Content uploaded by Gerardo E Soto
Author content
All content in this area was uploaded by Gerardo E Soto on Mar 24, 2016
Content may be subject to copyright.
1
Gayana 77(1): 1-9, 2013. ISSN 0717-652X
First description of the micro-habitat selection pattern of the island
endemic Juan Fernandez Tit-tyrant
Primera descripción del patrón de selección de micro-hábitat del Cachudito de Juan
Fernández
GERARDO E SOTO1,*, PABLO M VERGARA1, INGO J HAHN3, CHRISTIAN G PÉREZ-HERNÁNDEZ1, MARLENE E
LIZAMA1, JULIA BAUMEISTER3 & JAIME PIZARRO2
1Departamento de Gestión Agraria, Universidad de Santiago de Chile, Alameda Libertador Bernardo O’Higgins 3363, Estación
Central, Santiago, Región Metropolitana, Chile.
2Departamento de Ingeniería Geográfi ca, Universidad de Santiago de Chile, Alameda Libertador Bernardo O’Higgins 3363,
Estación Central, Santiago, Región Metropolitana, Chile.
3Institute of Landscape Ecology, University of Münster, Robert-Koch-Str. 28, D-48149 Münster, Germany
* E-mail: gerardo.soto@usach.cl
RESUMEN
Para las aves que son endémicas de islas oceánicas el nivel de especialización en el uso del hábitat puede ser un factor
importante de considerar en el incremento del riesgo de extinción de estas especies. En este estudio, usamos datos de redes
de niebla y radiotelemetría para determinar el patrón de uso de micro-hábitat por el Cachudito de Juan Fernández (Anairetes
fernandezianus), una especie de ave endémica del archipiélago de Juan Fernández (Chile). El individuo seguido por
radiotelemetría estableció su ámbito de hogar exclusivamente en áreas centrales del bosque nativo residual. Los modelos
regresivos de abundancia y las Funciones de Utilización de Recursos mostraron fuertes preferencias de micro-hábitat
por los cachuditos. Los valores de abundancia y utilización de recurso por los cachuditos decrecieron con la distancia
al matorral exótico y aumentaron con la distancia a los claros hechos por humanos dentro bosque. La heterogeneidad
natural en las condiciones del micro-hábitat también afectan el patrón de selección de micro-hábitat, con probabilidades de
uso y abundancia aumentando en las proximidades de lugares húmedos dominados por Gunnera peltata. Estos hallazgos
demuestran que los cachuditos no usan el bosque nativo al azar. Por lo tanto, la conservación de esta población de cachuditos,
que ha experimentado una fuerte declinación en las ultimas décadas, puede ser mejorada incorporando estas preferencias de
micro-hábitat en los programas de manejo.
PALABRAS CLAVE: Isla Robinson Crusoe, Anairetes fernandezianus, Micro-hábitat.
ABSTRACT
The level of specialization of endemic island birds to their native habitats could contribute importantly to increase their
extinction risk. We used abundance obtained from mist-netting and radio telemetry data from one individual to determine
the micro-habitat use pattern of Juan Fernandez Tit-tyrants (Anairetes fernandezianus) within native forest. The tracked
male tit-tyrant established its home range exclusively in core native forest areas. Regression models of abundance and
Resource Utilization Functions showed strong micro-habitat preferences of tit-tyrants. The abundance and utilization
probabilities of tit-tyrants declined with distance to exotic shrub and increased with distance to human created gaps.
Natural heterogeneity in micro-habitat conditions also affected the micro-habitat selection pattern, with use probabilities
and abundance increasing with proximity to the humid sites dominated by Gunnera peltata. These fi ndings demonstrate
that tit-tyrants do not use native forest sites at random. Therefore, the conservation and recovery of the declining tit-tyrant
population could be improved by incorporating such micro-habitat preferences in habitat management programs.
KEYWORDS: Robinson Crusoe Island, Anairetes fernandezianus, micro-habitat.
INTRODUCTION
Bird species endemic to oceanic islands tend to be highly
vulnerable to habitat loss and degradation (Owens &
Bennett 2000; Fordham and Brook 2010). Because their
populations have natural small sizes and have been exposed
to a long-term historical isolation, bird species endemic to
oceanic islands tend to have higher extinction rates than
their mainland relatives (Loehle and Eschenbach 2012).
In addition, the level of specialization of endemic island
Gayana 77(1), 2013
2
birds to native habitats which are being lost and disturbed
by humans could contribute importantly to increase their
extinction risk (Hahn et al. 2011a, Cofré 1999).
The Juan Fernandez Tit-tyrant Anairetes fernandezianus is
a small insectivorous species endemic to Robinson Crusoe
Island and considered as near threatened (IUCN 2011).
Previous studies suggest that this tit-tyrant species prefers
to use native forests instead of human-associated habitats
with their nests being located in the native canopy (Hahn
et al. 2005, 2006, 2011a, 2011b). The loss and degradation
of native forest by human activities and introduced
herbivores currently represents the major conservation
problem for species endemic to Robinson Crusoe like
tit-tyrants (Cuevas and van Leersum 2001; Ricci 2006).
This habitat loss has contributed to the reduction of the
population size of the tit-tyrant, which has experienced a
signifi cant decline of about 63 percent during the last 15
years (Hahn et al. 2011a).
Understanding how birds select and use their habitats
may be considered as a primary step for planning and
management of conservation areas like Robinson Crusoe
Island, which has been declared National Park since 1935
and an UNESCO Biosphere Reserve since 1977. However,
habitat selection is a scale-dependent process resulting from
spatial variation in how birds perceive habitat features and
the spatial distribution of habitat components (Kristan 2006;
Vergara & Armesto 2009). This implies that the ecological
responses of tit-tyrants to micro-habitat structures and
environmental gradients within the native forest vegetation
may be crucial for planning conservation activities for this
small tit-tyrant population.
In this study, we aim to report the fi rst home range estimate
of a Juan Fernandez Tit-tyrant and assess the micro-habitat
use pattern of this species. Although coarse scale habitat
selection has been described for the Juan Fernandez Tit-
tyrant (e.g., Roy et al. 1999; Hahn 2006; Hahn et al. 2005,
2006, 2011a, 2011b), their micro-habitat preferences within
the native forest still remain unknown. The native island
forest is naturally heterogeneous and surrounded mainly by
dense shrub formations of exotic species which have invaded
areas where native forest was cut (Hahn et al. 2010). The
unsuitable conditions in the exotic shrub matrix may spread
into the forest, thus affecting the habitat selection pattern of
tit-tyrants. In addition, the presence of forest gaps created
by the elimination of invasive plants within the forest, could
have further consequences on the habitat use pattern of tit-
tyrants.
METHODS
STUDY SITE
Robinson Crusoe Island (47.1 km2) is located ca. 600
km off the Chilean coast in the South-east Pacifi c Ocean
(33° 33’ 38’’ S and 78° 56’ 44’’ W). It was once largely
covered by native forest but since the European discovery
in the year 1574 habitats have been extensively degraded
and lost (Skottsberg 1956; Stuessy and Ono 1998). These
accumulated disturbances have resulted in a variety of
different human-induced habitats for native bird species
such as lowland scrub (290 ha), ridge scrub (599 ha),
settlement area (109 ha) and non-native grassland (2,946
ha), reducing native forest to less than 25 % (i.e., 990 ha) of
their original size (Hahn et al. 2005). Of these habitats, Juan
Fernandez Tit-tyrants tend to select for native forest while
marginally using exotic scrub, ridge scrub and settlement
areas (Hahn et al. 2011a). The structure and composition
of native forest, however, is naturally heterogeneous due
to the variation in topography, exposure and proximity of
watersheds (Hahn et al. 2010).
Our study area (10 ha) was located in the “Plazoleta del
Yunque”, a relatively fl at site appropriate to carry out
telemetry studies (Fig. 1). Vegetation and micro-habitat
features of the study area were classifi ed on the basis of
geo-referenced habitat plots and high-resolution Google
Earth imagery. We quantifi ed two types of micro-habitat
variables which could be important in explaining the micro-
habitat use pattern of tit-tyrants, including: i) Composite-
based variables, which characterize the micro-habitat used
directly by the animal; ii) Distance-based variables, defi ned
as the Euclidean distance from animal locations to habitat
features (e.g., Aebischer et al. 1993; Conner et al. 2003).
In order to estimate micro-habitat features, vegetation
of the study site was spatially classifi ed as: 1) Gunnera
shrub, defi ned as sites dominated by Gunnera peltata, an
endemic large-leaved plant. Gunnera shrubs occur in the
moister areas, including steep slopes sites as well as along
watersheds, being occasionally mixed with Robinsonia spp.
and Juania australis. 2) Native forest dominated by trees
such as Myrceugenia fernandeziana, Drimys confertifolia
and Fagara mayu. In native forest, tit-tyrant nests are usually
located in Myrceugenia trees. 3) Forest gaps ranging between
0.01 and 0.7 ha in size, which were artifi cially generated to
eliminate exotic shrub patches in core areas of the native
forest remnants by a Chilean government agency (CONAF;
see Fig. 2). Consequently, forest gaps lack vegetation cover
but most of the ground is covered by branches from cut
shrubs. 4) Exotic shrub formations composed by a mix of
Aristotelia chilensis, Rubus ulmifolius and Ugni molinae.
3
Micro-habitat selection of Juan Fernandez Tit-tyrant: GERARDO SOTO ET AL.
FIGURE 1. Map of the study site showing the main habitat types in Robinson Crusoe Island including telemetry fi xes (estimated locations)
of the sampled tit-tyrant and mist-nets used for validating kernel estimates (see text).
FIGURA 1. Mapa del sitio de estudio que muestra los principales tipos de hábitat en la isla Robinson Crusoe incluyendo los puntos de
telemetría (localizaciones estimadas) del Cachudito muestreado y de las redes de niebla usadas para validar las estimaciones del kernel.
FIGURE 2. (Left) Details of a human-caused forest gap including cut branches accumulated on the ground and the adjacent forest edge,
Plazoleta de Yunque (photo credit: Gerardo E. Soto). (Right) A male Juan Fernandez Tit-tyrant (photo credit: Ingo J. Hahn).
FIGURA 2. (Izquierda) Detalle de un claro de bosque hecho por humanos que incluye las ramas apiladas en el suelo y el bosque nativo
adyacente, Plazoleta del Yunque (Fotografía: Gerardo E. Soto). (Derecha) Cachudito de Juan Fernández macho (Fotografía: Ingo J. Hahn).
Gayana 77(1), 2013
4
CAPTURE DATA
Fourteen 10-m long and 2.6-m height mist-nets were randomly
established in the study area in order to determine the micro-
habitat use pattern by tit-tyrants within forest habitat. All
mist-nets were placed 1.5 m above the ground, a height that
is consistent with the height at which tit-tyrant move (Hahn et
al. 2009). Of these nets, 10 were located in native forest and
4 in forest gaps, these being separated by more than 100 m
(Fig. 1). Mist netting stations were opened for 6 hrs a day and
checked every half hour for about 26 days during February
and March 2011, totalizing a per-net effort (hours x net) of
156 hrs. All captured birds were ringed and then released.
Bird abundance was estimated for each mist netting station
by dividing the number of caught birds (without including
recaptures) by the total number of hours per net.
RADIO-TRACKING
During March 2011, corresponding to the post-breeding
season of tit-tyrants, a 9 g male tit-tyrant was caught in a
mist-net and color-ringed as part of a long-term monitoring
program (Fig. 2). Because of the threat status of this species,
and considering that transmitters could induce mortality of
small bird species, we tracked a single individual. A 0.6 g
backpack-mounted transmitter was attached to this tit-tyrant
using epoxy glue, which was later released ca. 20 min after
its capture. Two telemetry sessions were conducted daily
(once in the morning and another in the afternoon) every
other day for a week, totaling six sessions. Each session
involved a 40 min sampling period with bearings being
recorded at 1 min interval in order to obtain a fi ne scale time
series. During the last telemetry session we realized that the
transmitter was detached due to no movement was recorded.
HOME RANGE
Bearings were triangulated using LOCATE II (Nams 2000).
We tested for “time to independence” (Swihart and Slade
1985) using the Schoener’s (1981) index, which estimates the
time interval on which data become temporally independent
of each other, with values larger than 1 being considered
as acceptable (Kenward 2001). Once fi xes (estimated
locations) were fi ltered for autocorrelation, we used them
to determine a kernel home range for the male tit-tyrant.
Two different bandwidths (i.e., smoothing parameters) were
used to calculate the kernel: Least Squared Cross-Validation
(LSCV) and Plug-in (Kertson and Marzluff 2010, and
references therein). After fi tting the kernel density function
we selected the LSCV bandwidth because the Plug-in
estimation resulted in an overestimation of the kernel area
in one spatial axis. We used Animal Movement extension
in ArcView 3.3 (Hooge & Eichenlaub 1997) and the
KernSmooth package in R 2.13.0 (Wand 2006; Wand and
Jones 1995) for estimating LSCV and Plug-in bandwidths,
respectively. We validated the resulting kernel home range
by comparing its use distribution values (i.e., the amount
of use at the location relative to the other locations in the
home range) at locations where the individual was mist
netted and the site where the transmitter was recovered
with distribution values drawn from random locations.
Additionally, georeferenced observations of the individual
were graphically compared to the home range area.
HABITAT SELECTION
We assessed micro-habitat selection at the population
level using abundance estimates from mist netting capture
data. Previously we determined that covariates were
not correlated (r < 0.5, for all variables). We used linear
regression models to test the effects of composite-based
and distance-based variables. We used an information-
theoretic approach of model selection based in the Akaike’s
information criterion (AIC) and Akaike weights to evaluate
the support for competing a priori models explaining bird
abundance. Our selection procedure involved all possible
combinations of variables in the set of competing models.
Competing models were ranked by their differences in AIC
with the most parsimonious model (DAIC) and models with
ΔAICc values ≤2 were considered to be strongly supported.
We also computed the adjusted R2 statistic that represents
the proportion of variance explained by the model.
Averaged coeffi cients with their adjusted standard errors,
p-values and relative importance values were estimated
from the best supported competing models containing each
predictor variable (Burnham and Anderson 2002). Multi-
model inference analyses were carried out using the MuMIn
package (Barton 2010) of R program (version 2.13.1; R
Development Core Team 2011).
We examined third-order habitat selection (sensu Johnson
1980) of the tracked tit-tyrant by using a Resource
Utilization Function (RUF model implemented from the ruf
package in R 2.13.0; Marzluff et al. 2004). For this analysis
the 99 % kernel home range was rasterized at a cell size
of 8 m corresponding to mean distance error of telemetry
fi xes. Thus, for each cell we calculated the utilization values
rescaled from 0 to 99 (the heights of use distribution values
which represent the strength of use of each spatial location)
and micro-habitat covariates . A Matern covariance function
was included in order to account for spatial autocorrelation
between neighboring data points. We developed three model
subsets including the independent and combined effects of
composite-based and distance-based variables on the habitat
use of the tit-tyrant. An information-theoretical approach
based on the AIC (Akaike’s Information Criterion) was used
to evaluate the support for competing models (Burnham and
Anderson 2002).
5
Micro-habitat selection of Juan Fernandez Tit-tyrant: GERARDO SOTO ET AL.
RESULTS
From a total of 240 fi xes, only 32 were used to calculate the
Kernel home range of the male tit-tyrant (Fig. 1; Fig. 3). The
rest of fi xes were discharged due to the following reasons: 1)
we removed the 5% harmonic mean outliers; 2) we removed
fi xes with large error distances; 3) the bootstrapping curve of
the Schoener’s index reached an asymptotic value of ca. 1.4
at a 4 min interval, hence fi xes were sequentially removed
each 4 min in order to get temporal independence.
The 99 % home range size of the tit-tyrant was 0.56 ha,
being located entirely within native forest in the vicinity
of gaps. Native forest comprised the 81.3 % of the home
range area, with the rest of the area corresponding to forest
gaps (Fig. 3). The kernel home range was successfully
validated from independent observed locations (Fig. 3). The
utilization values at the locations where the transmitter was
recovered and where the individual was mist-netted were
36.2 and 41.3 %, respectively, whereas random locations
had a mean utilization value of 21.6 ± 3.4%. In addition,
all visual observations of the individual (n = 12 points) fell
within the home range.
Micro-habitat selection at the population level was
explained mostly by distance-based variables. Only two
candidate abundance models were supported by data
(DAIC<2). The best supported model had as predictors
the distance to Gunnera shrub and distance to exotic shrub
(Table 1). Akaike weights indicated that such a model was
about 2.3 times more likely than the second best model,
which only included the distance to exotic shrub (Table
1). The two best models explained, respectively, 54 and 27
percent of the variance in abundance (Table 1). Abundance
decreased with distance to Gunnera shrub and increased
with distance to exotic shrub, with both coeffi cients
being signifi cant (Table 2). According to their importance
values, distance to exotic shrub had a larger probability of
being included in the best model than distance to Gunnera
shrub (Table 2).
FIGURE 3. Kernel home range for the radio-tracked Juan Fernandez Tit-tyrant showing isopleths in increments of 5 %. The outer thick line
represents the 99 % kernel home range boundary. The point represents the location where transmitter was recovered whereas the cross is
the mist-net where the tit-tyrant was recaptured two times.
FIGURA 3. Ámbito de hogar del Cachudito de Juan Fernández muestreado usando telemetría mostrando isopletas en incrementos de 5 %. La
línea gruesa exterior representa la frontera del ámbito de hogar. El punto representa el lugar donde el transmisor fue recuperado mientras
que la cruz es la red en donde el Cachudito fue recapturado dos veces..
Gayana 77(1), 2013
6
Distance-based and mixed models were the most
parsimonious RUF models (i.e., DAIC ≤ 4) indicating that
micro-habitat use is dependent on both distance-based and
composite-based variables (Table 3). In addition, the Akaike
weights (W) of these models indicated that the best distance-
based model was only 1.6-fold more likely than the mixed
model (Table 3). The composite-based forest gap variable
affected negatively to resource utilization probabilities
(Table 4). Utilization values also decreased signifi cantly
with increasing the distance from the forest gaps and
Gunnera shrub, with the former standardized coeffi cient
showing a stronger effect (Table 4).
TABLE 1. A priori candidate models explaining the micro-habitat selection pattern within the native forest for Juan Fernandez Tit-tyrant
during the post-breeding season in Robinson Crusoe Island. Only models belonging to the 95% confi dence set of candidate models are
shown, with k representing the number of parameters in the model.
TABLA 1. Modelos candidatos que explican el patrón de selección del micro-hábitat dentro del bosque nativo para el Cachudito de Juan
Fernández durante la temporada post reproductiva en la isla Robinson Crusoe. Se muestran solamente el conjunto de los mejores modelos
que acumulan un 95% del peso en AIC sobre el total de modelos, con k representando el número de variables del modelo.
MODEL k AIC DAIC Weight R2
Distance to Gunnera shrub + Distance to exotic shrub 3 29.5 0.00 0.42 0.54
Distance to exotic shrub 2 31.2 1.66 0.18 0.27
Gaps 2 32.2 2.68 0.11 0.18
Gaps + Distance to Gunnera shrub 3 32.7 3.24 0.08 0.37
Distance to Gunnera shrub 2 33.0 3.54 0.07 0.00
Distance to Gaps 2 33.0 3.55 0.07 0.00
Distance to Gaps + Distance to Gunnera shrub + Distance to exotic shrub 4 35.1 5.61 0.03 0.56
Distance to Gaps + Distance to exotic shrub 3 35.4 5.93 0.02 0.17
Gaps + Distance to Gunnera shrub + Distance to exotic shrub 4 35.5 6.00 0.02 0.54
TABLE 2. Model-averaged coeffi cients, standard errors (SE), p values and Importance values from models explaining the micro-habitat
selection pattern within the native forest for Juan Fernandez Tit-tyrant during the post-breeding season in Robinson Crusoe Island (see
TABLE 1).
TABLA 2. Coefi cientes promediados de los modelos, error estándar (SE), valores de probabilidad (p) e Importancia de los modelos que
explican el patrón de selección del micro-hábitat dentro del bosque nativo para el Cachudito de Juan Fernández durante la temporada post
reproductiva en la isla Robinson Crusoe (ver TABLA 1).
VARIABLE COEFFICIENT SE p IMPORTANCE
Distance to exotic shrub 0.04 0.02 0.049 1.00
Distance to Gunnera shrub -0.01 0.00 0.030 0.70
TABLE 3. Selected Resource Utilization Functions (RUFs) for a Juan Fernandez Tit-tyrant during the post-breeding season in Robinson
Crusoe Island.
TABLA 3. Funciones de Utilización de Recursos (RUFs) elegidas para un Cachudito de Juan Fernández durante la temporada post
reproductiva en la isla Robinson Crusoe.
Model Covariates Log-lik p-value AIC ΔAIC* W
Distance-based Distance to Gunnera shrub + Distance to Gaps -395.7 < 0.001 801.5 0.00 0.58
Mixed Distance to Gunnera shrub + Distance to Gaps + Gaps -395.2 < 0.001 802.4 0.91 0.37
Composite-based Gaps -399.2 < 0.001 806.5 4.95 0.05
*Models with ΔAIC < 4 were considered as parsimonious models.
7
Micro-habitat selection of Juan Fernandez Tit-tyrant: GERARDO SOTO ET AL.
TABLE 4. Standardized RUF coeffi cients with their standard errors (SE) and p-values of the best Mixed RUF model for a Juan Fernandez
Tit-tyrant (Table 3).
TABLA 4. Coefi cientes estandarizados del RUF con su error estándar (SE) y valores de p del mejor modelo RUF mixto para un Cachudito
de Juan Fernández (Tabla 3).
Coeffi cient Value SE p-value
Distance to forest gaps -8.9 0.15 <0.001
Distance to Gunnera shrub -7.8 0.21 <0.001
Forest gaps -1.0 0.05 <0.001
Intercept 11.2 0.19 <0.001
DISCUSSION
The studied male Juan Fernandez Tit-tyrant established
its home range exclusively in core native forest areas with
presence of forest gaps. The preference of native forest by
tit-tyrants has been demonstrated in previous studies (e.g.,
Hahn 2006; Hahn et al. 2005, 2011b). Such studies suggest
that tit-tyrants maintain their preference for native forest in
spite of their population decline trend (Hahn et al. 2006,
2011a). If this population decline is associated to habitat
loss, then the preservation of the remnant native forest area
has got priority in landscape planning.
The studied tit-tyrant showed a home-range size similar to
that of birds with the same body size (Calder 1984). Bird
density estimated from mist netting data was about 1.3 ind/
ha (i.e., 13 individuals captured in a 10 ha plot). The above
density estimates are consistent with data collected between
1995 and 2009, which show a mean density of about 1.6 ind/
ha in native forest (Hahn 2006, 2011b). Although the home
range estimate was based on data taken during a relatively
short time period, our mist net sampling data and visual
observations demonstrated that the male tit-tyrant held
the same territory during the post-breeding season. These
fi ndings, together with the fact that during the breeding
season the tit-tyrants are concentrated in native forest areas,
corroborate that tit-tyrants tend to prefer native forest as a
primary habitat for most of the year.
Tit-tyrants showed strong micro-habitat preferences both,
at the population level (i.e., mist-netting abundance)
and within their home ranges (i.e., utilization values).
Although the tracked male tit-tyrant avoided using forest
gaps, the abundance of tit-tyrants and the utilization
probabilities declined with increasing distance to these
human created gaps.
This implies that tit-tyrants not only could have perceived
the gaps as an unsuitable micro-habitat, but also they could
respond to a habitat quality gradient from the forest edge
to the interior. Since this species should have evolved in an
environment dominated by native forests in a geographically
isolated area, its foraging behavior should have a high
degree of specialization in the use of forest resources.
Natural heterogeneity in micro habitat conditions also
affected the habitat selection pattern of the tit-tyrant. Our
results evidenced a positive effect of the proximity to
sites dominated by Gunnera peltata. It is probable that the
humid conditions that characterize such moist sites, located
along seasonal or permanent small watersheds, spread into
the adjacent upland native forest. Thus, local gradients
in humidity conditions, may be responsible for the larger
use probability by tit-tyrants of the sites located close to
Gunnera shrub formations (e.g., Hahn et al. 2010).
The micro-habitat use pattern showed by tit-tyrants could
contribute not only to understand the particular ecology
of this species but also to design effi cient guidelines for
species conservation at different spatial levels. At the habitat
stand level the preservation of native remnants in Robinson
Crusoe should be a priority since tit-tyrants fundamentally
use this habitat type for nesting and foraging (Hahn et al.
2006, 2011a). At the micro-habitat level, however, some
specifi c considerations should be adopted. First, the artifi cial
gap formation as a result of the exotic shrub eradication
should not be considered as a detrimental disturbance for
tit-tyrants because of their positive effects on micro-habitat
use. Second, Gunnera formations should be conserved as
buffer areas acting as food sources for tit-tyrants. Third,
the connectivity between Gunnera shrub formations and
upland native forest should be increased in order to maintain
moist native ecotones which are rich in invertebrates.
Although our results are consistent in showing how habitat
preferences vary within home-ranges, more detailed work
will be necessary to assess the effects of micro-habitat scale
variability, such as topography and exposure.
Our fi ndings demonstrate the importance of micro-habitat
on the distribution of tit-tyrants in this protected area. Since
our conclusions data were drawn from only one radio-
tracked individual, they could not necessarily represent the
expected habitat selection pattern at the entire population.
However, results from regression models of abundance and
Gayana 77(1), 2013
8
RUF models of telemetry data were consistent in showing
the same habitat selection pattern of tit-tyrants (e.g., Table
2, Table 4). Moreover, the radio-tracked individual was a
territorial male who did not changed its home range during
the studied season. Regarding the above reasons, it is highly
probable that the habitat selection pattern displayed by the
studied individual is representative to the set of reproductive
individuals within the population. In order to have more
complete information about individuals of different age,
gender and social status, we suggest the need to expand our
analysis in order to obtain robust conclusions for species
conservation. In addition, monitoring program studies are
urgently necessary given the sharp decline in population size
that this species has suffered in the last decades (Hahn et al.
2011a). This information could be used for implementing
habitat management programs at the landscape level, not
only for tit-tyrants but also for other endangered endemic
species.
ACKNOWLEDGEMENTS
The Chilean CONAF granted us our island and national
park work permits. We are grateful to M. Galvez, J. Reyes,
J. Mesa, G. Gonzalez, C. Diaz, E. Gonzalez and all park
rangers, with special thanks to Silvia Moreno and Ivan
Leiva for fi eld assistance and hospitality. The study was
supported by the Humboldt Foundation (AvH).
REFERENCES
AEBISCHER, N.J., ROBERTSON, P.A. & KENWARD, R.E. 1993.
Compositional analysis of habitat use from animal radio-
tracking data. Ecology 74:1313-1325.
BARTON K. 2010. MuMIn: multi-model inference, R package
version 0.13.17.
BURNHAM, K.P. & ANDERSON, D.R. 2002. Model Selection and
Multimodel Inference: A Practical Information-Theoretic
Approach. Second edition. Springer, New York. 488 pp.
CALDER, W.A. 1984. Size, Function, and Life History. Harvard
University Press, Cambridge. 431 pp.
COFRÉ, H. 1999. Patrones de rareza de las aves del bosque
Templado de Chile: implicancias para su conservación.
Boletín Chileno de Ornitología 6:8-16.
CONNER, L.M., SMITH, M.D. & BURGER, L.W. 2003. A comparison
of distance-based and classifi cation-based analysis of
habitat use. Ecology 84:526-531.
CUEVAS, J.G., & VAN LEERSUM, G. 2001. Project “Conservation,
restoration, and development of the Juan Fernández
islands”, Chile. Revista Chilena de Historia Natural
74:899-910.
FORDHAM, D. & BROOK, B. 2010. Why tropical island endemics
are acutely susceptible to global change. Biodiversity and
Conservation 19:329-342.
HAHN, I. 2006. First reproductive records and nest sites of
the endemic Juan Fernández Tit-tyrant Anairetes
fernandezianus (Philippi 1857) (Aves: Tyrannidae) from
Robinson Crusoe Island, Chile. Zoologische Abhandlungen
55:177-190.
HAHN, I., RÖMER, U. & SCHLATTER, R. 2005. Distribution, habitat
use, and abundance patterns of land bird communities on
the Juan Fernández Islands, Chile. Ornitología Neotropical
16:371-385.
HAHN, I., RÖMER, U. & SCHLATTER, R. 2006. Population numbers
and status of land birds of the Juan Fernández Archipelago,
Chile. Senckenbergiana biologica 86:109-125.
HAHN, I., RÖMER, U., VERGARA, P. & WALTER H. 2009. Biogeography,
diversity, and conservation of the birds of the Juan
Fernández I slands, Chile. Vertebrate Zoology 59:103-114.
HAHN, I., VERGARA, P. & RÖMER, U. 2010. Vegetation structure
affects bird community composition in an island ecosystem:
the role of nativity and introduction. Geo-Öko 31:1-10.
HAHN, I., VERGARA, P. & RÖMER, U. 2011A. Habitat selection and
population trends in terrestrial bird species of Robinson
Crusoe Island: habitat generalists versus forest specialists.
Biodiversity and Conservation DOI: 10.1007/s10531-011-
0109-x
HAHN, I., VERGARA, P. & RÖMER, U. 2011B. Importance of nest
attributes in the conservation of endemic birds of the
Juan Fernández Archipelago, Chile. Bird Conservation
International DOI: 10.1017/S0959270911000050
HOOGE, P. & EICHENLAUB, P. 1997. Animal movement extension
to ArcView ver 1.1. Alaska Science Center-Biological
Science Offi ce, Anchorage, AK, USA. US Geological
Survey.
IUCN. 2011. The Red List of threatened species v. 2011.1, Gland.
http://www.iucnredlist.org. Accessed 30 August 2011.
JOHNSON, D.H. 1980. The comparison of usage and availability
measurements for evaluating resource preference. Ecology
61:65-71.
KENWARD, R.E. 2001. A manual for wildlife radio tagging.
Academic Press, London. 311 pp.
KERTSON, B.N. & MARZLUFF, J.M. 2010. Improving studies
of resource selection by understanding resource use.
Environmental Conservation 38:18-27.
KRISTAN, W.B.III. 2006. Sources and expectations for hierarchical
structure in bird-habitat associations. The Condor 108:5-
12.
LOEHLE, C. & ESCHENBACH, W. 2012. Historical bird and
terrestrial mammal extinction rates and causes. Diversity
and Distributions 18(1):84-91 DOI: 10.1111/j.1472-
4642.2011.00856.x
MARZLUFF, J., MILLSPAUGH, J., HURVITZ, P. & HANDCOCK, M. 2004.
Relating resources to a probabilistic measure of space use:
forest fragments and Steller’s jays. Ecology 85:1411-1427.
NAMS, V. 2000. Locate II user’s guide. Pacer Computer Software,
Tatamagouche. 76 pp.
OWENS, I.P.F. & BENNETT, P.M. 2000. Ecological basis of extinction
risk in birds: habitat loss versus human persecution
and introduced predators. Proceedings of the National
Academy of Sciences of the United States of America
97:12144-12148.
RICCI, M. 2006. Conservation status and ex-situ cultivation efforts
of endemic fl ora of the Juan Fernandez Archipelago.
Biodiversity and Conservation 15:3111-3130.
ROY, M.S., TORRES-MURA, J.C. & HERTEL, F. 1999. Molecular
9
Micro-habitat selection of Juan Fernandez Tit-tyrant: GERARDO SOTO ET AL.
phylogeny and evolutionary history of the tit-tyrants (Aves:
Tyrannidae). Molecular Phylogenetics and Evolution
11(1):67-76.
SCHOENER, T.W. 1981. An empirically based estimate of home
range. Theoretical Population Biology 20:281-325.
SKOTTSBERG, C. 1956. The natural history of Juan Fernandez and
Easter Islands. 3 volumes. Almquist Wiksells Boktryckeri,
Sweden. 1920 pp.
STUESSY, T.F., ONO, M. (EDS). 1998. Evolution and Speciation of
Island Plants. Cambridge University Press, Cambridge.
358 pp.
SWIHART, R.K. & SLADE, N.A. 1985. Testing for independence of
observations in animal movements. Ecology 66:1176-1184.
VARGAS, R., REIF, A. & FAUNDEZ, M.J. 2011. Los bosques de la isla
Robinson Crusoe, Chile: hotspot de endemismo en peligro.
Bosque 32(2):155-164.
VERGARA, P.M. & ARMESTO, J.J. 2009. Responses of Chilean forest
birds to anthropogenic habitat fragmentation across spatial
scales. Landscape Ecology 24:25-38.
WAND, M. 2006. KernSmooth: functions for kernel smoothing
for Wand and Jones (1995). R package version 2.22-19. R
port by Brian Ripley [www document]. URL http://cran.r-
project.org/web/packages/KernSmooth/KernSmooth.pdf
WAND, M.P. & JONES, M.C. 1995. Kernel Smoothing. Chapman
and Hall, London. 212 pp.
Recibido: 10.01.12
Aceptado: 25.03.13
