and as numbers of individuals
(bottom).


1.2

1.0

0.8

0.6

0.4

0.2

0.0




























F M A M J JA S O N
MONTH


F M A M J J
MONTH


---t-0--


1984
1985


-**...." 1986


---+-0-













---


1987
1988















1984
1985
1986
1987
1988


Fig. 6. Reproductive data for Podomvs trapped on the SL
sandhill, 1983-1988: the proportion of adult
females who were pregnant (top) and the proportion
of males who were scrotal (bottom) are shown.









Layne (1966, 1968b). In my captive colony 8 females and 5

males produced 54 young in 24 litters; 11 litters were born

to a single pair. Litter size ranged from 1 to 4 (mean =

2.25). The distribution of births was: 8 in September; 4 in

July; 3 in October; 2 in March, August, and December; 1 in

January, February, and November. Four neonates have not

been sexed at this time; the remaining young consisted of 24

females, 19 males, and 7 infants who died before gender was

recognized.

Longevity of Podomys in the wild is unknown. Keim and

Stout (1987) reported that a male captured at approximately

three weeks of age lived in captivity for 7 years and 4

months. At SL, 12 females and 10 males (8.6% of the animals

marked) were present for more than 360 days (Table 3).

Seven females and one male were first marked as juveniles;

20 individuals were first found on the south side of SL and

two were marked on the north and later migrated south. In

captivity, two- and three-year-olds are continuing to repro-

duce.









TABLE 3. Longevity records exceeding 360 days for Podomvs
floridanus at SL. ID is the identification number, YR =
year marked, A = age when marked (Juvenile, Subadult, or
Adult), and P = persistence (in days) as indicated by both
trapping methods. Asterisk indicates trap mortality.


Females

ID YR

43 83

55 83

65 83

23 84

30 84

61 86

10 87

19 87

26 87

47 87

58 87

64 87


P

445

409

649

467

403

360

588

385

554

547

392

504


Males

ID YR

39 83

48 83

63 83

11 84

32 84

98 85

80 86

98B 86

42 87

54 87


P

905

570

410

920

620

401*

515

500

457

412













HOME RANGE


The concept of home range most widely accepted by

mammalogists is that defined by Burt (1943:351) as "that

area traversed by the individual in its normal activities of

food gathering, mating, and caring for young. Occasional

sallies outside the area should not be considered as

in part of the home range. The home range need not cover

the same area during the life of the individual." The

concept of territoriality, on the other hand, involves

domination by either overt or covert social interaction. As

defined by Kaufmann (1983:9), a territory is "a fixed por-

tion of an individual's or group's range in which it has

priority of access to one or more critical resources over

others which have priority elsewhere or at another time."

Among peromyscines, there is considerable variation in

descriptions of both home range sizes and degrees of ter-

ritoriality (Eisenberg, 1968; Stickel, 1968). For P. flori-

danus, home range and territoriality have not been des-

cribed. I examined the size of home ranges on the Ordway

sandhills with three methods: fluorescent powder, trapping

on grids, and trapping at burrows.










Fluorescent Powder

The use of fluorescent powder to track small rodents

and to determine movement patterns was introduced in the

sandhills of Nebraska (Lemen and Freeman, 1985, 1986). I

applied this technique to examine size of nightly activity

areas of Podomys on the SL sandhill.

I powdered five animals on 25 May 1985, three on 14

July 1985, and three on 15 June 1987. The first eight ani-

mals were trapped at burrows and held until 2100 EST the

following evening; the last three were removed from traps on

the grid at 0230 EST, powdered, and released immediately.

Each animal was placed into a plastic ziplock bag with a

fine, nontoxic fluorescent powder (green, pink, or yellow),

gently shaken until completely covered, and released at the

point of capture. At a later time, I returned with the

fluorescent light and marked the powder trails with forestry

flags. Trails then were mapped using a tape measure and

compass.

The mice showed no ill effects from this technique by

loss in weight or other signs of poor condition. One female

was so little traumatized by the procedure that she caught

and consumed a moth immediately after her release. However,

with the exception of the moth incident, I did not obtain

additional information regarding feeding.

Powdering was of some use in that it showed animals

utilizing fallen trees and branches as runways, as I had







42

observed in escape responses during trapping. Douglass and

Reinert (1982) showed that utilization of fallen logs by

Peromyscus leucopus was not an artifact of trapping but part

of the animals' normal movements. Additional evidence of

climbing was indicated by the powder trail of an adult male

(#39), which extended 36 cm up a turkey oak snag.

The powder indicated only one case of social interac-

tion. On 15 June 1987 females 10 and 47 were caught in the

same trap. Female 47 was powdered, released, and recaptured

later that morning. Again, she was found in the same trap

with #10, who also had become covered with powder. Female

47 was first marked as an adult on the SL grid on 19 April;

she remained at the grid and surrounding burrows until at

least October 1988. Number 10 was a younger female marked

as a juvenile on 12 April at a burrow located on the grid;

by September she had moved to an area west of the grid.

Possibly #10 was a daughter of #47. Use of radioisotopes,

DNA fingerprinting, and other techniques would be more

appropriate and efficient methods to examine family rela-

tionships and dispersal of young Podomys.

I Distances traveled by mice, as indicated by powder

trails, are shown in Table 4. Trails for two animals were

very short (< 10 m), but I believed that a single night's

excursion exceeded 10 m. By trapping on 14 July 1985 and 15

June 1987, I discovered powdered mice at locations to which

no powder trails led. Thus, the powder trail showed only a










TABLE 4. Activity data from ten Podomys dusted with fluo-
rescent powder at SL. ID indicates sex and identification
number of mouse; R indicates the burrow or grid station at
which an animal was caught and released; ORI indicates
initial direction from release point (in degrees); DIST is
the greatest distance (in meters) from R as indicated by
fluorescent powder; DEST is the burrow or grid station at
which an animal was found the following morning. The actual


distance


from R to DEST is shown in parentheses.


DATE

25 May 85

25 May 85

14 Jul 85

25 May 85

15 Jun 87

25 May 85

14 Jul 85

14 Jul 85

15 Jun 85

15 Jun 85


R

35

13

81

58

6,2

10

81

22

9,4

9,1


ORI

222

125

302


DIST

5.5

11.9


9.9


70 17.7

47 20.0


162

140

108

273

250


15.2

28.0

19.2

2.0

15.0


DEST


71 (36.0)


6,0

10


15 (45.0)

23 (15.0)

9,5 (10.0)


49 281 11.5


ID

F23

F30


F39

F47

Mil

M32

M39

M49

M54


M98 25 May 85









part of the night's activities. Because the fluorescent

powder yielded so little information regarding home range,

feeding behavior, or social activity, I concluded that it

was of little use with Podomys at Ordway.



Trapping on Grids

Results of trapping on the three grids are shown on

Fig. 7, which includes data from all animals, including

transients caught only once or twice. A total of 18 animals

was trapped on the AC grids. I calculated persistence on

the grid (P,) as the number of days from first to last

capture. Time spent by each individual on the grid is

represented by the horizontal bars; counting bars vertically

indicates the size of the known population per month. The

largest population occurred on the SL grid in the spring of

1987 and 1988, when there were 8 individuals or 8 ani-

mals/ha. Excluding animals caught three times or less, mean

persistence time on this grid was 195.9 days for females and

137.2 days for males.

For 20 mice captured four or more times, I plotted the

capture locations and number of captures. To calculate the

home range area, I used the Exclusive Boundary Strip Method

(Stickel, 1954), in which I mapped the capture sites and

drew a boundary strip (equal to half the distance between

traps). I selected this approach because Stickel (1954)

compared the exclusive boundary strip, inclusive boundary









M J J A S O M A M J J S N

ACI

TN 600 200 300 300 300 300 300 300 300 300 300 300


(21)
(75)


(3)
(21)
(21)
(21)
(21)


M94
M95
M96*


F85*


ACII


S(1)


300 200 300 300


300 300 300 300 300


F88
M89*
M92


(33)


M86


Fig. 7. Trapping data for Podomys on the ACI, ACII, and SL
grids. ID indicates sex and identification
number; asterisks indicate animals first captured
as juveniles; TN indicates trapnights per month;
and persistence on grid in days is shown in
parentheses.


1987


1988







1988


A M J J A S O M A M J J A S N

SL

Tr300 600 100 300 300 200 300 300 300 300 300 300 300


F10
F18*_
F26*
F47*
F64


(36)
(449)
(517)
(364)
(24)
(91)
(148)


M11


M16
M19
M49
M54
M42
M98
M58


(381)
(289)
(24)
(1)
(1)
(74)
(2)
(1)
(76)
(44)
(74)


M102


(61)
(1)


Fig. 7-- continued


1987










strip, and minimum area methods of calculating home range

area and she determined that the minimum area method under-

estimated home range size by 64%, and the exclusive and

inclusive boundary methods overestimated home range size by

2%, and 17%, respectively.

On ACII only one individual (female 87 on Fig. 7) was

captured at least four times. However, her last two cap-

tures occurred across the grid from the first three, and

this pattern suggests dispersal rather than a large home

range, so I did not include her in the analysis. Home range

data are summarized on Tables 5 and 6. Three animals were

marked on the grids as juveniles (gray pelage) and died or

dispersed off the grid at approximately 60 days of age. Two

more animals (female 26 and male 54 at SL) marked as juven-

iles did not leave the grid. The male shifted his home

range, but I could not distinguish between juvenile and

adult home ranges for female 26, who subsequently appeared

to have an activity area larger than that of other females.

Seventeen adults trapped on the two grids (ACI and SL) had

an average home range of 804 m2. In general, males (mean =

910 m2) appeared to have larger home ranges than females

(mean = 683 m2), but the difference was not significant

(Mann-Whitney test at p = 0.01).

To determine whether ranges of females and males over-

lapped, I plotted home ranges of Podomys caught during May

1987 on ACI and May 1987, September 1987, and May 1988 on SL









TABLE 5. Home range data for 20 Podomys trapped on two
grids. Id is the identification number, N is the number of
captures, PG is the persistence on the grid in days, A is
the area of the home range in m2 for juveniles (A,) and
adults (AA).


Grid Id Sex NG PG A A


ACI 89 F 6 21 700

ACI 91 F 7 75 500

ACI 96 F 4 21 400

ACI 94 F 6 21 916

ACI 95 M 5 21 750

ACI 96 M 5 21 800

SL 61 F 5 24 1000

SL 10 F 6 42 500

SL 18 F 6 36 600

SL 26 F 17 449 1350

SL 47 F 12 517 300

SL 54 F 6 74 300

SL 42 M 8 289 1850

SL 49 M 10 91 1200

SL 54 M 16 381 1093 200

SL 16 M 9 148 800

SL 53 M 6 76 600

SL 56 M 7 44 500

SL 59 M 8 74 1000

SL 102 M 10 156 400







49

TABLE 6. Capture data for 36 Podomys at Anderson-Cue. ID
indicates sex and identification number; YR is year of first
capture; AGE is age at first capture (juvenile, subadult, or
adult); N is number of captures at grids (NG), burrows (N,),
and total (NT); P is persistence, in days, on grids (PG),
burrows (P,), and total (PT). Asterisks indicate trap
mortality.


ID YR AGE NB PB NG PG NT PT


F83 88 J 1 1 1 1

F84 88 J 3 3 3 3*

F85 88 J 1 1 1 1

F86 88 A 7 109 7 109

F87 88 SA 2 2 5 33 7 33

F88 87 A 1 1 3 3 4 9

F89 87 A 6 21 6 21

F90 87 SA 1 1 1 1

F91 87 SA 7 75 7 75

F92 87 A 1 1 1 1

F93 87 A 1 1 1 1

F94 87 A 7 164 1 1 8 164

F95 87 A 1 1 3 3 4 21

F96 87 A 1 1 4 21 5 40

F97 87 A 9 171 3 3 12 193

F98 87 J 1 1 1 1

F99 87 A 1 1 1 1

F100 87 SA 4 124 4 124









TABLE 6---continued


ID YR AGE N P, NG PG NT PT


M80

M84

M85

M86

M87

M40

M88

M89

M90

M92

M93

M94

M95

M96

M97

M98

M99

M100


SA

A

SA

SA

SA

A

J

J

J

A

SA

A

A

J

A

?

SA

A


216

39

101



2

1

1



1

48


2 106


150

1

1

1


216

39

101

1

2

1

1

1

1

48

5*

106

21

21

150

1

21

1







51

(Fig. 8), the three months during which populations appeared

to be the largest (Fig. 7). Generally, ranges of males and

females overlapped, but ranges of adult females did not

(except occasionally with juveniles or young adults), a

pattern which suggests intra-sexual territoriality (Kau-

fmann, 1983). A female juvenile (female 18) overlapped with

the ranges of an adult of each sex (Fig. 8A). Female 26,

the one with the unusually large range, overlapped somewhat

with female 61 in September 1987. The only males that

overlapped were three males at SL (Fig. 8C), two of which

possibly were siblings.



Trapping at Burrows

With the data from trapping at burrows, I estimated

home ranges by mapping movements of adults trapped more than

four times in a year on AC and SL, using the exclusive

boundary method (Stickel, 1954) as I did for grids. For

burrows on the perimeter of the study site, I plotted a

boundary of 10 m (which I considered a conservative estimate

of distance moved beyond the burrow). All capture points

were weighted equally, but in a few instances I omitted

points that suggested unusual forays outside the usual home

range of the animal. I then measured the home range area

with a Lasico Compensating Planimeter.

Seventy-five home ranges for 62 animals who fit the

above criteria, including three females on AC, are shown in












A






* U

U I


* ,\ S I


I i.


* LJ_ I Ir
S
*~ U, *


I,.


9I I~ -


S,


0i I
*
* .

-J


B rr


I- I L

L.


o *


& 0 0 a 0



* t


/ /'


*



L--J1


1- I
0 U I I 11 S


Fig. 8. Home ranges of Podomys trapped on the ACI grid in
May 1987 (A) (n = 6 mice, 600 trapnights) and on
the SL grid in May 1987 (b) (n = 8, 600 trap-
nights), May 1988 (c) (n = 9, 300 trapnights), and
September 1987 (D) (n = 7, 300 trapnights). Dots
represent the outer boundary of grid. Solid lines
represent home ranges of females, dotted lines
those of males.


IJi
*.'
U. _









Fig. 9. Home ranges for adult males frequently overlapped,

but those of females did not; two SL females who appear to

overlap in 1988 actually were separated temporally. This

separation of home ranges, similar to that observed on

grids, suggests intra-sexual territoriality. Animals who

survived more than one year (e.g., SL male 11 in 1985-87)

showed fidelity to the same general area every year. Home

ranges varied in size from 210 m2 (for SL female 26, cap-

tured at a single burrow 6 times in 1987) to 11007 m2 (for

SL male 11 in 1986). The average home range size for 35

females was 2601 m2 (SD = 1315.1) and for 40 males was 4042

m2 (SD = 2269.4); home range size of the two sexes differed

significantly (Mann-Whitney U = 2.95, p = 0.01). For the SL

data I also compared areas in 1983-85 with those of 1986-88

to see whether home range size changed after the prescribed

burn in 1985, but I found no significant difference.

Capture data for the 62 mice whose home ranges were

mapped are shown on Table 7. Females 86, 94, and 97 (1987)

were from AC; the remaining animals were on SL. Each animal

was trapped from 5 to 22 times during the year in question,

and almost all mice utilized more than one burrow (mean

number of burrows = 3.8). Estimates of persistence (P,)

were calculated by counting the number of days between first

and last captures spent by the animal in the area shown on

the figure. These estimates of persistence are not com-

parable across years because trapping effort varied each































Fig. 9. Annual home ranges of 62 Podomys on the SL
sandhill, 1983-1988. Dots represent tortoise
burrows where mice were trapped. Solid lines
indicate home ranges of females, dashed line for
males; a) 1983, 52 burrows; b) 1984, 61 burrows;
c) 1985, 89 burrows; d) 1986, 103 burrows;
e) 1987, 131 burrows; f) 1988, 136 burrows. Bar
represents 20 m.



























--


0
S


*


53f-
.
* i '.
o S
* I


33


N


29
( *

S



48

S \



27


Fig. 9a.


49


I


j


4* -




























27,
SI


*4
*


*\ \
2
* '
I




23


* S
0




0



/
/
0


,- 48


25


38


29
*


3 9 .. "
39
85


Fig. 9b.


43


3






34

















100


*0
U ~


23


/ .


0
'0
I
I


97


39



0 0



-- 1 U


*.

* */
* --

/ 98
// 98


/ *


100


---39


Fig. 9c.


* 6

















97 .---
.* 995






11 --
S. 76






S73



11'
9 / -
.. 7 .,,




-*--
/. *.. \ ,...

A, '.' / *


0 98
*l / t ##

0 .





/



9 *
S80 0


Fig. 9d.


















17

65 --'




S 's *,
I I



S 68
47
/ .

*/4



S 7-* S. 0j


0 /
0 r




**26 *
----)~ri ~, \rr-

-----


* I*
* I *




72 --*


S *

* 9 8

.-'

** e l

57 *J
*
* 0


Fig. 9e.


* S


* *
**


. 80l -































S S
0
0


37


* 36
-
\ I \


49


-- z -I~


* *


20


10


Fig. 9f.


34


37


/
I
I
I

0

6 K


.
* /


102 '







61

TABLE 7. Capture data for the 62 adults whose home ranges
are shown on Fig. 9. ID is the sex and identification
number of each mouse; YEAR is the year for which the home
range was calculated; N is the number of captures, P, is
persistence (in days); # is the number of burrows where
each animal was caught; and asterisks denote animals present
when trapping ceased in November 1988.


ID YEAR N, P, #B


F5 1983 5 93 2

F8 6 71 3

F14 6 71 2

F27 5 114 2

F49 5 117 4

M29 5 79 2

M33 5 79 2

M47 5 99 2

M48 5 100 1

M53 5 79 2

F23 1984 8 125 2

F29 11 139 5

F34 5 56 2

F38 5 49 2

F43 15 222 3

M25 6 135 3

M27 5 61 3

M28 5 63 2


171 5


M39









TABLE 7--continued


ID YEAR N3 Ps #3


M48

M85

F23

F39

F30

F97

F100

M11

M32

M39

M98

M100

F89

F93

F94

F95

M11

M72

M73

M76

M80

M88

M97


1985


211

99

212

84

183

185

265

253

183

265

265

100

49

36

103

107

253

68

79

133

103

259

101


1986










TABLE 7--continued


ID YEAR N, P, #


M98

F26

F47

F61

F68

F72

F86

F94

F97

M11

M16

M17

M42

M57

M65

M80

M98

F10

F26

F36

F37

F46

F47


1987


171

78

113

243

170

125

109

164

171

143

233

193

31

242

213

201

218

227*

193

94

121

151

193


1988









TABLE 7--continued


ID YEAR N, PB #B


F49

F52

M20

M22

M25

M34

M37

M42

M51

M102


148*

151

151

151

132

104

148*

198*

151

198*









year; nor do they represent longevity because captures of

juveniles and subadults (or captures as adults in years

other than those shown) are excluded. The P, values are

useful as minimal estimates of persistence on the sandhill.


Conclusions

Fluorescent powder was not a useful method of determin-

ing the home range of Podomvs on Ordway. Powder trails did

not indicate the complete area traveled by mice even during

one night. Although some data regarding substrate use or

social interactions were gathered, considerable time and

effort were required. The technique proved successful with

smaller Peromyscus (Lemen and Freeman, 1985, 1986), so I

suspect the failure at Ordway was not related to the body

size or biology of Podomys, but possibly to density or

structure of vegetation.

Trapping on grids provided estimates of home range size

and persistence, as well as demonstrating the exclusivity of

home ranges of adult females. There also was evidence that

males had larger home ranges that occasionally overlapped

with each other and with females. However, estimates of

home range areas obtained from trapping at burrows suggest

that estimates from grids were too small. Evidently the

"standard" mammal-trapping grid (10000 m2 with 10-m trap

intervals) was insufficient to measure home ranges of Podo-

mvs in sandhills. Doubling the trap interval to 20 m would







66

produce a 10 row by 10 column grid covering 40000 m2, which

would encompass mean home ranges estimated at burrows for

females (2601 m2) and males (4042 m2), might be more useful.

No fewer than 100 trap stations should be used. However,

one should keep in mind that increasing grid size on some

sandhills will spread the grid into neighboring habitats,

and that trap success on the grid will be smaller than on

the burrows.

Trapping at burrows produced larger estimates of home

range size than I expected, but the means are not unreason-

ably large when compared with other values reported for

Peromyscus. Stickel (1968) reported that home ranges of

Peromyscus varied from 0.1 to 10 acres. Hoffmeister (1981)

reported mean home ranges for female (8290 m2 +/- 2688) and

male (10465 m2 +/- 4043) P. truei in pinon-juniper habitat,

also using the exclusive boundary strip method. Although

the range of home areas of Podomys varies considerably,

further research might show that my means are under-

estimates, based as they are on many individuals located on

the edges of my study sites. I also have no measure of

vertical habitat use; for example; it would be interesting

to determine whether the mice occupy a larger home range

volume in years when acorns are produced.

In analyses of both burrow and grid data, I omitted

animals that were recaptured less than three times. A pos-

sible consequence of this approach is that several animals







67

might occupy the same space. However, the fact that both

sets of data show abutment of home ranges of adult females

seems good evidence of mutually exclusive home ranges, which

suggests that females might be territorial.













DIET


Field Observations

Almost nothing is known about the natural diet of

Podomys. Merriam (1890) reported an observation that mice

in the southeast ate the seeds of "scrub-palmettoes." Layne

(1970) and Humphrey et al. (1985) suggested that acorns

might be a major food source during mating years, based on

the association of P. floridanus with habitats with Q.

laevis and other oaks. Milstrey (1987) reported that captive

Podomys ate engorged soft ticks (Ornithodoros turicata

americanus) that parasitize gopher frogs and tortoises,

raising the possibility that these mice prey on other resi-

dents of Gopherus burrows. Other foods presumably include

insects, seeds, nuts, fungi, and other plant material

(Layne, 1978).

Five field observations at SL provided additional

dietary information. An adult male caught and released on

the grid (5 May 1988) readily ate a cricket (Gryllinae) I

offered to him. On 15 June 1987 at approximately 0245 EST,

an adult female (who had been trapped, powdered, and re-

leased) caught a small moth and consumed all but the wings.

On 9 May, an adult female just released from a trap ate a









young shoot of Smilax auriculatus. On 20 July 1987, at

burrow #15, I also observed an adult male take a fruit of

the flag pawpaw (Asimina incarna) (Jones, 1989). After

being released from the trap, this mouse climbed 30 cm up a

plant on the apron of the burrow, and knocked a ripe pawpaw

to the ground. He then ran down, picked up the pawpaw, and

carried it 11.4 m to burrow #71. In July 1988 another mouse

took a pawpaw, again at SL. A female released from a trap

at a burrow hid in a hole at the base of a turkey oak ap-

proximately 5 m NW of the burrow. In a few minutes she left

the hole, picked up a small pawpaw fruit (A. incarna), and

carried it back to the hole and nibbled on it.

Predation and dispersal of Asimina fruits have not been

well studied. Willson and Schemske (1980) reported opossums

and raccoons sampling unripe fruits of A. triloba in Il-

linois. Norman and Clayton (1986) suggested that small

mammals probably help disperse Asimina in Florida and re-

ported an observation of Gopherus eating ripe fruits. Human

consumption of A. incarna was reported by Bartram (1791).

Whether Podomys disperse these fruits, and whether pawpaws

are a significant source of food, remains unknown. Willson

and Schemske (1980) measured a caloric content of 4.5 kcal/g

for fruits of A. triloba in Illinois; fruits ranged in size

from 2.1 to 48 g dry weight. Fruit set for A. obovata, A.

Dvgmaea, and A. triloba is less than 10% (Willson and

Schemske, 1980; Norman and Clayton 1986). To the best of my









knowledge, nutritional values and fruit set have not been

reported for A. incarna, the most abundant pawpaw on Ordway.

However, pawpaws certainly are the largest fruits produced

in the high pine; surely they represent a significant addi-

tion to the summer diet of Podomys.



Laboratory Observations

The maintenance diet of captive animals consisted of

rodent chow (Wayne Rodent Blox) and water provided ad libi-

tem. This diet was supplemented with lettuce, carrots,

apples, strawberries, sunflower seeds, mixed bird seed,

oatmeal, rice, mealworms, crickets, and acorns (Q.

michauxi).

In an attempt to determine what mice might eat under

natural conditions, I occasionally offered animals fruits

and seeds collected at Ordway. All of these offerings were

eaten: acorns (Q. chapmanii, Q. geminata, Q. hemisphaerica,

Q. laevis, Q. myrtifolia, and Q. nira), pine seeds (P.

elliottii and P. palustris), blueberries (Vaccinium mvr-

sinites), deerberries (V. stamineum), gallberries (Ilex

qlabra), blackberries (Rubus arqutus), gopher apples (Lica-

nia michauxii), pawpaw fruits (A. incarna and A. yvcmaea),

flowers of Stillingia sylvatica, and seed pods of Crotalaria

rotundifolia and Galactia elliotti. Seeds and stems of

unidentified grass species were shredded and incorporated









into the cotton nesting material, and probably eaten as

well.

Predation on vertebrates appears limited to stressful

conditions. Eisenberg (unpubl. data) trapped an adult male

on SL in 1983 who detached the posterior part of a juvenile

red rat snake (Elaphe quttata) that was caught half way in

the trap. The mouse ate about 3.5 cm of the snake to the

bone. An adult male caught on AC in October 1987 ate the

viscera of a juvenile also caught in the trap. I also lost

a litter of three young in captivity on the only occasion

that I had a male and two pregnant females in the same

aquarium.

In general, feeding behavior follows that described for

Peromyscus by Eisenberg (1968), in which the animal crouches

and manipulates the food with its forepaws. Larger items

such as pawpaws are held against the substrate. Seed pods

were opened by grasping the pod vertically, resting one end

on the substrate, chewing off one end of the pod, and open-

ing the pod longitudinally along a suture. Larger foods,

such as turkey oak acorns, are dragged with the incisors;

smaller items are carried in the mouth. Some individuals

consistently took favored foods (such as acorns and sunflow-

er seeds) to the corners of the aquarium and covered them

with kitty litter; I considered this caching behavior. Food

items and remains also commonly are found underneath nests.









This ready acceptance of a wide variety of foods im-

plies that P. floridanus is an opportunistic feeder, like

Peromvscus. As King (1968) pointed out, we lack systematic

studies of taste discrimination and food preferences. For

Podomys we also need basic studies (stomach and fecal ana-

lyses) to learn more about the natural diet.



Acorn Preference Tests

I performed preference tests to determine whether

acorns of turkey oaks (Q. laevis) are preferred over acorns

of other species that are less abundant or produce fewer

acorns in the high pine. Such a preference might be a

partial explanation for the distribution of Podomys in

sandhills at Ordway.

For feeding tests, an individual was placed in an

aquarium with clean kitty litter, nesting material, water,

and rat chow at least 24 hours before beginning the test.

To start a trial the chow was removed and three bowls (each

containing five acorns of a single species) were added at

about 1900 hours. Each acorn was marked and weighed and I

measured its height and width with dial calipers. No effort

was made to ascertain that all acorns in a bowl were identi-

cal in size. Acorns that had weevil holes were not used,

and in a single trial all acorns either had caps or lacked

them. Approximately 12 hours later I removed the bowls,

acorns, and acorn fragments. I recorded whether acorns were










removed from bowls and whether they were opened and eaten,

opened and evidently not eaten, gnawed, or apparently un-

touched. I controlled for location effects by shifting the

relative positions of the three types of acorns in each

trial.

Two unexpected difficulties arose. The first was the

unpredictability of the acorn supply, which made it impos-

sible for me to run tests with identical species in 1988 and

1989. Additionally, the majority of acorns on the ground

already contained weevil larvae or were otherwise spoiled.

Secondly, I discovered that some of the acorns I tested were

spoiled (by insects or by mold) even though they lacked

holes. Possibly this spoilage is the reason that some hulls

were opened but the meat not removed. For this reason I

simplified analysis of results by examining whether acorns

were unopened or opened, regardless of whether any of the

meat appeared to have been removed. These data were then

ranked (rank 1 to the smallest number opened) and the Fried-

man test (Conover, 1980) was used to test the hypothesis

that species of acorns were opened in equal numbers.

In 1988, I concentrated on determining which acorns

were eaten by Podomys. I presented acorns from six species

of oaks (Chapman's (Q. chapmanii), live (Q. geminata),

turkey (Q. laevis), laurel (2. hemisphaerica), myrtle (Q.

mvrtifolia), and water oaks (Q. niara)) to four captive

animals. Five of these species belong to the red-oak group,









which generally contains three or four times more tannin

(Briggs and Smith, 1989) than species of the white-oak group

(which includes Q. chapmanii). Each mouse was presented

with turkey oak acorns in combination with acorns from two

other species; two to four trials were run per animal for a

total of 12 trials. Although the sample is inadequate for

statistical analysis, I noted that the mice opened acorns of

all species, and in all but one trial turkey oak acorns were

opened in the smallest numbers. I then planned to test

animals with a combination of acorns from Chapman's, live,

and turkey oaks, the most prevalent Quercus species on the

SL sandhill.

In 1989 Chapman's and live oak acorns were not avail-

able, so I gathered acorns from Q. laevis and two different

trees of Q. hemisphaerica, one from the old pasture near

Ross Lake Bridge and the second from a hammock past Ander-

son-Cue. I expected that Q. laevis, the predominant acorn-

producing species on SL, would be preferred. I tested nine

animals, two trials each. For analysis I used results of

the first trial only; there was no difference in ranks of

first and second trials for any of these mice. Results

indicated that acorns were not opened in equal numbers

(Friedman test, T = 26.75, significant at p = 0.01). When

the Friedman test indicates rejection of the null hypothe-

sis, multiple comparisons may be made to determine which

treatment is significantly different from the others







75

(Conover, 1980). For multiple comparisons at a significance

level of p = 0.01, acorns from Ross Lake (Q. hemisphaerica)

were opened significantly more often; differences between

the acorns from the hammock and from Q. laevis were not

significant. These results indicate not only that the mice

preferred laurel oak acorns over turkey oak acorns, but that

mice distinguished acorns from two different individuals of

Q. hemisphaerica.

The apparent preference for Q. hemisphaerica was sur-

prising in that its acorns contain more tannic acid than Q.

laevis (Harris and Skoog, 1980). The laurel acorns were

smaller than the turkey oaks', so possibly the preference is

due to differences in handling time. Another possibility is

selection according to chemical composition other than

tannin content. Halls (1977) reported that Q. laurifolia

(= Q. hemisphaerica) had higher fat and carbohydrate content

than Q. chapmanii and Q. incana; unfortunately, values for

Q. laevis were not listed. Possibly my results were some

artifact of using captive animals. However, in the only

published study of acorn preference in Peromyscus, Briggs

and Smith (1989) found that five P. leucopus captured in

habitats lacking oaks consumed equal amounts of acorns from

species of the red and white oak groups. Peromvscus caught

in areas containing oak trees selected acorns of Ouercus

species found in their habitat, independent of fat, protein,









and tannin content. These results lead me to believe that

there is some real difference between the acorns I used.

Except for smaller acorns that occasionally were split

in half, acorns were opened consistently at the hilum (basal

scar). Although acorns typically were carried by the point

at the distal end, the mice never chewed open the hull

there. For small, round acorns a neat incision was made

around the scar; on more elongate nuts (Q. geminata and 2.

laevis) the hull might be nibbled farther down the sides.

Caps, if present, were removed; there was no difference in

the numbers of capped and capless Q. laevis acorns opened.

Occasionally I found acorns opened in this manner on aprons

or near burrow entrances. Hulls of Q. laevis were found in

excavated burrows (Jones and Franz, in prep.) and E. Mil-

strey vacuumed remains of what appear to be Q. qeminata and

Q. laevis out of burrows on Roberts Ranch, Putnam County;

all of these hulls were opened in a manner consistent with

what I observed with captive animals. Small piles of emptied

hulls occasionally are found at the base of turkey oaks on

Ordway; however, it is possible that Glaucomys or birds

might make these piles. Of the 450 acorns presented in 30

trials to 13 mice, 66% were removed from bowls, whether

opened or not.

I avoided use of acorns that appeared spoiled because

mice did not appear to eat blackened nutmeats. I did not

test preferences of sound acorns vs. acorns with weevil









larvae. On one occasion an adult female immediately ate a

larva from an acorn opened but not eaten by another in-

dividual. Semel and Andersen (1988) suggested that such

differences in behavior might be due to mice being unable to

detect larvae in unopened acorns, or that individuals can

detect infestation but some individuals avoid infested nuts.

They thought that tooth marks on the hull and movement of

acorns might represent assessment.

Based on these results, I suggest that acorns of Q.

laevis are eaten but other species are preferred if avail-

able. Turkey oaks are an unreliable food supply. Umber

(1975) noted acorn production of Q. laevis was low and

variable on his study site in Citrus/Hernando counties, and

Kantola and Humphrey (in prep.) found that trees on low

slopes at Ordway produced significantly more acorns than

trees on higher slopes. Layne (in press) correlated the

relatively greater abundance of Podomys in scrub and scrubby

flatwoods with higher and more consistent acorn production

than in the high pine. Proximity to Q. geminata and other

acorn-producers might be a partial explanation for the

relatively higher and more constant population of Podomys at

SL than at AC and BP. If oak species in hammocks are more

reliable producers, the restriction of P. gossypinus and

Ochrotomys nuttali to hammocks at Ordway might be partly due

to the food supply. This leads to the hypothesis that

competition, possibly linked with other factors such as








78
predation, forces Podomys into the high pine when the other

rodents are present.



Supplemental Feeding Experiment

Are Podomys limited by food supplies? If food is a

limiting factor, mice should respond to extra food with an

increase in numbers, increased individual mass, and an

extended breeding season. A food limitation could partly

explain population sizes and distributions at Ordway. I

attempted an experiment of food supplementation using foods

favored by captive Podomys. The only supplementation ex-

periment that has been done with the Florida mouse was that

by Young (described in Young, 1983 and Young and Stout,

1986) on two grids in sand pine scrub in Orange County.

Other rodents responded to the additional food, but Podomys

rarely appeared on Young's grids and failed to establish a

permanent population during the experiment, although the

species had been abundant on both grids before the study

(Young, 1983). She concluded that Florida mouse populations

were limited by factors other than food.

My experimental design consisted of three grids, two on

the AC and one on the SL sandhill. Trapping procedure was

described under Home Range. The last grid (SLII) was set on

the SL sandhill west of Breezeway Pond. Prior trapping at

burrows in each of these areas indicated that mice were

present. In April, May, and June of 1988 I trapped SLI for








79

800 trapnights to determine how many nights of trapping were

needed to obtain all animals on the site. Consequently,

SLI, ACI, and ACII were trapped for three consecutive nights

per month for a total of 300 trapnights monthly. On ACI a

mixture of sunflower seeds, mixed bird seed, and oatmeal was

provided for one year in seven chick feeders fitted with

glass jars. To eliminate non-target species, plastic buck-

ets with two one-inch holes cut near the rims were upended

over the feeders and weighted down with a cinder block. I

checked feeders periodically and refilled as needed.

The results of this experiment are shown in Fig. 10.

Capture rates declined sharply on AC and were highly vari-

able on SL. I had intended to provide food on a second grid

(SLII, west of Breezeway Pond). However, 208 trapnights on

this grid and surrounding burrows in May, July, and August

resulted in the capture of one Podomys. The lack of success

on ACI led me to abandon SLII although I continued to pro-

vide food on ACI until the end of the 1988 trapping year.

On the AC sandhill I continued to trap mice in small numbers

at burrows west and east of ACI. Clearly, mice were not

obtaining the food or some other factor kept them out of the

grid area in spite of obtaining food.

Maximum trapping success for a single night was 9%, but

success was usually much lower. Additionally, I captured

non-target species, including Sceloporus undulatus, Coluber

constrictor, and Glaucomys volans. In spite of the low trap




















10 4
On V


0


0
to!




4-)


'-4)


W00
Ow

VI







400


r.4-4
0 0
-a'


U)
tt
ri











r-4



>


-440
411













U









.4 0
C;t
00


















V-4
l'b















r34
.r
r&4 t



















CC
2 oe
0c
O O 0



4~


-4
' o









U.


-*u C








0
-, 1

SO








Sr


0 N 0 N
,- 1'


T N 0







82

success I attempted to estimate the persistence (i.e., the

time between first and last captures) of mice on the three

grids (Table 8). Because these were not closed grids,

disappearance of an animal might be due to death, dispersal,

or to not entering traps.







83

TABLE 8. Persistence data for animals on grids. Data
presented are minimum number known alive (MNA), mean +/- SD
days spent on a grid (maximum values in parentheses), and
total numbers of trapnights (TN).


Grid MNA PERSIST TN


ACI 12 16.00 +/- 20.86 (75) 4200

ACII 6 7.17 +/- 12.68 (33) 3000

SLI 24 122.75 +/- 155.7 (517) 4100













ANALYSES OF VEGETATION


Sampling Technique

Can the local distribution of Podomys be predicted on

the basis of vegetation structure or species composition?

Vegetation was sampled in circular quadrats (radius = 10 m,

area = 314.159 m2) set with tortoise burrows as center

points. In the absence of large numbers of trees, circular

quadrats are laid out easily using a center pole at the

burrow and a rotating radius line; forestry flags were used

to delineate the quadrat. Quadrat size was based on prelim-

inary estimates of activity areas obtained by dusting Podo-

mys with fluorescent powder. Although one animal ranged at

least 28 m from his burrow in one night (see Home Range

above), I decided a radius of 10 m was sufficient to contain

significant numbers of plants and small enough to allow me

to feasibly obtain counts and identifications of plant

species. When I began these samples in 1985 I was uncertain

that burrows were distributed randomly, so point-quarter

sampling seemed inappropriate. Without knowing the shape of

Podomys activity areas, I felt that line-intercept sampling

and belt transects would be less representative of areas

encountered by mice than circular quadrats.








85
Sandhills on the Ordway Preserve have been subjected to

logging, grazing, suppression of fire, and prescribed burn-

ing on different rotations. Cessation of fire, or the

frequency of prescribed burns and other activities, affect

plant species composition and density on sandhills (e.g.,

Myers, 1985; Gates and Tanner, 1988). To determine how

burrow use changes with successional changes in vegetation,

I examined quadrats at SL in 1985, 1987, and 1988. SL has

been burned and the pines bear turpentine scars. Trapping

effort was equal among these burrows within each year but

not among years. Initially 29 quadrats were examined; this

number was reduced to 27 after a large oak fell over one

burrow and a clearcut was performed at a burrow under a

powerline. To see how Podomys behaved at burrows on a

sandhill where logging occurred, I set quadrats at 15 bur-

rows at BP in 1988.

I selected burrows so as to include burrows of all

three classes of tortoise activity (active, inactive, or

old, as defined by Auffenberg and Franz, 1982). Within each

quadrat, I counted stems of all woody species (Table 9) in

three size categories (tree = dbh > 7.6 cm, sapling = height

> 1 m and dbh < 7.6 cm, and seedlings = height < 1 m). I

noted the presence or absence of woody debris, including

fallen branches or trees, or rotting logs. I estimated the

per cent ground cover of bare ground, wiregrass, and gopher

apple using five categories (I = 0-10%, II = 10-25%,









TABLE 9. Habitat variables measured in each quadrat. Prime
signs indicate variables used in discriminant analyses.


Variable (abbreviation)


Location (LOC)'
Year (YR)'

Burrow (BUR)'

Orientation (ORI)'


Sand pine (SP)'
Slash pine (HP)
Longleaf pine (LP)'

Loblolly pine (YP)
Bluejack oaks (BO)'
Turkey oaks (TO)'
Laurel oaks (LA)
Post oaks (PO)
Live oaks (LO)'
Persimmon (PER)'
Pawpaws (PAW)'
Rosemary (RO)'
Sumac (SUM)'

Blueberry (VAC)'
Debris (DEB)'


Description


BP or SL sandhill
1985, 1987, or 1988
Identification number

Compass reading away from
entrance

Total stem count (P. clausa)

(P. elliotti)
(P. palustris)

(P. taeda)
(Q. incana)
(Q. laevis)

(Q. laurifolia)
(Q. stellata)
(Q. geminata)
(D. virginiana)
(Asimina spp.)
(C. ericoides)
(R. copallina)

(V. myrsinites)
Presence or absence of fallen

logs or branches in each

quarter of quadrat (4 indi-
cates presence in all 4 quar-

ters)










TABLE 9--continued


Variable (abbreviation)


Bare ground (GRO)'






Bare ground (ZGRO)'




Wiregrass (WIR)'






Wiregrass (ZWIR)'




Gopher apple (GA)'






Gopher apple (EGA)'




Legumes (LEG)'


Herbs (HERB)'


Description


Average of categorical es-

timates of % of bare ground

coverage in each quarter of

quadrat

Summation of categorical

estimates of bare ground

coverage per quadrat
Average of categorical es-

timates of % of coverage by
wiregrass of each quarter of

quadrat

Summation of categorical

estimates of wiregrass cover-

age per quadrat
Average of categorical es-
timates of % coverage by
Licania of each quarter of
quadrat

Summation of categorical esti-

mates of Licania cover per

quadrat
Total number of species of

Fabaceae
Total number of all non-woody

species







88

III = 25-50%, IV = 50-75%, V = 75-100%). I grouped species

of grasses into five categories (I = Andropoqon spp.;

II = Panicum hemitonum, III = Panicum (Dicanthelium) spp.,

IV = Paspalum setaceum, V = Gymnopogon spp.) and sedges into

two ("small", such as Bulbostylis warei, and "large", such

as Cyperus retrorsus), but did not identify each species. I

identified and counted the number of herbaceous species

(excluding grasses and sedges), but I did not count stems.

All plant species (Appendix) were identified using the plant

list compiled for the preserve by David W. Hall and Richard

Franz.

I plotted a species-area curve (cumulative number of

species vs. cumulative number of samples) for the 1985 data

from SL; the curve levelled off after the fifteenth sample,

and I concluded that the number and size of quadrats were

adequate. I believe smaller sample areas would be inap-

propriate for two reasons: Hansen (1964) noted that only 8

of 24 burrows he examined had turkey oaks and pines within a

radius of approximately 305 cm around the burrow entrance,

so larger areas are required to represent trees in the

vicinity. Secondly, some quadrats I examined were located

at burrows with large aprons, which are disturbed areas that

might be favored by certain animals such as Tantilla relic-

ta, Eumeces eqregius, and Neoseps reynoldsi (Campbell and

Christman, 1982); it is unknown whether these species affect

plant distribution on the apron. I suspect that burrow