Skip to main content

Full text of "Distribution and habitat affinity of the swamp rabbit (Sylvilagus aquaticus:Lagomorpha:Leporidae) on the Edwards Plateau of Texas"

See other formats

Occasional Papers 

Museum of Texas Tech University 

NUMBER 167 1 JULY 1997 



John T. Baccus and Michael W. Wallace 

The historical distribution of Sybrilagus aquations 
is primarily in east Texas, with a western boundary that 
extends from Cooke, Wise, and Palo Pinto counties in 
north-central Texas south to Brown County, just north of 
the Edwards Plateau, thence to Travis and Bexar counties, 
juxtaposed to the Balcones Escarpment in central Texas, 
thence south to Aransas County in the southern coastal 
plains (McCarlcy, 1959; Davis, 1966; Schmidly, 1983; 
Dalquest and Homer, 1984; Gamer et al., 1990; Davis 
and Schmidly, 1994). In central Texas, Allen (1895) first 
reported the species in riverbed drift piles and old fallen 
trees of the San Antonio and Medina rivers south of the 
Balcones Escarpment. Later, the species was found east 
of the escarpment in riparian habitat on the Colorado River 
in Travis County (Davis, 1966), Although riparian 
communities of the Colorado and Medina rivers extend 
westward and northward onto the Edwards Plateau, no 
specimens or observations of sw r amp rabbits have been 
reported for the Edwards Plateau. 

The swamp rabbit inhabits swamps and marshes 
of bottomland forests and shrub-dominated wetlands, 
where fallen and hollow trees, brush piles, and dense 
herbaceous vegetation meet year-round habitat require¬ 
ments of food and cover for the species (Blair, 1936; 
Lowe, 1958; Hunt, 1959; Terrel, 1972; Chapman and 

Fcldhamer, 1981; Allen, 1985; Kjolhaug et al., 1987). 
In eastern Texas, S. aquaticus inhabits four vegetational 
areas of the state. In the east Texas Pincywoods, swamp 
rabbits are common in dense thickets of shrubs, trees, 
and vines of hardw'ood bottomland drainage systems and 
are usually the most common lagomorph in second- 
growth timber (Taylor and Lay, 1944; Schmidly, 1983). 
On the Gulf Prairies and Marshes, this lagomorph is 
common in thick, marsh vegetation and briar bushes of 
marshes and canals (McCarley, 1959; Schmidly, 1983). 
They prefer meadows of tall grass, interspersed with 
scrub oaks, in the ecotone between bottomlands and prai¬ 
rie in the Post Oak Savannah (Schmidly, 1983). In the 
Blackland Prairies region, the species is virtually re¬ 
stricted to floodplain corridors of rivers and creeks (Hunt, 
1959; Schmidly, 1983). 

We initiated a study in 1991 to answer tliree ques¬ 
tions. What is the current distribution of swamp rabbits 
along mesic, riverine, corridor habitats of the Edw'ards 
Plateau? What are the characteristics of habitats used 
by swamp rabbits? Are criteria of the habitat suitabil¬ 
ity model (sensu Allen, 1985) applicable to riparian 
habitats inhabited by swamp rabbits along rivers of the 
Edwards Plateau at the westernmost boundary of the 




The westernmost boundary of the study area was 
the Nueces River. The continental north-south line for 
moisture sufficicnt/defieient regions proposed by 
Thomthwaite (1948) divides the Edwards Plateau into 
two regions. The eastern one-half is dry, subhumid, 
mcsothermal with 99.8 to 114 cm of annual precipita¬ 
tion; whereas, the western one-half is semiarid, 
mesothermal with 85.5 to 99.8 cm of annual moisture. 
Based on available climatic, floral (Diamond et al., 
1987), and topographic (Hatch ct al., 1990) informa¬ 
tion, we assumed the area west of the Nueces River was 
too arid to support the continuity of mature riparian for¬ 
ests necessary for suitable swamp rabbit habitat. 
Whereas swamp rabbits need permanent water in their 
habitat (Terrel, 1972; Allen, 1985), we focused our field¬ 
work in riparian forests of river systems of the eastern 
Edwards Plateau. 

The composition of the riparian flora of rivers of 
the eastern Edwards Plateau contrasts with upland veg¬ 
etation. For example, the floral composition and distri¬ 
bution of the riparian community along the Guadalupe 
River contrast with the upland shrub-forest of the 
Edwards Plateau. There is a severe elevational and hori¬ 
zontal compression of the riparian community with a 
loss of mcsic-adaptcd species on thinner, xeric, upslope 
soils. Plants with an Austroriparian affinity occur pri¬ 
marily in the streamside riparian forest; several species 
reach the western boundary of their distribution in this 
riverine zone (Ford and Van Auken, 1982). 


From September 1991 to June 1995, a reconnais¬ 
sance of major river systems of the eastern Edwards 
Plateau was conducted at highway crossings and on pub¬ 
lic and private lands to locate sites with potential swamp 
rabbit habitat. We surveyed riparian plant communities 
of the Edwards Plateau (Diamond et al, 1987) for habitat 
similar to that described in the habitat suitability model 
for swamp rabbits (Allen, 1985). We searched sites 
identified as potential habitat for swamp rabbits and for 
fecal droppings. Swamp rabbits, unlike their congeners 
S.floridanus (eastern cottontail rabbit) and S. auduhonii 
(desert cottontail rabbit), mark territories by depositing 
large, discoidal fecal droppings on elevated objects such 

as tree stumps, fallen logs, rocks and similar perches. 
This behavior is a useful field indicator that has been 
used to corroborate the presence of the species and qual¬ 
ity of the habitat (Low-e, 1958; Hunt, 1959; Terrel, 1972; 
Kort and Fredrickson, 1977; Whitaker, 1980; Dailey et 
al., 1993). Since fecal pellets of the eastern and desert 
cottontail rabbits and white-tailed deer (Odocoileus 
virginianus ) could be mistaken for those of the swamp 
rabbit, we became proficient in die recognition of known 
fecal samples from each species at die beginning of the 
fieldwork. If we found no signs of rabbits on the initial 
visit to a site, we searched die site a minimum of two 
additional times. In addition, potential suitable habitat 
was identified using 1:20,000 or 1:40,000 ASCS (Ag¬ 
ricultural Stabilization and Conservation Service) black- 
and-white aerial photographs taken during 1984 to 1990. 

At all sites, the continuity of suitable habitat was 
characterized and ranked. Sites with a continuous 
breadth (about 8 km) of upstream and downstream suit¬ 
able habitat were assigned a rank of 1. Sites widi frag¬ 
mented habitats and patches < 5 km in breadth were 
given a rank of 0.5. A rank of 0 designated areas with 
small, isolated patches of unsuitable habitat. The wddth 
of the riparian zone at each site was determined by mea¬ 
suring die distance from the edge of the stream to the 
outermost edge of the riparian community. Width val¬ 
ues were obtained for only one side of the stream at a 
site; however, die symmetry of the plant community on 
both sides of the stream w r as noted. If a floodline was 
present, the width of the riparian zone with adequate 
cover for rabbits above the floodline was also measured. 

The woody and herbaceous vegetation of the ri¬ 
parian zone were sampled and characterized using ran¬ 
domly placed 50 m 2 (5 m x 10 m) quadrats at all sites. 
A species richness value was obtained by listing all plant 
species in each quadrat. We used dominant woody spe¬ 
cies at each site to identify the floral series level (Dia¬ 
mond et al., 1987). Frequencies (%) of plant species at 
a site were calculated by dividing the number of sites 
with occurrence by the number of total sampling sites, 
A value for canopy cover was derived by ocular estima¬ 
tion of canopy cover (Allen, 1985). The total canopy 
cover for each site was calculated by adding the mean 
area coverage (%) values for trees, shrubs, herbaceous 
vegetation, and ground debris. Multilayer canopies had 
values > 100%. Herbaceous vegetation at each site was 



sampled seasonally by four 0.25 m 2 quadrats placed at 
the corners of the 50 m 2 quadrat. A species list of her¬ 
baceous vegetation was developed for each site. We 
included identifiable dead or dormant annuals in the 
species list. In each 0.25 m 2 quadrat, the percent area 
coverage of dominant herbaceous plants and the amount 
of bare ground were estimated. The presence and per¬ 
cent ground cover by green herbaceous vegetation dur¬ 
ing winter were recorded. Land use practices on areas 
next to sampling sites were characterized as to tillage, 
grazing intensity by livestock, and land development. 
At sites grazed by livestock, the height of ground cover 
was randomly measured at 10 sites. 

Escape cover in the 50 m 2 quadrat at each site 
was classified into five categories: vines, rock piles and 
crevices, drift debris, dense grasses and sedges, and 
brush and deadfalls. The frequency (%) of each escape 
cover type was calculated for each site. Vines were clas¬ 
sified in cover classes based on growth form. 

Statistical analyses were performed with the 
StatView 4.01 statistical analysis package. Variables 
measured in integers were analyzed using the Mann- 
Whitncy U test to determine the probabilities of differ¬ 
ences in the distribution of measurements between sites 
with swamp rabbits present or absent. Variables ranked 
into two categories (swamp rabbits present or absent) 
were tested for goodness of fit (Zar, 1984) using contin¬ 
gency tables and G 2 probabilities (log likelihood ratios). 
Habitat continuity data ranked into three categories had 
X 1 probabilities determined for the likelihood of differ¬ 
ences among sites with swamp rabbits present or ab¬ 


Intensive searches for swamp rabbits or their sign 
were conducted at 54 sites in 22 counties of the eastern 
Edwards Plateau. Fourteen new Edwards Plateau and 
three peripheral county records for S. aquations in 
Bandera, Blanco, Bosque, Burnet, Caldwell, Comal, 
Gillespie, Guadalupe, Hays, Kendall, Kerr, Lampasas, 
Llano, Medina, Mills, San Saba, and Williamson coun¬ 
ties were documented at 39 sites along the Blanco, 
Bosque, Colorado, Guadalupe, Lampasas, Little Blanco, 
Llano, Medina, Pedernales, San Gabriel, San Marcos, 
and San Saba river systems (Fig, 1). Several sightings 

of swamp rabbits and fecal records on Johnson Creek 
(3 km S Mountain Home), a tributaiy of the Guadalupe 
River in Kerr County, extend the distribution of the spe¬ 
cies about 125 km north and west of Bexar and Travis 
counties. This is the westernmost distributional record 
for the species in North America (Hall, 1981). No signs 
of swamp rabbits were found on the Colorado River in 
Brown, Coleman, and McCulloch counties below Stacy 
Reservoir; Frio River in Real and Uvalde counties; Ll¬ 
ano River in Kimble and Mason counties; Nueces River 
in Edwards, Kinney, and Uvalde counties; or Sabinal 
River in Bandera and Uvalde counties. 

Swamp rabbits inhabited three series-level plant 
communities along rivers of the eastern Edwards Pla¬ 
teau. The frequency (%) of occurrence by this lago- 
morph at sites in these communities was Sugarberry- 
Elm (77%), Bald Cypress-Sycamore (31%) and Pccan- 
Sugarberiy (15%). The westernmost extent of die dis¬ 
tribution of swamp rabbits along rivers of the Edwards 
Plateau corresponded with the upstream distances and 
continuity of these communities. The dominant trees in 
these communities were sycamore, Texas sugarberry, 
pecan, cedar elm, and black willow. The dominant un¬ 
derstory shrubs were possumhaw, roughleaf dogwood, 
waxleaf ligustrum, and false indigobush (Table 1). 

The Sugarberry-Elm community provided the best 
habitat for sw'amp rabbits. The species inhabited a thick 
understory of brush, deadfalls, and briars in a Sugar- 
berry-Elm community on Walnut Creek, a tributary of 
the Colorado River in Travis County. However, swamp 
rabbits were found at only three isolated sites in a Sug- 
arberry-Elm community along the Colorado River be¬ 
tween Travis and San Saba counties. Two sites, Sandy 
Creek near Lake L B J in Llano County and the conver¬ 
gence of Penny and Hamilton creeks in Burnet County, 
were at the confluence of creeks where the river was 
permanently flooded by water impoundments (Joe 
Wallace, Texas Game Warden, pers. comm.). The other 
site was at Colorado Bend State Park. Swamp rabbits 
inhabited a diverse Sugarberry-Elm community' on the 
Lampasas River, a tributary of the Brazos River, near 
the northeast comer of Burnet County. Steep slopes 
along the river limited livestock grazing in the riparian 
zone, and consequently, a lush growth of vines and un¬ 
derstory brush combined with boulders and drift piles 
provided an excellent, yet narrow habitat for swamp 

Figure 1. The distribution of Sylvilagus aquaiicus on the eastern Edwards Plateau 
compared to the historical distribution in Texas. 



rabbits. The Sugarberry-Elm community (and swamp 
rabbits) continued west along the San Gabriel River to 
about 16 km east of Burnet, Burnet County. Here, on 
small creeks, the Sugarberry-Elm community integrated 
with plateau live oak, cedar elm and Ashe juniper trees. 
Hie Sugarberry-Elm community occurred along the 
Guadalupe River from the Blackland Prairie south of 
the Edwards Plateau into Kerr County where it became 
patchy near Kerrville. 

The Bald Cypress-Sycamore community was the 
only riparian community present on some rivers. In 
expansive river valleys with a multiple-community ri¬ 
parian forest along the river, the Bald Cypress-Sycamore 
community occupied the immediate strearaside. Either 
a Sugarberry-Elm or Pccan-Sugarberiy community or 
both occurred on upslopes to this community. A Bald 
Cypress-Sycamore community occurred only along the 
San Gabriel River in Williamson County, but in Burnet 
County the upper watershed riparian forests became 
patchy and swamp rabbits were absent. A substantial 
Bald Cypress-Sycamore community (and swamp rab¬ 
bits) extended along the Guadalupe River from the 
Blackland Prairie south of the Balcones Escarpment 
west to Kerr County at the river’s headwaters. A Bald 
Cypress-Sycamore community extended west along the 
Medina River to its headwaters in Bandera County. 
However, habitat suitable for swamp rabbit was dis¬ 
junct with swamp rabbits occurring at only one site west 
of Medina Lake in Bandera County. 

Sw'amp rabbits inhabited only one site with a lush 
Pecan-Sugarberry community on the San Saba River in 
San Saba Comity'; however, this community was patchy, 
with unsuitable habitat south of U. S. Highway 190. 
This lagomorph inhabited the ecotone between the ri¬ 
parian and upland community series. Fecal materials 
were found consistently in ecotones near the riparian 
zone. Here, swamp rabbits were sympatric with east¬ 
ern cottontails, a species common on xeric upslopes of 
ectones. The species did not inhabit the Sycamore-Wil- 
low series along the upper Nueces River in Edwards 
and Real counties. This community occurs on gravely 
soils of flood-scoured streambeds of the Edwards Pla¬ 
teau and consists of small trees, representing growth 
since the last catastrophic flood, and a poorly devel¬ 
oped shrub layer of disturbance-type species. No sign 
of S. aquations was found in the Plateau Live Oak- 

Netleaf Hackbcrry community, the dominant riparian 
community in xeric streambeds of the Sabinal, Frio, and 
Nueces rivers. 

The width of the riparian zone at sites inhabited 
by swamp rabbits ranged from 16 to 220 m 
(X = 61.6 m, SE — 0.31). The width of the riparian 
zone at sites with swamp rabbits absent ranged from 
0 to 175 m (X - 54.7 m, SE = 13.9). Swamp rabbit 
habitat did not occur above floodline at two of 39 sites 
with swamp rabbits present and seven of 15 sites with 
the species absent. The maximum width of swamp 
rabbit habitat above floodline at sites with swamp 
rabbits present was 120 m (X = 27,3 m, SE = 4.2). The 
width of swamp rabbit habitat above floodline at sites 
with this lagomorph absent ranged from 0 to 100 m 
(X - 22.1 m, SE = 8.81). 

Significantly more plant species occurred at sites 
with swamp rabbits present than sites with swamp rab¬ 
bits absent (% 7 = 18.7, d. f. = 1, p <0,001). Seventy- 
seven plant species were identified at sites with swamp 
rabbits present. The mean number of plant species for 
sites with swamp rabbits present was 17.7 (SE = 0,7), 
whereas sites with the species absent had a mean of 
11.7 (SE = 1.2). Six of 13 common plants at sites with 
swamp rabbits absent had an affinity with upland plant 
communities. Vines were an important component of 
swamp rabbit habitat. Saw greenbriar w'as the most 
common vine associated with sw'amp rabbit habitat, 
occurring at 82% of sites w'ith this lagomorph present. 
Other common vines associated with swamp rabbit habi¬ 
tat were dewberry', grapevine, poison ivy and Carolina 
snailsecd (Table 1). 

Forbs and tall grasses composed a canopy of her¬ 
baceous vegetation at a height about 1 m. Important 
forbs and grasses of this canopy layer in suitable habi¬ 
tat were giant ragweed, frostweed, common goldcnrod, 
switchgrass, Johnsongrass, and Indiangrass (Table 1). 
During winter significantly more green herbaceous veg¬ 
etation occurred at sites with swamp rabbits present 
(95%) than sites with this lagomorph absent (40%) (G^ 
18.6 , d. f. = 1, p < 0.001). Excessive grazing in the 
riparian zone was common at several sites. There was 
a substantial, but insignificant, difference in the amount 
and height of vegetation (G 2 - 3.1, d. f. = 1, p = 0.08) 
at sites with sw amp rabbits present compared to sites 



Table 1. Percent frequency of woody vegetation, understory shrubs and 
vines, and herbaceous ground vegetation at sites with Sylvilagus aquaticus along 

rivers of the eastern Edwards Plateau, 1993. ______ 


Canopy Woody Species 
Acer negundo 
Carya illinoensis 
Celtis laevigata 
Celtis reticulata 
Diospyros texana 
Fraxinus pennsylvanica 
Fraxinus velutina 
Juglans nigra 
Juglans microcarpa 
Juniperus ashei 
Maclura pomifera 
Metia azedarach 
Morns rubra 
Platanus occidentals 
Populus deltoides 
Quercus buckleyi 
Quercus macrocarpa 
Quercus shumardii 
Quercus fusiformis 
Salix nigra 
Sambucus canadensis 
Taxodium distichum 
Ulmus americana 
Ultttus crassifolia 

Understory Shrubs and Vines 
A morph a fruticosa 
Ampelopsis arborea 
Baccharis neglecta 
Bumelia lanuginosa 
Cocculus carolinus 
Comus drummondii 
Ilex decidua 
Ligustrum lucidum 
Lonicera japonica 
Parlhenocissus quinquefolia 
Ptelea trifoliata 
Rhamnus caroliniana 
Rubus trivialis 
Smilax bona-nox 
Toxicodendron radicans 

Box Elder 




Texas Sugarberry 


Nct-lcaf Hackberry 


Mexican Persimmon 


Green Ash 


Arizona Ash 


Little Walnut 


River Walnut 


Ashe Juniper 


Osage Orange, Bois d'Arc 




Red Mulberry 




Eastern Cottonwood 


Texas Oak 


Bur Oak 


Shumard Oak 


Plateau Live Oak 


Black Willow 


American Elderberry 




American Elm 


Cedar Elm 


False Indigobush 




Roosevelt Weed 


Wooly Bumelia 


Carolina Snailseed 


Rough-leaf Dogwood 




Wax-leaf Ligustrum, Privet 


Japonese Honeysuckle 


Virginia Creeper 




Carolina Buckthorn 






Poison Ivy 


Mustang Grape 


without this species. Although herbaceous vegetation 
was reduced by livestock grazing at 92% of sites with 
swamp rabbits present compared to 73% of sites with 
swamp rabbits absent, the amount of reduction was sub¬ 
stantially greater at sites with the species absent. 

Woody vegetation canopy cover and closure at 
sites with swamp rabbits present were compared to sites 
with no swamp rabbits (Table 2). An incomplete canopy 
closure was characteristic of most sites. This openness 
in the canopy enhanced development of a dense under- 



Table 1. continued 




Herbaceous Ground Vegetation 
Ambrosia trifida 

Giant Ragweed 


Arundo donax 

Giant Reed 


Carex spp 



CUasmanthium latifolium 

Creek Oats 


Cynodon dactylon 



Elymus canadensis 

Canada Wild Rye 


Pan icum virgatum 



Ricinus communis 



Schizachyrium scoparium 

Little Bluestem 


Solidago altissima 

Tall Goldenrod 


Sorghastrum nutans 



Sorghum halepense 



Typha latifolia 



Verbesina virgin ica 

Fros tweed 


Vemonia baldwinii 

Western Ironweed 


Xanthium spinosum 

Spiny Cocklebur 


story of shrubs and ground vegetation. Although only 
one canopy cover type (debris ground cover) was sig¬ 
nificantly different in a comparison of sites with swamp 
rabbits present versus sites with the species absent, there 
was a highly significant difference in total canopy cover 
between sites where this lagomorph was present or ab¬ 
sent. Habitats with swamp rabbits present had denser 
foliage and a more complex stratification of canopy lay¬ 

Vine tangles, brush and deadfalls, and drift de¬ 
bris w r cre identified as the most important types of es¬ 
cape cover for swamp rabbits (Table 3). The amounts 
of all escape cover types except dense grass at sites 

Table 2. Mean canopy cover and closure 
(%) for sites with Sylvilagus aquaticus present or 
absent along rivers of the eastern Edwards Plateau, 




(n = 39) 


ID _ 

z 2 * 

Tree Canopy Cover 






Shrub Canopy Cover 






Herb. Canopy Cover 






Debris Ground Cover 






Total % Canopy Cover 






♦degrees of freedom = 1 

with swamp rabbits present were significantly greater 
than the amounts at sites with the species absent. Drift 
debris and brush piles or fallen tree tops occurred at 
87% of sites with swamp rabbits present and only 64% 
of sites with lire species absent. Rock piles and crev¬ 
ices provided escape cover at 33% of sites with swamp 
rabbits present. 

There was a significant difference in swamp rab¬ 
bit presence based on habitat continuity' ranks ( x 7= 33.5> 
d. f.= 2, p < 0.0001). Swamp rabbits inhabited all sites 
(32) with a continuity rank of 1.0. The species occurred 
at only seven of 17 sites with a continuity rank of 0.5. 
Neither swamp rabbits or their signs wore found at sites 
with a continuity rank of 0. 

Table 3. Percent frequency of escape cover 
types at sites with Sylvilagus aquaticus present or 
absent along rivers of the eastern Edwards Plateau, 




G 3 

Vine Tangles 




Brush/ Deadfalls 




Dense Grass/ Sedges 




Drift Debris 




Rocks/ Crevices 





Past Distribution 

Holocene fossil remains of swamp rabbits from 
Levels 5 and 6 of Kinkaid Shelter (Uvalde County, late 
Rancholabrcan) have been dated between about 8,000 
to 6,000 yBP (Tamers et al., 1964). Specimens from 
Holocene Straia TIa and I lb of Eagle Cave and the Up¬ 
per Zone of Centipede Cave (Val Verde County) with 
dates between about 5,000 to 400 yBP were assigned 
to S. aquaticus by discriminate analysis; however, be¬ 
cause of an inherent 8% error in the analysis, Hulbert 
(1984) concluded that the species was a doubtful mem¬ 
ber of the fauna . Fossil remains from several sites of 
small mammals {Sorex vagrans , Blarina brevicauda , 
Zapus hudsonius, Synaplomys cooperi, Microtus 
pennsylvanicus , Lepus townsendii) with present dis¬ 
tributions far north and east of central Texas indicate a 
moist and cool environment existed during the late Pleis¬ 
tocene-early Holocene in central Texas (Lundclius, 
1967; Dalquest ct al., 1969; Scmkcn, 1983; Hulbert, 
1984; Winkler, 1990; Wilkens, 1992; Hairier, 1993). 
Hafner (1993) considered the presence of Zapus 
hudsonius in Schulze Cave (Edwards County) as con¬ 
firmation that an eastern grassland mammalian assem¬ 
blage influenced the Wisconsinan fauna of central Texas. 

The Kinkaid Shelter site is located about 45 km 
west of swamp rabbit records on the Medina River 
(Medina County). Our findings, when combined with 
the fossil evidence, suggest that swamp rabbits have 
probably inhabited the eastern Edwards Plateau for sev¬ 
eral thousand years. Thus, the occurrence of swamp 
rabbits along riparian corridors of the eastern Edwards 
Plateau is not considered a recent biological event, but 
rather, the result of intensive collecting effort. Resi¬ 
dents of Kerr Comity reported that swamp rabbits were 
more common in river bottoms from 1930 to 1950 (Al 
Kanz, pers. comm.; Frank Syfan, pers. comm.). How¬ 
ever, it is interesting that Lacey and Bailey collected 
mammals along the Guadalupe River (Bailey near 
Ingram, one of our collection sites) in Kerr County in 
the late 1800s, but neither reported collecting swamp 
rabbits (Allen, 1895; Bailey, 1905). The fragmented 
distribution of the species and large tracts of inacces¬ 
sible private land limited our ability to document all 
possible occurrences of the species on the eastern 
Edwards Plateau, 

Primary Habitat 

Community diversity.-- Species richness is a 
characteristic of suitable swamp rabbit habitat. The 
significant difference in the mean number of plant spe¬ 
cies at sites with swamp rabbits present compared to 
sites with swamp rabbits absent indicates that riparian 
community complexity benefits swamp rabbits. Ford 
and Van Auken (1982) reported that the greatest total 
density, total dominance, and species richness for plants 
occurred at the water’s edge along Edwards Plateau riv¬ 
ers. hi a comparison of woody plants of uplands and 
crcck bottoms with associated Quaternary deposits of 
the southern Edwards Plateau, Van Auken et al. (1979) 
found 49% of the species had a riparian affinity, 14% 
occurred only on uplands, and 37% were common to 
both habitats. Creek bottoms had three times greater 
densities of woody plant species than the adjacent Buda 
formation of uplands. Periodic flooding may be impor¬ 
tant in maintaining the diversity of plant communities 
inhabited by swamp rabbits. Floods improve the habi¬ 
tat for swamp rabbits in riparian communities by de¬ 
positing topsoil, piling up flood debris, stimulating forb 
growth and thinning the forest. Bald Cypress-Sycamore 
forests of the eastern Edwards Plateau require distur¬ 
bance for growth of seedlings and maintenance of flo¬ 
ral diversity and inlcrspcrsion (Van Auken, 1993). 

Canopy.-- Shrubs are a source of food and an 
important structural component of habitat used for es¬ 
cape cover by swamp rabbits (Allen, 1985). The mean 
percent shrub crown canopy closure for all sites in this 
study was well below the optimum (50%) reported for 
the suitability model. Sites with swamp rabbits present 
had a higher mean shrub canopy closure. Only six of 
54 sites had > 50% shrub crown canopy closure value; 
values ranged from 50% to 98%. Swamp rabbits in¬ 
habited fiv e of these sites. Compared to its importance 
in the habitat suitability model for the eastern United 
States, shrub canopy closure apparently is not as an 
important component of swamp rabbit habitat on the 
eastern Edwards Plateau. 

In the habitat suitability model (Allen, 1985), tree 
canopy closure ranges from 25% to 60%. Tree canopy 
closure values for Edwards Plateau sites with swamp 
rabbits present fit the model. This optimum level of 
tree canopy closure occurs in riparian communities be¬ 
cause trees grow in narrow r strips, and the edge is near 



any point within the forest. Even sites with swamp rab¬ 
bits absent had values slightly below the optimum suit¬ 
ability value. 

Herbaceous vegetation dominates open areas 
along rivers of the eastern Edwards Plateau. Herba¬ 
ceous vegetation provides critical escape cover and is 
the primary source of food for swamp rabbits (Bailey, 
1969; Smith, 1982; Allen, 1985). In the habitat suit¬ 
ability model, the optimum coverage for herbaceous 
canopy is 75%. Although the mean canopy cover 
(45.3%) for herbaceous vegetation at sites with swamp 
rabbits present in our study was substantially below the 
optimum value of the suitability model, this value for 
herbaceous canopy cover would have an index value of 
0.6 in the suitability model (Allen, 1985). However, 14 
of 39 (36%) sites with swamp rabbits present had a > 
75% herbaceous plant cover. Green vegetation occurred 
throughout the year at 95% of sites w ith swamp rabbits 
present compared to only 40% at sites with swamp rab¬ 
bits absent. 

Ground cover.— Diversity and interspersion of 
cover types are an important aspect of swamp rabbit 
habitat (Chapman ct al., 1982). Dense, low-growing 
tangles of vines provide escape cover for swamp rab¬ 
bits, In the habitat suitability model, Allen (1985) 
grouped vines with trees, shrubs, or herbaceous veg¬ 
etation depending on their physiognomy, but he sug¬ 
gested that vines should be a separate canopy cover type 
in future habitat studies. Vines were analyzed as a sepa¬ 
rate escape cover type in our study. Dense vine tangles 
of saw 7 greenbriar, dew'berry or Japanese honeysuckle 
provided the most important type of escape cover at 97% 
of sites with swamp rabbits present. 

Ground debris was not included as a separate 
cover type in the habitat suitability model (Allen, 1985). 
This cover type was an important cover component of 
habitat in the eastern Edwards Plateau. Although a small 
percent of the surface area at sites with swamp rabbits 
present was ground debris, it covered a larger area at 
sites with swamp rabbits present than sites with the spe¬ 
cies absent. However, when identitying types of es¬ 
cape cover at sites with swamp rabbits present, ground 
debris composed one of three important cover types. 
Observations and/or collection of swamp rabbits often 
occurred at drift piles or fallen tree lops. 

Disturbance. — Swamp rabbit habitat is enhanced 
by disturbance, and areas with intermittent flooding have 
a higher habitat suitability index (Allen, 1985). Streams 
of the eastern Edw r ards Plateau because of topography 
have substantial periodic flooding after torrential rams. 
Flooding displaces swamp rabbits from the riparian zone 
onto upland areas, but as the water recedes, rabbits usu¬ 
ally return to previously occupied areas (Conaway et 
al., 1960). The intensity of land use near streams may 
determine swamp rabbit survival during floods. Since 
most suitable habitat for swamp rabbits on the Edwards 
Plateau occurs in relatively narrow 7 strips, compression 
of or loss of riparian zones during flooding may cause 
severe survival problems for swamp rabbits (J. K. Jones, 
Jr., pers. comm.). Blair (1939) studied the adverse ef¬ 
fects of flooding on populations of small, terrestrial 
mammals inhabiting river floodplains in eastern Okla¬ 
homa and found terrestrial species confined to flood- 
plains may be virtually extirpated in parts of their range 
by severe floods. After such a flood, one swamp rabbit 
w as the only small mammal observed in the floodplain. 
When sw'amp rabbits arc forced out of suitable habitat 
by floods, predation probably increases. 

Fragmentation of habitat— Several factors con¬ 
tributed to fragmentation of habitat for swamp rabbits 
along rivers of the eastern Edwards Plateau. A thin, 
sandy substrate on the Pedemales River in Blanco 
County cast of Johnson City at the Robinson Ranch and 
Pedemales Falls State Park caused an unsuitable edaphic 
habitat for the plant species of riparian communities. 
Riverbed physiography caused a natural break in suit¬ 
able habitat on the Medina River in Bandera County, 
wdicre flooding had deposited debris up to the Ashe Ju¬ 
niper-Live Oak community. Similarly, the riparian com¬ 
munity along the Llano River is very narrow and patchy. 
The westernmost rivers of the eastern Edwards Plateau, 
such as the upper Nueces River, have poorly developed 
riparian zones juxtaposed to a Live Oak-Midgrass Sa¬ 
vannah community. These rivers with narrow strips or 
small isolated patches of riparian habitat next to in¬ 
tensely used land are poor swamp rabbit habitat (Kort 
and Fredrickson, 1977). 

Water impoundments on the rivers cause substan¬ 
tial breaks in the continuity of sw r amp rabbit habitat 
along Edwards Plateau rivers. A series of dams im¬ 
pound the Colorado River between Travis Comity and 



San Saba County for a distance of about 222 km. At 
Lake Georgetown on the San Gabriel River, the water 
level cxtnds up to the Ashe Juniper-Live Oak and Live 
Oak-Midgrass communities for 8 km. Canyon Lake 
covers 16 km of former riparian habitat on the 
Guadalupe River, and 24 km of former riparian habitat 
on the Medina River is inundated by Medina Lake. 
Neither swamp rabbits nor their signs were found along 
the shores of these reservoirs 

Conversion of bottomland hardwood forest to ag¬ 
ricultural use has been the primary cause of swamp rab¬ 
bit habitat loss in eastern parts of its range (Terrel, 1972; 
Schmidly, 1983; Kjolhaug et ah, 1987). Since 98% of 
the Edwards Plateau is rangeland (Hatch ct ah, 1990), 
timber harvesting and farming arc not major causes of 
habitat degradation. Farm use of bottomlands near study 
sites was pecan orchards and hay fields. The major 
agricultural use of land that is detrimental to the habitat 
of swamp rabbits is the reduction of available food and 
ground cover by intense grazing of livestock. Swamp 
rabbits seem to tolerate a moderate reduction of veg¬ 
etation, but severe overgrazing results in an unsuitable 

Swamp rabbits mhabited undisturbed lots between 
houses in subdivisions along Edwards Plateau rivers. 
The rate of development and density of houses are im¬ 
portant factors in the degradation of habitat. However, 
often the purchase of land for a subdivision usually re¬ 
sults in the removal of livestock, fallowing of land, suc¬ 
cession of the plant community', and an improvement in 
the continuity of an otherwise fragmented riparian habi¬ 
tat. The continuity of riparian habitat improves, and 
depending on the intensity of vegetation management 
by landowners and density of houses, the subsequent 
development of the subdivision may not be as detrimen¬ 
tal as livestock grazing on the sw'amp rabbit popula¬ 

Swamp rabbits occupied undeveloped riparian 
habitat at some state parks and natural areas. The in¬ 
tensity of vegetation management seems to determine 
their presence. Picnic grounds at Blanco State Park on 
the Blanco River have a closely mowed bermudagrass 
lawn. Swump rabbits w'ere not observed in this part of 
die park. However, sw r amp rabbits inhabited undevel¬ 
oped land downstream and upstream from the picnic 

grounds. Sw r amp rabbits were observed in tall grass 
and deadfalls along the hiking trail upstream from the 
campgrounds at Colorado Bend State Park. Cover was 
patchy because the vegetation had been severely dam¬ 
aged by foraging feral hogs. This lagomorph also in¬ 
habited a Bald Cypress-Sycamore community at Honey 
Creek State Natural Area. 

Recruitment and dispersal are important biologi¬ 
cal processes of rabbit populations, and when suitable 
riparian habitats are fragmented on a large portion of a 
river, the dispersal corridor for recruits is lost (Chapman 
et al., 1982). Kjolhaug et al. (1987) found rabbits in 
small areas connected by narrow habitat corridors along 
Illinois rivers, however smaller, isolated areas located 
on small tributary streams were unsuitable habitat. 
Riparian forests of the eastern Edwards Plateau are of¬ 
ten discontinuous “islands” of habitat. Eighteen per¬ 
cent of sites with swamp rabbits were classified as 
patchy, unsuitable habitat, suggesting the species is vul¬ 
nerable to extirpation in these areas. Swamp rabbits 
inhabiting riparian corridors along rivers of the eastern 
Edwards Plateau live in environments that would be 
classified as marginal to poor habitat when compared 
to the habitat suitability model of the species. Swamp 
rabbit populations tend to fluctuate, but with a discon¬ 
tinuous riparian habitat, isolated populations are unlikely 
to be reestablished once extirpated (Kjolhaug et al., 

Decreasing the fragmentation of plant communi¬ 
ties inhabited by swamp rabbits would be the most ef¬ 
fective method of managing for sw amp rabbits in ripar¬ 
ian corridors. In the eastern Edw ards Plateau, the best 
habitat for sw'amp rabbits often occurs in right-of-ways 
under highway bridges, vacant lots in minimally devel¬ 
oped subdivisions, undeveloped areas of state parks and 
natural areas and ungrazed riparian areas. Improve¬ 
ment of swamp rabbit habitat along rivers and streams 
of the eastern Edwards Plateau could be accomplished 
by expanding the patch size of these habitats by allow¬ 
ing the growth of vines, grasses and forbs, selectively 
thinning mature forests, excluding livestock grazing (at 
least until slirub thickets form), controlling feral hog 
populations, and leaving piles of brush or fallen tree 
lops for structure. An essential component of sw amp 
rabbit habitat is adequate cover above the floodline. 
Disturbances to the Sugarberry-EIm and Bald Cypress- 



Sycamore community series along rivers of the eastern 
Edwards Plateau should not be closer than 100 m above 
the floodline. Habitat corridors to allow dispersal of 
swamp rabbits could be created around large reservoirs 
and sections of rivers that naturally have a poorly de¬ 
veloped riparian forest by planting or allowing the 
growth of native species. 


Wc gratefully acknowledgment the cooperative 
assistance of Roy Welch, Ray Aguirre, and personnel 
at the Kerr Wildlife Management Area, Colorado Bend 
State Park, Blanco State Park, Honey Creek Natural 
Area, and Heart of the Hills State Fish Hatchery of the 
Texas Parks and Wildlife Department in supplying 
specimens or study sites. Kevin Schwausch, Scott Sum¬ 
mers, and Todd Pelcik assisted in field collection. We 
acknowledge the following landowners for allowing us 
to collect specimens and habitat data on their land: 
Harvey Ferris, Mike Miller, Scott Summers, and Mrs. 
Marley of Bend, Texas, an advocate of habitat manage¬ 
ment for sw r amp rabbits. 


Allen, A. W. 1985. Habitat suitability index models: 
swamp rabbit. U. S. Fish Wildl. Scrv. Biol. 
Rep., 82 (10.107). 20 pp. 

Allen, J. A. 1895. Descriptions of new American mam¬ 
mals. Bull. Amer. Mus. Nat. Hist., 7 (10): 327- 

Bailey, J. A. 1969. Exploratory study of nutrition of 
young cottontails. J. Wildl. Mgmt., 33: 346- 

Bailey, V. 1905. Biological survey of Texas. N. Amer 
Fauna, 25: 1-222. 

Blair, W. F. 1939. Some observed effects of stream- 
valley flooding on mammalian populations in 
eastern Oklahoma. J. Mamm,, 20: 304-306. 

Chapman, J. A., and G. A. Feldhamer. 1981. Mamm. 
Species. 151: 1-4. 

Chapman, J. A., J. G. Hockman, and W. R. Edwards. 
1982. Cottontails Sylvilagus floridemus and 
allies. Pp 83-123 in Wild mammals of North 
America: biology, management, and econom¬ 
ics. (J. A. Chapman and G. A. Feldhamer, eds.). 
Johns Hopkins Univ. Press, Baltimore. XIII + 
1147 pp. 

Conaway, C. H., T. S. Baskett, and J. E, Toll. 1960. 
Embryo resorption in the swamp rabbit. J. 
Wildl. Mgmt., 24: 197-202. 

Dailey, T. V., T. M. Vangildcr, and L. W. Burger, Jr. 
1993. Swamp rabbit distribution in Missouri. 
Proc. Southeastern Assoc. Fish Wildl. Agen¬ 
cies, 47: 251-256. 

Dalquest, W. W., and N. V. Homer. 1984. Mammals 
of north-central Texas. Midwestern State Univ. 
Press, Wichita Falls. 261 pp. 

Dalquest, W. W., W. E. Roth, and F. Judd. 1969. The 
mammal fauna of Schulze Cave, Edwards 
County, Texas. Bull. Florida State Mus,, Biol. 
Sci., 13: 205-276. 

Davis, W. B. 1966. The mammals of Texas. Texas 
Parks and Wildl. Dept., Bull. 41: 236-244. 

Davis, W. B., and D. J. Schmidly. 1994. The mam¬ 
mals of Texas. Texas Parks and Wildl. Press, 
Austin. x+338 pp. 

Diamond, D. D„ D, H. Riskind, and S.L. Orzcll. 1987. 
A framework for plant community classifica¬ 
tion and conservation in Texas. Texas J. Sci., 
39: 203-221. 

Ford, A. L., and O. W. Van Auken. 1982. The distri¬ 
bution of woody species in the Guadalupe River 
floodplain forest in the Edwards Plateau of 
Texas. Southwestern Nat., 27: 383-392. 

Gamer, W. H., J. G. Joy, K. D. Huckabee, R. M. Pitts. 
1990. Additional records of mammals along 
the boundary of the Kansan-Tcxan biotic prov¬ 
inces in north-central Texas Texas J. Sci., 42: 

Hafner, D. J. 1993. Reinterpretation of the Wisconsin 
mammalian fauna and paleoenviromnent of the 
Edwards Plateau Texas. J. Mamm., 74: 162- 



Hall, E. R. 1981. The mammals of North America. 
Second ed. John Wiley & Sons, New York, 1: 
XV + 1-600 + 90 and 2: VI + 601-1181 +90. 

Hatch, S. L., K. N. Gandhi, and L. E. Brown. 1990. 
Checklist of the vascular plants of Texas. MP- 
1655. Texas Agric. Expcr. Stn., College Sta¬ 
tion. 158 pp. 

Hulbert, R. C., Jr. 1984. Latest Pleistocene and Ho¬ 
locene leporid faunas from Texas: their com¬ 
position, distribution and climactic implica¬ 
tions. Southwestern Nat., 29: 197-210. 

Hunt, T. P 1959. Breeding habits of the swamp rabbit 
with notes on its life history. J. Mamm., 40: 

Kjolhaug, M. S., A. Woolf, and W. D. Klimstra. 1987. 
Current status and distribution of the swamp 
rabbit in Illinois. Trans. Illinois Acad. Sci., 
80: 299-308. 

Kort, P. A., and L. H. Fredrickson. 1977. Swamp rab¬ 
bit distribution in Missouri. Trans. Missouri 
Acad. Sci., 10 and 11: 72-77. 

Lowe, C. E. 1958. Ecology of the swamp rabbit in 
Georgia. J. Mamm., 39: 116-127. 

Lundclius, E. L., Jr. 1967. Late Pleistocene and Ho¬ 
locene faunal history of central Texas. Pp 287- 
319 in Pleistocene extinctions: the search for 
a cause (P. S. Martin and H. E. Wright, Jr., 
eds.). Yale Univ. Press, New Haven. 453 pp, 

McCarley, H. 1959. The mammals of eastern Texas. 
Texas J. Sci., 1: 385-426, 

Schmidly, D. J. 1983. Texas mammals east of the 
Balconcs fault zone. Texas A&M Press, Col¬ 
lege Station. 400 pp. 

Semkcn, H. A,, Jr. 1983. Holocene mammalian biogc- 
ography and climatic change in the Eastern and 
Central United States. Pp 182-207 in Late- 
Ouartemary environments of the United States, 
the Holocene (H. E, Wright, Jr., ed.). Univ. 
Minnesota Press, Minneapolis. 277 pp. 

Smith, M. W. 1982. Effects of selected silvicultural 
practices on swamp rabbit (Sylvilagus 
aqnaticus ) habitat. Unpublished M. S. thesis., 
Mississippi State Univ., State College. 176 pp. 

Tamers, M. A., F. J. Pearson, and E. M. Davis. 1964, 
University of Texas radiocarbon dates II. 
Radiocarbon, 6: 138-159. 

Taylor, W. P, and D. W. Lay. 1944. Ecologic niches 
occupied by rabbits in eastern Texas. Ecol¬ 
ogy, 25: 120-121. 

Terrel, T. L. 1972. The swamp rabbit {Sylvilagus 
aquaticus) in Indiana. Amer. Midi. Nat., 87: 

Thornthwaite, C.W. 1948. An approach toward a ra¬ 
tional classification of climate. Geog. Rev., 
38: 55-94. 

Van Auken, O. W. 1993. Size distribution pattern and 
potential population change of some dominant 
woody species of the Edwards Plateau region 
of Texas. Texas J. Sci., 45: 200-210. 

Van Auken, O. W., A. L. Ford, and A. Stein. 1979. A 
comparison of some woody upland and ripar¬ 
ian plant communities of the southern Edwards 
Plateau. Southwestern Nat., 24: 165N80. 

Whitaker, J. O., Jr. 1980. The Audubon Society field 
guide to North American mammals. Alfred A. 
Knopf Publ., New York. 745 pp. 

Wilkins, K. T. 1992, Mammalian paleofaunas of cen¬ 
tral Texas from the late Wisconsinan glacial 
period to the latest Holocene. Texas J. Sci. 
44: 263-281. 

Winkler, A. J. 1990. Small mammals from a Holocene 
sequence in central Texas and their palco- 
enviromental implications. Southwestern Nat., 
35: 199-205. 

Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, 
Inc., Englewood Cliffs. 718 pp. 

Address of Authors: 


Department of Biology, Southwest Texas State 

University, San Marcos, Texas 78666 

e-mail: jb02@swt. edu 




Specimens examined or swamp rabbits ob¬ 
served: All collected materials are housed in the South¬ 
west Texas State University Mammal Collection. All 
counties listed are in Texas, Bandera: Medina River, 
13 km E Medina; Bexar: Medina River, 16 km S San 
Antonio; Blanco: Little Blanco River, 21 km W Fischer; 
Cypress Creek, 21 km SE Johnson City; Pedemales 
River, 5 km NE Johnson City; Pedemales River, 5 km 
N Johnson City; Blanco River, 5 km S Blanco; Blanco 
River, 5 km W Blanco; Bosque : Bosque River, Valley 
Mills; Barnet : Colorado River, 11 km E Marble Falls; 
Lampasas River, 40 kin NE Burnet; San Gabriel River, 
13 km NE Bertram; San Gabriel River, 11 km N 
Bertram; San Gabriel River, 5 km SE Bertram; 
Confluence of Penny and Hamilton Creeks, 5 km E 
Marble Falls; Caldwell : San Marcos River, Martindale; 
Comal: Guadalupe River, 3 km N Sattler; York Creek, 
13 km E New Braunfels; Honey Creek, Honey Creek 
State Natural Area; Gillespie: Pedemales River, 8 km 

E Stonewall; Pedemales River, 5 km E Stonewall; 
Pedemales River, 8 km W Stonewall; Guadalupe: San 
Marcos River, Staples; Hays: Blanco River, 5 km N 
San Marcos; San Marcos River, San Marcos; Kendall: 
Guadalupe River, 21 km NE Bocme; Kerr: Guadalupe 
River, 19 km W Hunt (SWTSU 107); Johnson Creek, 
Ingram; Johnson Creek, 4 km SE Mountain Home; 
Guadalupe River, 3 km W Ingram (SWTSU 85); 
Guadalupe River, 5 km W Ingram; Guadalupe River, 5 
km W Center Point; Johnson Creek, 3 km S Mountain 
Home (SWTSU 92); Lampasas: Colorado River, 32 
km W Lampasas (SWTSU 102, 103); Llano: Sandy 
Creek, 17 km SE Kingsland; Llano River, Kingsland; 
Medina: Medina River, 8 km N Castroville; Mills: Colo¬ 
rado River, 13 km S Goldthwaite; San Saba: Colorado 
River, 21 km SE San Saba; Colorado River, 16 km E 
San Saba; San Saba River, 8 km W San Saba; Travis: 
Cypress Creek, 40 km SW Austin; Walnut Creel, Aus¬ 
tin; Williamson: San Gabriel River, 3 km W 
Georgetown; San Gabriel River, 5 km NW Georgetown, 
San Gabriel River, 13 km NW Georgetown. 


It was through the efforts of Horn Professor J Knox Jones, as director of Academic Publications, that Texas 
Tech University initiated several publications series including the Occasional Papers of the Museum. This and 
future editions in the series are a memorial to his dedication to excellence in academic publications. Professor 
Jones enjoyed editing scientific publications and served the scientific community as an editor for the Journal of 
Mammalogy, Evolution, The Texas Journal of Science, Occasional Papers of the Museum, and Special Publica¬ 
tions of the Museum. It is with special fondness that we remember Dr. J Knox Jones. 

Institutional subscriptions are available through the Museum of Texas Tech University, attn: NSRL Publica¬ 
tions Secretary, Box 43191, Lubbock, TX 79409-3191. Individuals may also purchase separate numbers of the 
Occasional Papers directly from the Museum of Texas Tech University. 

ISSN 0149-175X 

Museum of Texas Tech University ,; Lubbock , TX 79409-3191