HABKTAT FAJORS ASSOCiéTED WW MGRTAUW OF
SOUTHERN MiCHiGAN WELGUFE ON AN INTERSTATE
HEGHWAY

Thesis for the Degree of M. S.

MiCH‘éGAN STATE UNEVERSETY

RICHARD WJRENCE KASUL
1976

' \|fl\\\\\l1\\\i||fl\fl\\\lWM\lH W \\ \l ' 1/-

1 5531

 

 

HABITAT FACTORS ASSOCIATED WITH MORTALITY OF
SOUTHERN MICHIGAN WILDLIFE ON AN INTERSTATE HIGHWAY

BY

Richard Lawrence Kasul

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Fisheries and Wildlife

1976

ABSTRACT

HABITAT FACTORS ASSOCIATED WITH MORTALITY OF
SOUTHERN MICHIGAN WILDLIFE ON AN INTERSTATE HIGHWAY

BY

Richard L. Kasul

There were 372 individuals from 25 mammalian species
and 142 individuals from 14 avian species known to be killed
by vehicles from August 20, 1975 to March 15, 1976 on a
15.5 mile (25 Km) length of Michigan Interstate 96 in a
rural-agricultural area of Ingham County. An estimated
40% of the mammals which were mink size (0 68 Kg and 60.9 cm
overall body length) or larger, and birds which were screech
owl size (55.5 cm wingspan) or larger, and at least 80% of
smaller mammals and birds were not recovered.

The group of larger mammals and birds included 379
individuals of the 19 species in the general size range
of those normally removed daily from the highway by the
Michigan Department of State Highways and Transportation.
They were large enough to be easily visible to motorists,
and as such, were considered to be potential direct or
indirect threats to motorist safety. Among this group,
the most commonly observed were; cottontail rabbit (67),
raccoon (64), opossum (50), muskrat (46), and ring-necked
pheasant (38). Larger mammals and birds were killed in
the greatest numbers (59) from August 20 to 31, 1975, and

the kill continually declined through the next February 15

ii

iii
(none during January 1 - 16), when it again increased after
the snowpack melted.

Analysis of variance on the number of large and medium
sized animals killed per mile indicated that the three
habitat factors on which this variable was measured, road-
side vegetation, median vegetation, and median width, all
affected the distribution of vehicle-killed animals on the
highway. In general, the largest number of animals killed
per mile occurred where the roadsides were wooded, and
especially where woods also occurred in the median. This
was probably because the kill was composed largely of
animals which use wooded habitats for food and/or cover.

An average kill of 33.4 individuals per mile occurred
where there were woods on at least one roadside. This was
significantly greater than a mean animal kill of 17.8 per
mile on highway sections where only old field and/or cr0p-
land occurred along both roadsides. The mean animal kill
of 29.4 per mile on highway sections with a mowed median
less than 100 feet (30.5 m) wide was significantly greater
than the mean animal kill of 20.8 per mile found on highway
sections with an unmowed median less than 100 feet wide.
The largest number of animals killed per mile (76.6) oc-
curred where there were woods in the median in addition
to at least one roadside. There did not appear to be any
seasonal change in the percentage of animals killed in
relation to wooded and non-wooded highway segments during

the study period.

iv

These results suggest that movement of animals, in
particular the movement of medium-sized forest mammals,
may be affected by the interstate highway. Progressively
wider medians result in an overall wider right-of-way,
which reduced the incidence of vehicle-caused wildlife
mortality possibly by inhibiting movement in some species.
Wooded medians appear to facilitate animal movements, either
by l) reducing the size of the highway as a barrier to
movement, or 2) by attracting animals which would normally
cross the highway at other locations. Some management

implications of this study are also discussed.

ACKNOWLEDGEMENTS

I wish to express my sincere graditude to the
Michigan Department of State Highways and Transporta-
tion. In particular I am indebted to Mr. Bud Bekaert
for cooperation given in the field, and Mr. Jan Raad
and Mr. John Baker for helping to initiate the study.

I am especially grateful to my major professor,

Dr. Leslie Gysel, and the members of my guidance commit-
tee, Drs. Donald Beaver and Walter Conley, for sugges-
tions concerning experimental design, for encouragement
throughout the study, and for helpful criticism of the
thesis manuscript.

Appreciation is extended to Mr. Joel Young of the
USDA Agricultural Stabilization and Conservation Service
for the loan of aerial photographs which were essential
during several phases of the study. Additionally, I
wish to acknowledge Dr. John Gill for statistical advice,
and Mr. Glenn Dudderar for suggestions concerning the
interpretation of my results. Use of the Michigan State
University computer facilities was made possible through
support, in part, from the National Science Foundation.

I am particularly indebted to Arlene Wong Hestbeck,
for critically reviewing and typing the thesis manuscript.

V

vi

Finally, I wish to thank fellow graduate students,
Jay E. Hestbeck, J. Edward Gates, Philip B Davis, Gary P.
Friday, and Thomas M. Butynski, for their continual advice

and encouragement.

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . viii
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . ix
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . 1
STUDY AREA . . . . . . . . . . . . . . . . . . . . . 4
METHODS. . . . . . . . . . . . . . . . . . . . . . . . 8
RESULTS. . . . . . . . . . . . . . . . . . . . . . . . 13
Seasonal Variation. . . . . . . . . . . . . . . . 16
Roadside Vegetation and Median Width. . . . . . . 18
Roadside and Median Vegetation. . . . . . . . . . 24
Average Daily Probabilities . . . . . . . . . . . 26
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 31

LITERATURE CITED 0 O O O O O O O O O O O O O O O O I O 38

vii

Number

LIST OF TABLES

Experimental design and distribution of
sampling units among roadside vegetation,
median vegetation, and median width.

Numbers of vehicle-killed mammals and birds
observed during each two-week period from
August 20, 1975 to March 15, 1976 along
15.5 miles of Interstate 96 between the
Interstate 496 and M-52 interchanges.

Analysis of variance of animal kill per
mile in relation to roadside vegetation
and median width.

Mean animal kill per mile in relation to
roadside vegetation, median width, and
median vegetation.

Mean kill per mile for the more important
species of mammals and birds in relation
to roadside vegetation, median vegetation,
and median width.

Analysis of variance of animal kill per
mile in relation to roadside and median
vegetation.

Average probabilities of finding at least
one large and/or medium sized mammal or
bird per sampling unit per day from August
20 through November 30, 1975.

Average probabilities of finding at least
one large and/or medium-sized mammal or
bird per sampling unit per day from
December 1, 1975 through March 15, 1976.

viii

Page
12

14

20

21

23

25

28

29

Number

1

LIST OF FIGURES

Location map of Interstate 96 study
area with generalized vegetative
features.

Numbers of vehicle-killed mammals and
birds observed during each two-week peroid
from August 20, 1975 to March 15, 1976,
along 15.5 miles of Interstate 96 between
the Interstate 496 and M-52 interchanges.

ix

Page

17

INTRODUCTION

The U.S. interstate highway system is ultimately ex-
pected to consist of 41,000 miles of divided 4- or 6- lane
highways encompassing over 1 million acres of land. In fol-
lowing the principles of the complete highway concept (Anon.
1966), the modern interstate highway system often includes
large acreages in median and roadside vegetation that may
be suitable for development as wildlife habitats. Contin-
ued declines in the amount of native wildlife habitat has
brought recommendations to manage highway rights-of-way for
wildlife (Egler, 1953, 1954, 1957). Such developments may
supplement local native habitat or may serve as refuges for
plants and animals. Additionally, they may be aesthetically
desirable to motorists. The possibilities for management
of roadside wildlife cover have been explored in relation
to nesting cover for waterfowl in North Dakota (Oetting and

Cassel, 1971) and ring-necked pheasant (Phasianus colchicus)

 

in Illinois (Joselyn gt 31., 1968).

Highway officials concerned for motorist safety have
been reluctant to allow wildlife habitat development along
highway rights-of-way because they believe that development
of roadside cover to attract wildlife.may result in a greater
number of vehicle-animal collisions (Joselyn, 1969). While

1

2
this argument may be valid, present highway design and main-

tenance practices, including the planting of trees and other
cover and the use of vegetation for screening, may unwitting-
ly encourage the high rates of animal mortality that highway
officials would like to avoid. Unfortunately, there is lit-
tle data available to relate wildlife mortality to the veg-
etative and natural features along highways.

In eastern and mid-western states, the most serious wild-
life-motorist problems are caused by white-tailed deer (gggf

coileus virginianus). Accident reports from toll roads and

 

turnpikes for years 1964 through 1966 indicated that deer
caused an average of 5.7% of all reported vehicle accidents
(Joselyn, 1969). An additional 0.8% of accidents were caused
by animals other than deer; these included ring-necked phea-

sant, raccoon (Progyon lotor), and striped skunk (Mephitis

 

mephitis).

Because deer have caused a significant number of vehicle
accidents, several studies have examined the behavior of deer
in relation to highways and the nature of deer-vehicle col-
lisions in relation to natural features associated with high-
ways (Carbaugh 33 31., 1975; Puglisi 33 31., 1974; Vaughn,
1970; Bellis and Graves, 1968). Similar data have not been
obtained for animal species other than deer. The large num-
bers of these mostly medium sized mammals and birds has been
well documented. These animals account for far more vehicle-
wildlife collisions than deer, may cause vehicle accidents,

and in Michigan, are regularly removed from highways at

3

considerable expense in response to the wishes of the motor-
ing public.

This study was conducted to examine the distribution of
highway mortality of medium— and large-sized animals in rela-
tion to vegetation adjacent to the highway right-of-way, veg-
etation of the highway median, and width of the highway me-
dian. My intent was to provide data to aid decision making
processes regarding the manipulation of major vegetation pat-
terns in the design, maintenance, and use of highway rights-
of-way to benefit wildlife, and motorist safety and aesthe-

tics.

STUDY AREA

The study area was a 15.5 mile (25 Km) segment of Inter-
state 96 that traverses rural-agricultural landscape in Ing-
ham County, Michigan, between the Interstate 496 interchange
and the M-52 interchange (Figure l). The study area was con-
structed during 1960-61 and opened to traffic in summer, 1962,
and includes a 2-lane eastbound travelway and a 2-lane west-
bound travelway separated by a vegetated median strip which
varies from 10 to 200 yards (9.1 to 182.9 m) in width. The
right-of-way extends an average of 26.7 yards (24.5 m) from
the outside edge of each travelway to a 4.5 foot (1.38 m)
vertical fence.

The surrounding area is characterized by flat to moder-
ately hilly topography and a complex of well-drained to some-
what poorly drained loam soils of glacial origin. Approxi-
mately 50% of the land directly adjacent to the right-of-way
(hereafter called roadside vegetation) was planted with crops;

mainly alfalfa (Medicago sativa L.), corn (Zea mays L.), hay,

 

(mostly Phleum prantense L.), and soy beans (G1ycine max L.).

 

 

A small portion of these and other fields were used as pas-
ture. Remaining roadside vegetation consisted of woods and
fallow or abandoned agricultural land. Abandoned fields

varied in age from recently abandoned fields containing a

4

ransom
as

assays.
NT:

 

A
J

 

   

Fog»...

D
.32....

 

   

mousumwm w>fiumumom> pmufiawumcmm nufi3 mmnm hpsuw mm mum» HmucH mo awe :ofiumooq

 

 

.
i - «I. a r
i .1 — . .Jfi . . 2.. : .
I — .
I 1 . .
v.3 I -3. . . .. .

 

 

 

 

 

 

. 8e _
.‘O Q
\ . .
\ . o . . ..
w a: a, .a. .
IC- . .. 1.
‘ .ifl ‘ {.1 i ‘ 3 .u. b
1‘ {l 1 .4. o‘\ "y‘..
. I... . . a. . 1.1% 1.. .
i‘. .s .5: . {.1 . {-34 ..
$3.... . . 1. . O .
. ,4 1. x .s. i. .
. . i . r c358
fl.

mucuHooHHmm flHu

33... So E
£503 a

mpcom pom .mmnouflp wmmcwwnp Home .mfimmuum.ll.ul
mmaumpcson m>aumuwmm> paw .mmmcmsoumucH

is... .3 I 3.3...
a

+2—

 

.mmcma Hw>muu mmzsmwm

.H mnsmwm

6

few early successional shrubs or sapling trees to more de-
veloped field communities containing extensive (up to 60%)
development of shrubby cover. The 45 to 450 yard (41.1 to
411.9 m) long segments of woods occurred along the highway
primarily on low-lying areas not suited for agriculture and
consisted mainly of young stands (20-38 cm dbh) of American

beech-sugar maple (Fagus grandifolia Ehrh.-Acer saccharum

 

 

Marsh.) woods mixed with aspen (Populus tremuloides Michx.),

 

and lesser amounts of hickory (Carya spp. Nutt.), red oak

(Quercus rubra L.), white oak (Quercus alba L.), and on wet-

 

 

ter sites, silver maple (Acer scharinum L.). Several small

 

streams and drainage ditches crossed or paralleled the right-
of-way at irregular intervals. A generalized map of these
patterns is shown in Figure 1.

Control of vegetation on the right-of-way consisted of
mowing and herbicide treatment to control vegetation around
guard rails. A 5 yard (4.57 m) strip of vegetation adjacent
to the driving surface was mowed each June and August by
personnel from the Michigan Department of State Highways and
Transportation. In addition, highway median which is less
than 100 feet (30.5 m) wide was completely mowed. Approxi-
mately 40% of the median on the study area was completely
mowed while the remainder was unmowed except for the 5 yard
strips adjacent to the driving surface.

Vegetative cover on unmowed portions of the right-of-

way consisted of smooth brome (Bromus inermus Leyss.) inter-

 

sPersed with scattered patches of less common herbaceous

7
cover including alfalfa, red clover (Trifolia pratense L.),

 

and other locally cultivated crops. Roadside ditches con-

tained common cattail (Typha latifolia L.) and scattered

 

patches of reed (Phragmites australis (Cav.) Steudel). Ex-
cept for 11 small woodlots found in the median and scattered
ornamental trees planted.on approximately 2 miles (2.51 Km)
of the highway, little woody vegetation occured on the right-

of-way.

METHODS

Wild and domestic animals killed by vehicles on the
study area were collected 4-5 days each week from August 20,
1975 through March 15, 1976, except for the one week period
from December 20 through 27. Vehicle-killed animals were
observed while driving the right lane of the highway at the
minimum 45 mph (72.4 Km/hr) legal speed. Each individual
which I observed was buried or removed from the right-of-way
so that it would not be counted again on subsequent days.
Approximately once each week I made observations while driv-
ing along the emergency shoulder to detect animals lying
among the roadside vegetation. A yellow flashing caution
light mounted on top of my vehicle was operated while driving
the emergency shoulder or stopping to remove vehicle-killed
animals from the highway. In addition, a fluorescent orange
vest was worn at all times to ensure maximum safety for my-
self and other motorists. Observations were generally made
during the morning hours at off-peak traffic times.

Permission to st0p a vehicle and trap animals within
the highway right-of-way of the study area was given in wri-
ting by the Michigan Department of State Highways and Trans-
portation. The permit was valid only during daylight hours
and off-peak traffic times. For the duration of the study,

8

9
the Highway Department suspended daily collection and dis-

posal of vehicle-killed animals on the study area so they
would not bias my observations. The Ingham County Animal
Control Department and the Michigan State Police were noti-
fied of this study, and they agreed to refer motorist com-
plaints regarding vehicle-killed animals to me.

Data recorded on prepared field sheets for each indi-
vidual were species, location, distance to nearest permanent
water, distance to nearest woodlot and, where possible, sex
and age. The location of each observation was recorded in
relation to type of roadside vegetation, type of median veg-
etation, and width of median. These features were surveyed
using vehicle reconnaissance and aerial photographs, classi-
fied and mapped for use on field sheets, prior to initial
collection of vehicle-killed animals. Several areas were
later reclassified as errors in the initial classification
were encountered.

Roadside vegetation was classified as: 1) both sides
wooded; 2) one side wooded, one side old field; 3) one side
wooded, one side agricultural; 4) both sides old field;

5) one side old field, one side agricultural; and 6) both
sides agricultural. All roadside classified as wooded was
composed of closed stands of aspen or beech-maple woods as
previously defined. Old fields consisted of abandoned agri-
cultural land dominated by naturally occurring herbaceous
vegetation. Many segments of roadside classified as old

field also contained varying percentages of early

10

successional woody cover which was visually estimated.
Naturally vegetated fields subject to light grazing pressure
were generally classified as old field while heavily used
pastures and land supporting all crop types were classified
as agricultural.

Vegetative composition of the median was described as
herbaceous or wooded. Width of the median was described as
less than 100 feet or greater than 100 feet. Herbaceous
median less than 100 feet was completely mowed each June and
August. Herbaceous median greater than 100 feet was unmowed.
Altogether, there were 84 variable-sized sampling units des-
cribed relative to 6 types of roadside vegetation, 2 types
of median vegetation, and 2 types of median width. Sampling
units were delineated by natural boundaries between roadside
vegetation, boundaries between mowed and unmowed median, and
woodlots found on the median. The shortest sampling unit
extended 89 yards (81.4 m) along the right-of-way. Shorter
segments of vegetation were not classified or mapped.

In order to detect animals not observable from the ve—
hicle and to examine the distribution of animals missed from
the vehicle in relation to roadside vegetation, median veg-
etation and median width, 14 segments of right-of-way total-
ing 1.76 miles (11.3%) of the study area were searched on
foot. A systematic ground search of both roadsides and me-
dian was conducted 5 times for each area.

The distribution of vehicle-killed mammals and birds

observed from the vehicle were analyzed relative to roadside

ll

vegetation, median vegetation, and median width (Table 1).

The total number of animals killed for selected species was
used to compute the variable animal kill per mile for each

sampling unit as

(number of animals killed)(l760_yards/mile)
length of sampling ufiit in yards

where the length of each sampling unit was measured from
aerial photographs to the nearest 10 yards. Analyses of
variance were used to test null hypotheses of no difference
between animal kill per mile in relation to roadside vege-

tation, median vegetation, and median width.

Table

1.

12

Experimental design and distribution of sampling
units among roadside vegetation, median vegetation,
and median width.

 

 

Vegetation on
opposite sides
of right—of-way

Number of sampling units (% of study area)

 

Herbaceous median

 

(100 ft. wide 2100 ft. wide

Woodedimedian
£fl00 ft. wide

 

Woods
Woods
Woods
Field
Field

Agric

woods
field
agric
field
agric

agric

 

 

(6.1)
(4.0)
(2.7)
(10.6)
(13.9)
(12.9)

{'4- (‘2703' " "

(6.0)

— — ’5' '("2'.§)'1
l
2 (0-9)l

__ _3 (4.5)}

 

 

 

 

 

 

r"fi Design for ANOVA of
. ' I O
L__J roadSide vegetation

‘Design for ANOVA of
roadside vegetation

animal kill per mile in relation to
x median size.

animal kill per mile in relation to
x median vegetation.

RESULTS

There were 364 vehicle-killed individuals from 23 mam-
malian species and 133 individuals from 13 avian species that
were observed during the daily surveys of the study area.
Combined, a minimum of 0.20 animals per mile per day were
killed during the 7 month study period. Of these, 379 (76.0%)
were used in the analysis of animal mortality in relation
to roadside vegetation and median characteristics. These
consisted of 14 species of mammals having overall body sizes
averaging larger than 24 inches (60.9 cm) overall length
and 1 1/2 1b. (0.68 Kg) in weight and 5 species of birds with
wingspans of 22 inches (55.5 cm) or larger (Table 2). These
species were within the general size range of those animals
normally removed daily from the highway by Highway Depart-
ment personnel. They were large enough to be easily visible
to motorists, and, as such, were considered to be potential
direct or indirect threats to motorists' safety. They were
also the animals for which the most accurate data was ob-
tained. Field observation suggested that smaller animals,
particularly small rodents and passerine birds, were not
readily observable from the vehicle; they were quickly Oblit-
erated by vehicle traffic.

There were 17 vehicle-killed individuals observed in

13

l4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NH NAonm msuovH3O commuom
mmEuouHmHumum Hmpuo H Hmumummuo musHmocouymHOE mmOCHMHm
NH AmsoHumsmn msmonomvaOE cumummm
N HmHummH9> macadamvmcHHumum muo>HuommcH “mono
m hmsomumeOU ummmmmv3ouummm mmsom
N AmHuoumumHE mucusBNchon cmoHHmfid
H HmHHmmHon mDHDHmMHvumn pom
H ma mcapumo mHHmcHUHmovachumu wumumouanu “mono
H HMHSQHM msumuovaHoHHo cumnuuoz .
m Hmcmme mHHmcuopmwxumHzocmmE cumummm
N HmHsomHnw msHmomHSvaonmum COEEOU «m HHHouoH comwoumvcoooomm
meHOMHummmmm umpno HH HHmHanmweH mHuHmmwzvxcsxm ommHHum
m HAGOmH> mepmszvxcHz
mm HmDOHnoHoo mscmHmmnmv m moon m m m: wmmm3 Hm mco
N .c...wmm.w.flomm-.mwm. m r . A..wx.w “.mwwwzczwm.w. w...“
m.HH 0 U o H HHmemv mwcmeevuwmomm
H HcQNmemmHuwo mmmzvuwanmmcHx pmunm m Amdumo mHHmmvumo UHDmOEOQ
mmEHOMHHomHou umpuo H H HHm>Hsm mmmw5>vxom pom
H stmucmmumowumcHo cowoouovxom hone
NH NAmH>HH MMEwwowvm>ow xwwmu OH H AmsoHumeOU chmovap UHummEoo
m. QESH U U o H muo>chmu umpuo
H NHmmcomm devxosc @003 n AmscchHme> msmHHooocov
v NamoaocmdumumHm mmcflvou6HHmz H Hoop pmHHmunmanz
mmEHOMHmecd umpuo meuompoHuud umpuo
wawwx HwEmc OHMHucmHomvamc QOEEOU pwwwmx Hmsmc OHMHucmHomvawc COEEOU
mpuHm mHmEEmz

 

 

.mmmcmnoumucH lez can mmv wuwumnmucH mzu
:wmsumn mm mumumumucH mo mmHHa m.mH mcon mHmH .mH scum: on mHaH .om umsms¢ scum
UHoumm xmmslosp comm mcHusp Uw>ummno mpan Ucm mHmEEME UOHHHxlmHoH£m> mo mumnfidz .N mHnma

15

 

.Eo m.mm mo cmmm

mcHz mmmum>m mm mm: SOHBB .H30 nommuom on» mmB umeHmEm one .moHumHuouowumso GMHme cam

coHHMHmam> onmcwou ou coHuMHmH CH mHHE mom HHHx HmEHcm mo mHmmHmcm GH poms mmHowmm cmH>¢

.numcmH HHmum>o Eu m.oo paw uanws cH mm mo.o
mommum>m BOHaz .xcHE map was umeHmEm one .onumHumuomumno cMHUwE cam COHumummm> monomou

0p coHumHoH cH mHHE Mom HHHx HmEHcm mo mHmmHmcm on» aH poms mHmEEmE UmNHmuadHomE cam momma

 

«N

wv
om

om

hm

 

(HmecoE muoEmmzvxosnooooz
H.Qmm msommaoummvmmsoe owHOOMImuHBB

HmSHMHHum msusHommHEvaHmuuHsvm pom
HmDOHmw>Hoc msuummvuwu mm3uoz
HHMOHnquHN muumpcovumuxmsz

HummH: mousHomNkuHHDUw xOH cumummm

HmsumHHum mmHEvaxcsEmHno suwummm
mHucmoom pruo

 

 

 

 

 

H

 

HHmHHMHmmmHmE mHanproonmmomo
wHHmHmsmHmE pruo

 

HAmscmpHuon mamMHH>HNmV

qubmu HHmucouuov
msmuosomMH umpuo

 

. ADHGOOV N GHQMB

16

median and roadside vegetation during the 5 ground searches
of 1.76 miles of highway, conducted to provide a check on
the accuracy of the visual counts from the vehicle. These
included 3 species which were not observed from the vehicle:

1 meadow vole (Microtus pennsylvanicus), 1 house mouse (Mus

 

 

musculus), and 2 common crows (Corvus brachyrhynochos).

 

Only 7 were medium-sized mammals or birds of concern to mo—
torist aesthetics or safety. The ratio of the number of
animals counted during the ground searches to the total num-
ber of animals counted (during ground searches and from the
vehicle) suggests that approximately 40% of large and medium-
sized mammals and birds and at least 80% of small mammals

and birds were not recovered during the surveys from the ve-
hicle. Because only 7 large- and medium-sized mammals and
birds were counted during the ground surveys, the distri-
bution of these animals in relation to roadside vegetation,

median vegetation and median width could not be evaluated.

Seasonal Variation

 

For the 7 month period during which observations were
made, vehicle-caused mortality of animals was highest during
fall and early spring (Figure 2). Most of the spring obser-
vations were the result of an apparent pre-breeding muskrat
dispersal. Mortality was lowest during the winter months.
Only 9.8% of all observations occurred from November 16

through February 15 when the ground was snow covered. A

 

 

 

up:

ea
3.32 $1.23. 3-23 5.88— mzsa was: 3.38 cm 252 m3 8: filo—.8 2.28 8....ng 9.28 3.83

    
    

p. as
0....
O ‘
.:
.0...
O O
U C 0
VIC...
v.....
I...
0.

 

 

 

 

l7

mUHHQ ocm mHmEEmE uwHHmEm HHa a

HmwumH
no AcmmmmcH3 E0 m.mmv wNHm H30 nommuom mpan Ucm HmmHmH
.Ho EumcmH HHmum>o EU m.oo cam gm No.3 mNHm xcHE mHmEEmz I

.mwmcmnonmusH lez pom mmv mumumumucH map cmmzumn mm mumum

InmucH mo mmHHE m.mH mcon .mhmH .mH noun: 0» man .ON umsmsd Eoum UHOHOQ
xmm3no3u some msHHso pw>ummn0 mpan cam mHmEEME pmHHHxImHoHnw> mo mHOQEdz

2.2

 

S

8

on

2..

.N musmHm

SWWINV (IJ‘IIIN-J'IOIHJA 10 839mm

18

deep snowpack, which reached a depth of 14 inches (35 cm)
in mid-January, probably inhibited the movement of normally
active animals. Animals killed during or prior to the many
winter snowstorms were removed from the highway by snOWplows
and lost to observation. Because inclement weather appeared
to inhibit movement of most animals, removal by snowplows
was considered insignificant.

Movement of animals from old fields to the dense cover
of wooded areas was hypothesized to occur as the season changed

from fall to winter, and a concomitant shift in animal mor-

 

tality was expected to occur. However, when the total num-

 

ber of animals killed per mile was divided into two groups,
those occurring adjacent to wooded roadsides and those that
did not, the percentage of animals killed adjacent to woods
was nearly constant (approximately 70%) in each two week
period. Although there was no data for the summer months,
there appeared to be no seasonal shift in overall animal
activity with respect to wooded versus non-wooded sections

of highway.

Roadside Vegetation and Median Width

 

Animal kill per mile was calculated for each sampling
unit using the total number of vehicle-killed individuals
from the 19 largest Species of mammals and birds (Table 2).
Analysis of variance on animal kill per mile in relation to

roadside vegetation and median width was conducted on the

19

77 sampling units having an herbaceous median (Table l).
The distribution of residuals for animal kill per mile was

not normally distributed (Kolmogorov-Smirov Zn = 1.46,

=75
POV2 (0.028); thus, the square root of each value of animal
kill per mile was computed to normalize the distribution of
>o.560) and

residuals (Kolmogorov-Smirov Z = 0.79,

n=75 Pd/z
produce a homogeneous variance. At the = 0.05 level of
acceptance, there were significant effects due to roadside
vegetation and median width and no significant interaction
(Table 3).

The mean animal kill of 29.4 per mile on highway with
a median width less than 100 feet was significantly larger
than the mean animal kill of 20.8 per mile found on highway
with a median width greater than 100 feet. The mean animal
kill per mile for each combination of roadside vegetation
and median width is given in Table 4. Pairwise comparisons
among means for classes of roadside vegetation were not sig-
nificantly different at the 0(= 0.10 level of acceptance
using the conservative Scheffe's multiple comparison test
for group means. This does not necessarily indicate that
there were no actual pairwise differences. Small sample
sizes and wide structural heterogeneity among sampling units
may have contributed to large within class variation which
masked actual pairwise differences in animal kill per mile
among classes of roadside vegetation. Scheffe's multiple

comparison test indicated that a mean animal kill of 33.4

per mile on segments of highway with woods occurring on at

'1.” 1r. 4hK'.- - .

 

.fi

 

20

Table 3. Analysis of variance of animal kill per mile in
relation to roadside vegetation and median width.

 

 

 

Source of df Sums of Mean F Probability
Variation Squares Square of a larger F
Roadside vegetation 5 59.47 11.89 2.73 .053 P2. .025
Median vegetation 1 17.96 17.96 4.13 .05) p) .025
Roadmée X media“ 5 33.95 6.79 1.56 .25) P), .10
interaction
Residual 63 274.14 4.35

Total 75 381.90

 

21

Table 4. Mean animal kill per mile in relation to roadside
vegetation, median width, and median vegetation.

 

 

Vegetation on Number of sampling units (% of study area)
opposite sides Herbaceous median Wooded median
of right-of-way 100 ft. wide 100 ft. wide 100 ft. wide

 

 

 

Woods - woods 39.5 24.6 96.0
Woods - field 37.6 31.8 68.4
Woods - agric 52.0 19.5 56.3
Field - field 24.6 11.5
Field - agric 21.2 17.8

Agric - agric 11.9 15.0

 

22

least one side of the right-of-way was significantly greater
(P((0.05) than a mean animal kill of 17.8 per mile on high-
way segments where there were no woods.

A large portion of the total animal kill associated
with wooded roadside was composed of raccoon, opposum, fox
squirrel, and cottontail rabbit (Table 5). These were among
the species killed in the largest numbers (Table 2), and,
except for the cottontail rabbit, they are generally asso-
ciated with wooded habitats.

Sampling units, on which animal kill per mile was cal-
culated, varied in length from 89 — 1489 yards (81.4 -
1361.5 m). Wooded areas had the shortest average sampling
units (186.5 m) and the largest mean killrate (33.4 animals
per mile) while agricultural areas had the longest average
sampling units (491 m) and the lowest killrate (13.6 animals
per mile). Since type of roadside vegetation and size of
sampling unit varied together, the significant effect due to
roadside vegetation indicated in the analysis of variance
may have been due, at least in part, to variation in size
of the sampling units. Smaller sampling units would indi-
cate greater interspersion of vegetative types along the
highway. Greater interspersion of habitats may result in
greater animal abundance (Leopold, 1933; Allen, 1962;
Dasmann, 1964), and concomitantly, greater mortality on those
sections of highway. However, regression analysis conducted
for each class of roadside vegetation did not show any sig-

nificant variation in killrate over the range of sampling

23

.ousuHsoHummumHsuHsoHumm

w .musuHSOHHmMIpHmHm

m .UHOHmtonHm u v

 

 

 

 

 

 

 

.wuspHSOHummlmoooa n m .cHwHMImpooa u N .mpooslmpoo3 u H "coHumuwmm> prmpmou mom mmoooH
o o o o m.o o o o 0 £096 @003

o o o o m.o o o o m.o xozcooooz

o o m.N m.o m.o o m.o o m.o Home pmHHmuuwuHsz

o o m.m m.o m.o m.o 0.0 N.H o xcsxm emmHuum
H.n o o.h m.o m.o m.o o m.H m.H H30 commuom

o o o m.o m.o >.H o m.o m.H m>oo xoom

o o o m.~ m... m.o e.m e.~ m.o ucememse emxomaumeHm
N.hH m.m o.vH m.N m.H n.H m.m m.N m.n coouomm

o o m.Nm m.N m.o v.0 m.H N.m m.m Edmmomo

o o o m.o o.m b.H m.m m.> w.N pmuxmsz

o o o o m.o v.o o o m.o xcHz

o o o o o o v.0 o v.N UHMHHmz

o o o o m.o o o o o xom smug
e.m o H.HH m.o m.o o e.m m.H m.n HmuuHsem xom
>.m m.mH m.mH o.N v.w H.m N.m m.v m.m uHQQmu HHmucouuou

o o o m.o o o o o o Hmmpmm

m N H m m w m N HH
mcH3 .um OOHM.CMHomE omBOEcs prB .um OOHV CMHme @6306 mmHommm
fiuH3 comumummw> wwmmpmom mcm>mn coHumummm> mommpmom .
mHHE mom HHHx com:
.nupH3 QMchE paw .coHumummm> cmemE .coHumuwmw> mpHmcwou ou QOHumHmn
cH mpuHQ pom mHmEEmE mo meowmm ucmuuomEH muoE map How mHHE mom HHHx cmmz .m mHQma

24
unit sizes present on the study area. This suggests that

the significant variation in killrate per mile in relation
to the different types of road side vegetation was not in-

fluenced by size of the sampling units.

Roadside and Median Vegetation

 

Where the highway was routed through a wooded area, the
woods were either left in the median or the median was com-
pletely cleared of trees and maintained in herbaceous growth.
Only where woods occurred on at least one side of the right-
of-way was there a possibility of woods on the median. An-
alysis of variance was used to test the hypothesis of no
difference between animal kill per mile in relation to wooded
roadsides and median vegetation using the 29 sampling units
having woods on at least one side of the right-of-way and
a median greater than 100 feet wide (Table 1). Data from
sampling units with a median less than 100 feet wide were
not used because of the significant effect of median width
on animal kill per mile found previously.

The distribution of residuals for animal kill per mile
was not normally distributed (Kolmogorov-Smirov Zn=29 = 1.31,
P092(0.064); thus the square root of each value of animal
kill per mile was computed to normalize the distribution
of residuals (Kolmogorov—Smirov Zn=29 = 0.52, Pay2)0-950)
and produce a homogeneous variance. A significant effect

(P‘(0.05) due to median vegetation was detected (Table 6).

25

Table 6. Analysis of variance of animal kill per mile in
relation to roadside and median vegetation.

 

 

 

Source of df Sums of Mean Probability
Variation Squares Square of a larger F
Roadside vegetation 2 17.85 8.93 1.82 .25 a pg .10
Median vegetation 1 80.08 80.08 16.32 (.001
Road5199 X media“ 2 2.86 1.43 0.29 >.999
interaction
Residual 23 112.83 4.91
Total 29 222.19

 

26

There were an average of 76.6 animals killed per mile where
the median was wooded, and an average of 25.4 animals killed
per mile where the median was herbaceous. The average kill
per mile for each combination of roadside and median vege-
tation is given in Table 4. Although segments of highway
with a wooded median composed only 8.2% of the study area,
22.25% of the observed animal mortality occurred in those

portions.

Average Dai1y Probabilities

 

Average daily probabilities of finding at least one
vehicle-killed animal on a sampling unit were calculated
on highway sections for every combination of roadside and
median characteristics during two time periods, August 20
through November 30, 1975, and December 1, 1975 through
March 15, 1976. The calculations utilized data only from
those days where a survey was conducted on the immediately
preceding day. These probabilities are based on the for-
mula S

ZN
P=__];.=.i___
(S)(D)
where:

P = the average probability of finding at least one
vehicle-killed animal each day on a single sampling
unit;

N = total number of days for which at least one ve-

hicle-killed animal was observed on each sampling
unit;

U)
ll

# of sampling units;

27

D = total observation days, 65 for the period August 20
through November 30, 1975, and 62 for the period
December 1, 1975 through March 15, 1976.

For the fall months, the probability of finding at
least one vehicle-killed animal on any day was lowest
(P = .012) on sampling units having a wide, herbaceous
median, and both roadsides wooded or one wooded and the
other in agriculture. It was highest (P = .108) on sam-
pling units having a wooded median and one wooded roadside
with the other in agriculture (Table 7). For the winter
months, the probability of finding at least one vehicle-
killed animal was much lower than for fall months in all
cases except where the median was wooded and the roadside
consisted of wooded areas and field (P = .032 in winter
vs. P = .023 in fall). The highest probability value
(P = .048) occurred when the sampling unit contained a
wooded median and one wooded roadside with the other in
agriCulture (Table 8).

These probabilities were calculated using sampling
units of different lengths, and thus the values do not have
a common distance denominator. As previously indicated,
sampling units averaged the shortest for wooded areas,
intermediate for old field areas, and longest for agricul-

tural areas. As expressed relative to a standard length

for a sampling unit according to the formula

8
2

 

th.omvaH.
AhH.NNvmmo.

HmN.wVHNo.

Hmm.NVNvo.
Hmm.vvmmo.
Hmm.NVNHo.
Amm.HVNHo.
Aom.mvao.

HmH.mvHNo.

Amw.vaNo.
Hoo.vaHo.
Amm.Hv0No.
Hom.mvao.
Hon.mvhHo.

Ahm.thmo.

UHHmm
UHHmm

UHme

UHHmm

pHmHm

mpooz

UHumd
pHmHm
UHmHm
moooz
mpooz

mpooz

 

mpH3 .um OOHN
cmemE Umpooz

033 .0.“ boHN

mUH3 .fim bOHV

cmemE msomommuwm

 

AuHcalmcHHmEom
mo mNHm Mom vcmEpmsflpm umumm muHHHnmnoum

ummon Hm>o wuHHHnmnoum CH ommmuocH Houommv muHHHQmQOHm

mm3lm0IunoHH mo
mmon wuHmommo

so COHumuwmm>

 

 

.om Honfim>oz
nmsounu om unsoom Eoum hop Mom uHcs mcHHmEMm mom UHHQ Ho HmEEmE

omNHmIEsHme H0\pcm mman moo ummoH um mchch we mOHuHHHQwQOHm mmmuo>¢

.h mHQMB

29

 

loo.HvHoo. lem.vaHo. oHnoo I oHnoH

 

 

 

loo.mvmmo. lom.HcNHo. oHnoe I oHoHn

AIIIIV oo.o Hmo.NVNHo. pHmHm I pHmHm

Ae.chmHo. loo.HVooo. Amm.mvmoo. oHnoo I nooos

Hmm.vmvmmo. Avo.mVOHo. Hmv.mvNHo. cHme I mpooz

Av.mVoHo. Avm.mvvoo. Hmm.mvao. mpoo3 I mpooz
oUH3 .um oodM osz .um OOHM opH3 .um opHv

cmwmoe mooooz cmmmmE msomomflumm mmsIHOIuann mo

HuHcs mcHHmew mopHm wuHmommo

mo oNHm mom ucoeumsflpm “mums haHHHnmnoum so coHumummm>

umo30H Hm>o MHHHHQMQOHQ GH mmmouocH Houommv muHHHnmnoum

 

 

.mhmH .mH noun:
smsouzu mbmH .H Hmnfimomo Eoum moo Mom uHcs mcHHmem mom UHHQ no HMEEME

pmNHmIESHowE u0\©cm moHMH wco ummoH um msHUch mo mOHuHHHnmnonm wmmum>4 .m OHQMB

30
P

 

average length of sampling unit

 

P

 

average length of sampling unit
lowest
non-zero value
the factor increases in probability over the lowest non—
zero probability after adjustment for sampling unit length
differences show that, per mile of highway, the highest
probability of finding at least one vehicle-killed animal

occurred where the roadside was wooded; it was particularly

high where the median was also wooded.

DISCUSSION

The higher overall mortality of mostly medium-sized
mammals and birds that occurred on highway sections with
wooded roadsides is related to several factors: the spe-
cies which composed the bulk of the observed mortality,
the habitat preferences of these species, and their move-
ment patterns and home range characteristics. Wooded areas
adjacent to the highway often occurred along small streams
or on wet soils where temporary or permanent sources of wa-
ter were available nearby. These areas provided the major
source of habitat for raccoons, opossums, fox squirrels,
screech owls, and the other medium-sized woodland animals
which composed the bulk of the observed wildlife mortality.
It appeared that these animals were killed in numbers di-
rectly related to their abundance along the highway; the
overall observed kill being highest where the local habitat
allowed the highway population densities. In addition to
habitat specificity, home range and movement characteristics
of these animals also appears to be an important factor
which influences the higher rate of vehicle-c sed mortality
on wooded sections of highway.

In similar studies involving white-tailed deer (Allen
and McCullough, 1976; Puglisi 33. 31., 1974), there were

no differences in mean deer kill per mile among the same

31

32

three types of roadside ocver difined in this study, which
were woods, old field and agriculture. Michigan deer have
a home range averaging 3.5 square miles (Dahlgerg and
Gottinger, 1956); and because they are highly adaptable
browsers and grazers, deer utilize the food and cover pro-
vided by the habitats which are typical of mixed-agricul-
tural areas in southern Michigan. Due to their more gen-
eral habitat requirements, vehicle-caused deer mortality
could not be related to any specific habitat characteris-
tics. In contrast, species observed in this study general
ly have much smaller home ranges associtated with more spe-
cific habitat requirements. For example, raccoons and opos-
sums, which are predominantly woodlot or forest animals,
have home ranges averaging 100 - 200 acres (Ellis, 1964)
and 10 - 40 acres (Lay, 1942) respectively. Because these
and other medium-sized species are mainly found in wooded
areas and have similarly small home ranges, vehicle-caused
mortality was highwst where woods occurred adjacent to the
highway.

While, in general, variation in animal kill per mile
between types of roadside vegetation appears to have been
a function of the abundance of medium-sized animals in these
different habitats, the very high rate of animal mortality
along highway sections with a wooded median was probably
not entirely related to wildlife numbers, but to the movement
patterns of amimals relative to wooded medians. The woods

which occurred along highway roadsides in areas where the

33
median was also wooded did not appear to provide better

wildlife habitat than the woods which occurred along highway '
sections where there was an herbaceous median. Also, the
wooded medians, which were approximately 25 yards (22.9 m)
wide and 25 - 100 yards (22.9 - 91.4 m) long, could not pro-
vide sufficient additional habitat along the highway to in-
crease animal numbers and produce the large number of dead
animals observed along these sections of highway. My invest-
igations of the wooded medians indicated heavy browsing on
seedling trees by cottontail rabbits, light browsing by deer,
and minimal use of oak acorns by fox squirrels. However, these
areas did not appear to provide permanent ocver for any of
these species. Thus, while wooded medians nay be utilized
somewhat for food, I suggest that they are most important as
travel cover for animals crossing the highway. Higher rates of
animal mortality would be expected in these areas if animals
were to concentrate along highway with wooded medians in order
to use the median as travel cover to access food or cover

on the other side of the highway. It may not necessarily

be the tree canopy which provides attractive travel cover.

The forest understory of wooded medians was usually dense,
providing excellent cover while the forest floor was rela-
ively open, providing ease of passage. Median vegetation,
which can provide similar advantages such as an extensive
development of shrubs without an overstory tree cover, may
affect animal movement as wooded medians appeared to do

in this study.

34

Wooded medians may affect animal movements in one of
two ways. They may merely redistribute existing mortality
by attracting animals which would normally cross the high-
way at some other point. Or, if the highway acts as an
actual barrier to movement of these animals, the wooded
median may facilitate the crossing of more animals than
would cross the highway otherwise.

Oxley 35. 31. (1974) observed that progressively wider
highway rights-of-way reduced the overall movement of small
and medium-sized forest mammals across the highway. In
this study, there was an average of 8.6 more animals killed
per mile on highway with a narrOW'(4(30.5 m), mowed median
than the wider (.Z30.5 m), unmowed median. This may have been
been due to reduced movement of animals along the highway
wider medians, a result of either the home range character-
istics of the animals involved or the highway acting as a
barrier to movement. Clarification of this point could be
obtained from trapping or telemetry observations of animal
movement patterns in relation to highway of different widths
and median characteristics.

The potential for control of vehicle-killed animals
through highway siting, design, and maintenance practices
is dependent on the presently unknown behavioral and move-
ment patterns of animals in relation to the interstate
highway. There is currently one mamagement practice that
is beneficial and does not result in a large number of

vehicle-killed animals. Where a highway is routed through

35

a wooded area, particularly where near the edge of a large
woodlot, the main body of the woods is left to one side of
the right-of-way and a narrow strip of woods may be left on
the other side. The parctice of leaving this narrow strip
of woods is encouraged in present highway design and mainte-
nance policies to screen against vehicle headlamps and to
help maintain a natural setting in relation to the highway
landscape (Anon., 1966; Noyes, 1969). My data indicate that
wildlife mortality is no higher on areas where screening of
this type is used than where the screening is not used and
woods occurs only along one side of the right-of-way.
Disposal ov vehicle-killed animals may vary from one
locality to another depending on the extent of the problem,
and the policies determined by each maintenance garage
responsible for care specific areas of highway. The Highway
Department maintenance garage at Williamston, Michigan is
responsible for 150 miles of roads, including my 15.5 mile
study area. During 5 days each week the entire length of
these roads is surveyed for damage to the road surface,
highway signs, etc.; in addition to being surveyed for
vehicle- killed animals. Although diSposal of vehicle-
killed animals is currently included with other duties, dif-
ferent management personnel have, in the past, commissioned
twice weekly trips by a crew and truck specifically to dispose
of vehicle-killed animals. My results indicate that the tot-~
a1 effort given to disposal of these animals on interstate

highways may be minimized in locatities where, because of

 

36

voluminous complaints or the large number of animals killed,
a special and regular effort is required to adequetely con-
trol an animal problem. Per mile of highway, the probability
of finding at least one vehicle-killed animal per day is high-
est where the roadside is wooded, particularly where the
median is also wooded. For fall and winter months, there is
a 5 times greater probabilty of finding an animal on high—
way with wooded roadsides than of finding an animal on high-
way with non-wooded roadsides. In order to maintain the same
degree of control over vehicle-killed animals, non-wooded
sections ofhighway need to be checked only one-fifth as often
as wooded sections. The savings in labor and equipment would
result from minimizing the effort of animal disposal on sec-
tions of highway where, because of the small number of wood-
lots located near the highway, the probability of finding
vehicle-killed animals is low.

Some caution must be exercised when interpreting these
results. My data was collected for only a seven month period.
Nbasurements were not made during the spring breeding season
and summer season when many young are normally killed. It
is possible that because of seasonal shifts in the activity
of some animal species, mortality patterns may change
seasonally between habitat types along the roadsides. Allen
(1939) and Haugen (1942) suggested that cottontail rabbits
may shift their activity from fields to wooded areas as
winter approaches. In this study, most of the rabbits killed

on the highway were accosicated with wooded areas. Similar

37

movements may occur with pheasants in anticipation of severe
winter storms (Madson, 1962), such as the ones that occur-

red during this study. Any shift in animal activity from

one habitat to another would probably result in a similar
shift in animal mortality patterns on highways. A large
portion of my observations were made during the fall after

a heavy mid-September frost had killed much of the herbaceous
cover. Animal movements toward woody cover might be expected
to occur at this time. Although there was no relative shift
in animal kill per mile between wooded and non-wooded sections
of highway during the study, late spring and summer data would
be necessary to make certain that the seasons not included in

this study did not affect the results.

LITERATURE CITED

LITERATURE CITED

Allen, R.E., and D.R. McCullough. 1976. Deer-car acci-
dents in southern Michigan. Journal of Wildlife Man-
agement. 40(2): 317-325.

Allen, D.L. 1939. Michigan cottontails in winter. Jour-
nal of Wildlife Management. 3(4): 307-322.

 

. 1962. Our Wildlife Legacy. Funk and
Wagnalls, New York. 422p.

Anonymous. 1966. The art and science of roadside develop-
ment; a summary of current knowledge. Highway Research
Board Special Report No. 88: 1-81.

Bellis, E.D., and H.B. Graves. 1971. Deer mortality on
a Pennsylvania interstate highway. Journal of Wildlife
Management. 35(2): 232-237.

Carbaugh, B., J.P.Vaughn, E.D. Bellis, and H.B. Graves.
1975. Distribution and activity of white-tailed deer
along an interstate highway. Journal of Wildlife Man-
agement. 39(3): 570-581.

Dahlberg, B.L., and R.C. Gottinger. 1956. The white-
tailed deer in Michigan. Wisconsin Cons. Dept. Tech.
Bulletin No. 14. 282p.

Dasmann, R.F. 1964. Wildlife biology. John Wiley and
Sons, Inc., New York. 231p.

Egler, F.E. 1953. Our disregarded rights-of-way--ten
million unused wildlife acres. Trans. N. Am. Wildlife
Conf. 18: 47-157.

 

——~ . 1954. Vegetation management for rights—
of-way and roadsides. Pp. 299-322. In: Annual Report
of the Board of Regents of the Smithsonian Institution,
1953. Publ. 4149. U.S. Gov't. Printing Office.

 

——’ ' 1957. Rights-of-way and wildlife habitat:
a progress report. Trans. N. Am. Wildlife Conf.
22:133-142.

38

39

Ellis, R.L. 1964. Tracking raccoons by radio. Journal
of Wildlife Management, 28(2): 363-368.

Hougen, A.O. 1942. Life history studies of the cotton-
tail rabbit in southwestern Michigan. Am. Midl.
Nat. 28: 204-244.

Joselyn, G.B., J.E. Warnock, and S.L. Etter. 1968. Man-
ipulation of roadside cover for nesting pheasants--
a preliminary report. Journal of Wildlife Management.
32: 217-233.

. 1969. Wildlife--an essential considera-
tion determining future highway roadside maintenance
policy. Highway Research Record. 280:1—14.

 

Lay, D.W. 1942. Ecology of the opposum.in eastern Texas.
Journal of Mammalogy. 23: 147-159.

LeOpold, A. 1933. Game management. Charles Scribner and
Sons, New York. 481p.

Madson, J. 1962. The ring-necked pheasant. 01in Mathieson
Chemical Corporation, East Alton, I11.

Noyes, J.H. 1969. Woodlands, highways, and people. Plan-
ning and Resource Development Series No. 9, Publ.
No. 33. Coop. Ext. Service, University of Massachu-
setts, Amherst. 21p.

Oetting, R.B., and J.B. Cassel. 1971. Waterfowl nesting
on interstate highway rights-of-way in North Dakota.
Journal of Wildlife Management. 35(4): 774-781.

Oxley, D.J., M.B. Fenton, and G.R. Carmondy. 1974. The ef-
fects of roads on populations of small mammals. Jour-
nal of Applied Ecology. 11: 51-59.

Puglisi, M.J., J.S. Lindzey, and E.D. Bellis. 1974. Fac-
tors associated with highway mortality of white-tailed
deer. Journal of Wildlife Management. 38(4): 799-801.

Vaughn, J.P. 1970. Influence of environment on the activ-
ity and behaviour of white-tailed deer (Odocoileus
virginianus) along an interstate highwayiin an agricul-
turaI'area of Pennsylvania. D. Ed. Thesis. Penn.
State University, University Park. 73p.

 

"11111 111111 1111717 T