ArticlePDF Available

Daily Activity Patterns and Habitat Use of the Lowland Tapir (Tapirus terrestris) in the Atlantic Forest

Authors:
  • Instituto de Biología Subtropical, nodo Iguazu. Universidad Nacional de Misiones - CONICET
  • CONACYT-Guyra Paraguay

Abstract and Figures

We studied the daily activity pattern and habitat use of the lowland tapir Tapirus terrestris and their relationship with environmental and anthropic variables. We used photographic records of tapirs obtained during five camera-trap surveys conducted in three areas of the Atlantic Forest of Argentina that differ in their protection against poaching. The daily activity pattern was analyzed with circular statistics and linear regression. The effect of protection against poaching and environmental variables on habitat use of tapirs was analyzed using occupancy modeling. Tapir were nocturnal all year round, with 89% of the records between 1800 h and 0700 h. The proportion of nocturnal records and the recording rate did not change with mean daily temperature. The daily activity pattern of tapirs was not affected by the sex of the individuals, the lunar cycle or the protection level of the area. The probability of detecting tapirs increased with the distance to the nearest access points for poachers and decreased with the abundance of bamboo in the understory and increasing trail width. The probability of use of an area by tapirs increased with increasing protection against poaching and distance to the nearest access points for poachers. These results suggest that poaching is one of the factors with significant effect on habitat use by tapirs but not on their daily activity patterns.
Content may be subject to copyright.
Mammalian
Biology
79
(2014)
376–383
Contents
lists
available
at
ScienceDirect
Mammalian
Biology
jou
rn
al
h
om
epa
ge:
www.elsevier.com/locate/mambio
Original
Investigation
Daily
activity
patterns
and
habitat
use
of
the
lowland
tapir
(Tapirus
terrestris)
in
the
Atlantic
Forest
Paula
Cruza,b,,
Agustín
Pavioloa,b,
Roberto
F.
c,
Jeffrey
J.
Thompsond,
Mario
S.
Di
Bitettia,b,e
aInstituto
de
Biología
Subtropical
(IBS)
Nodo
Iguazú,
Universidad
Nacional
de
Misiones,
Consejo
Nacional
de
Investigaciones
Científicas
y
Técnicas
(CONICET),
Moisés
Bertoni
85,
3370
Puerto
Iguazú,
Misiones,
Argentina
bAsociación
Civil
Centro
de
Investigaciones
del
Bosque
Atlántico
(CeIBA),
Puerto
Iguazú,
Misiones,
Argentina
cGrupo
de
Investigaciones
en
Ecología
de
Humedales
Laboratorio
de
Ecología
Regional
(GIEH),
Departamento
de
Ecología
Genética
y
Evolución,
Facultad
de
Ciencias
Exactas
y
Naturales,
Universidad
de
Buenos
Aires,
Ciudad
Autónoma
de
Buenos
Aires,
Argentina
dGuyra
Paraguay,
Gaetano
Martino
Nro.
215
esq.
Tte.
Ross,
Asunción,
Paraguay
eFacultad
de
Ciencias
Forestales,
Universidad
Nacional
de
Misiones,
Argentina
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
29
November
2013
Accepted
15
June
2014
Handled
by
Heiko
G.
Rödel
Available
online
4
July
2014
Keywords:
Tapirus
terrestris
Camera-traps
Occupancy
modeling
Daily
activity
Poaching
a
b
s
t
r
a
c
t
We
studied
the
daily
activity
pattern
and
habitat
use
of
the
lowland
tapir
Tapirus
terrestris
and
their
rela-
tionship
with
environmental
and
anthropic
variables.
We
used
photographic
records
of
tapirs
obtained
during
five
camera-trap
surveys
conducted
in
three
areas
of
the
Atlantic
Forest
of
Argentina
that
differ
in
their
protection
against
poaching.
The
daily
activity
pattern
was
analyzed
with
circular
statistics
and
linear
regression.
The
effect
of
protection
against
poaching
and
environmental
variables
on
habitat
use
of
tapirs
was
analyzed
using
occupancy
modeling.
Tapirs
were
nocturnal
all
year
round,
with
89%
of
the
records
between
1800
h
and
0700
h.
The
proportion
of
nocturnal
records
and
the
recording
rate
did
not
change
with
mean
daily
temperature.
The
daily
activity
pattern
of
tapirs
was
not
affected
by
the
sex
of
the
individuals,
the
lunar
cycle
or
the
protection
level
of
the
area.
The
probability
of
detecting
tapirs
increased
with
the
distance
to
the
nearest
access
points
for
poachers
and
decreased
with
the
abundance
of
bamboo
in
the
understory
and
increasing
trail
width.
The
probability
of
use
of
an
area
by
tapirs
increased
with
increasing
protection
against
poaching
and
distance
to
the
nearest
access
points
for
poachers.
These
results
suggest
that
poaching
is
one
of
the
factors
with
significant
effect
on
habitat
use
by
tapirs
but
not
on
their
daily
activity
patterns.
©
2014
Deutsche
Gesellschaft
für
Säugetierkunde.
Published
by
Elsevier
GmbH.
All
rights
reserved.
Introduction
The
lowland
tapir
(Tapirus
terrestris)
is
considered
a
keystone
species
as
a
result
of
its
role
as
a
seed
disperser
and
ecosystem
engi-
neer
(Bodmer,
1991;
Fragoso,
1997;
Chalukian
et
al.,
2009).
Globally
it
is
categorized
as
vulnerable
(VU)
with
a
decreasing
population
trend
(IUCN,
2010),
while
in
Argentina
it
is
considered
endangered
(EN;
Chalukian
et
al.,
2012).
Habitat
loss,
fragmentation
and
poaching
(illegal
hunting)
have
been
identified
as
the
main
factors
responsible
for
population
declines
of
lowland
tapirs
throughout
their
range
(Medici
et
al.,
Corresponding
author
at:
Instituto
de
Biología
Subtropical
(IBS)
Nodo
Iguazú,
Universidad
Nacional
de
Misiones,
Consejo
Nacional
de
Investigaciones
Científicas
y
Técnicas
(CONICET),
Moisés
Bertoni
85,
3370
Puerto
Iguazú,
Misiones,
Argentina.
Tel.:
+54
3757
423511.
E-mail
address:
policruz@hotmail.com
(P.
Cruz).
2007;
Taber
et
al.,
2008).
In
the
Upper
Paraná
Atlantic
Forest
(UPAF)
poaching
of
wild
animals
is
culturally
accepted
and
is
widely
prac-
ticed
(Giraudo
and
Abramson,
2000;
Cullen
et
al.,
2001;
Paviolo
et
al.,
2009a).
In
the
province
of
Misiones,
Argentina,
the
hunting
of
tapirs
is
illegal,
which
makes
it
difficult
to
assess
its
magnitude
(Di
Bitetti
et
al.,
2008).
Most
animals
do
not
use
their
habitats
randomly
but
select
areas
that
provide
the
resources
and
conditions
necessary
to
sat-
isfy
their
daily
requirements,
while
avoiding
areas
of
high
predation
risk
and
human
threats
(Manly
et
al.,
1993).
All
tapir
species
show
discernible
habitat
selection
patterns
although
differences
in
habi-
tat
preferences
have
been
observed
among
studies
(Salas,
1996;
Lizcano
and
Cavelier,
2000a;
Foerster
and
Vaughan,
2002;
Medici,
2010).
Tapirs
generally
prefer
riverine
forests
and
avoid
open
areas,
such
as
grasslands
and
crops
(Salas,
1996;
Medici,
2010),
but,
depending
on
the
study
site,
may
show
preferences
for
either
secondary
forests,
primary
forests
or
both
(Lizcano
and
Cavelier,
2000a;
Foerster
and
Vaughan,
2002;
Medici,
2010).
Water
bodies
http://dx.doi.org/10.1016/j.mambio.2014.06.003
1616-5047/©
2014
Deutsche
Gesellschaft
für
Säugetierkunde.
Published
by
Elsevier
GmbH.
All
rights
reserved.
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
377
are
important
as
resting
places,
defecation
sites,
to
avoid
ectopar-
asites,
to
facilitate
movement
among
foraging
sites
and
for
cooling
off
during
the
hottest
hours
of
the
day
(Padilla
and
Dowler,
1994;
Foerster
and
Vaughan,
2002).
In
the
UPAF
some
environments
have
been
shown
to
be
avoided
by
the
lowland
tapir,
such
as
areas
with
high
density
of
bamboo
(Chusquea
ramosissima),
probably
due
to
the
difficulty
of
movement
(Gallardo
et
al.,
2008).
Furthermore,
anthropic
factors
can
also
potentially
influence
habitat
selection
by
tapirs
(Licona
et
al.,
2011).
Lowland
tapirs
have
been
described
as
nocturnal
or
crepus-
cular
(Padilla
and
Dowler,
1994;
Noss
et
al.,
2003;
Tobler,
2008;
Medici,
2010;
Wallace
et
al.,
2012);
however,
several
factors
have
been
shown
to
affect
the
activity
patterns
of
all
tapir
species.
For
lowland
tapirs,
increased
diurnal
activity
was
observed
in
humid
months
(Foerster
and
Vaughan,
2002;
Medici,
2010).
Mountain
tapirs
(Tapirus
pinchaque)
showed
higher
nocturnal
activity
dur-
ing
full
moon
in
primary
forest
with
dense
canopy
(Lizcano
and
Cavelier,
2000b),
while
lowland
tapirs
showed
the
opposite
pat-
tern
(Medici,
2010).
No
response
to
the
lunar
phase
has
been
reported
around
natural
salt
licks
for
the
latter
species
(Coelho
et
al.,
2008;
Link
et
al.,
2012).
Finally,
differences
in
the
activity
patterns
between
sexes
have
been
reported
in
lowland
tapirs,
with
females
being
active
during
a
wider
hourly
range,
probably
as
a
result
of
their
higher
metabolic
requirements,
especially
during
late
pregnancy
and
peak
lactation
(Medici,
2010).
Considering
the
extensive
gestation
period
(13–14
months;
Barongi,
1993),
and
pro-
longed
lactation
phase
(8–9
months;
Padilla
and
Dowler,
1994)
of
tapirs,
it
is
possible
that
females
have
higher
year-round
energetic
requirements
than
males.
The
use
of
remote
cameras
is
a
noninvasive
technique
that
allows
the
study
of
nocturnal,
rare
or
elusive
animals,
or
those
that
inhabit
areas
of
low
visibility
(Karanth
and
Nichols,
1998;
OConnell
et
al.,
2011).
Furthermore,
remote
cameras
are
an
efficient
tool
to
study
the
distribution,
abundance,
activity
patterns
and
habitat
use
of
large-
and
medium-sized
mammals
(Di
Bitetti
et
al.,
2006,
2008;
Tobler,
2008;
OConnell
et
al.,
2011).
The
elusive
habits
of
low-
land
tapirs,
combined
with
the
relatively
dense
understory
of
most
habitats
within
the
UPAF
make
camera-trapping
an
appropriate
methodology
to
study
the
species.
We
used
records
of
lowland
tapir
obtained
during
five
intensive
camera-trap
surveys
conducted
in
the
Atlantic
Forest
of
Argentina
to
analyze
habitat
use
and
daily
activity
patterns
of
the
species.
We
employed
circular
statistics
(Batschelet,
1981)
and
occupancy
modeling
(MacKenzie
et
al.,
2006)
to
test
hypotheses
on
how
envi-
ronmental
(natural)
and
anthropic
factors
affect
daily
activity
and
the
pattern
of
habitat
use
and
detection
of
lowland
tapirs
in
areas
with
differing
levels
of
protection
against
poaching.
We
hypothe-
sized
that
increasing
distance
from
access
points
and
higher
levels
of
protection
against
poaching
would
positively
affect
habitat
use
by
tapirs
due
to
higher
abundance
and/or
reduced
avoidance,
while
tapirs
would
avoid
areas
with
increasing
amounts
of
bamboo.
In
addition
we
expected
that
decreasing
distance
from
bodies
of
water
would
positively
increase
habitat
use
by
tapirs.
This
study
is
aimed
at
elucidating
the
principal
drivers
of
habitat
use
while
incorporat-
ing
incomplete
detection
through
occupancy
modeling.
Materials
and
methods
Study
area
We
conducted
this
study
in
three
areas
of
the
Green
Corri-
dor
of
Misiones,
Argentina
(Fig.
1).
The
Green
Corridor
constitutes
one
of
the
largest
remnants
of
UPAF
that
still
contains
the
complete
regional
native
mammal
assemblage
(Giraudo
and
Abramson,
2000),
and
is
considered
one
of
the
few
areas
in
the
Fig.
1.
Map
of
the
Green
Corridor
of
Misiones,
Argentina
including
the
main
pro-
tected
areas
and
the
study
sites.
ecoregion
with
high
potential
for
long-term
persistence
of
tapirs
(Paviolo
et
al.,
2008;
Taber
et
al.,
2008).
The
area
has
a
wet
subtrop-
ical
climate
with
seasonality
in
temperature
and
day
length,
and
with
peak
availability
of
resources
for
herbivores
and
frugivores
during
the
spring-summer
months
(October–March,
Agostini
et
al.,
2010).
Mean
monthly
temperatures
vary
between
17
and
22 C
and
annual
rainfall
is
about
2000
mm
with
no
distinctive
rainy
season
(Agostini
et
al.,
2010).
One
of
the
surveys
was
conducted
during
2003
in
the
upper
basin
of
the
Urugua-í
river
(2558S,
5406W;
Fig.
1),
comprising
a
portion
of
the
privately
owned
Urugua-í
Wildlife
Reserve
(UWR),
a
portion
of
the
Urugua-í
Provincial
Park
(UPP),
and
a
section
of
a
private
property
(Campo
Los
Palmitos)
belonging
to
a
timber
com-
pany.
In
both
UWR
and
UPP
timber
was
selectively
extracted
until
the
late
1980s,
although
native
forest
remained
in
relatively
good
condition
(Di
Bitetti
et
al.,
2006).
Another
study
site
was
the
Yabotí
Biosphere
Reserve
(YBR),
located
in
the
southeastern
portion
of
the
Green
Corridor
(2655
S,
5400W;
Fig.
1),
where
we
conducted
a
camera-trap
survey
in
2005.
This
multiple
use
reserve
contains
strictly
protected
areas,
private
lands
with
selective
logging,
and
areas
inhabited
by
indige-
nous
people
of
the
Mbya-Guaraní
ethnic
group
who
practice
small
scale
agriculture
and
subsistence
hunting.
At
this
site
camera-trap
stations
were
located
along
old
logging
roads
within
the
Esmeralda
Provincial
Park
(EPP)
and
some
private
properties
(Di
Bitetti
et
al.,
2008;
Paviolo
et
al.,
2008).
The
third
study
site
was
Iguazú
National
Park
(INP,
2540S,
5430W;
Fig.
1),
located
in
northern
Misiones,
bordering
UPP
in
the
south
and
Iguac¸
u
National
Park
of
Brazil
in
the
north.
INP
was
selectively
logged
until
1934
when
the
park
was
created.
Three
surveys
were
conducted
in
this
area
(in
2004,
2006
and
2008).
The
first
survey
(2004)
was
conducted
in
the
central
part
of
INP,
while
378
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
the
second
survey
(2006)
also
included
the
Iguazú
National
Reserve
and
the
privately
owned
San
Jorge
Forest
Reserve.
The
2008
survey
in
INP
included
the
previous
surveyed
areas
but
extended
south-
ward
to
the
northern
part
of
the
Urugua-í
survey
site
(Di
Bitetti
et
al.,
2010).
Protection
against
poaching
Poaching
of
wildlife
is
a
relatively
common
recreational,
although
illegal,
activity
performed
by
both
rural
and
urban
peo-
ple
in
Misiones.
The
most
common
hunting
technique
in
the
area
is
with
firearms,
either
at
stands
(fruiting
trees
or
artificial
salt
licks)
or
with
the
aid
of
dogs.
Most
poaching
occurs
during
day-
light
hours
and
the
first
hours
of
the
night.
We
used
indirect
evidence
to
estimate
the
level
of
poaching.
First,
we
recorded
signs
of
poaching
activities
(poaching
campsites,
spent
cartridges,
artifi-
cial
saltlicks,
etc.)
during
our
field
activities.
Second,
we
conducted
informal
interviews
with
park
rangers,
biologists,
and
inhabitants
of
rural
areas
near
our
study
sites
about
the
modality
and
inten-
sity
of
poaching
in
the
area
and
the
principal
access
points
used
by
poachers
to
enter
the
study
areas
(roads,
forest
borders,
rivers,
etc.).
Finally,
we
quantified
the
resources
(personnel,
infrastruc-
ture,
etc.)
invested
in
anti-poaching
activities
in
each
study
area
as
an
index
of
protection.
Poaching
pressure,
as
ascertained
by
the
evi-
dence
found
and
interview
information,
was
negatively
associated
with
the
resources
invested
in
protection
and
with
the
distance
to
the
nearest
access
points
for
poachers
(DNAPP).
Thus,
we
catego-
rized
the
study
areas
into
three
relative
levels
of
protection:
“good”
(Iguazú)
“intermediate”
(Urugua-i)
and
“poor”
(Yabotí)
(see
details
in
Di
Bitetti
et
al.,
2008;
Paviolo
et
al.,
2008,
2009b).
We
also
used
the
DNAPP
for
each
survey
station
as
another
proxy
of
poaching
intensity
(Hill
et
al.,
1997;
Peres
and
Lake,
2003).
We
defined
as
“access
points”
all
routes
(vehicular
roads,
borders
of
forest
and
rivers)
known
to
be
used
by
poachers
based
on
interview
results
and
our
field
assessments.
Camera-trap
surveys
Camera-trap
surveys
were
designed
to
estimate
jaguar
densities
using
capture-recapture
models
(Paviolo
et
al.,
2008).
Each
camera-
trap
station
consisted
of
two
camera-traps
facing
each
other
on
both
sides
of
rarely
used
unpaved
roads
(2–6
m
wide)
and
small
trails
opened
with
a
machete
(1–2
m
wide).
These
small
trails
were
specifically
opened
for
this
study
and
do
not
constitute
access
roads
used
by
poachers.
Each
survey
had
a
preliminary
or
trial
period
and
a
systematic
sampling
period.
During
the
preliminary
period
the
location
of
the
stations,
the
number
of
cameras
per
station
and
the
duration
the
station
was
active
did
not
follow
standard
protocols
established
for
a
capture-recapture
survey
(see
Paviolo
et
al.,
2008).
The
systematic
period
was
divided
into
two
phases
to
increase
the
area
surveyed
with
the
camera-traps
available.
During
the
first
phase
camera-traps
were
located
in
half
of
the
stations
and
were
relocated
to
the
remaining
half
during
the
second
stage
of
the
survey
(Paviolo
et
al.,
2008).
The
systematic
period
lasted
90–96
days
(45–48
days
each
phase)
during
which
we
deployed
34–47
camera-trap
stations
(Table
A.1).
The
mean
distance
between
near-
est
camera-trap
stations
and
its
standard
deviation
during
the
systematic
periods
was
1960
±
767
m.
Cameras
were
active
24
h,
except
in
stations
with
relatively
frequent
vehicular
traffic
during
the
day;
these
cameras
were
active
only
at
night
when
traffic
was
almost
nil
(16
percent
of
the
2006
stations,
and
15
percent
of
the
2008
stations).
For
each
camera-trap
station
we
characterized
the
relative
abundance
of
bamboo
in
the
understory
as
being
high,
medium
or
low
(following
Di
Bitetti
et
al.,
2006).
The
location
of
each
camera-
trap
station
was
geo-referenced
and
incorporated
into
a
geographic
information
system
(GIS)
using
ArcView
3.2
(ESRI,
Redlands,
U.S.A.).
Using
the
GIS
we
measured
the
linear
distance
from
each
camera
station
to
the
closest
body
of
water
(rivers,
streams
and
lakes)
and
the
DNAPP.
Daily
activity
patterns
We
used
the
photographic
records
of
tapirs
from
both
the
pre-
liminary
and
systematic
periods
only
from
those
stations
that
were
active
24
h
to
document
the
daily
activity
pattern
of
tapirs
using
the
time
printed
on
the
photographs.
To
avoid
pseudo-replication
>1
h
had
to
pass
for
two
successive
photographs
of
tapirs
to
be
consid-
ered
independent
records,
resulting
in
703
independent
detections
(48
in
Urugua-í,
324
in
Iguazú
2004,
105
in
YBR,
99
in
Iguazú
2006
and
127
in
Iguazú
2008).
We
classified
photographs
as
diur-
nal
or
nocturnal
based
on
day-to-day
information
on
sunrise
and
sunset
for
this
location
(obtained
at
The
Weather
Channel
web
page;
http://espanol.weather.com/climate/sunRiseSunSet-Iguazu-
ARMS3370?month=2,
accessed
15.01.12).
To
assess
changes
in
the
daily
activity
pattern
in
relation
to
the
annual
photoperiod
we
compared
the
observed
activity
pattern
during
the
months
with
nights
of
shorter
duration
(10.15–11.22
h,
November–February),
the
period
where
the
availability
of
leaves
and
fruit
that
comprise
tapir
diet
is
highest
(Agostini
et
al.,
2010),
with
that
observed
dur-
ing
the
months
with
longer
nights
(12.25–13.28
h,
May–August)
where
the
availability
of
food
resources
is
lowest.
We
analyzed
the
variation
in
the
proportion
of
nocturnal
records
as
a
function
of
mean
daily
temperature
to
evaluate
whether
tapirs
are
more
nocturnal
in
days
with
higher
temperature.
We
also
ana-
lyzed
the
variation
of
the
mean
rate
of
tapir
records
(number
of
records/number
of
days)
in
relation
with
the
mean
daily
tempera-
ture
to
evaluate
whether
tapirs
are
less
active
during
days
of
higher
temperatures.
We
only
used
the
values
of
temperatures
(one
degree
intervals)
that
had
at
least
fifteen
days
representing
each
interval
(between
15 C
and
27 C,
N
=
493
records).
Temperature
records
were
obtained
from
the
National
Meteorological
Service
station
located
at
the
Iguazú
International
Airport
and,
consequently,
for
this
analysis
we
only
used
camera
trapping
data
from
the
INP
surveys
due
to
its
proximity
to
the
meteorological
station.
To
eval-
uate
whether
there
was
a
relationship
between
the
mean
rate
of
tapir
records
and
mean
daily
temperature,
and
also
between
the
proportion
of
nocturnal
records
and
mean
daily
temperature,
we
conducted
linear
regression
analyses
with
the
program
InfoStat
2010
(Di
Rienzo
et
al.,
2010).
To
analyze
the
daily
activity
pattern
in
relation
to
the
lunar
phase
we
used
a
moon
calendar
(Serrano,
2011)
and
considered
the
new
moon
day
to
be
the
first
day
of
each
lunar
cycle
and
the
full
moon
day
as
the
fifteenth
day.
We
divided
the
lunar
cycle
in
two
periods
of
ten
or
eleven
days;
the
first,
centered
on
the
new
moon,
comprised
days
1–5
and
25–29
or
25–30
and
the
second
period,
centered
on
the
full
moon,
comprised
days
10–20.
Transitional
days
(6–9
and
21–24)
were
omitted
from
the
analysis
(following
Di
Bitetti
et
al.,
2006).
To
evaluate
the
effect
of
photoperiod,
sex,
lunar
phase
and
pro-
tection
against
poaching
(with
two
levels;
good
vs.
poor
protection)
on
the
daily
activity
pattern
we
used
the
Mardia–Watson–Wheeler
test,
which
assess
whether
two
circular
distributions
differ
(Batschelet,
1981).
Habitat
use
We
used
occupancy
modeling
to
estimate
habitat
use
since
it
provides
an
unbiased
estimate
of
the
probability
that
a
species
occupies
an
area
( )
when
detection
probability
(p)
is
<1
while
allowing
modeling
of
the
effects
of
covariates
on
p
or
(MacKenzie
et
al.,
2002,
2006).
Since
the
distance
between
camera-trap
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
379
stations
did
not
ensure
that
detections
were
spatially
indepen-
dent
(we
detected
the
same
individual
at
two
different
stations)
should
be
interpreted
as
the
probability
of
habitat
use
by
tapirs
rather
than
occupancy
(MacKenzie
et
al.,
2006).
We
defined
habi-
tat
as
an
n-dimensional
space
which
is
a
function
of
environmental
and
anthropic
variables.
To
compile
a
sufficiently
large
number
of
detections
to
pro-
duce
robust
estimates
of
and
p
we
pooled
the
survey
data
from
Urugua-í
2003,
Iguazú
2004
and
Yabotí
2005
(N
=
117
stations;
total
effort
=
5205
camera-trap
days)
in
a
single
analysis.
A
fundamental
assumption
of
occupancy
analysis
is
that
populations
are
closed
(MacKenzie
et
al.,
2006),
though
this
assumption
could
be
relaxed
if
changes
in
a
population
occur
at
random
(Kendall,
1999).
In
comparing
data
collected
from
three
sites
over
approximately
two
years
we
assumed
that
the
tapir
population
was
stable
during
our
sampling
period
with
no
violation
of
the
closure
assumption.
We
believe
this
is
valid
since
longevity
and
generation
time
is
long
in
tapirs
(Barongi,
1993)
and
there
were
no
unusual
climatic
events
during
the
survey
period
that
could
have
potentially
altered
tapir
populations
or
behavior
sufficiently
to
bias
our
results.
Given
the
importance
of
poaching
on
other
tapir
populations
(Cullen
et
al.,
2001;
Peres,
2000,
2001),
comparing
the
three
sites
in
one
anal-
ysis
(instead
of
performing
one
analysis
per
survey)
allowed
us
to
explicitly
model
the
effect
of
the
level
of
legal
protection
on
tapir
occurrence.
Data
from
Iguazú
2006
and
2008
surveys
were
not
included
in
the
analysis
to
avoid
pseudo-replicates,
since
sev-
eral
stations
were
located
at
the
same
sites
during
the
three
Iguazú
surveys
(Di
Bitetti
et
al.,
2010).
To
reduce
variance
in
parameter
estimates
it
is
necessary
to
maximize
detection
probabilities
(MacKenzie
and
Royle,
2005).
Thus,
we
combined
every
four
consecutive
trap
days
into
trapping
occasions,
resulting
in
10
sampling
occasions.
Since
a
fundamental
assumption
of
the
single-season
occupancy
model
is
that
hetero-
geneity
is
accounted
for
through
the
inclusion
of
the
effects
of
covariates
(MacKenzie
et
al.,
2002,
2006),
we
incorporated
the
effects
of
several
covariates
on
and
p
which
we
deemed
important
based
upon
our
experience
and
the
literature.
We
included
as
covariates
affecting
the
distance
to
the
nearest
access
point
for
poachers
(DNAPP,
continuous
variable),
the
relative
abundance
of
bamboo
(BA,
ordinal
variable:
low,
medium,
high),
the
distance
to
the
nearest
water
body
(DW,
continuous
variable),
and
the
level
of
protection
against
poaching
(PROTECT,
ordinal
vari-
able:
good,
intermediate,
poor).
For
p
we
included
as
covariates
DNAPP,
BA,
PROTECT,
and
the
width
of
the
trail
or
road
where
the
trap
station
was
located
(TW,
categorical
variable:
narrow
vs.
wide).
Since
the
survey
was
designed
to
maximize
records
of
large
felids,
which
make
ample
use
of
roads
and
wide
trails
(Harmsen
et
al.,
2010;
Tobler
and
Powell,
2013),
a
large
proportion
(50%
on
average)
of
camera-traps
were
placed
on
wide
(2–6
m)
roads.
We
included
TW
as
a
variable
in
the
analyses
because
trail
width
may
affect
the
detectability
of
a
species
(Weckel
et
al.,
2006).
Variables
were
z-
transformed
so
their
means
were
equal
to
0
(Donovan
and
Hines,
2007).
We
expected
that
DNAPP
and
PROTECT
would
positively
affect
detection
because
tapirs
would
be
less
elusive
as
these
covariates
increased
in
value:
for
example;
in
sites
with
high
level
of
poaching
animals
may
avoid
camera-traps
due
to
perceived
signs
of
human
activity,
such
as
odors
(Séquin
et
al.,
2003;
Séquin
Larrucea
et
al.,
2006).
In
areas
with
greater
BA
we
expected
that
detection
of
tapirs
would
be
reduced
due
to
the
density
of
vegetation,
while
we
also
expected
that
trail
width
(TW)
would
negatively
affect
detection
due
to
avoidance
of
wider
roads
that
are
more
used
by
humans.
Since
the
covariates
could
plausibly
affect
either
or
p
sepa-
rately
or
in
combination,
we
modeled
all
combinations
of
covariates
using
a
single-species
single-season
occupancy
model
in
program
PRESENCE
3.1
(Hines,
2006),
resulting
in
a
candidate
set
of
128
models.
Models
were
ranked
using
Akaike’s
Information
Criterion
(AIC)
model
weights.
We
employed
model
averaging,
selecting
those
models
with
weights
within
10
percent
of
the
highest
ranked
model
(Burnham
and
Anderson,
2002).
The
significance
of
the
effect
of
covariates
on
and
p
was
deter-
mined
using
the
model
averaged
parameter
estimates.
Where
the
95
percent
confidence
intervals
of
the
parameter
estimates
did
not
include
zero
the
effect
of
the
covariate
was
considered
to
be
significative
(MacKenzie
et
al.,
2006).
One-tailed
tests
were
used,
establishing
an
overall
alpha
level
of
0.05
for
committing
a
type
one
error.
Results
Daily
activity
patterns
Tapirs
were
mostly
nocturnal,
with
89
percent
of
records
occur-
ring
between
sunset
and
sunrise
(Fig.
2).
Activity
began
around
1800
h
and
finished
around
0700
h,
with
reduced
nocturnal
records
between
2200
h
and
0200
h
(Fig.
2).
The
proportion
of
nocturnal
records
did
not
increase
significantly
with
mean
daily
tempera-
ture
(R2=
0.07,
P
=
0.37),
and
the
mean
rate
of
tapir
records
did
not
decrease
significantly
with
mean
daily
temperature
(R2=
0.03,
P
=
0.61).
The
daily
activity
pattern
was
not
affected
significantly
by
the
annual
photoperiod
(Mardia–Watson–Wheleer
test,
2=
3.02,
df
=
2,
P
=
0.22),
the
lunar
cycle
(2=
1.15,
df
=
2,
P
=
0.56),
the
sex
of
the
individuals
(2=
1.16,
df
=
2,
P
=
0.56),
or
the
degree
of
protec-
tion
of
the
area
(2=
2.13,
df
=
2,
P
=
0.34).
Patterns
of
habitat
use
Nineteen
models
were
selected
for
model
averaging
and
all
covariates
used
in
the
analysis
were
included
in
the
best
model
set
(Tables
1
and
A.1).
The
and
p
constant
model
( (*),
p(*)),
with
an
AIC
weight
of
0
and
a
AIC
of
69.52
was
not
included
within
these
19
top
models.
Model
fit
was
good
with
a
value
of
c-hat
for
the
global
model
of
1.
Based
upon
the
model
averaged
parameter
esti-
mates,
both
DNAPP
and
PROTECT
had
a
significant
positive
effect
on
,
since
zero
was
not
included
within
the
95%
confidence
interval
of
the
parameter
estimates
(Table
2,
Fig.
3).
Based
upon
the
com-
posite
model
from
the
model
averaging,
was
estimated
to
be
0.83
(±0.025)
for
the
good
protection
level
(Iguazú),
0.50
(±0.025)
for
the
intermediate
protection
level
(Urugua-í)
and
0.40
(±0.062)
for
the
poor
protection
level
(YBR).
Although
DW
and
BA
demonstrated
the
expected
directional
response,
the
effect
of
those
covariates
on
was
not
significant
(Table
2,
Fig.
3).
There
was
a
significant
negative
effect
of
TW
and
BA,
and
a
significant
positive
effect
of
DNAPP
on
p
(Table
2).
Discussion
Daily
activity
patterns
In
accordance
with
previous
research,
we
found
that
lowland
tapir
activity
is
mostly
nocturnal,
where
peak
hours
of
movement
are
after
sundown
and
before
sunrise
(Noss
et
al.,
2003;
Tobler,
2008;
Medici,
2010;
Wallace
et
al.,
2012).
This
pattern
has
also
been
described
for
other
tapir
species
(Lizcano
and
Cavelier,
2000b;
Foerster
and
Vaughan,
2002;
Holden
et
al.,
2003).
Tapir
evidently
remain
active
throughout
the
night,
but
reduce
their
displacements
around
midnight
because
it
is
the
period
when
they
concentrate
their
foraging
activity
(Tobler,
2008;
Medici,
2010).
Differential
use
of
habitats
between
day
and
night
has
been
documented
for
tapirs,
with
individuals
resting
in
dense
vegetation
by
day,
and
visiting
foraging
sites
from
sunset
until
sunrise
(Padilla
and
Dowler,
1994;
380
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
Fig.
2.
Daily
activity
pattern
of
the
lowland
tapir
in
the
Atlantic
Forest
of
Argentina
(N
=
703).
The
proportion
of
records
was
calculated
using
the
number
of
records
per
hour
divided
by
the
number
of
totals
records.
The
data
belong
to
the
surveys
conducted
between
2003
and
2008.
Table
1
PRESENCE
output
with
the
top
ranking
models
with
2
AIC
value.
Models
were
ranked
using
Akaike’s
Information
Criterion
(AIC).
AIC
and
AIC
weight
are
calculated
from
AIC.
ModelaAIC
AIC
weight
Model
likelihood
#
Parameters
(DNAPP,
BA,
PROTECT),
p(DNAPP,
TW,
BA)
0
0.1475
1.000
8
(DNAPP,
DW,
BA,
PROTECT),
p(DNAPP,
TW,
BA)
0.61
0.1087
0.7371
9
(DNAPP,
PROTECT),
p(DNAPP,
TW,
BA)
0.98
0.0903
0.6126
7
(DNAPP,
DW,
PROTECT),
p(DNAPP,
TW,
BA)
0.98
0.0903
0.6126
8
(DNAPP,
BA,
PROTECT),
p(DNAPP,
TW,
BA,
PROTECT) 1.96 0.0553 0.3753
9
(BA,
PROTECT),
p(DNAPP,
TW,
BA)
2.04
0.0532
0.3606
7
(DNAPP,
BA,
PROTECT),
p(DNAPP,
TW)
2.07
0.0524
0.3552
7
a :
probability
of
habitat
use;
p:
detection
probability;
BA:
relative
abundance
of
bamboo;
DNAPP:
distance
to
the
nearest
access
points
for
poachers;
DW:
distance
to
closest
water
body;
PROTECT:
relative
level
of
protection
against
poaching;
TW:
trail
or
road
width.
Tobler,
2008;
Medici,
2010).
It
is
possible
that
the
activity
pattern
we
observed
reflects
this
same
behavior.
Unlike
most
large
mammals
(>100
kg),
which
need
to
be
active
both
diurnally
and
nocturnally
to
meet
energetic
needs
(van
Schaik
and
Griffiths,
1996),
tapirs
at
our
sites
are
strictly
nocturnal.
Tapirs
may
be
nocturnal
because
their
low
surface
to
volume
ratio
makes
body
heat
dispersal
difficult
(Randall
et
al.,
1998;
Foerster
and
Vaughan,
2002;
Medici,
2010).
If
thermoregulation
were
the
driv-
ing
factor
in
determining
daily
movements,
as
for
other
mammal
species
(Morão
and
Medri,
2007),
we
would
expect
a
seasonal
shift
in
activity
patterns
in
our
study
area
given
the
variation
in
mean
temperature
between
winter
and
summer.
However,
we
did
not
find
a
relationship
between
either
the
proportion
of
nocturnal
detections
or
the
recording
rate
and
the
mean
daily
temperature.
Contrary
to
what
has
been
reported
by
Lizcano
and
Cavelier
(2000b)
and
Medici
(2010),
but
in
concordance
with
Coelho
et
al.
(2008)
and
Link
et
al.
(2012),
lunar
phase
did
not
affect
tapir
activ-
ity
in
our
study
area.
The
differences
among
study
sites
may
be
due
to
higher
levels
of
predation
risk
during
the
full
moon
in
some
regions
(e.g.
Medici,
2010).
At
our
study
sites
predation
risk
is
min-
imal
since
jaguar
densities
(probably
the
only
important
predator
of
tapirs
besides
humans
in
most
of
the
Neotropics)
are
relatively
low
(Paviolo
et
al.,
2008)
and,
consequently,
the
potential
effect
of
predation
on
activity
patterns
is
reduced.
We
also
did
not
observe
an
effect
of
the
degree
of
protection
from
poaching.
Tapirs
may
fail
to
modify
behavior
regardless
of
poaching
pressure
because
of
a
low
plasticity
to
alter
their
circadian
activity
or
because
this
conduct
is
a
pre-adaptation
(exaptation,
sensu
Futuyma,
1997),
with
the
observed
daily
activity
pattern
being
the
best
strategy
to
Table
2
Model
averaged
beta
values,
standard
errors,
and
95%
confidence
intervals
for
the
variables
that
affect
the
probability
of
habitat
use
( )
and
the
probability
of
detection
(p)
of
lowland
tapirs
in
the
Atlantic
Forest
of
Misiones,
Argentina.
VariableaParameter
estimates
Standard
error
95%
confidence
intervalb
Lower
limit
Upper
limit
DNAPP
0.437
0.250
0.027
0.848
DW
0.116 0.123
0.317
0.085
BA
0.238
0.177
0.529
0.053
PROTECT
0.881
0.387
0.246
1.515
pDNAPP
0.430
0.089
0.284
0.576
TW
0.783
0.203
1.117
0.450
BA
0.154
0.082
0.289
0.020
PROTECT
0.004
0.027
0.040
0.048
aBA:
relative
abundance
of
bamboo;
DNAPP:
distance
to
the
nearest
access
points
for
poachers;
DW:
distance
to
closest
water
body;
PROTECT:
relative
level
of
protection
against
poaching;
TW:
trail
or
road
width.
bSince
the
hypotheses
are
one-tailed
(directional),
the
error
is
concentrated
at
one
extremity.
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
381
Fig.
3.
The
estimated
probability
of
habitat
use
of
tapirs
( )
in
relation
to
the
distance
to
the
nearest
access
points
for
poachers
(DNAPP),
the
distance
to
closest
water
body
(DW),
the
relative
level
of
protection
against
poaching
(PROTECT),
and
the
relative
abundance
of
bamboo
(BA).
was
estimated
based
on
the
composite
model
and
the
parameters
estimates
resulted
from
the
model
averaging.
minimize
the
risk
of
being
killed
by
poachers.
The
similarity
in
the
daily
activity
pattern
among
populations
of
lowland
tapirs
across
the
distribution
of
the
species,
as
well
as
with
other
tapir
species,
suggests
a
lack
of
flexibility
in
this
behavioral
trait
despite
changing
environmental
conditions.
It
also
suggests
that
the
daily
activity
of
tapirs
may
be
phylogenetically
constrained.
Some
morphological
characteristics,
such
as
poor
vision
and
an
acute
sense
of
smell
and
hearing,
indicate
that
tapirs
are
primarily
adapted
for
nocturnal
activity.
Patterns
in
habitat
use
The
occupancy
modeling
suggests
that
poaching
is
one
of
the
factors
with
the
greatest
effect
on
habitat
use
by
tapirs
since
both
the
level
of
protection
and
the
distance
to
nearest
access
points
for
poachers
(DNAPP)
had
significant
positive
effects
on
the
presence
of
tapirs.
Similar
results
were
reported
by
Licona
et
al.
(2011),
who
observed
lower
habitat
use
of
tapirs
in
sites
with
higher
human
accessibility.
We
found
no
effect
of
protection
on
detectability,
suggesting
that
there
were
no
large-scale
effects
on
tapir
behavior
and
any
factors
that
affect
detectability
occur
at
more
localized
scales.
This
is
supported
by
the
observed
reduction
in
detection
related
to
increased
accessibility
and
trail/road
width
which
could
result
from
active
avoidance
of
areas
frequently
used
by
people
or
from
eva-
sive
behaviors
in
response
to
camera-traps
(probably
associated
to
human
activities).
Lowland
tapirs
have
been
shown
to
avoid
closed
understory
areas
with
high
density
of
the
bamboo
Chusquea
ramosissima
in
the
Upper
Paraná
Atlantic
Forest,
which
potentially
hinders
their
movements
(Gallardo
et
al.,
2008),
as
well
as
associated
with
areas
close
to
water;
a
tendency
that
has
been
related
to
thermoregu-
lation,
ectoparasite
avoidance
and
resting
behaviors
(Padilla
and
Dowler,
1994;
Foerster
and
Vaughan,
2002).
For
our
study
area,
however,
the
model
averaged
estimates
for
these
covariates
did
not
indicate
a
significant
effect
of
either
bamboo
density
or
distance
to
water
on
site
use
by
tapirs.
As
expected,
however,
we
found
a
nega-
tive
relationship
between
detection
and
bamboo
abundance
which
382
P.
Cruz
et
al.
/
Mammalian
Biology
79
(2014)
376–383
we
potentially
attribute
to
the
dense
bamboo
layer
decreasing
the
ability
of
cameras
to
detect
tapirs.
The
lower
probability
of
habitat
use
of
tapirs
in
less
protected
areas
suggests
a
population
response
to
poaching.
The
increase
in
habitat
use
and
detection
with
the
distance
to
access
roads
indicate
differences
in
abundance
and
behavioral
responses
respectively
in
relation
to
differences
in
human
activity.
There
is
evidence
that
animals
tend
to
avoid
areas
perceived
as
being
of
higher
risk
of
pre-
dation
or
hunting
(Laundre
et
al.,
2010).
Furthermore,
if
occupancy
is
interpreted
as
a
surrogate
for
abundance
(MacKenzie
et
al.,
2004)
we
can
infer
that
the
negative
relationship
of
distance
to
access
points
with
the
use
of
an
habitat
is
a
result
of
decreasing
abun-
dance
of
tapirs,
which
is
consistent
with
previous
research
where
tapir
abundance
and
poaching
level
were
negatively
related
(Cullen
et
al.,
2001;
Peres,
2000,
2001).
Medici
et
al.
(2007)
conducted
a
population
and
habitat
viabil-
ity
analysis
for
lowland
tapirs
and
found
that
poaching
was
one
of
the
major
threats
for
the
species.
Tapirs
may
be
especially
vul-
nerable
to
hunting,
compared
to
other
ungulates,
because
of
their
relatively
low
population
density
coupled
with
their
life
history
traits
(low
reproductive
rate;
Bodmer
et
al.,
1997).
Poaching
may
not
only
negatively
affect
tapir
populations
but
may
also
indi-
rectly
affect
the
composition
and
structure
of
the
forest,
since
tapirs
play
an
important
ecological
role,
affecting
the
structure,
composition,
growth
and
regeneration
of
the
vegetation
(Bodmer,
1991;
Dirzo
and
Miranda,
1991;
Fragoso,
1997;
Wright
et
al.,
2000).
The
Green
Corridor
of
Misiones
is
one
of
the
few
areas
of
Upper
Paraná
Atlantic
Forest
with
high
probabilities
of
main-
taining
lowland
tapir
populations
in
the
long
term
(Paviolo
et
al.,
2008;
Taber
et
al.,
2008).
However,
a
large
portion
of
the
Green
Corridor
suffers
from
medium
to
high
poaching
pressure;
which,
as
our
study
suggests,
negatively
affects
the
abundance
and
the
patterns
of
habitat
use
by
tapirs.
Thus,
as
a
conse-
quence
of
poaching
most
of
the
largest
remnants
of
Upper
Paraná
Atlantic
Forest
contains
no
or
very
low
densities
of
tapirs
(Cullen
et
al.,
2001;
Paviolo
et
al.,
2009a).
To
preserve
an
important
and
well
preserved
area
of
the
Upper
Paraná
Atlantic
Forest,
where
the
ecological
and
evolutionary
processes
involv-
ing
lowland
tapirs
persist,
Argentinian,
Brazilian
and
Paraguayan
authorities
should
improve
forest
protection
and
reduce
levels
of
poaching.
Acknowledgements
We
are
grateful
to
all
the
volunteers
and
park
rangers
that
helped
us
with
fieldwork.
We
acknowledge
the
support
and
permits
provided
by
the
Ministry
of
Ecology,
Natural
Resources
and
Tourism
of
Misiones
Province
(MERNRT)
and
the
National
Parks
Adminis-
tration
of
Argentina.
We
thank
A.
Ricieri
and
A.
Bertrand
for
their
help
and
permission
during
survey
at
Iguac¸
u
National
Park.
We
are
grateful
to
Fundación
Vida
Silvestre
Argentina
(FVSA)
and
prop-
erty
owners
for
their
support
and
permission
to
conduct
this
work.
Financial
support
was
provided
by
CONICET,
FVSA,
WWF
USA,
WWF
International,
WWF
Switzerland,
Lincoln
Park
Zoo,
Fun-
dación
Antorchas,
Wildlife
Conservation
Society,
Idea
Wild,
Rufford
Small
Grant
Foundation,
and
Eden
Project
through
a
grant
from
the
Darwin
Initiative.
Appendix
A.
Supplementary
data
Supplementary
data
associated
with
this
article
can
be
found,
in
the
online
version,
at
http://dx.doi.org/10.1016/j.mambio.
2014.06.003.
References
Agostini,
I.,
Holzmann,
I.,
Di
Bitetti,
M.S.,
2010.
Are
howler
monkey
species
ecolog-
ically
equivalent?
Trophic
niche
overlap
in
syntopic
Alouatta
guariba
clamitans
and
Alouatta
caraya.
Am.
J.
Primatol.
72,
173–186.
Barongi,
R.A.,
1993.
Husbandry
and
conservation
of
tapirs.
Int.
Zoo
Yearb.
32,
7–15.
Batschelet,
E.,
1981.
Circular
Statistics
in
Biology.
Academic
Press,
New
York.
Bodmer,
R.E.,
1991.
Strategies
of
seed
dispersal
and
seed
predation
in
Amazonian
ungulates.
Biotropica
23,
255–261.
Bodmer,
R.E.,
Eisenberg,
J.F.,
Redford,
K.H.,
1997.
Hunting
and
the
likelihood
of
extinction
of
Amazonian
mammals.
Conserv.
Biol.
11,
460–466.
Burnham,
K.P.,
Anderson,
D.R.,
2002.
Model
Selection
and
Inference:
A
Practical
Information-theoretic
Approach,
2nd
ed.
Springer-Verlag,
New
York.
Chalukian,
S.,
de
Bustos,
S.,
Lizárraga,
L.,
Quse,
V.,
Paviolo,
A.,
Varela,
D.,
2009.
Plan
de
acción
para
la
conservación
del
tapir
(Tapirus
terrestris)
en
Argentina.
Grupo
de
especialista
de
Tapires,
IUCN.
Chalukian,
S.,
de
Bustos,
S.,
Di
Bitetti,
M.,
De
Angelo,
C.,
Paviolo,
A.,
2012.
Orden
Perissodactyla.
In:
Ojeda,
R.A.,
Chillo,
V.,
Díaz
Isenrath,
G.B.
(Eds.),
Libro
Rojo
de
los
Mamíferos
Amenazados
de
la
Argentina.
SAREM,
Argentina,
p.
116.
Cullen
Jr.,
L.,
Bodmer,
E.,
Valladares-Padua,
C.,
2001.
Ecological
consequences
of
hunting
in
Atlantic
forest
patches,
Sao
Paulo,
Brazil.
Oryx
35,
137–144.
Coelho,
I.P.,
Oliveira,
L.F.,
Oliveira,
M.E.,
2008.
Does
Moonlight
Affect
the
Use
of
Natu-
ral
Licks
by
Lowland
Tapir
(Tapirus
terrestris
Linnaeus,
1758)
in
the
Northeastern
Brazilian
Pantanal?,
vol.
17.
Newsletter
IUCN/SSC
Tapir
Specialist
Group,
pp.
10–14.
Di
Bitetti,
M.S.,
De
Angelo,
C.,
Di
Blanco,
Y.E.,
Paviolo,
A.,
2010.
Niche
partitioning
and
species
coexistence
in
a
Neotropical
felid
assemblage.
Acta
Oecol.
34,
403–412.
Di
Bitetti,
M.S.,
Paviolo,
A.,
De
Angelo,
C.,
2006.
Density,
habitat
use,
and
activity
pat-
terns
of
ocelots
Leopardus
pardalis
in
the
Atlantic
Forest
of
Misiones,
Argentina.
J.
Zool.
270,
153–163.
Di
Bitetti,
M.S.,
Paviolo,
A.,
Ferrari,
C.,
De
Angelo,
C.,
Di
Blanco,
Y.E.,
2008.
Differ-
ential
responses
to
hunting
in
two
sympatric
species
of
brocket
deer
(Mazama
americana
and
Mazama
nana).
Biotropica
40,
636–645.
Di
Rienzo,
J.A.,
Casanoves,
F.,
Balzarini,
M.G.,
González,
L.,
Tablada,
M.,
Robledo,
C.W.,
2010.
InfoStat
versión.
Grupo
InfoStat,
FCA,
Universidad
Nacional
de
Córdoba,
Argentina.
Dirzo,
R.,
Miranda,
A.,
1991.
Altered
patterns
of
herbivory
and
diversity
in
the
forest
understory:
a
case
of
the
possible
consequences
of
contemporary
defaunation.
In:
Pric,
W.P.,
Lewinson,
M.,
Wilson,
G.,
Woodruff,
W.B.
(Eds.),
Plant–Animal
Interactions:
Evolutionary
Ecology
in
Tropical
and
Template
Regions.
John
Wiley
and
Sons,
New
York,
USA,
pp.
273–447.
Donovan,
T.M.,
Hines,
J.,
2007.
Exercises
in
Occupancy
Modeling
and
Estimation.
Foerster,
C.R.,
Vaughan,
C.,
2002.
Home
range,
habitat
use,
and
activity
of
Baird’s
Tapir
in
Costa
Rica.
Biotropica
34,
423–437.
Fragoso,
J.M.,
1997.
Tapir-generated
seed
shadows:
scale-dependent
patchiness
in
the
Amazon
Rain
Forest.
J.
Ecol.
85,
519–529.
Futuyma,
D.J.,
1997.
Evolutionary
Biology,
3rd
ed.
Sinauer,
Sunderland,
MA.
Gallardo,
A.,
Montti,
L.,
Bravo,
S.P.,
2008.
Efectos
del
tacuarembó
(Chusquea
ramosis-
sima,
Poaceae)
sobre
el
proceso
de
dispersión
de
semillas
en
la
Selva
Misionera.
Ecol.
Austr.
18,
347–356.
Giraudo,
A.R.,
Abramson,
R.R.,
2000.
Diversidad
cultural
y
usos
de
la
fauna
silvestre
por
los
pobladores
de
la
selva
misionera:
¿Una
alternativa
de
conservación?
In:
Bertonatti,
C.,
Corcuera,
J.
(Eds.),
La
situación