TECHNICAL REPORT NO. 22

STATUS SURVEY OF FIVE FLORIDA MAMMALS
by
Stephen R. Humphrey,
Robert W. Repenning,
and Henry W. Setzer

Florida State Museum
University of Florida
Gainesville FL 32611

Supported by:

Jacksonville Endangered Species Office
U.S. Fish and Wildlife Service
2747 Art Museum Drive
Jacksonville, FL 32207

through the
Florida Cooperative Fish and Wildlife Research Unit
117 Newins-Ziegler Hall
University of Florida
Gainesville FL 32611

Research Work Order No. 16

February 1, 1986

TABLE OF CONTENTS

Introduction . . . . . .

Acknowledgments . . . . . .

Homosassa shrew, Sorex longirostris eionis . . .

Sherman's short-tailed shrew, Blarina carolinensis shermani

Pine Island rice rat, Oryzomys palustris planirostris .

Sanibel Island rice rat, Oryzomys palustris sanibeli .

Chadwick cotton mouse, Peromyscus gossypinus restrictus .

Conclusions . . . . . .

Literature Cited . . . . . .

2

3
. . 2
. . 3

. . 10

. . 15

. . 19

S . 28

. . 35

S . 38

STATUS SURVEY OF FIVE FLORIDA MAMMALS

Introduction

The purpose of this project was to determine the population status and

geographic distribution of five subspecies of mammals that occupy small areas

of coastal west-central Florida: Sherman's short-tailed shrew (Blarina

carolinensis shermani), the Homosassa shrew (Sorex longirostris eionis), the

Pine Island rice rat (Oryzomys palustris planirostris), the Sanibel Island

rice rat (Oryzomys palustris sanibeli), and the Chadwick cotton mouse

(Peromyscus gossypinus restrictus). A secondary goal was to evaluate the

subspecies' taxonomic status.

Concern about the future of these mammals is justified by their

restricted distributions, their genetic uniqueness, their habitat

requirements, and conversion of their habitats to human uses. Four of these

five subspecies are listed as Species of Special Concern by the State of

Florida (Florida Game and Fresh Water Fish Commission 1985); only the Pine

Island rice rat is not listed. All five subspecies are being considered for

possible listing under the federal Endangered Species Act (U.S. Fish and

Wildlife Service 1985). Additional research is needed to ascertain whether

the status of any of these taxa justifies listing and to provide biological

support for any management actions that might be proposed to assure the

subspecies' survival.

Acknowledgments

This study was made possible by the leadership of David Wesley and

Michael Bentzein of the U.S. Fish and Wildlife Service, the facilitation of

Franklin Percival of the Florida Cooperative Fish and Wildlife Research Unit,

and the cooperation of Don Wood of the Florida Game and Fresh Water Fish

Commission. We appreciate the interest and cooperation of numerous

individuals in the field, including Ralph Lloyd and Ron Hight of the J.N.

"Ding" Darling National Wildlife Refuge; Steve Phillips of the Sanibel-Captiva

Conservation Foundation; Ellison Hardee of the Department of Natural

Resources; Jeff Lincer of the Sarasota County Office of Environmental

Management; and J.P. Garner of the Homosassa Springs Attraction.

Homosassa shrew, Sorex longirostris eionis

Methods.--The type locality of the Homosassa shrew was identified from

the published description and verified by conversation with a long-time

employee of the Homosassa Springs tourist attraction. Sampling was done in

forest along the Homosassa and Crystal rivers (Figs. 1, 2). The latter site

was on the Williams Tract--State land managed by the Department of Natural

Resources, in habitat similar to the type locality. Sampling was done

continuously with five 25-m-long metal drift fences at each site, beginning

variously from 18 December to 9 March and checked through April 14. Drift

fences were buried several cm deep in the soil; to trap animals, a screen-wire

funnel trap was placed at each end and two pitfall traps (perforated coffee

cans) were buried flush with the surface along the middle. Additional

sampling was done intermittently at both sites with Sherman livetraps placed

in grids and lines from 18 December 1984 to 24 January 1985, for a total of

880 adjusted trapnights. Adjusted trapnights were computed as total

trapnights minus one-half the number of traps found sprung in the morning,

under the assumption that the average trap was sprung midway through the night.

Results.--The drift fences produced three Homosassa shrews, one from each

of three fences near the type locality (Table 1). None were found in the

Crystal River site. Twelve Blarina carolinensis also were captured, some from

each site. Shrews were captured only immediately after rain, presumably

because saturation of the soil forced shrew activity to the surface.

Therefore it is possible that capture success might have been higher during

the summer rainy season. The livetrapping produced 22 Peromyscus gossypinus,

2 Ochrotomys nuttalli, and 2 Sigmodon hispidus.

Taxonomy.--The Homosassa shrew was described on the basis of 10 specimens

Citrus County,
Florida

0 1o 2e
Kilometers

Fig. 1. Distribution of trapping sites at Homosassa Springs.

Crystal River, Florida

O 1 2
Kilometers

Fig. 2. Distribution of trapping sites at Crystal River.

Table 1. Shrews captured in hydric hammock near Homosassa
Springs and Crystal River, 18 December 1984 to 14 April 1985.

Fence Sorex Blarina Cryptotis Fence
Site nights longirostris carolinensis parva success

Homosassa Springs

0.025
0.008
0.043
0.020
0.010

Fence A
Fence B
Fence C
Fence D
Fence E

Crystal River

Fence F
Fence G
Fence H
Fence I
Fence J

0.000
0.024
0.000
0.027
0.027

Total 772 3 11 1 0.019

3 11

1 0.019

Total 772

7

taken in the swamp forest immediately adjacent to Homosassa Springs, Citrus

County (Davis 1957). At-the time, only 6 specimens of the southeastern shrew

from other areas of Florida were available to the author for comparison. We

examined 11 specimens of S. 1. eionis and 12 of S. 1. longirostris from

Florida. In our opinion, differentiation of the Homosassa subspecies is

valid. It has slightly larger teeth and the supporting bone structure is more

robust.

Survival Status.--The Homosassa shrew persists at the type locality.

Though considerable sampling effort was required to confirm its presence, the

data are inadequate to judge abundance of the population.

Distribution and Habitat.--The type locality is a very small circular

area contained within gravel walkways near the old entrance to the tourist

attraction. The new specimens reported here were taken near the perimeter of

the parcel of land containing the tourist attraction. The habitat of this

parcel, hydric hammock, resembles its original condition except that some of

it has been destroyed (Fig. 3). The type locality is now owned and operated

as a tourist attraction by Citrus County. The site was added to the CARL

(Conservation and Recreational Lands) acquisition list by the Florida Cabinet

in April 1985. The city and vicinity of Homosassa Springs have been developed

by conversion of hydric hammock, and this process is continuing, but

significant amounts of undeveloped forest remain.

A continuous tract of hydric hammock occurs along the Gulf coast from the

southern boundary of Pasco County north almost to Crystal River. After a

small discontinuity of slash pine flatwoods, the hydric hammock continues

north to Gulf Hammock and west nearly to Cedar Key. However, most of this

northern area was cleared and converted to pine plantation in the 1960's.

Fig. 3. ASCS aerial photos (upper, DCP-16C-58, November 1944; lower, A20
18017-374-75, April 1974) of Homosassa Springs. The town is on the right, and
the tourist attraction is on the left, where the road bends around the two
spring runs. The obvious habitats are hydric hammock on the left and slash
pine flatwoods on the right.

Like most terrestrial mammals, Sorex longirostris is not particularly

habitat-specific. The most detailed study on the ecology of this species,

conducted in the Dismal Swamp of Virginia, showed that the shrews occurred in

any type of forested wetland and avoided only uplands and marshes (Robert

Rose, unpublished). Comparably broad habitat use in Florida could mean that

the distribution of the Homosassa shrew might be much larger than documented.

If so, it might be separated from the range of S. 1. longirostris only by

upland topographic features like the Archer and Brooksville ridges.

Management.--Protection of habitat clearly is desirable. The obvious

target of such protection at present is the tourist-attraction parcel.

Secondary targets could be inferred but not supported by scientific evidence.

Research Needs.--Much more field work must be done to determine whether

the geographic range of this subspecies is as small as documented, or as large

as possible. Depending on the answer, the current rate of habitat loss would

be judged either alarming or not.

Sherman's short-tailed shrew, Blarina carolinensis shermani

Methods.--Because two subspecies were targeted in the Fort Myers area,

sampling was done in a variety of habitats ranging from marsh to pine

savanna. The vicinity of the type locality has undergone extensive suburban

and urban development, and we were unable to pinpoint its exact location. We

sampled several remnants of undeveloped oak woodland and drainage ditches

within one-half mile of the published coordinates, as well as similar habitats

far to the east and west (Fig. 4). Sherman livetraps were placed in lines or

grids with an inter-trap distance of 10 m. Rolled oats was used as bait.

Results.--Field work in North Fort Myers and Cape Coral failed to produce

any new specimens of Sherman's short-tailed shrew (Table 2).

Taxonomy.--Sherman's short-tailed shrew was described on the basis of 27

specimens (Hamilton 1955a). Because the original description made detailed

comparisons with the other Florida subspecies of short-tailed shrew, and

because few museum specimens of shermani have been collected since then, a

taxonomic review at this time would yield no new insights on the genetic

uniqueness of this population.

Survival Status.--Although we could not confirm the presence of the

population in the vicinity of the type locality, two specimens were collected

9 miles east of Fort Myers by staff of the Florida State Museum in 1981.

Therefore the population appears to survive, but there is no basis to judge

its status.

Distribution and Habitat.--Sherman's short-tailed shrew was described on

the basis of 27 specimens taken from a single locality near Fort Myers, Lee

County (Hamilton 1955a). The site was 1 mile north of the Edison Bridge

spanning the Caloosahatchee River and 0.25 mile east of U.S. Highway 41 (see

11

co

W
0

z

e-
LL

*0

SCO
z LL

Fig. 4. Distribution of trapping sites at North Fort Myers, Cape Coral,

Little Pine Island, and Pine Island.
Little Pine Island, and Pine Island.

Table 2. Small mammals captured on Pine Island, Little Pine Island, and the
adjacent mainland, 6 December 1984 to 18 January 1985.

Adjusted
trap Sigmodon Oryzomys Peromyscus Blarina Trap
Site nights hispidus palustris gossypinus carolinensis success

Pine Island

0.034
0.094

Grid A
Grid B

Little Pine Island

Grid C
Grid D
Line E

127.5
111
32

0.031
0.162
0.438

mainland (Cape Coral

Grid F
Line G
Grid H
Grid I
Line J
Line K
Grid L
Line M
Line N
Line 0

and North Fort Myers)

124
112
127
127.5
19.5
111.5
63.5
18
67
41.5

0.008
0.098
0.008
0.071
0.154
0.009
0.047
0.000
0.194
0.120

62 30 5

Total 1298

0 0.075

Fig. 5). Shrews were caught in drainage ditches and in tunnels made by

moles. Hamilton's sampling totaled 4,000 trapnights. Judging from the

unsuccessful attempts to collect more specimens at the type locality in this

study and in 1956 (Layne 1978), habitat conversion may be severe enough at the

type locality to have extirpated the population there. Collection of two

specimens 9 miles east of Ft. Myers in 1981 demonstrates that the subspecies

is distributed beyond the type locality, but no insight is available on the

extent of this range.

Like most terrestrial mammals, the short-tailed shrew has broad habitat

tolerances. In Florida it is known from several types of closed-canopy

forest, from hydric hammock upslope to mesic hammock.

Management.--Not enough information exists to know whether management is

needed nor what actions might help Sherman's short-tailed shrew.

Research Needs.--Much more field work must be done to document the

geographic range of this subspecies.

Fig. 5. ASCS aerial photo (DCT-1LL-152, February 1970) of North Fort Myers,
including sample sites H, I, and J. The Edison bridge and the older route of
U.S. Highway 41 are on the right.

Pine Island rice rat, Oryzomys palustris planirostris

Methods.--Sampling was in marshes and drainage ditches. The type

locality on Little Pine Island was near our grids C and D (Fig. 4). Because

we were unable to pinpoint the mainland site, we sampled several drainage

ditches within one-half mile of its coordinates (published in Hamilton 1955a

rather than in the type description [Hamilton 1955b]), as well as similar

habitats and marshland far to the east and west. Sherman livetraps were

placed in lines or grids with an inter-trap distance of 10 m. Rolled oats was

used as bait.

Results.--We captured 28 rice rats on Little Pine Island and 8 on the

mainland (Table 2, Fig. 4).

Taxonomy.--The Pine Island rice rat was described by Hamilton (1955b).

Though the locality for 12 of his specimens is given as Pine Island, Lee

County, details of the description actually refer to Little Pine Island.

Another specimen taken on the mainland, 2 miles north of Fort Myers, was

included in the type series. This treatment of the mainland specimen suggests

a wider distribution than implied by the common name of the subspecies. The

Little Pine Island specimens were captured in a garbage dump and an adjoining

Spartina patens marsh. Persistence of the insular population was confirmed by

staff of the Florida State Museum in 1981, when 3 specimens were captured.

We examined 15 male specimens of 0. p. planirostris and compared them

with 135 males of other subspecies from Florida and southern Georgia. Females

also were examined, but they were not used for statistical comparison because

of greater variation in body size than males. Canonical discriminant analysis

showed that planirostris is distinct from the mainland subspecies (0. p.

coloratus) but not from the Sanibel Island subspecies (0. p. sanibeli). In

the original description (Hamilton 1955b), the two island forms were

distinguished only by color. Color is seldom adequate alone as a basis for

distinguishing mammal taxa, because pelage color is highly variable and

affected strongly by chemicals and light in the environment, and by fumigants

in museum cabinets. The two island populations of rice rats live in

contrasting environments--freshwater marshes on Sanibel Island and

brackish-to-salt water marshes in the Pine Island area. Thus the two island

populations should be treated as one. The Gulf islands rice rat is slightly

smaller than the mainland coloratus and has a flat rostral profile rather than

a convex one. Although the original description (Hamilton 1955b) includes

mainland animals and hence indicates a non-insular distribution for the Pine

Island population, our analysis shows that the mainland population is

referable to 0. P. coloratus.

Survival Status.--This population is abundant in appropriate habitat.

Distribution and Habitat.--The entire center of Little Pine Island is a

mosaic of marsh and salt flats (see Fig. 6) that is suitable habitat for rice

rats. Presumably the entire area is occupied. In contrast, no Pine Island

rice rats have been collected on Pine Island proper. On Pine Island, suitable

habitat is a relatively narrow strip of marsh and salt flats between the pine

savanna and mangrove (see Fig. 6). Before it was altered by mosquito ditches,

this marsh was maintained by the combination of high tides in spring, summer

rains, and high tides again in autumn. This long perimeter strip is

inaccessible, and it has been inadequately sampled for small mammals by both

Hamilton (1955b) and this study.

Management.--Little Pine Island is owned by the state of Florida and

managed by the Department of Natural Resources as Charlotte Harbor State

Fig. 6. ASCS aerial photo (DCT-1LL-64, February 1970) of Little Pine Island
(right) and central Pine Island, including sample sites B, C, D, and E.

19 -- -

18

Reserve. The marshland of Little Pine Island is being invaded extensively by

paperbark tree (Melaleuca quinquenervia). Perhaps this is a result of

hydrologic changes resulting from mosquito ditches dug in the 1960's. Unless

this plant succession can be reversed, habitat quality for rice rats on Little

Pine Island may diminish considerably in the future. Much of the marsh

perimeter of Pine Island was ditched in the 1960's, and the extent of marsh

habitat was reduced by both direct drainage and subsequent invasion of woody

plants. The ditches themselves are now occupied by white mangrove

(Laguncularia racemosa), and the spoil banks are covered by Australian pine

(Casuarina equisetifolia). Probably habitat could be improved on both islands

by filling the existing mosquito ditches.

Research Needs.--The primary research needs are sampling adequate to

determine if rice rats live on Pine Island and finding a habitat manipulation

that will reverse the woody succession on Little Pine Island.

Sanibel Island rice rat, Oryzomys palustris sanibeli

Methods.--Sampling was done at a series of sites extending along the

interior marshland of Sanibel Island (Figs. 7, 8). Sites were selected

according to accessibility, habitat character, and property ownership.

Sherman livetraps were placed in lines or grids with an inter-trap distance of

10 m. Rolled oats was used as bait.

Results.--The Sanibel Island rice rat was widespread in the interior

wetlands of Sanibel Island (Table 3).

Survival Status.--This population is abundant in appropriate habitat.

However, the absolute amount of habitat has been reduced by drainage and

development.

Taxonomy.--The Sanibel Island rice rat was described on the basis of 11

specimens captured 4 miles west of the lighthouse on Sanibel Island (Hamilton

1955b), at a site now known as the Bailey Tract of the J. N. "Ding" Darling

National Wildlife Refuge (see Figs. 8, 9). All were taken in freshwater

marshes dominated by cattails.

We examined 7 male specimens of 0. E. sanibeli (plus 1 from Captiva

Island and presumably referable to this subspecies) and compared them with 142

males of other subspecies from Florida and southern Georgia. Females also

were examined, but they were not used for statistical comparison because of

greater variation in body size than males. Canonical discriminant analysis

showed that sanibeli is distinct from the mainland subspecies (0. E.

coloratus) but not from the Pine Island subspecies (0. p. planirostris). The

Captiva Island specimen was not distinct from sanibeli. In the original

description (Hamilton 1955b), the island forms were distinguished only by

color. Color is seldom adequate alone as a basis for distinguishing mammal

Fig. 7. Distribution of trapping sites on Sanibel Island.

Fig. 8. Detail of trapping sites on the Bailey Tract, Ding Darling National
Wildlife Refuge.

Table 3. Rodents captured on Sanibel Island, 29 October to 14 November 1984.

Adjusted
trap Sigmodon Oryzomys Mus Rattus Trap
Site nights hispidus palustris musculus rattus success

Line A 20.5 1 0 2 0 0.146
Line B 107 2 3 0 0 0.046
Line C 20 0 1 1 0 0.100
Grid D 24 0 2 3 1 0.250
Grid E 247.5 1 5 15 1 0.089
Grid F 96 2 0 10 0 0.125
Grid G 187.5 1 5 3 0 0.048
Grid H 47 0 0 2 0 0.043
Line I 21 0 0 0 0 0.000
Line J 24 1 0 0 0 0.042
Line K 21.5 1 0 0 0 0.047
Line L 24 0 0 0 0 0.000
Grid M 174.5 7 0 5 0 0.069
Grid N 60 3 0 1 0 0.067
Grid 0 185.5 8 4 13 4 0.156
Grid P 190.5 3 6 17 2 0.147
Line Q 22 0 1 2 0 0.136
Line R 56 0 0 0 0 0.000
Line S 19 3 2 0 0 0.263
Line T 30 3 2 2 0 0.233
Line U 87.5 5 2 6 0 0.149
Line V 32 0 0 >0* 0 0.000

Total 1697 41 33 82 8 0.097

* Present, recording error.

Fig. 9. ASCS aerial photo (DCT-ILL-70, February 1970) of central Sanibel
Island. The wildlife drive on the major dike through Ding Darling National
Wildlife Refuge is prominent in the upper left. The Bailey Tract is just
south of the heavily subdivided area at right. Just west of this area, the
channalized Sanibel Slough bisects the island, and a network of mosquito
ditches with adjacent spoil piles is evident. Marshlands are the lighter gray
areas between ridges of coastal hammock, which appear as parallel striations.

taxa, because pelage color is highly variable and affected strongly by

chemicals and light in the environment. The two island populations live in

contrasting environments--freshwater marshes on Sanibel Island and

brackish-to-salt water marshes in the Pine Island area. The pelage of

specimens we examined varied from silver to brown. Thus the two island

populations should be treated as one. The Gulf islands rice rat is slightly

smaller than coloratus and has a flat rostral profile rather than a convex one.

Distribution and Habitat.--This population is widely distributed on

Sanibel Island. In 1960, staff of the Florida State Musuem collected two rice

rats on Captiva Island, 1.5 and 2 miles north of the Sanibel-Captiva bridge,

in coastal hammock on the Gulf shore.

The current distribution on Sanibel Island includes a portion of the Ding

Darling National Wildlife Refuge. Other portions of the interior wetlands are

owned by private conservation organizations, including the Sanibel-Captiva

Conservation Foundation, The Nature Conservancy, and the Center for

Rehabilitation of Wildlife. However, the majority of the habitat is in other

private ownerships, and it could be expected to be developed for human

residences as time passes.

Sanibel Island is unique among barrier islands of the Gulf coast of the

United States in containing a meandering slough (the Sanibel "River", see Fig.

9), which drains the interior during high-water periods. Except for low,

wooded ridges, this interior basin supported marsh vegetation (Fig. 10)

dominated by cordgrass (Spartina patens) and sawgrass (Cladium jamaicensis).

This habitat was maintained by yearly flooding of rainwater and irregular

invasion by salt water from either extremely high tides or storm surges.

However, the character of this marshland has radically changed in the last 25

Fig. 10. Control-burned cordgrass marsh in the interior of Sanibel Island, on
property of the Sanibel-Captiva Conservation Foundation, near sample sites
G-L. This is prime habitat for the Gulf islands rice rat.

Fig. 11. Former marshland along the channalized Sanibel Slough, severely
impacted by altered hydroperiod. Whatever the climax association of the
obvious hardwood succession, this is no longer good habitat for rice rats.

years, as a result of three factors (U.S. Fish and Wildlife Service 1984).

First, in 1951, local residents plugged the outlet of the Sanibel Slough on

the east end of the island, impounding water in the eastern end of the basin

and isolating the impounded area from new inundations of salt water. Second,

in the 1950's and early 1960's, the Lee County Mosquito Control program

channalized the Sanibel River and built a system of ditches (see Fig. 9) and

stop-log water control structures that reduced the water table and completed

the impoundment of the interior wetlands. Third, the dewatered wetlands were

invaded by a combination of salt-intolerant native plants and aggressive

exotic plants, to form cattail marshes in the wettest places and dense

woodlands of buttonwood (Conocarpus erectus) and Brazilian pepper (Schinus

terebinthifolius) in the drier areas (Fig. 11). This habitat change has

drastically reduced the extent of wetlands and its carrying capacity for

wetlands wildlife.

Management.--A significant conservation action taken by the Sanibel City

Council on 25 July 1984 was establishment of the Interior Wetlands

Conservation District as a section of the Sanibel Comprehensive Land Use

Plan. This District is a 3,000-acre tract within the island's interior

wetlands. The legislation prohibits development within 200 feet of either

bank of the Sanibel River, limits the intensity of use of areas less than 3

feet in elevation, and provides guidelines for reducing environmental damage

where development is permitted. However, the area in which development would

be precluded is a narrow strip within the District. In 1985, the City of

Sanibel was moving to acquire some parcels in this zone.

If the original flora and fauna are to be retained, the interior wetlands

need to be burned to kill woody plants and flooded to favor the marsh

association. Although it can be debated whether fires caused by natural

agents were normal events, now they are the only means of setting back plant

succession that can be applied over large areas at low cost. Flooding should

simulate both the historic water levels and the fluctuating hydroperiod of the

slough. Human needs for water may preclude use of groundwater for this

purpose, but the alternative of actively restoring some of the pre-existing

drainage pattern by filling mosquito ditches should be carefully studied.

Burning, spraying with herbicide, and hand-removal of exotic woody plants on

Ding Darling NWR, on the Bailey Tract and at Grid E north of the highway

S-curve, have improved the habitat there for rice rats. The most recent

efforts suggest that burning alone may be the best approach. Similarly,

marshland at the Sanibel-Captiva Conservation Foundation has been burned with

good results.

Research Needs.--Before the highly modified marshlands of these islands

can be managed with known consequences to rice rats, experiments need to be

conducted to learn how their populations respond to habitat manipulation.

More field work is needed to determine if Gulf islands rice rats occur on

other islands around Charlotte Harbor, such as on North Captiva, Cayo Costa,

and Gasparilla islands.

Chadwick cotton mouse, Peromyscus gossypinus restrictus

Methods.--Sampling for the Chadwick cotton mouse was conducted on

Manasota Key in Sarasota and Charlotte Counties during 1-19 October 1985 and

16-23 March 1985. Sherman livetraps were placed in lines or grids with an

inter-trap distance of 10 m. Rolled oats was used as bait. Seven sites (Fig.

12) were sampled along the Key from a city park in Venice (Grid A) south to

Englewood Beach (Grid K). Other public lands sampled were Casperson Beach

County Park (Grid B) and Blind Pass Beach County Park (Grid I). Three

mainland sites (Grids C-E) also were trapped to obtain comparative data.

Results.--During the first sampling period, a total of 980.5 adjusted

trapnights on the Key resulted in no captures of Peromyscus gossypinus, though

three species of non-target rodents were captured (Table 4). On the mainland,

two of the three sites yielded P. gossypinus. During the second sampling

period, an additional 741 trapnights on the Key, distributed among Grids A, I,

and J, resulted in capture of only a single spotted skunk (Spilogale putorius).

Taxonomy.--The Chadwick cotton mouse was described on the basis of 15

specimens taken at Chadwick Beach, near Englewood, Sarasota County (Howell

1939). They occurred in the sea oats along the beach and in the adjacent

cabbage palm forest. No specimens of the Chadwick cotton mouse appear to have

been collected since the type series was taken. However, a taxonomic review

of this population might be interesting in view of the qualitative nature of

the comparisons in the type description. Because so few specimens of this

subspecies are available, such a review would require a major statistical

comparison with other subspecies in the region. In view of the apparent

extinction of the taxon, such an exercise would be strictly academic.

Survival Status.--The Chadwick cotton mouse appears to be extinct.

Manasota Key, Florida

0 1 2 34 5
I Kilometers
Kilometers

Fig. 12. Distribution of trapping sites on Manasota Key.

Table 4. Rodents captured on Manasota Key and the adjacent mainland, 1-19
October 1984 and 16-23 March 1985.

Adjusted
trap Mus Rattus Sigmodon Oryzomys Peromyscus Spilogale Trap
Grid nights musculus rattus hispidus palustris gossypinus putorius success

A

B

C

D

F

F

G

H

I

J

K

Total

127.0

121.0

90.5

45.5

122.0

121.5

171.5

120.0

318.5

638.0

104.0

1979.5

0.031

0.050

0.033

0.286

0.123

0.000

0.000

0.050

0.009

0.000

0.029

0.027

Distribution and Habitat.--The place name "Chadwick Beach" no longer

occurs on maps, but conversation with local residents and real estate

businessmen revealed its location to be near the southern end of the Key in

Englewood Beach. Manasota Key is a peninsula, not an island. However, the

primary habitat of this subspecies is essentially insular, because the coastal

forest is very narrow at the neck of the peninsula. The maritime forest is a

very narrow strip at the northern end of Manasota Key, becoming wider

southward. The forest edge is affected by the harsh coastal environment and

is composed only of cabbage palms, which are relatively tolerant of salt and

wind. At Grid A, only this narrow edge occurs. Where the forest widens

farther south, its interior reflects a successional history and is dominated

by live oaks (Quercus virginiana) and southern red cedar (Juniperus

silvicola). The greatest tree-species and structural diversity now occurs at

the southern end of Sarasota County, at and near Grid J (Fig. 13).

Cotton mice prefer closed-canopy forest, but as noted in the type

description of P. g. restrictus, cotton mice also occupy grassland such as sea

oats adjacent to coastal forest (Fig. 14). The sea oats strip along the coast

of Manasota Key has been reduced to a few remnants, mostly in publicly owned

parks dedicated to beach recreation. Landward, numerous remnants of the

forest remain, divided into many, small, privately owned parcels developed for

residences and vacation homes (Fig. 15). Some of the larger lots in Sarasota

County were developed by clearing vegetation for a house site but leaving

forest to buffer neighboring properties. The southernmost portion, in

Charlotte County, has been completely destroyed by human construction. In

view of the amount of forest remaining in Sarasota County, we were surprised

that no cotton mice appear to remain on Manasota Key. As has been suggested

Fig. 13. Coastal hammock
is most extensive.

on Manasota Key, at sample site J, where the forest

Fig. 14. Sea oats grassland seaward
Blind Pass County Park, Manasota Key.

of the coastal hammock at sample site I,

Fig. 15. ASCS aerial photos of Manasota Key. Upper row, Sarasota County
between sample sites H and I. Lower row, Charlotte County between sample
sites J and K. Left column, March 1957. Right column, February 1970. Photo
numbers, clockwise from upper left: DEW-2T-182, A20 12115 274-261, A20 12115
274-194, DEW-2T-174.

34

for other mouse populations on heavily settled coastlines, predation by the

large number of continuously distributed house cats (Felis cattus) associated

with the human residences may render the remaining woodlots uninhabitable by

cotton mice. It is conceivable but unlikely that a remnant population of

Chadwick cotton mice persists somewhere on Manasota Key.

Research Needs.--The only imaginable need for further research is

additional survey work in the hope of reversing our conclusion.

Conclusions

The conclusions of this study are summarized as follows:

Taxonomic Survival
Subspecies status status Distribution Habitat

Homosassa shrew Confirmed Not threatened Possibly insular Forested
but extensive wetlands

Sherman's short- Valid Not ascertained Extent unknown, Mesic to
tailed shrew may be declining hydric forest

Pine Island rice =sanibeli Not threatened Gulf islands Marsh
rat

Sanibel Island =planirostris Threatened Gulf islands, Marsh
rice rat declining

Chadwick cotton Moot Extinct Formerly insular Closed-canopy
mouse forest

The Homosassa shrew persists in the type locality but was not found in a

nearby tract of similar habitat. The immediate vicinity of the type locality

is buffered from development pressures by public ownership. The most

important information need is to extend the field search for other

populations. Although this subspecies is known only from a single site, the

species has broad habitat tolerances, so possibly its distribution could be

quite large. The subspecies is a valid taxon.

Sherman's short-tailed shrew was not located in the fragments of habitat

remaining around the type locality, so its distribution may be declining. The

area around the type locality has changed to suburban and urban habitat, and

further development of an infilling nature is continuing there. This shrew

was recorded east of Fort Myers in 1981, confirming survival of the population

and raising the possibility that its distribution could be much larger than

documented. The taxon appears to be valid, and few new specimens have been

taken since the original description. The existing information on the

subspecies' distribution and population status remains inadequate.

The Pine Island and Sanibel Island rice rats are distinctive from the

mainland form but not from each other, so they should be pooled as a single

subspecies. It also did and still may occur on Captiva Island. More field

work is needed to document the full distribution of this population, which may

include all islands with marsh or salt-flat habitat in the Charlotte Harbor

area. Probably the population should be termed the "Gulf islands rice rat",

and its survival status should be considered from this larger perspective.

Experiments are needed to determine how population densities respond to

habitat manipulation, and ways need to be found to reverse the successional

impacts of historical mosquito-ditching, channelizing, and diking of marshland.

The population on Pine Island is abundant in the marshes and salt flats of

Little Pine Island. This land is under State ownership, but the habitat is

rapidly being degraded by invasion of Melaleuca. Management action will be

required if this succession is to be reversed. Rice rats have never been

recorded on Pine Island. The potential habitat there has been severely

impacted by mosquito-ditching, and it has not been adequately sampled for rice

rats.

Rice rats remain abundant in the interior wetlands of Sanibel Island, but

this habitat has been severely impacted by drainage and plant succession.

Carrying capacity and probably distribution have been reduced considerably.

Portions of the population occur in Ding Darling National Wildlife Refuge and

on private land dedicated to conservation, so the potential exists to sustain

this segment of the population in perpetuity with appropriate management

programs. These should focus on burning to kill woody invaders and flooding

37

to restore historic hydrologic conditions.

The Chadwick cotton mouse appears to be extinct. Its taxonomic status

appears to be valid, but the point is moot. Its distribution was insular, in

the coastal forest and adjacent grasslands of Manasota Key. Its habitat has

been heavily impacted by development of human residences and thoroughly

invaded by house cats, which may make the remaining woodlots uninhabitable for

cotton mice. It is unlikely that further searches will reveal a remnant

population of this mouse.

Literature Cited

Davis, J.A., Jr. 1957. A new shrew (Sorex) from Florida. Amer. Mus.

Novitates 1844:1-9.

Florida Game and Fresh Water Fish Commission. 1985. Official lists of

endangered and potentially endangered fauna and flora in Florida.

Tallahassee. 24 p.

Hamilton, W.J., Jr. 1955a. A new subspecies of Blarina brevicauda from

Florida. Proc. Biol. Soc. Washington 68:37-40.

Hamilton, W.J., Jr. 1955b. Two new rice rats (Genus Oryzomys) from Florida.

Proc. Biol. Soc. Washington 68:83-86.

Howell, A.H. 1939. Descriptions of five new mammals from Florida. J. Mammal.

20:363-365.

Layne, J.N. 1978. Sherman's short-tailed shrew. Pp. 42-43 in J.N. Layne

(ed.), Rare and endangered biota of Florida. Vol. I. Mammals. Univ.

Presses of Florida. 52 p.

U.S. Fish and Wildlife Service. 1985. Endangered and threatened wildlife and

plants; review of vertebrate wildlife; notice of review. Federal

Register 50:37958-37967.

U.S. Fish and Wildlife Service. 1984. Bailey Tract management plan. File

document, Ding Darling National Wildlife Refuge. 20 p.