You are on page 1of 12



Author(s): Sharon Reid, Ivn A. Daz, Juan J. Armesto, and Mary F. Willson
Source: The Auk, 121(2):515-525. 2004.
Published By: The American Ornithologists' Union
DOI: 10.1642/0004-8038(2004)121[0515:IONBFU]2.0.CO;2

BioOne ( is an electronic aggregator of bioscience research content, and the online home to over
160 journals and books published by not-for-profit societies, associations, museums, institutions, and presses.
Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of
BioOnes Terms of Use, available at
Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries
or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research
libraries, and research funders in the common goal of maximizing access to critical research.
The Auk 121(2):515525, 2004


,1,2,5 I A. D,1,3 J  J. A
,1,2   M F. W 4
Center for Advanced Studies in Ecology and Biodiversity (CASEB), Ponticia Universidad
Catlica de Chile, Casilla 114-D, Santiago, Chile;
Laboratorio de Sistemtica y Ecologa Vegetal, Facultad de Ciencias, Departamento de Biologa,
Universidad de Chile, Casilla 653, Santiago, Chile;
Department of Wildlife Ecology and Conservation, University of Florida, 110 Newins-Ziegler Hall,
P.O. Box 110430, Gainesville, Florida 32611-0430, USA; and
5230 Terrace Place, Juneau, Alaska 99801, USA

A.In South American temperate rainforests, ve endemic understory birds (four

Rhinocryptidae and one Furnariidae) are oen associated with the main understory plant,
the native bamboo Chusquea valdiviensis (Poaceae: Bambusoideae). We studied the eects of
bamboo cover on species abundance and richness of those understory birds and explored the
functions of bamboo as food resource and escape cover. In Chilo Island (42S), southern Chile,
we selected four old-growth forest patches >100 ha and in each patch conducted bird surveys
in six plots with >70% understory cover. Three plots were dominated by native bamboo and
three plots had a sparse bamboo cover. Bird abundance (point counts) was signicantly corre-
lated with both total understory cover and percentage of bamboo cover but was not correlated
with other kinds of understory plant cover. Bird species richness was positively correlated with
bamboo cover and negatively correlated with other kinds of understory cover but unrelated
to total understory cover. Leaf-gleaners Magellanic Tapaculos (Scytalopus magellanicus), Ochre-
anked Tapaculos (Eugralla paradoxa), and Des Murss Wiretails (Sylviorthorhynchus desmursii),
and the ground-gleaner Chucao Tapaculos (Scelorchilus rubecula) were more abundant in
high-bamboo plots; but the ground-gleaner Black-throated Huet-huet (Pteroptochos tarnii) was
recorded more times in plots with low-bamboo cover.
Availability of invertebrates per unit of understory dry mass did not dier between high- and
low-bamboo plots; but plant biomass was greater in high-bamboo plots, so total invertebrate
abundance per plot was higher there. Ground-li er invertebrate abundance was similar in all
plots. To examine escape-cover preferences, nine captured Chucao Tapaculos were released in
front of two dierent understory scenarios (high-bamboo cover or bamboo-free understory);
88% of released birds moved into bamboo cover. We suggest that the structure of native bam-
boo understory is critical for the maintenance of four of those species, and retaining bamboo
cover in managed stands may help minimize the eect of logging on understory birds. Received
9 May 2003, accepted 19 January 2004.


.En bosques templados de Sudamrica, cinco especies de aves endmicas de
sotobosque (cuatro Rhinocryptidae y un Furnariidae) se encuentran frecuentemente aso-
ciadas a la principal planta de sotobosque, el bamb nativo Chusquea valdiviensis (Poaceae:
Bambusoideae). Estudiamos los efectos de la cobertura de Chusquea sobre la abundancia y
riqueza de especies de estas aves de sotobosque, y exploramos las funciones del bamb como
fuente de recursos (alimento) y cobertura de escape. En la isla de Chilo (42 S), sur de Chile,
seleccionamos cuatro parches de bosque primario mayores a 100 ha y en cada parche real-
izamos censos de aves en seis estaciones con ms de 70% de cobertura de sotobosque. Tres
estaciones estaban dominadas por Chusquea y tres tenan una escasa cobertura de bamb. La
abundancia de aves (en estaciones de escucha) estaba positivamente correlacionada con la
cobertura total de sotobosque y el porcentaje de cobertura de bamb, pero no presentaba cor-
relacin con la cobertura de otras especies de plantas en el sotobosque. La riqueza de especies
de aves estaba correlacionada positivamente con la cobertura de bamb y negativamente con
la cobertura de otras especies del sotobosque, pero no tuvo relacin con la cobertura total del
sotobosque. Las especies que se alimentan en el follaje, Scytalopus magellanicus, Eugralla para-


516 R

 . [Auk, Vol. 121

doxa y Sylviorthorhynchus desmursii, y una especie que forrajea en suelo, Scelorchilus rubecula,
fueron ms abundantes en las estaciones dominadas por bamb, pero otra especie de suelo,
Pteroptochos tarnii, fue ms frecuente en estaciones con escasa cobertura de bamb.
La disponibilidad de invertebrados por unidad de biomasa seca de sotobosque no diri
entre estaciones con dominancia o escasez de bamb, pero la biomasa vegetal fue mayor en las
primeras. Por lo tanto, la abundancia total de invertebrados en stas fue mayor. La abundancia
de invertebrados en la hojarasca fue similar en todas las estaciones. Para examinar posibles
preferencias de cobertura de escape, se liberaron nueve individuos de S. rubecula frente a dos
tipos de sotobosque (alta cobertura de bamb o sin bamb); 88% de las aves eligieron cober-
tura de bamb. Se sugiere que el sotobosque de bamb es crtico para la mantencin de cuatro
especies de aves, por lo que la retencin de cobertura de bamb en bosques manejados puede
reducir el impacto de la alteracin del hbitat sobre las aves del sotobosque.

   structure on avian popu- patches (fragmentation and isolation) and verti-
lations and mechanisms involved in producing cal stratication within patches are predicted
those eects have been considered a primary to aect avian species abundance and richness
research area for avian conservation biology (Wiens 1989).
(Morrison et al. 1992, Rozzi et al. 1996, Walters Previous studies in south-temperate forests
1998). Vegetation structure may provide escape have suggested that certain forest elements
cover against predators, safe nesting sites, and (such as big trees, logs, and understory cover)
food resources for birds (Feinsinger et al. 1988, are important for the presence of a number
Lima 1993). Understanding how understory of species of native birds (Willson et al. 1994,
vegetation structure inuences habitat selection 1996; Sieving et al. 2000; Reid et al. 2002; I. A.
by understory birds could help elucidate the Daz et al. unpubl. data). In particular, bamboo
mechanistic bases of community organization understory apparently is a critical habitat for
and structure (Moermond 1990). ve endemic species (four Rhinocryptidae and
South American temperate rainforests cover one Furnariidae) of understory birds on Chilo
a narrow area of the Andes between 35S Island, southern Chile. Bird surveys suggest
and 55S in southern Chile and westernmost that the presence and abundance of Chucao
Argentina and sustain over 44 bird species; 64% and Ochre-anked tapaculos and Des Murss
are found only in southern South America and Wiretails (scientic names in Table 2) is com-
30% of those are endemic to south-temperate monly associated with bamboo cover (Sieving
forests (Vuilleumier 1985, Rozzi et al. 1996). et al. 1996, 2000; Daz 1999; McPherson 1999; De
Native bamboo (Poaceae: Bambusoideae), typi- Santo et al. 2002). Those understory birds are
cally Chusquea valdiviensis (hereaer bamboo), nonmigratory invertebrate-eaters that have poor
dominate the understory, especially under light ying abilities, and their abundance and species
gaps (Armesto et al. 1996, Ma hei 1997). Where richness decreases in small forest fragments
forests have been burned or selectively logged, in Chilo Island (Willson et al. 1994). A high
bamboo typically dominates the understory in proportion of endemic understory birds in the
treefall gaps, forming extensive thickets (Veblen family Rhinocryptidae are considered to be
1982). In industrial exploitations bamboo is globally endangered (Vuilleumier 1985, Glade
oen removed from the understory together 1988, Collar et al. 1992), so understanding the
with other woody residues. habitat requirements of those birds is important
Only in the last decade have studies begun to for their conservation.
document the relationship between forest struc- Considering that rainforests on Chilo Island
ture and habitat use by avian species in south- and other areas in southern Chile continue
temperate forests (Willson et al. 1994, 2001; to be logged and structurally degraded (Lara
Sieving et al. 1996, 2000; Daz 1999; Estades et al. 1996, Willson and Armesto 1996), we
and Temple 1999; Morrison and Phillips 2000; designed a study to assess and quantify the
De Santo et al. 2002; I. A. Daz et al. unpubl. ecological role of native bamboo in determining
data). Both the landscape conguration of forest the distribution and abundance of the ve
April 2004] Importance of Bamboo for Understory Birds 517

understory bird species that inhabit temperate size of the forest patches was large enough to avoid
lowland rainforests in northern Chilo Island. possible eects associated with small fragments (see
Here we ask specically whether bamboo cover Willson et al. 1994) and also allowed us to nd areas
is associated with the presence and abundance with a dense bamboo understory as well as areas with
sparse bamboo cover, because of the vast heterogene-
of understory birds in southern forests, or
ity of microhabitats (Freemark and Merriam 1986). In
whether other types of understory vegetation each patch, we established six 50-m-radius sample
(e.g. tree saplings) will suce. In addition, we plots with >70% of the area of the plot covered with
examined the following hypotheses about the understory vegetation (estimated visually by two
ecological role of bamboos: (1) bamboo habitat observers). Bamboo patches within forest fragments
provides more food resources to understory varied in size and shape, and it was impossible to
birds than bamboo-free understory, and (2) establish pure bamboo plots (i.e. no bamboo patch
bamboo cover may be selected by birds as a was large enough to cover 100% of a 50-m-radius
potential escape cover, which could be safer in sample plot). The high frequency of bamboo patches
comparison to other types of understory cover. did not allow us to establish bamboo-free plots.
Therefore, we established three plots averaging 75%
bamboo cover (hereaer high-bamboo plots), and
 three matching plots with dominance of tree seedlings
and saplings and a sparse bamboo cover (on average
Study area.The study was conducted in a rural 14% bamboo cover; hereaer low-bamboo plots).
landscape in northeastern Chilo Island in southern Canopy tree density and composition was kept as
Chile (4150S, 7340W). The region has been subject similar as possible between plots, despite the fact that
to a continuous environmental degradation because high-bamboo plots were located near or under small
of a century of forest logging and land clearing for tree-fall gaps (Table 1).
agricultural use, which has caused intense fragmenta- Bird surveys.Bird surveys followed the point-
tion and reduction of forest area (Armesto et al. 1994, count methodology used in a similar study by Willson
1998; Lara et al. 1996; Willson and Armesto 1996). et al. (1994), described in Ralph et al. (1993), and were
Climate is wet-temperate with a strong oceanic carried out in each of the six 50-m-radius plots in each
inuence (Di Castri and Hajek 1976). Annual precipi- patch. All understory birds seen or heard within the
tation is 2,090 mm with a mean annual temperature plots were recorded for an 8-min period. Plots were
of 12C (Senda Darwin Biological Station, ve- separated from each other by at least 200 m to mini-
year record). Forests are dominated by evergreen, mize the risk of counting the same individual twice
broad-leaved trees, and some narrow-leaf conifers (Hu o et al. 1986). Bird surveys were performed from
(Saxegothaea conspicua, Podocarpus nubigena). Main 0730 to 1000 hours EST on no-rain days (Willson et al.
emergent trees are Nothofagus nitida, Drimys winteri, 1994). Each plot was surveyed ve times during the
and Eucryphia cordifolia. Understory is dominated breeding season, from December 2000 until late March
by bamboo under light gaps, and by shade-toler- 2001 (i.e. austral spring and summer). We concentrated
ant myrtaceous seedlings and saplings, such as our sampling eort in the breeding season because of
Amomyrtus luma, Amomyrtus meli, and Myrceugenia the higher vocalization activity and detectability of
parviora in shaded areas under the closed canopy. those birds. On Chilo Island, a marked decrease in
Study design.We selected four old-growth forest activity of resident birds in the nonbreeding season
patches >100 ha using 1:20,000 orthophotos from 1993, increases the chances of underestimating bird species
named hereaer as F1, F2, F3, and F4. The abundance (S. Reid and I. Daz pers. obs.).

1. Vegetation variables and food resources associated with high- and low-bamboo plots in two forest
patches (F1 and F3) in Chilo Island, southern Chile (mean SD).

Vegetation variables High-bamboo plots Low-bamboo plots P
Understory height 2.37 0.27 2.3 0.48 0.783
Total understory cover (%) 83.83 12.3 69.8 21.7 0.207
Bamboo cover (%) 74.5 14.6 14.3 11.9 <0.001
Cover of other understory plants (%) 10.7 5.9 55.8 21.1 <0.001
Canopy cover (%) 37.5 27.4 68.8 15.3 0.034
Number of logs intersecting transects 8.2 3.4 7.7 3.4 0.806
Foliage invertebrates (individuals per plot) 64.4 20.6 105 17.4 3.4 105 <0.001
518 R

 . [Auk, Vol. 121

That methodology did not allow us to estimate model C305 to estimate number of invertebrates per
absolute values of abundance and richness of birds; unit of understory biomass. To estimate invertebrate
so for patch comparisons, we considered abundance abundance per plot, we rst calculated the biomass of
as the mean number of birds per plot per day and 1 m3 of bamboo (n = 5) and bamboo-free understory
richness as the mean number of species per plot per (n = 5) by removing that volume (dened by a 1 m3
day. We also analyzed how the maximum abundance PVC cube) from the understory with pruning scissors
and maximum richness of species diers between and weighing the oven-dried samples. To estimate
high- and low-bamboo plots and how they relate to total understory biomass in the 50-m-radius plots, we
understory and bamboo cover. Maximum abundance rst calculated the total understory volume (bamboo
was calculated by summing the maximum number of and bamboo-free understory separately) with the
individuals recorded per species in the ve surveys following equation r2 h (C/100), where =
and maximum richness was calculated by summing 3.14, r is the plot radius (i.e. 50 m), h is the average
all the species recorded in the ve surveys in each understory height, and C is bamboo and bamboo-free
plot. We chose that method because we were inter- cover, respectively, for each case. Biomass of bamboo
ested in the maximum number of birds and bird spe- and bamboo-free understory per plot was estimated
cies that potentially use the understory vegetation in by multiplying the result of that equation with the
the plots, and that allowed comparisons among plot estimated biomass of 1 m3 of bamboo and bamboo-
types on the same basis. free understory, respectively (see above). The sum of
Understory characterization.We characterized the biomass of bamboo and bamboo-free species gave us
amount and composition of understory cover in two an estimate of the total understory biomass per plot.
of the four forest patches, F1 and F3 (i.e. 12 plots in That result was then multiplied by the mean number
total). In each plot, we established two 50-m-long of invertebrates in 1 g of biomass of bamboo and
transects perpendicular to each other, centered in the bamboo-free species. Abundance of invertebrates
plot and oriented along cardinal points. We then set associated with understory foliage of high- and low-
up ve regularly spaced points along those transects bamboo plots was analyzed by a one-way ANOVA.
(one in the center and the other four in each of the For invertebrates in ground li er, we searched for
four cardinal points). At each point, two independent invertebrates in a rectangle of 7 20 cm (area = 210 cm2)
observers visually estimated understory height, bam- during an 8-min period and recorded all counted
boo cover, other understory cover (i.e. percentage of invertebrates (see Willson and Comet 1996 for a
cover of species other than bamboo) and total under- similar method). That was repeated randomly eight
story cover within a 5-m-radius area. All understory times per plot in the same two forest patches where
and canopy species were recorded and canopy cover understory foliage was sampled. Invertebrate abun-
above the plots was estimated using four intervals, dance was expressed as number of invertebrates per
025%, 2550%, 5075%, and 75100%. The presence area of ground li er, and that was extrapolated to the
of fallen logs >15 cm diameter was recorded when area of the plot. Comparisons between high- and low-
logs intersected the transects; this diameter for logs bamboo plots were done by a one-way ANOVA.
was chosen because logs with smaller diameters are Bamboo as escape cover.To test bamboo as a poten-
seldom used as a nesting site by understory birds (S. tial escape cover (hypothesis 2), we conducted experi-
Reid et al. pers. obs.). ments with nine Chucao Tapaculos captured using a
Food availability.All ve understory bird species 50 40 cm wooden box-trap. We placed a petri dish
studied are invertebrate-eaters, although Chucao with bait beneath the box to a ract the birds a en-
Tapaculos and Black-throated Huet-huets also con- tion. Common li er invertebrates (such as crickets,
sume fruits (Armesto et al. 1987). To test whether larvae, and grubs) were used for bait. Individuals
bamboo habitat provides more food resources than were kept in captivity in the dark inside a 20-cm
low-bamboo understory (hypothesis 1), we performed wide 20-cm tall 100-cm long tube, with both ends
time-constrained searches for invertebrates on under- closed for 510 min, for acclimatization. Then, one of
story foliage and ground li er in plots with high- and the ends was opened, and released birds were faced
low-bamboo cover in the same 12 plots where veg- with two dierent and equally distanced understory
etation variables were measured. Invertebrate abun- scenarios, one with >80% bamboo cover (visually
dance was used as a substitute for food availability estimated) and the other with >80% of tree seedlings
(Griths 1975). and saplings. The direction and rst perching place
For invertebrate abundance on foliage, we ran- aer release was recorded. The small number of birds
domly chose rst eight 1-m-long branches and counted tested reects the low capture rates.
all invertebrates on the branch during a 3-min period. Data analyses.We tested for normality of the data
Then, we removed the target branch at its base and using one-sample Kolmogorov-Smirnov tests, with
later dried it in a Shellab multi-purpose oven model Lilliefors distribution (i.e. standard normal distribu-
1350 FX (Sheldon Manufacturing, Cornelius, Oregon), tion). Survey data were not normally distributed,
at 50C for 48 h and weighed it in a Ohaus balance so comparisons of mean bird abundance among
April 2004] Importance of Bamboo for Understory Birds 519

forest patches were done using the nonparametric

Kruskal-Wallis test. Our data did not show signicant


dierences between forest patches, so we grouped all

high-bamboo plots and all low-bamboo plots sepa-
rately and compared mean abundance of each species,
maximum total abundance, and maximum species

2. Mean number of individuals per point per day of understory bird species in four forest patches in Chilo Island (range in parentheses).
richness between those two groups. All comparisons

KW a



were done by the nonparametric Mann-Whitney
U-test. To relate maximum species abundance and
maximum richness with bamboo cover, we performed
regression analyses in the 12 plots where vegetation

1.83 (0.82.2)
variables were characterized. To test for bamboo as
a major food source, we compared the amount of

0.23 (02)

1.47 (04)

0.73 (02)
invertebrates between high- and low-bamboo plots by

a one-way ANOVA. To test for bamboo as a preferred


cover selection site, we used a sign test on the number
of captured birds selecting bamboo. All statistical
tests as described in Zar (1984) were performed using

1.67 (0.42.8)

0.27 (01)
0.23 (02)
0.97 (02)
0.07 (01)
0.47 (02)

Understory characterization.All four for-
est patches presented abundant tree-fall gaps

1.97 (12.8)
and were similar in structure and tree species
0.37 (01)
0.23 (02)
1.63 (03)
0.03 (01)
0.57 (02)
composition. Understory composition in all


patches was characterized by a mosaic of areas

dominated by bamboo and areas dominated
by myrtaceous tree saplings. Two patches, F1
1.97 (1.23.2)

and F4, presented a more dispersed canopy as

a result of old logging practices and re; in the
0.33 (02)
0.17 (02)
1.47 (03)
0.10 (02)
0.67 (02)

la er, we did not record two of the ve bird spe-



cies studied, the Black-throated Huet-huet and

Ochre-anked Tapaculo.
High- and low-bamboo plots presented simi-
lar total understory cover and number of logs
Sylviorthorhynchus desmursii

(Table 1). As expected, high-bamboo plots were

KW = Kruskal-Wallis statistic for dierences among forest fragments.
Scytalopus magellanicus

characterized by a signicantly higher bamboo

Scelorchilus rubecula

cover (t = 7.80, df = 10, P < 0.001), whereas low-

Pteroptochos tarnii

Eugralla paradoxa

bamboo plots presented a higher cover of other

Scientic name

understory vegetation (i.e. myrtle saplings) and

a denser canopy cover (t = 5.04, df = 10, P < 0.001;
t = 2.45, df = 10, P = 0.034, respectively).
Bird abundance and species richness.Mean
abundance of each species was similar among
forest patches (Table 2). Chucao Tapaculos were
Black-throated Huet-huet

the most abundant bird species in the under-

Ochre-anked Tapaculo

story of all four forest patches, followed by

Total species richness
Magellanic Tapaculo
Des Murss Wiretail

Magellanic Tapaculos and Des Murss Wiretails.

Chucao Tapaculo

Total abundance

Ochre-anked Tapaculos were rare in all patches

Common name

(Table 2). Des Murss Wiretails and Chucao and

Magellanic tapaculos were signicantly more
abundant in high-bamboo plots (U = 17, P =
0.001; U = 25.5, P = 0.007; and U = 31.5, P = 0.016,

respectively). Ochre-anked Tapaculos were

520 R

 . [Auk, Vol. 121

never recorded in low-bamboo plots, although

recorded individuals were too few to observe a
signicant dierence between plots (U = 54, P =
0.071), and Black-throated Huet-huets showed a
similar abundance in both plot types (U = 77, P =
0.75; Fig. 1).
Maximum abundance of birds was sig-
nicantly greater in high-bamboo plots than in
low-bamboo plots (U = 36.5, P = 0.036; Fig. 2A).
Species richness was also signicantly greater in
high-bamboo plots in three of the four patches
(U = 16.0, P = 0.024; Fig. 2B). The most degraded
forest (F4) was treated separately for species-
richness comparisons because two species were
absent despite the presence of dense bamboo
F. 1. Total number of birds counted in high
In the 12 plots where vegetation variables
and low bamboo plots during the breeding season.
were measured, understory bird abundance
Abbreviations: CT = Chucao Tapaculo (n = 166), MT
= Magellanic Tapaculo (n = 73), DM = Des Murss was signicantly correlated with total under-
Wiretail (n = 36), BH = Black-throated Huet-huet (n = story cover (Pearsons r2 = 0.34, P = 0.045; Fig.
19), and OT = Ochre-flanked Tapaculo (n = 6). 3A) and bamboo cover (r2 = 0.43, P = 0.021; Fig.
3B), but was not correlated with other kinds of
plant cover (r2 = 0.18, P = 0.17). Bird-
species richness was positively correlated with

F. 2. Relationship between (A) maximum abun-

dance in 12 low- and high-bamboo plots for four
forest patches and (B) maximum species richness (N F. 3. Relationship between maximum abundance
spp.) in 9 low- and high-bamboo plots (three forest and (A) percentage of total understory cover and (B)
patches) and in a degraded forest patch (shaded bar). percentage of bamboo cover.
April 2004] Importance of Bamboo for Understory Birds 521

Number of invertebrates in ground li er

showed signicant dierences between forest
patches (ANOVA: F = 27.70, df = 1 and 10, P =
0.001); therefore, comparisons between high-
and low-bamboo plots were done separately
for each patch. No signicant dierences were
observed between high- and low-bamboo plots
in either of the two forests (ANOVA: F = 0.024,
df = 1 and 4, P = 0.88; F = 1.923, df 1 and 4, P =
0.24 for F1 and F3, respectively). The most com-
mon invertebrates in ground li er were ants,
moths, and spiders.
Cover selection sites.In experiments with nine
dierent Chucao Tapaculos released from live
traps, seven chose bamboo understory as the rst
perching place aer release, one chose bamboo-
free understory vegetation, and one ew straight
ahead so fast that we were unable to record its
rst perching site. Although a marked prefer-
ence for bamboo understory was observed, that
preference did not yield a statistically signicant
dierence with a sign test (P = 0.07).


Importance of bamboo for understory birds.

Four of the ve understory bird species were
F. 4. Relationship between maximum richness always more abundant in high-bamboo plots,
and (A) percentage of bamboo cover and (B) percent- supporting, in part, our initial hypothesis of
age of bamboo-free understory.
a strong anity between all those birds and
bamboo understory. Many related understory
bamboo cover (r2 = 0.40, P = 0.028; Fig. 4A) and species of tropical rainforest birds in the fami-
negatively correlated with other plant cover lies Rhinocryptidae and Formicariidae are also
(r2 = 0.34, P = 0.046; Fig. 4B) but unrelated to total strongly associated with understory bamboo
cover in the understory (r2 = 0.10, P = 0.33). thickets (Ridgely and Tudor 1994). Our ndings
Bamboo as food resource.Number of inver- are in accord with recent studies of McPherson
tebrates per gram of understory showed no (1999), who suggests that Ochre-anked
signicant dierences between forest patches Tapaculos depend on bamboo cover, and with
(ANOVA: F = 1.92, df = 1 and 10, P = 0.20), so we Sieving et al. (2000), who suggest that bamboo is
pooled all plots with high bamboo and all plots an important predictor of bird use of forest cor-
with low bamboo from the dierent patches for ridors on Chilo Island. Data of all bird species
the comparisons. No signicant dierences in recorded in the bird surveys here (n = 24) also
the number of invertebrates per gram of under- showed a signicant higher bird-species abun-
story were found between bamboo and other dance in high-bamboo plots (ANOVA: F = 28.54,
understory plants (ANOVA: F = 2.56, df = 1 and df = 1 and 118, P < 0.01), but the importance of
10, P = 0.15). However, bamboo biomass per forest elements for those species is discussed
cubic meter was marginally higher than other in I. A. Daz et al. (unpubl. data). In addition,
understory biomass (1.4 0.8 kg dry mass and many generalist species in those forests use
0.5 0.3 kg, respectively; t = 2.27, df = 8, P = understory for nesting, which reinforces the
0.053, n = 5). As a result, estimated number of importance for the maintenance of a diverse
invertebrates per plot was signicantly higher forest avifauna; examples are the vulnerable
in high-bamboo plots (t = 5.501, df = 10, P < Chilean Pigeon (Columba araucana), the Green-
0.001; see Table 1). backed Firecrown (Sephanoides galeritus), the
522 R

 . [Auk, Vol. 121

endemic Patagonian Tyrant (Colorhamphus par- obs.). Reported predators for understory spe-
virostris), and the White-crested Elaenia (Elaenia cies in Chilo Island are Rufous-legged Owl
albiceps) (Johnson and Goodall 1967, Chesser (Strix rupes), Austral Pygmy-Owl (Glaucidium
and Marn 1994; S. Reid et al. pers. obs.). We nanum), Bicolored Hawk (Accipiter bicolor), the
propose that stands in a forest corresponding to native forest cat (Oncifelis guia), the Darwins
the 50-m-radius sample plots in our study, hav- fox (Pseudalopex fulvipes), the mustelid quique
ing >75% of the area covered by bamboo, will (Galictis cuja), and possibly an introduced
help to ensure a signicantly high understory mustelid (Mustela vison) if its range continues
bird abundance and species richness in simi- to increase (Martnez and Jaksic 1996, Medina
lar forest patches in Chilo Island. Although 1997, Sieving et al. 2000, Willson et al. 2001). All
landscape conguration (fragmentation and of those, except for the Austral Pygmy-Owl (20
isolation) must also be taken into account when cm; Johnson and Goodall 1967), are larger than
predicting a species presence or abundance in the understory bird species and the spaces in
a patch. We hypothesize that the absence of the branching pa erns of bamboo, so escaping
Ochre-anked Tapaculos and Black-throated into bamboo thickets could prevent the a ack
Huet-huets in F4 may be explained in part by of those predators. Our preliminary result sug-
the higher isolation of F4 to continuous forest, gests a bamboo selection by Chucao Tapaculos,
in addition to the higher disturbance because of and bamboo is likely to serve as a more eec-
past logging practices and re. tive refuge from predators in comparison to
Ecological role of native bamboo.An innovative other types of understory cover. Further studies
nding here was that bamboo habitat oers a should be conducted in the eld to examine that
greater potential food supply to invertebrate hypothesis.
eaters because of the higher density of bamboo Management and conservation implications.
foliage. The higher invertebrate abundance in Habitat heterogeneity of forests, together with
high-bamboo plots supports our rst hypoth- the size and connectivity of forest patches, is of
esis, contributing to the observed preferences paramount importance for the maintenance of a
of understory bird species. Particularly, bamboo diverse forest avifauna (Freemark and Merriam
could be an important food supply for three 1986, Wiens 1989, Wenny et al. 1993, Faaborg
leaf gleaners (Ochre-anked and Magellanic et al. 1998, I. A. Daz et al. unpubl. data). In
tapaculos and Des Murss Wiretails) that for- Chile, most industrial logging practices tend to
age preferentially in branches of the under- simplify forest structure, and the understory is
story. However, invertebrate availability on the usually burned or removed (Estades 1994, 1997;
ground did not dier between plots with high Donoso and Lara 1999). A dense understory, par-
or low bamboo and may not aect birds that ticularly bamboo thickets, has traditionally been
forage on the ground (Chucao Tapaculo and considered a factor that signicantly inhibits the
Blacked-throated Huet-huet). The observed establishment and growth of tree seedlings, and
similar abundance of Black-throated Huet-huet hence forest regeneration (Veblen 1982, Veblen
in both plot types means that there are other et al. 1996); but a recent study in a Nothofagus
variables (not considered here) that explain forest in south-central Chile demonstrated that
habitat use in that species, such as patch size the removal of understory has no signicant
because of their large home ranges (Willson et advantage to overstory wood production (Lusk
al. 1994). and Ortega 2003). Therefore, it could be possible
Concerning bamboo as a preferred escape to maintain some level of understory cover in
cover (hypothesis 2), our preliminary explora- managed forests, enhancing understory bird
tion may support the preference of understory conservation. Industrial management practices
birds for bamboo cover (low number of birds contrast with traditional small-scale, selective
tested could explain the lack of statistical logging by landowners in Chilo Island. In the
signicance). Bamboo cover would function former, extensive homogeneous stands with a
as a potential refuge from predators because sparse understory are generated; whereas in the
bamboo thickets are oen very dense, with second, understory heterogeneity is maintained
intricate branching that impedes the rapid and the development of bamboo cover may
movements of any organism taller than 30 even be enhanced by removal of some canopy
cm through the vegetation (S. Reid et al. pers. trees. According to our results, those two forms
April 2004] Importance of Bamboo for Understory Birds 523

of management will have contrasting conse- Biodiversity, and by an Endowed Presidential Chair
quences for the persistence of understory birds in Science (to J.J.A.). This is a contribution from the
in the regional landscape. research program of Senda Darwin Biological
Southern temperate understory birds present Station, Chilo Island, Chile.
several characteristics that make them very sen-

sitive to land-cover changes: (1) they have poor
ying abilities, (2) they are ground- and thicket- A
, J. J., R. R, P. M ,   C. S.
dwellers, (3) they are particularly sensitive to 1987. Plant/frugivore interactions in South
changes in habitat structure, and (4) they tend American temperate forests. Revista Chilena
to decrease in abundance and species richness de Historia Natural 60:321336.
from large to small patches (Willson et al. 1994; A
, J. J., R. R, C. S -R 
Sieving et al. 1996, 2000; Cofr 1999; Estades and M. T. K. A. 1998. Conservation targets
Temple 1999; Cornelius et al. 2000). Endangered in South American temperate forests. Science
forest pheasants (Phasianidae), which inhabit
, J. J., C. V ,   M. K. A.
bamboo understory in southwestern China 1996. Ecologa de los Bosques Nativos de
and share similar characteristics with the rhi- Chile. Editorial Universitaria, Universidad de
nocryptids in our study, are being successfully Chile, Santiago.
conserved in areas where bamboo patches and A
, J. J., C. V ,   C. D . 1994.
corridors are being set up, in the provinces of Desde la era glacial a la industrial. La historia
Shaanxi, Gansu, and Sichuan (Reid and Jien del bosque templado Chileno. Ambiente y
1999). For southern temperate forest birds, Desarrollo 10:6672.
bamboo density, corridor width, and presence C

, R. T.,   M. M . 1994. Seasonal dis-
of streams are the main predictors of bird use tribution and natural history of the Patagonian
Tyrant (Colorhamphus parvirostris). Wilson
of corridors in Chilo Island (Sieving et al.
Bulletin 106:649667.
2000). Our results suggest that bamboo not only C, H. 1999. Patrones de rareza de las aves
functions as protective cover during travel but del bosque Templado de Chile: Implicancias
also plays an important role in determining the para su conservacin. Boletn Chileno de
distribution and abundance of four of the ve Ornitologa 6:816.
understory birds studied; it provides abundant C, N. J., L. P. G , N. K
, A.
food resources to invertebrate gleaners and M N
, L. G. N !, T. A. P
may serve as a be er refuge from predators. We   D. C. W

. 1992. Threatened Birds of the
propose that the maintenance of bamboo patch Americas: The ICBP Red Data Book, part 2, 3rd
heterogeneity and the re-establishing of bam- ed. International Council for Bird Preservation,
Cambridge, United Kingdom.
boo corridors in managed forests in Chilo
 , C., H. C,   P. A. M"
. 2000.
Island will help minimize the eect of logging Eects of habitat fragmentation on bird species
on understory birds. in a relict temperate forest in semiarid Chile.
Conservation Biology 14:534543.

S , T. L., M. F. W , K. E. S
 ,   J.
J. A
. 2002. Nesting biology of tapaculos
We thank C. Jones for his support and superb (Rhinocryptidae) in fragmented south-temper-
enthusiasm in interpreting data and suggesting ideas; ate rainforests of Chile. Condor 104:482495.
F. Oyarzn and T. Darnell for eld assistance; D. D, I. 1999. Exito reproductivo de
Pavlacky and M. Carmona for constant encouragement Sylviorthorhynchus desmursii en bosques
and assistance in experimental designs; P. Chacn and fragmentados de la Isla Grande de Chilo. M.S.
F. Daz for assistance with data analysis; M. Baeza, C. thesis, University of Chile, Santiago.
Marticorena, and M. Muoz for botanical collaboration; D C, F.,   E. H!
. 1976. Bioclimatologa
and K. E. Sieving, T. Contreras, J. Davis, M. Reetz, and de Chile. Ediciones Universidad Catlica de
M. Milleson for valuable comments on the manuscript. Chile, Santiago.
We thank the landowners that gave us permission to D , C.,   A. L. 1999. Silvicultura de
work on their properties. Funding for this research Los Bosques Nativos de Chile. Editorial
was provided by the Millennium Center for Advanced Universitaria, Santiago, Chile.
Studies in Ecology and Research on Biodiversity P99- E
, C. F. 1994. Impacto de la sustitucin
103-FICM, by grant FONDAP-FONDECYT 1501-0001 del bosque nativo por plantaciones de Pinus
to the Center for Advanced Studies in Ecology and radiata sobre una comunidad de aves en la
524 R

 . [Auk, Vol. 121

Octava Regin de Chile. Boletn Chileno de M

, O. 1997. Las especies del gnero Chusquea
Ornitologa 1:814. Kunth (Poaceae: Bambusoideae), que crecen
, C. F. 1997. Habitat fragmentation, pine en la X regin, Chile. Gayana Botnica 54:
plantation forestry and the conservation of for- 199220.
est bird communities in central Chile. Ph.D. dis- MP
 , H. 1999. Landscape eects on the
sertation, University of Wisconsin, Madison. distribution of an endemic rhinocryptid, the
, C. F.,   S. A. T
. 1999. Deciduous- Ochre-anked Tapaculo, Eugralla paradoxa, in
forest bird communities in a fragmented land- fragmented south-temperate rainforests. M.S.
scape dominated by exotic pine plantations. thesis, University of Florida, Gainesville.
Ecological Applications 9:573585. M
 , G. 1997. A comparison of the diet and
F, J., F. R. T # III, S. K. R  , T. M. distribution of southern river o er (Lutra
D  , D. R. W

,   J. D. B . provocax) and mink (Mustela vison) in south-
1998. Understanding fragmented midwestern ern Chile. Journal of Zoology (London) 242:
landscapes: The future. Pages 193207 in Avian 291297.
Conservation Research and Management (J. M. M
  , T. C. 1990. A functional approach
Marzlu and R. Sallabanks, Eds.). Island Press, of foraging: Morphology, behavior, and the
Washington, D.C. capacity to exploit. Studies in Avian Biology
, P., W. H. B , K. G. M , J. H. 13:427430.
, W. Z. P ,   Y. B. L . 1988. M , J. L.,   L. M. P#. 2000. Nesting
Mixed support for spatial heterogeneity in spe- habitat and success of the Chimango Caracara
cies interactions: Hummingbirds in a tropical in southern Chile. Wilson Bulletin 112:
disturbance mosaic. American Naturalist 131: 225232.
3357. M , M. L., B. G. M,   R.W. M  .

, K. E.,   H. G. M
 . 1986. 1992. WildlifeHabitat Relationships. University
Importance of area and habitat heterogeneity of Wisconsin Press, Madison.
to bird assemblages in temperate forest frag- R#, C. J., G. R. G
, P. P
, T. E. M ,
ments. Biological Conservation 36:115142.   D. F. D
. 1993. Handbook of Field
, A. 1988. Red Book of Chilean Terrestrial Methods for Monitoring Landbirds. U.S.
Vertebrates. Corporacin Nacional Forestal, Department of Agriculture, Forest Service
Santiago, Chile. General Technical Report PSW-GTR 144.
G, D. 1975. Prey availability and the food R
, D. G.,   G. J
. 1999. Giant panda con-
of predators. Ecology 56:12091214. servation action plan. Pages 241245 in Bears
H , R. S., S. M. P

,   P. H
. Status Survey and Conservation Action Plan
1986. A xed-radius point count method for (C. Servheen, S. Herrero, and B. Peyton, Eds.).
non-breeding and breeding season use. Auk International Union for the Conservation of
103:593602. Nature, Cambridge, United Kingdom.
J  , A. W.,   J. D. G. 1967. The Birds R
, S., C. C
 , O. B, C. M
of Chile and Adjacent Regions of Argentina, C. S-G,   P. A. M"
. 2002.
Bolivia and Per, vol. 2. Pla Establecimientos Conservation of temperate forest birds in Chile:
Grcos S. A., Buenos Aires, Argentina. Implications from the study of an isolated for-
L, A., C. D ,   J. C. A
. 1996. est relict. Biodiversity and Conservation 11:
La conservacin del bosque nativo en Chile: 19751990.
Problemas y desafos. Pages 335363 in Ecologa R
, R.,   G. T . 1994. The Birds of South
de Los Bosques Nativos de Chile (J. J. Armesto, America, vol 2. University of Texas Press, Austin.
C. Villagrn, and M. K. Arroyo, Eds.). Editorial R, R., D. M
, M. F. W ,   C.
Universitaria, Universidad de Chile, Santiago. S. 1996. Avifauna de los bosques templa-
L , S. L. 1993. Ecological and evolutionary dos de Sudamrica. Pages 135150 in Ecologa
perspectives on escape from predatory a ack: de Los Bosques Nativos de Chile (J. J. Armesto,
A survey of North American birds. Wilson C. Villagrn, and M. K. Arroyo, Eds.). Editorial
Bulletin 105:147. Universitaria, Universidad de Chile, Santiago.
L , C. H.,   A. O
. 2003. Vertical structure S
 , K. E., M. F. W ,   T. L. D
S .
inuences basal area development of second- 1996. Habitat barriers to movement of under-
growth Nothofagus stands in south-central story birds in fragmented south-temperate
Chile. Journal of Applied Ecology 40:639645. rainforest. Auk 113:944949.
, D.,   F. M. J. 1996. Habitat, S
 , K. E., M. F. W ,   T. L. D
S .
abundance and diet of the Rufous-legged Owls 2000. Dening corridor functions for endemic
(Strix rupes) in temperate forest remnants of birds in fragmented south-temperate rainfor-
southern Chile. Ecoscience 3:259263. est. Conservation Biology 14:11201132.
April 2004] Importance of Bamboo for Understory Birds 525


, T. T. 1982. Growth pa erns of Chusquea W , M. F.,   J. J. A
. 1996. The natural
bamboos in the understory of Chilean history of Chilo: On Darwins trail. Revista
Nothofagus forests and their inuences in for- Chilena de Historia Natural 69:149161.
est dynamics. Bulletin of the Torrey Botanical W , M. F.,   T. A. C
. 1996. Bird
Club 109:474487. communities of northern forests: Ecological

, T. T., T. K

, B. R. B  ,   A. J. correlates of diversity and abundance in the

 . 1996. Perturbaciones y dinmica de understory. Condor 98:350362.
regeneracin en bosques andinos del sur de W , M. F., T. L. D
S , C. S,   J.
Chile y Argentina. Pages 169198 in Ecologa J. A
. 1994. Avian communities of frag-
de Los Bosques Nativos de Chile (J. J. Armesto, mented south-temperate rainforest in Chile.
C. Villagrn, and M. K. Arroyo, Eds.). Editorial Conservation Biology 8:508520.
Universitaria, Universidad de Chile, Santiago. W , M. F., T. L. D
S , C. S,   J. J.

, F. 1985. Forest birds of Patagonia. A
. 1996. Avian communities in temper-
Ornithological Monographs 36:255304. ate rainforests of North and South America.
, J. R. 1998. The ecological basis of avian Pages 228247 in High-latitude Rainforests and
sensitivity to habitat fragmentation. Pages Associated Ecosystems of the West Coast of the
181192 in Avian Conservation Research and Americas (R. C. Lawford, P. B. Alaback, and E.
Management (J. M. Marzlu and R. Sallabanks, Fuentes, Eds.). Springer, New York.
Eds.). Island Press, Washington, D.C. W , M. F., J. L. M , K. E. S
 , T. L.
, D. G., R. L. C , S. L. S
,   J. D
S , L. S 
 ,   I. D. 2001.
F. 1993. Population variation, habitat Pa erns of predation risk and survival of bird
selection, and minimum area requirements nests in a Chilean agricultural landscape.
of three forest interior warblers in central Conservation Biology 15:447456.
Missouri. Condor 95:968979. Z, J. H. 1984. Biostatistical Analysis, 2nd ed.
, J. A. 1989. The Ecology of Bird Prentice-Hall, Englewood Clis, New Jersey.
Communities. Processes and variations, vol.
2. Cambridge University Press, Cambridge,
United Kingdom. Associate Editor: P. Escalante