An analysis of movements of the wood mouse Apodemus sylvaticus in its home range

Behavioural

Processes,

0 1990 Elsevier

BEPROC

235

22 (1990) 235-250

Science Publishers

B.V. 0376.6357/90/$03.50

00326

of movements

An analysis

Apodemus

CNRS, Laboratoire

in its home range

syhticus Simon

de Neurosciences

of the wood mouse

Benhamou

Fonctionnelles,

13402 Marseille

Chdex 9, France

(Accepted 30 August 1990)

Abstract The aim of this study was to quantify the movements of the wood mouse Apodemus sylvaticus in its home range. The movements were recorded using a new tracking method: it consisted of tracking a visually tagged wood mouse by a telescope from an observation tower and automatically recording the successive positions of this telescope (corresponding to the successive positions of the wood mouse in the field). Statistical analysis of 51 recorded movements, involving 10 wood mice, was carried out. The characteristics of the general movement patterns fitted a first order correlated random walk model. Furthermore, the detailed structure of paths was quantified. An analysis of variance showed that an individual factor was involved in metric

and temporal

movement

characteristics

but not in the statistical

structure

of

the paths. Lastly, examination of all the movements of 6 wood mice, with for which at least 4 different paths were recorded, revealed their space use patterns within their home ranges.

Key words:

Wood mouse; Movement;

Space use pattern; Tracking

Introduction Foraging behaviour largely determines how animals use available space. Precise spatio-temporal analysis of movement patterns in the field therefore constitutes a fundamental approach to the study of this behaviour: in a given sedentary species, this analysis provides characteristic parameters on strategies used by individuals foraging in their home ranges. Although spacing patterns have been studied extensively in mammals and birds (review in Brown & Orians, 1970; Waser & Wiley, 19BO), only a few field studies, such as those by Cody (1971,1974) and Smith (1974a,b) on

236 birds

and

centrated study

Siniff

and

Jessen

(1969)

on spatio-temporal

aimed

to

and

analysis

conduct

Macdonald

(1980)

of movement

an analysis

of

this

on

patterns

kind

on

mammals,

in the

the

have

field.

wood

The

mouse

con-

present

Apodemus

s ylvaticus. A close-cropped In a previous their

home

ranges

The

present

study

characteristics (1)

lawn

study

was

Bovet

& Benhamou

to

movements

space

the

in which

to

mice

the

tracks

son,

1977;

the

weight

enough

& Benhamou,

wood of

random

walk

mouse’s

wood

in

1989).

movement

be

from

one

developed for

(Banhamou

& Benhamou,

might

mice

model

to be useful

mechanisms

model

home

patterns shown

Bovet

& Bovet,

1990).

by

theoretical 1989;

Here

useful

for

analysing

of the

wood

mice

it

was

actual

ranges.

of the

movements

of movement

located

characteristics of

of

animals

in the

of the use

data about

field

lawn

wood

patterns

1985),

the animal

1978;

were mice

between

considered. are individually

exhibited

to not

by the

animals

imprecision

disturb

or marked

1966

wood

faeces,

1984;

of this

method

of the

recordings

& Freeman, distances,

used

to locate

Mineau

is limited

& Madi-

because

when tracking

does

continuous

large

been

1977;

a new

or locating

Lemen over

has

mice’s of small

1975a,b)

Recording

path

Madison,

movements

wood tracks

in Twigg,

paths.

et al.,

it. Consequently,

mice’s

and

animals’

1964;

the

punctual

radio-tracking

& Hartline,

wood

record

an individual’s

Although

the

to

Recording

Duplantier

the practicability

and

record

necessary

the

to record

(Rawson

of transmitters

was

in Sanderson,

information.

1983,

it

of toe clipped

(Jones,

in the field

to precisely

space

of precision.

(review

animals

in the

the

view, level

precise

static

Wolton,

from

developed

(Jamon

the

methods). to move

ranges. point

to use

rodents

already

1990;

this

characteristics

sufficiently

only

al,

movement

home

by dyed

was

structure parts

correlated

orientation

et

as the footprints

is tricky

small

area (see observed

are not.

labelled

and affords

and

a satisfactory

such

left

model

in their

the

practical

provide

1985)

the

with

radioactive not

as study

mice were

quantify

movement

order

whether

mice

only

their

movements

use

which

analysis within

mammals,

used

wood

to another

to precisely

This

detailed

and which

From

was

field, shrub

general

Benhamou

of wood

to determine

(4)

the

determine

to analyse

controlled

shrubs

one

fit the first

(1988).

1989;

attempted

(3)

would

some

Benhamou,

shrubs,

from

designed

whether

to another

(2)

with

in a similar

by piloting

to determine

of

out

in order:

shrub studies

sprinkled

carried

working method

of far was

in the field.

Met hods

Principle The used

of the recording

recording by Jamon

an observation angular (0).

method

was

& Benhamou tower

positions

Corresponding

with

method adapted (1989).

a telescope

of the ball-and-socket positions

of the

(and

considerably

It consisted placed joint animal

improved)

of tracking

on a tripod in the

from

each wood

and recording

horizontal

can be determined

(6)

a method

mouse

from

the successive

and vertical by trigonometry,

planes on

237 condition that it moves on a flat area. Each position of the wood mouse on the field (X, Y) then corresponded to a recorded position of the joint: X = OM . cos 6 and Y = OM . sin 6, with OM = OT . tan 8 where OT is the observation

height (at which the telescope

is positioned)

and OM is

the distance from which the animal is observed. Thus, an animal’s path can be perfectly reconstructed from the successively recorded positions of the joint. A similar recording principle was used by Tyack (1981) to determine the position of whales using a theodolite

from the top of a cliff.

Visual tagging of the wood

mouse

As established by Brown (1956a), Baumler (1975), Greenwood (1978), and Wolton (1983), the wood mouse is a nocturnal rodent. Direct observations were conducted at night using Buchler’s (1976) technique: the animal was tagged with a small bulb (0.4 ml) containing cyalume (l/4 reagent, 3/4 solvent), a chemiluminescent product that allows observations for 4 hours at distances over 100 m. To facilitate changing of bulbs, the male part of a press-stud was sewed (under anesthesia) onto the neck of the wood mouse and the female part was glued onto the bulb. The sewing held for about 15 days. It was then possible to sew another press-stud on the animal. The bulb was filled with cyalume and afterwards set on the animal, just before its movements were recorded. The bulb weighed about 1 g (< 5% of the body weight). Each bulb was used only once. The cyalume technique did not disturb the behaviour of the animal, as previously noted by Buchler (1976), Jamon & Bovet (1987) and Jamon & Benhamou (1989). The telescope (15 X 50) used to track the tagged wood mouse was equipped with a radio-luminescent nocturnal movements.

Recording

reticle,

so that one could focus on the animal

during

its

technique

The telescope was set on the joint and placed on a steady tripod at the top of an observation tower, strongly guyed to the soil, at an observation height of 10.84 m. Successive

positions

of the joint

corresponding

to the tracked wood mouse’s

succes-

sive positions in the field were recorded by two precision potentiometers fixed to the joint (Fig. la): the angle B was measured by the vertical potentiometer and the angle S was measured by the horizontal potentiometer. The mechanical multiplication (by 6) of the vertical movements of the joint considerably improved the precision of the vertical angular measurement which is of major importance because of the “perspective effect”, while the mechanical demultiplication of the horizontal movements (by 0.875) ensure to avoid the blind area of the potentiometer (between 355 and 360 degrees). The angular values of 13and 6, as measured by the potentiometers, were digitalized by a 12-bit analogical-numerical converter. The time factor was given in seconds by an electronic clock. The data were stored in the central memory of a microcomputer. Every time the wood mouse was centered with the telescope, its spatio-temporal position was coded, using a switch (Fig. 1 b). The electronic working of the system was under the general control of a computer program. After each recording session, the

238

data were cassette The

conveyed

recorder.

maximum

around

of 100

central system

error

position

on other

at a distance

the

whole

electronic

the actual

depended

from

The

memory was

storage

such

to a radius

observed

as the sighting,

on the

in the field

(1 bit) corresponded

of an animal

factors,

for

supplied

from

magnetic

by three

tape of a

12 V batteries.

of approximately

a distance

of 100

so that the actual

error

25 cm

m. Accuracy

was about

1 m

m.

0b Fig.

1. Scheme

of the tracking

reticle;

WI,

vertical

and horizontal

sion

WI-W2,

W2,

W3, W3-W4,

analogical-numerical

W4:

(a) and recording

notched

potentiometers; respectively); converter;

EC:

wheels VB, 5:

HB:

96, 16,

vertical

switch

electronic

micro-computer;

(b) system.

with

T: telescope; 21,

24

teeths,

and horizontal

controlling clock;

CR: cassette

VIA:

the versatile

recorder.

R: radio-luminescent respectively;

notched recording interface

belts

VP,

HP:

(transmis-

system;

ANC:

adapter;MC:

239

Study

area and timing

The field work was conducted, from July to December 1984, at the Biological Station of Tour du Valat, in the Camargue (Southern France). The field was a flat flat horizontal close-cropped lawn sprinkled with shrubs (Phillyrea angustifolia; density: 85 shrubs/ha). This lawn was delimited by meadows and cultivated fields to 180 m in width (N-S) and 300 m in length (E-W). Wood mice caught by Sherman live traps were individually marked by ear-tagging. Trapping began in July using a 200 by 160 m trapping grid and recording began in August, Only sedentary wood mice were studied. Sedentariness was established when one week in a definite area a wood mouse was trapped for at least (Andrzejewski

& Wierzbowska,

1961; Bovet, 1963).

During the recording session, only a few traps were used. They were checked in the morning. One wood mouse was selected from the trapped wood mice. Because it is more important in movement analysis to have numerous data available on the same animal than to have only some data for a large number of animals, priority was always given to the wood mouse whose movements had been recorded most often. The selected wood mouse was kept in its trap until night. A cyalume bulb was then tagged onto its neck, and it was released at the same place where it was trapped. The procedure then consisted of tracking the wood mouse until 4 hours had elapsed (lifetime of cyalume). Recordings were made during sive days of a lunar month, to ensure good conditions mouse.

moonless nights, for 15 succesfor detecting the tracked wood

Data processing Recorded data were first transferred from magnetic tape to the micro-computer, and afterwards, to a powerful computer by means of a communication interface (RS232).

A computer

program

was

then

used

to decode the

recordings,

i.e. to

reconstitute the paths as sequences of points (X, Y, t). With each recording, path length (L), travel time (T), time spent in shrubs (TS), percentage of time spent in shrubs (%TS = 100. TS/T) and mean speed of movements between shrubs

(V = L/(T-TS))

were computed.

As explained above, two analysis levels of wood mice’s paths were used, because of the heterogeneity of the environment: -Level 1: analysis of the general movement patterns of wood mice from one shrub to another. It was tested if these patterns, represented by the locations of the successively visited shrubs, could fit the first order correlated random walk model. Movements line Bovet

between

which

& Benhamou

successive mean

successively

segments, steps

(1988). are

and a standard

S = o/Jm(P),

where

than 1.2 rad (Bovet corresponding tion

(a)

and

The

shrubs

were

to as steps, model

therefore

according

assumed

that

randomly

drawn

in a wrapped

deviation

u. Then

the sinuosity

m(P) is the mean step length, & Benhamou,

percentage mean

visited

are referred

vector

of path length

1988). length (r,

Step

changes normal of the

number expresses

(N), N

of

mean

the

given

direction

distribution paths

. m(P)/L),

by straight

terminology

step

with

a null as

that u is less length

and standard

concentration

in

between

can be computed

under the condition

(%L = 100.

which

approximated to the

of

(m(P)), deviaangular

240 values

around

changes -Level

the angular

of direction 2: analysis

only

the

of

were

were

after

rediscretization

analysed

by

Bovet

correlated movements

in the

Note

that

were

it is possible

always

the

the

of the distributions

However, made

obviously

shrub

account

length

level

apparatus

lengths,

of

at this

second

oriented,

during

These

the recording.

of direction to

use

level

which

to

obtained

the

the

method

first

analysis,

order

because

constituted

an external

model.

deviation

(a)

the dispersion but

not

of an angular of angular

a wrapped

analyis,

in which

considered.

according

attempt

rad for

the shrubs,

were

no

in the

(r) when

to u < 1.2

second

lawn

of changes

step

was

into

corresponding

case for

1981) between

the

distribution

to use the standard

of the mean vector

large (r > 0.5,

(1988).

is not taken

paths

on

constant

model

lawn

of the

by the tracking

the

different

& Benhamou

which

Batschelet,

located

discretized

with walk

1972;

structure

movements

by computing

random

constraint instead

the

arbitrarily

They given

Mardia,

computed.

of the detailed

parts

movements

mean;

were

normal

for

the

distribution

values

is not too

distribution);

first

level

it was

analysis

(see

below). Finally, tics

by examining

which

space

all the

are individually

use

patterns

were

movements

of the wood

controlled

were

identified

determined

from

the

mice,

movement

by analyses

time

spent

characteris-

of variance,

by the

wood

and the

mice

in the

shrubs.

Results The (C,

analysis

F, G, H,

given (i.e.

was

in Fig. most

2. Wood

usual)

out

mice

speed

greater

than

1). This

difference

the

carried

on 51 recordings

I, and J) and four

the

females

spent

speed

ranged

between

56% from

shrubs

was due to irregular

of IO wood

B, D, and

on average

of movement

mean

(A,

of their which

time was

movements

including

examples

0.5 to 2 m/s,

(V)

slow

mice,

E). Three

in shrubs. that

males

The

much

IO m/min mice

are

modal

is therefore

about

the wood

six

of paths

(Table

performed

in

lawn.

First level analysis Movements wood

between lengths

successively

1972;

Batschelet,

than

43

another

shrubs

constitute

the

general

of movements

accounts

paths.

1981),

distribution

only

the successive

5. Using it was

differ

for

67%

of changes

(given

changes

that

from

in

framework

by straight average

the

were

(1960;

line

of

the

of direction,

test

between

of changes

of

the

segments total

path

distribution

of the observed of

to show whereas

Jupp & Mardia

significant 37 paths

successive

of direction

see Stephens, of changes

normal

to that

the

found

also

distribution

a wrapped

using

of direction

the number

test

by r) equal

Furthermore, 2 paths

where

Kuiper’s

found

at p = 0.10

a concentration

Batschelet, between

of the

1981),

not significantly of

to

schematization

out on 43 recordings

to or greater

out

The

visited

analysis

was carried

and with

shrub

1).

equal

38

one

movements.

(Table

Statistical steps

from

mice’s

1965;

of direction with

null

distribution, (1980;

correlation did not show

was

Mardia, did mean with also

see

at p = 0.05 any such

Fig. 2. Three

examples

of paths of wood

mice C, E, and H. Shrubs

circles and the small square in the center represents delimited

in the north and the south

to a width

about 100 m beyond the limits

correlation that

the

centered

at p = 0.10. changes

on 0, that

However,

the

of all recordings

this

was How,

then,

changes

a null

next

step

correlation shrub

mean, existed

and

the

preferentially

Second

level

Since

or less

chose

the first

length: 240 m).

possible

to reject

in a wrapped

order

was

too

sinuosity

choose

differ

correlated large

their

of

with

to

it.

the

i.e.

hypothesis distribution

random

(r -C 0.5,

paths

the

normal

sinuosity.

the

next

significantly

be said in the

at p = 0.10

distance

that

the

direction

shrub?

from shrub

of the

walk (T > 1.2

model. rad)

In the cases

was

in

where

computed

cannot

chosen

therefore

visited

shrubs

Because

a wrapped

last

41 out of 43 paths It

the successively

to

be

However,

between

that

on the basis

to make

its

no significant

the size

said

of

distribution

by a mouse

step. be

the distribution normal

of the choosen the

of their

wood

mice

apparent

size.

analysis

the schematization

visited

mice

not

it can first

more

seem

drawn

values

the

by small

The study area was

‘12.

did the wood

was

with

rad),

rad/m did

not

us to be able to compute

(a < 1.2

and 0.44

of direction

with

sively

for

possible 0.22

it does

of angular

tower.

of 180 m, but went on to the east and the west

drawn here (represented

are randomly

is compatible

dispersion

66%

between

Therefore,

of direction

are represented

the observation

shrubs

of movements

accounts

for

67%

by straight in

average

line of

segments

the

total

between path

succes-

lengths,

the

242 difference line

between

therefore

movements

the actual

in the

to 5 m) in order nhamou,

1988).

observed

after

lawn

This

walks,

movements

did

decrease

with

the existence

equal

TABLE

a step

is

the

case search

length

than

of the

5. The

A

L

2

T

195.0

2 4

to 0. This

goal

distributions root

order the

of

expected,

values

oriented

the

correlated

was

However,

1

& Be-

of

(ok

movements,

component. value

distribution did

given

were

of the standard of changes

the number

mean

chart

of direction

analysed

deviations

not differ

1981)

was

of direction

significantly

in Batschelet,

after a, are

of direction

of changes

from

with

0 at

43 out

of

r

%L

02

59.0

0.36

(91.9)

(22.1)

(55.1)

(35.4)

(0.26)

422.5

195.0

40.0

(4.2) 3.9

53.5

0.28

0.79

(13.9)

(4.2) 23.3

(1.3) 7.7

(13.4)

(0.01)

(0.11)

57.0

0.27

0.74

(14.0)

(1.3) 9.0

(6.9) 71.9

(0.08)

(0.13)

0.45

0.56

(6.9) 13.8

(17.6)

(0.17)

(0.23)

69.5

0.52

0.58

(3.5) 75.5

(0.43)

(0.11)

175.1 (40.8)

0.56 (0.19)

213.5

82.9

55.4

(51.6)

(21 .O)

196.0

27.2

46.0 (14.1)

558.5

(8.9) 137.2

26.5

(0.8) 8.1

0.48

0.57

(318.9)

(71.7)

(8.1) 5.7

(3.5) 59.2

(0.08) 0.31

(0.00) 0.74

(2.7) 7.7

(17.7)

(0.26)

(0.24)

71.6

0.41

0.57

(14.2)

(0.26)

(0.14)

(26.9) 2

182.7

85.4

(10 6) 54.2

(139.0)

(68.3)

(16.7)

337.7

145.7

61.6

6 7

(59.0)

(19.3)

99.7

40.2

84.3

(5.1) 19.6

69.2

0.56

0.47

(34.4)

(19.4)

(5.7) 15.8

(4.7) 65.7

(0.19)

(0.14)

78.7

(9.4) 67.8

0.38

0.68

(60.3)

(25.7)

(14.9)

(7.3)

(0.15)

(0.22)

(247.2) 6 6

137.5 (88.7)

wood mouse;

RN: number

percent of time spent in shrubs;

of paths recorded;

general

changes of direction

L: path length (m); T: travel time (mn); %TS:

V: mean speed in the lawn (m/min);

by general movement for

V

(150.4) 2

direction

%TS

6.6

967.5

14

represented

characteristics

49.0

(143.6)

R=2m.

square

Bovet

93.8

(150.6)

of

these

the

R (from

1

RN

WM:

strictly

where

angular

interval

Mean values (and SD) of the main movement WM

with

(see

first

The

lengths.

lengths

uR of with

close

oriented.

R = 2 m. The

analysis

the confidence

remained goal

of changes

45 out of the 51 recordings

to or greater

(from

as

were

of a random

1. Statistical

p = 0.05

means

shrubs

step

than

happen

and a straight

path

of direction

slowly

what

lawn

total

different

deviations

more to

on the

these

of changes

of the distributions

given was

angular

taken of

five

standard

R, contrary

R

with

out with

the

33%

with

increased

the

rediscretization in Table

to

distribution

that

between

characteristics

carried

the

length

whereas

of the paths

average

rediscretized

showed

step

because

The

were

rediscretization

random

indicating

on

to analyse

rediscretization

not

structure

corresponded

patterns;

movement

for movements

%L: percent of path length

r: mean vector length of the distribution

patterns;

a,: standard

deviation

of

in the lawn (rad) after rediscretization

the with

of changes

distribution

of

a step length

243

45 paths, but the normal distribution assumption was rejected at p = 0.05 paths, although the distribution shapes were symetrical and unimodal. deviations ranged from 0.26 to 1.15 rad, which points to the existence of random component. Furthermore the assumption that no correlation existed successive changes of direction was rejected at p = 0.05 with 11 paths, which high as might be expected with strictly goal oriented movements lead to an alternation of rediscretization step length) which appears as successive changes of direction. This also points component

Movement

with 28 Standard a major between is not as

oriented movements: indeed, such right and left turns (whatever the an important correlation between to the existence of a random search

in the path taken in the lawn.

variability

analysis

The results of analysis of the variability of given movement characteristics allowed us to test the hypothesis that the values of these characteristics obtained with individual value.

wood mice are drawn from

parent populations

Taking each wood mouse as an independent showed the existence of an “individual wood

with

the same expected

data group, an analysis of variance mouse” factor contributing to the

metric and temporal movement characteristics: indeed the homogeneity hypothesis was rejected at p = 0.05 for L, T, %TS, and V. On the other hand, this factor could not be shown at p = 0.10 in the case of characteristics depending on the statistical structure of paths: %L, r and a2 (Table 2).

Space use pattern The home ranges of the six wood mice on which at least four recordings were obtained were represented by the elliptical home range model at the 0.95 probability level (Jennrich & Turner, 1969; Mazurkiewicz, 1969,197O; see also Benhamou 1988), from the times spent by these wood mice in the shrubs (Table 3).

& Jamon,

TABLE 2 Variance

analysis

of the main

movement

characteristics.

L

T

%TS

V

%L

r

F

11.75

4.07

3.40

2.23

1.00

0.83

1.13

df

9-41

9-41

9-41

9-41

9-41

9-33

9-35

P

-C 0.0005

< 0.001

-=C0.005

< 0.05

> 0.10

> 0.10

> 0.10

0,

F

20.10

4.69

5.30

3.25

1.29

1.24

1.51

df

5-37

5-37

5-37

5-37

5-37

5-29

5-31

P

< o.ooo5

-C 0.005

-=z0.001

< 0.025

> 0.10

> 0.10

> 0.10

Upper with ratio;

part gives variance the 6 wood df: degrees

analysis

mice on which of freedom;

results

on the 10 wood

at least four

p: probability

recordings level.

mice,

were

whereas

obtained

lower

part deals

only

(see Table 1). F: variance

244

L Fig. 3. Home

ranges and paths of wood mice C, D, C, ti, I, and ). Home

by the elliptical

home range model

at the 0.95 probability

level

The

range limits size

are drawn

of the represented

field is 240 m bv 180 m.

It can be seen the

bivariate

ellipitical able

home home

were

this

representation

distribution

range model

heterogeneity

The They

that

normal

range similar

of

was

areas size

were to

(Fig.

between those

1700

obtained

ranges

is accurate

‘position

not confirmed,

of the environment is

of home animals

in

in particular

space

enough, assumed

because

even by

if

the

of the consider-

3). and 14000 in

the

mL, with

same

a mean

environment

of 6500 by

m’.

trapping

‘(%p)

UO!l3aJ!p

3-M

ayl

0)

UO!PlaJ

U!

S!XL’

JO!PLU

aL#

40

UO!lPKla!JO

:a0

Sl@JJl

Nkl

!asnow

S!Xe-!UlaS

:WM

syled

WM

3iavi

a8ueJ aluoH

POOM

q3e.3 Aq uayel

y(W)

:Ny

~00~

se sys!JaPeJey3

11

40 Jaqwnu

(asnow

I I

1

a~# Aq paleur!lsa

11

sqied

9

H

Sll le>!ld!lla

pue 3X

9

3

:XA

‘258

L

auioq

pue 11 ‘11 !P~PJOMJ

lelol

865

9

3

a

$0 saleu!p~oo>

((~1) sJwJ.33

L’ZLP

P9f z

PL

:Sll

‘(~1) y@~al

6’8SZ

S’ LPZ

9601

P

Al!h!l3e

lelol

:VdH

61

5’605

Z’Z 1s

6’6 LO 1

686 z

away

PL

Z’6LZ

O’fL9

OL8f

pue (u!u~) awy

LZ

CL

L‘OOL

1’1911

eaJe ague

51

LP-

S’O81

8’6P9

au~11 leloj f69

EL-

LL-

Ly !(mu1)

LZZP

ZL-

1s -

pue

6f

LSOP

8158

S-

8f-

u! wads

PP

L6fL

:Za

PL

6P

9s

L08fL

sqnJc(s

PLOf

LS

pue Jo!ew

EL LP LZ

86

lie u! (u!w)

Jou!w

ZL 99Lb

VtlH

&ueJ

3X

SP 18

3A

FS Zi?l

~6.0 ayi

ie lapolu

9s-

t10

‘lahal &!l!qeqoJd

F

246 (jamon,

1982,

p = 0.02

between

1987).

Furthermore,

the mean

range area. A similar by Wolton reflect

(1985)

result

amount

looks

at all

portions

overlap.

network

of pathways.

genity

was

paths

might The

networks

the

figure

of

that

found

to

and the size

by Brown

suggested wood

that

(1956b)

exist

at

of its home

by trapping

the home

the shrubs all

mouse,

wood

of this

showing

had common

each

indicate

structure

of the environment,

Furthermore,

obtained

which

was

mouse

range

and

area might

of movement.

the

This

correlation

of each wood

previously

by radio-tracking,

the mean

If one

a significant

path length

network

51

move

may

corresponding the

it can

mice

be seen more

be favoured

together

(Fig.

several

less

along

by the

to the junctions

paths

that

or

a

hetero-

of the network.

4) points

that

these

portions.

Discussion The

hypothesis

trails

has

O’Farrel,

put

1965;

Rathbun,

1979;

move

on to

with

1967;

as

with

Frame,

pattern

suggested

1982),

Barry

pathways

Peters

has

by

or

in

clearly study

mice

piloting

and

1979;

been this

wood

by visual

(1980)

1959;

Peters,

always

mice

whether

move

scent-marked

(Pearson,

1975;

not

wood

however,

& Francq

or

species

& Mech,

but

exhibited

pathways by

of

numerous

,1975;

It is not clear,

of scentmarked

another,

a network

Thompson,

use

hypothesis.

a network

on

connection

1982;

space

this

move

in

& Davis,

Pulliainen, The

compatible

mammals

forward

Adams

demonstrated.

shrub

that

been

from

Drickamer

is

actually one

& Stuart

(1984). The

general

associated spent

(56%

for wood

on average) against

of movements from

one

on the lawn

shrub

findings

exhibited home they for

ranges were but

centres

(Jamon

mice

was

quite

different

In most dispersion of only

at the

of

patterns

cases,

paths.

that

order

might

mice

a significant

of

changes

of

to another

showed

correlated

random

possible normal

of the distributions, This

to

from

as possible

were

ranges

speed

time

in the

were

associated

of the released

same

the

same

towards

All with

movements within

not explore

extent

lawn.

environment

using did

0.5 to

their when

procedure their

home

their

activity

in

general

1989).

one shrub

statistically

home

case, the wood

mean

ranged

lawn

in the

the

The as fast

organization

exhibited

and

between

lawn

of their

they

places etc.)

speed

the

in the

where

of their

distributions

in a wrapped

of most

from

movements

of the first

it was

drawn some

from

44%

taken

favourable

lengths.

were of time

arthropods,

in the lawn

path on

the general study,

(seeds,

the usual

travel

on average

this

periphery

& Benhamou,

analysis

in

in the latter

their

not

movements

Furthermore,

wood

assumptions

randomly

lawn.

the

oriented

Statistical basic

spent

and tracking:

ranges,

movement

They

that these

total

whereas

to another

proportion

are indeed

movements

of these did

one shrub

by the high

of energy

hand,

was 10 m/mn,

to another.

released

release

to 33% mice

of the

by

the other

wood

from

Shrubs

sources

the

indicate

an exploitation

On

mice

as indicated therein.

various

on average

Consequently,

these

animals

both

predators.

of the wood

of shrubs,

by the

providing

corresponded

2 m/s.

patterns

the exploitation

mice,

protection shrubs

movement

with

to show

distribution however,

that

direction walk

characteristics

model

changes

centered I was

observed

that their

fairly

of direction

on 0. Because

able to compute

be due to the heterogeneity

well:

fit the indeed, appeared

of the great the sinuosity

of the environment

which

Fig. 4. Drawing of ali the 51 recorded paths taken by all 10 wood represented Field is 240 by 180 m.

mice. The size of the

led the wood mice to walk great distances between successively visited shrubs, m on average. Wood mice therefore need to make large turns in order to restrict

i.e. 9 their

movements within their home ranges (6500 m2 in average). On the other hand, although movements in the lawn were shrub directed, they contained a iarge residual random search component, probably due to the exploitation of the lawn. The results of movement variability analysis showed that the metric and temporal characteristics of the wood mice’s movements, but not the characteristics depending on the statistical path structure, were sensitive to an “individual factor”. This confirms the hypothesis put forward in Covet & Benhamou (‘l988) that the random component of paths might constitute a general adaptation of movement behaviour to the stochasticity of the environment: unlike metric and temporal movement characteristics, this component did not depend on an individual factor but was imposed by the environment. Various habitats have been colonized by the wood mouse. This study showed how the movements can be influenced by a “semidesertic” environment (i,e. an open area sprinkled with shrubs). The movement and space use patterns exhibited by the waod mice in this study are similar to those of other rodents species living in this type of environment. They are ~robabi~ different, however, mice living in a more homogenous environment.

from patterns

exh~b;ted by wood

248

Acknowledgements I thank field

Pierre

work

Bovet,

Marc Jamon

was carried

out

with

and Marilyn

the kind

Cranjon

permission

for their

invaluable

help.

du Valat”

Biological

of the “Tour

The

Station.

References Adams,

L. & Davis, SD.,

Andrzejewski, small

rodents

Theriol., Barry,

in a defined

anatomy of home range. J. Mamm.,

T., 1961. An attempt at assessing

forest

48: 529-536.

the duration

area and the rate of interchange

between

of residence individuals.

of

Acta

5: 1533172

R.E. & Francq,

Peromyscus

Batschelet, Baumler,

1967. The internal

R. & Wierzbowska,

E.N.,

1980.

E., 1981. Circular W., 1975. Activity

Benhamou,

S., 1989.

ranges. J. theor. Benhamou, Anim.

Ortentation

J. Mamm.,

leucopus.

statistics

Biol.,

in biology.

of some small

An olfactory

to

landmarks

within

the

preferred

habitat

by

61: 292-303. Academic press,

mammals

orientation

in the field.

model

London. Acta Theriol.,

for mammals’

20: 365-377.

movements

in therr

home

139: 379-388.

S. & Bovet,

P., 1989.

How

animals

use their

environment:

a new

look

at kinesis.

Behav., 38: 375-383.

Benhamou,

S. & Jamon, M., 1988. Analysing

spatial relationships

from trapping

data. J. Mamm.,

69: 828-831. Benhamou,

S., SauvP, J-P. & Bovet, P., 1990. Spatial memory

and limitation

of the egocentric

Bovet, J, 1963. Observations mus

Bovet,

coding process.

sur la sedentariti.

en Camargue. Terre

sylvaticus)

P. & Benhamou,

S.,

1988.

random walk model. J. theor. Bovet, P. & Benhamou,

Biol.,

et le domaine

efficiency

145: 1-12.

vital du mulot

sylvestre

(Apode

et Vie, 17: 266-279.

Spattal Biol.,

in large scale movements:

J. theor.

analysis

of animals’

movements

using

a correlated

131: 419-433.

S., 1990. Optimal

sinuosity

in central place foraging

movements.

Anim.

Behav., in press. Brown,

J.L. & Orians,

C.H.,

1970.

Spacing patterns

field

experiments

in mobile

animals.

Ann.

Rev. Ecol. Syst.,

1:

2399262. Brown,

L.E.,

1956a.

and Microtus.

Clethrionomys,

Brown, Buchler,

L.E., 1956b. Movements E.R.,

1980.

Evidence

on

Proc. Zool.

the Sot.

of some British

activity

of

the

small

mammals

Apodemus,

Lond., 126: 549-564.

small

mammals.

for the use of a scent

post

J. Anim.

by Myotis

Ecol., 25: 54-71. J. Mamm.,

lucifugus.

61:

525-528. Cody, M.L., 1971. Finch flocks Cody, M L., 1974. Optimization Drickamer,

L.C. & Stuart,

navigation. Duplantier, rescentes

Am. Midl.

in the Mohave desert. Theor.

Pop. Biol.,

2: 142-158.

in ecology. Science, 183: 1156-1164.

J., 1984.

Peromyscus:

snow

tracktng

and

possible

cues

used

for

Nat., 111: 202-204.

J.M., Cassaing,

J., Orsini,

P. and Croset,

pour I’analyse des d&placements

H.,

1984.

des petits rongeurs

Utilisation

des

poudres

dans la nature. Mammalia,

fluo48:

293-298. Frame, L.H.,

1975. Mouse

Greenwood,

P.J., 1978.

wood mouse

runways.

Timing

Apodemus

Pactfic Discovery,

of activity

sylvaticus

Jamon, M , 1982. Capacites d/orientation en Camargue. These

28: 12-15.

of the bank vole

in a deciduous

Clethrionomys

woodland.

et exploitation

de 3eme cycle en Neurosciences,

Oikos,

glareolus,

de I’espace chez Apodemus Universite

and the

31: 123-127.

Aix-Marseille

sylvaticus

II

249 Jamon, M., 1987. populations. I.N.R.A.,

Le partage de I’espace chez In:

R. Campan & F. Spitz

le mulot

(editeurs),

et son

rBle

Organisation

dans

sociale

la dynamique chez

des

les vertebr+s.

Paris, pp. 167-180.

Jamon, M. & Benhamou,

S., 1989. Orientation

Jamon, M. & Bovet, P., 1987. Possible wood mice. Behav. Processes, Jennrich,

and movement

released inside and outside

mus sylvaticus)

R.I. & Turner,

a familiar

patterns

of wood

mice (Apode-

area. J. Comp. Psychol.,

use of environmental

gradients

103: 54-61.

in orientation

by “homing”

15: 93-107.

F.B., 1969. Measurement

of non-circular

home range. J. theor.

Biol.,

22:

227-237. Jones, N.B.,

1978. The use of a fluorescent

rodents.

Int. Biodeter.

Bull.,

pigment

Jupp, P.E. & Mardia, K.V., 1980. A general correlation regression Kuiper,

problems.

N.H.,

Biometrika,

1960. Tests

for tracing the movements

of commensal

14: 61-64. coefficient

for directional

data and related

67: 163-173.

concerning

random points

on a circle.

Konikl.

Nederl.

Akad. Wet.

A,

63: 38-47. Lemen, C.A., & Freeman,

P.W.,

1985. Tracking

animals

with

fluorescent

pigments:

a new tech-

nique. J. Mamm., 66: 134-136. MacDonald,

D.W.,

Madison,

1980. The red fox, Vulpes vulpes,

Z. Tierspsychol.,

terrestris.

D.M.,

1977. Movements

revealed by radiotelemetry. Mardia, K.V., 1972. Statistics Mazurkiewicz,

as a predator upon earthworms,

Lumbricus

42: 171-200. and habitat use among interacting

Can. FieldNat., of directional

Peromyscus

as

leucopus

91: 273-281,

data. Academic press,

M., 1969. Elliptical

modification

M., 1970.

of home

London.

of home range pattern.

Bull.

Acad. Pol. Sci., 17,

427-431. Mazurkiewicz, method. Mineau,

Bull.

Analysis

range directions

based on the catch-mark-release

Acad. Pol. Sci., 18: 465-468.

P&Madison,

D.,

1977.

Radiotracking

of

Peromyscus

leucopus.

Can.

J. Zool.,

55:

465-468. O’Farrel,

M.J., 1965. Home

range and ecology of snowshoe

hares in interior

Alaska.

J. Mamm.,

46: 408-418. Pearson,

O.P.,

169-l

1959.

A traffic

survey

of Microtus-Reithrodontomys

Peters, R.P., 1979. Mental maps in wolf territoriality. and ecology of wolves. Peters,

runways.

J. Mamm.,

40:

80.

R.P. & Mech, L.D.,

Pulliainen,

Z. Saugetierk.,

P.H.,

Science, 146: 1596-l C.C.,

1966.

in wolves.

(editor),

The behavior

pp. 119-152. Am. Scient.,

63: 628-637.

in the pine marten Martes martes in finnish

G.B., 1979. The social structure

Sanderson,

In: E.K. Klinghammer

New-York,

forest

lapland in

47: 91-99.

20 PP. Rawson, K.S. & Hartline, omyscus.

Press,

1975. Scent-marking

E., 1982. Scent marking

winter. Rathbun,

Garland STPM

The

1964.

and ecology of elephantshrews.

Telemetry

of homing

behavior

Advances in ethology, by the deermouse,

Per-

597.

study

of mammal

movements-a

review.

J. Wildl.

Manage.,

30:

215-235. Siniff,

D.B. & Jessen,

CR.,

1969. A simulation

model

of animal

movement

patterns.

Adv. Ecol.

Res., 6: 185-219. Smith,

J.N.M.,

analysis Smith,

1974a. The food searching behaviour

of search paths. Behaviour,

J.N.M.,

1974b. The food searching

ness of search patterns. Stephens,

M.A.,

1965.

The

Behaviour,

of two European thrushes.

I Description

and

48: 276-302. behaviour

of two European

thrushes.

II The adaptive-

49: l-61,

goodness-of-fit

statistic

VN:

distribution

and significance

point.

Biometri ka, 52: 309-321. Thompson, lepida

S.D.,

1982. Spatial utilization

lepida.

J. Mamm., 63: 570-581.

and foraging behavior of the desert woodrat

Neotoma

250 Twigg,

C.I., 1975a. Finding

Twigg,

C.I., 1975b. Marking

mammals,

Tyack,

P., 1981.

Interactions

P.M. &Wiley,

R.H.,

1979.

Marler & J.C. Vandenbergh behavior and communication. Wolton,

Mammal

between

nearby. Behav. Ecol. Sociobiol., Waser,

their signs or remains.

mammals.

singing

Review,

Mammal

Review,

5: 71-82.

5: 101-106.

Hawarran

humback

whales

and conspecrfics

8: 105-116. Mechanism

(editors), Plenum

and evolution

Handbook Press,

R.J., 1983. The activity of free-ranging

of

of behavioral

New-York,

spacing

in

neurobiology,

animals. vol

in:

P.

3 Social

pp. 159-223.

wood mice Apodemus

sylvaticus.

J. Anim.

Ecol.,

52: 781-794. Wolton,

R.J., 1985. The rangrng behaviour

radro-tracking.

J. Zool.

A, 206: 203-224.

of wood

mice, Apodemus

sylvaticus,

as revealed by