the role of seed dispersers in the population dynamics of the

Ecology, 83(9), 2002, pp. 2617-2629
@ 2002 by the Ecological Society of America
THE ROLE OF SEED DISPERSERS IN THE POPULATION DYNAMICS
OF THE COLUMNAR CACTUS NEOBUXBAUMIA TETETZO
HÉCTOR GODÍNEZ-ALVAREZ,I
J
ALFONSO
VALIENTE-BANUET,
AND ALBERTO
ROJAS-MARTÍNEZ
Instituto de Ecología, Universidad Nacional Aut6noma de México, Apartado Postal 70-275, Coyoacán,
Distrito Federal, C6digo Postal 045/0 México
Abstract. We examined the effect of frugivorous bats and birds on the population
dynamics of the giant columnar cactus Neobuxbaumia tetetzo in the Tehuacán Valley,
Mexico. Because successful seedling establishment occurs only beneath the canopies of
shrubs and trees, we hypothesized that seed dispersal is a key process in the maintenance
of its populations. We determined the dispersal effectiveness of different frugivores, conside¡:ing the quantity and quality components of seed dispersa!. We algo evaluated the
potential effects of each frugivore species on the finite cate of increase of N. tetetzo populations by modifying the fecundity values of a Lefkovitch matrix mode!. The bat Leptonycteris curasoae had the highest effectiveness whereas the bird Carpodacus mexicanus
had the lowest. The estimated finite cates of increase calculated to evaluate the effects of
frugivores on the population dynamics of the cactus differ marginally from unity, except
",h..n
.. u_u
th..
>u- ..l'f
_u_n t "f
--
th..
hgt
-_o -.T
N.~no~~n
-_o
nmo
n --
gng!"7..A
_U_'J
--_o
("\n~
~-'
...onl+o
._~_.>~
rn~~"r'
~-óó-~>
'kg'
>u_>
.k..
>u-
kg'
u_>
Leptonycteris curasoae could be considered the legitimate dispersal agent of N. tetetzo,
dispersing seedsdirectly to gafe sites, and thus representing a key species in the ecology
~"
-.
,"':~
_~1
--'-"'U_'
---,..-
~.
Key words: arid tropical scrub; bats; birds; frugivory; Leptonycteris curasoae; matrix models;
sale sités; seed dispersal; seed germination; seedling establishment; Tehuacán Valley, Mexico.
INTRODUCTION
One central aspect in ecology is to understand the
evolution and ecology of mutualistic systems (Jordano
1992, Bronstein 1994, Bohning-Gaese et al. 1999). In
the case of seed dispersal, animal vectors that disperse
seedsmight playa key TOlein the maintenance of plant
populations becausethey act as the link between plant
reproduction and the subsequent recruitment of new
individuals (Herrera et al. 1994, Schupp and Fuentes
1995). Particularly, the physiology and activities offrugivores might play an outstanding TOlein this respect
because they determine the moment, the site, and fue
conditions in which seedsare deposited (Jordano 1992,
Schupp 1993, Loiselle and Blake 1999).
The contribution of frugivores to the recruitment of
the dispersed plants depends on both the quantity of
seeds dispersed and the quality of the dispersal provided to each seed (Schupp 1993, Jordano and Schupp
2000). The independent evaluation of these two components of seed dispersal is ecologically important because it allows identification of the legitimate dispersal
agents for particular plant species. Often, frugivores
that disperse the majority of seeds do not always provide the best quality of dispersal; and in contrast, frugivores that provide the best quality of dispersal are
not necessarily those that disperse the highest number
of seeds (Reid 1989, Schupp 1993, Loiselle and Blake
Manuscript received 25 February 2001; revised 5 December
)m!. , ~rr,.nt,.tI
)- __n__"
r~nn~rv )m)
_n-r------I E-mail: [email protected]
1999). The product of these two components determines dispersal effectiveness (sensu Schupp 1993), indicating whether a particular animal species is a legitimate dispersal agent for a certain plant species.
Although effectiveness of dispersal is an important
element in the evaluation of the effects of frugivores
on plant recruitment, at present, the determination of
the actual contribution of seeddispersers to plant population dynamics is the primary goal in the study of
seeddispersal (Schupp and Fuentes 1995). Understanding the relevance of one or various frugivores in relation to the quantity or quality of dispersed seeds as
well as its effects on the probability that a dispersed
seed becomes a reproductive adult is basic to identify
the frugivores that may be playing a key role in the
maintenance of plant populations in natural communities (Schupp and Fuentes 1995). Despite this fact, a
great amount of information on seed dispersal considers only fruit selection, seed handling, and habitat selection of frugivorous animals and ver y few refer to
the effects of these animals and their activities on the
subsequent stages of dispersed seeds such as germination, seedling establishment, and sapling growth and
survival (but see Reid 1989, Herrera et al. 1994, JordaDo and Herrera 1995, Jordano and Schupp 2000,
W,.nnv
.
7000)
In this paper, we propose that one approach towards
a general understanding of the role of seed dispersers
in the demography of plants is to combine the assessment of the effectiveness of each disperser with demographic tools to integrate the quantity and quality
2617
2618
HÉCTOR GODÍNEZ-ALVAREZ
components of seed dispersal in the fecundity values
of reproductive size categories. This approach would
permit us to simulate different scenarios in which the
presenceor absenceof particular frugivore speciesand
their potential effects on the population dynamics of
plants could be analyzed.
In deserts, the successful establishment of c.olumnar
cacti occurs only in gafe sites beneath the canopies of
perennial nurse plants which ameliorate environmental
conditions, enhancing seed germination and seedling
survivorship (Shreve 1931, Turner et al. 1966, Steenbergh and Lowe 1969, Valiente-Banuet and Ezcurra
1991). Seed dispersal may strongly influence the population dynamics of these cacti, since the survival of
those individual s associated with trees and shrubs has
been reported to be the most important life history
parameter affecting the finite rate of increase of Neobuxbaumia tetetzo (Coulter) Backeberg populations
(Godínez-Alvarez et al. 1999), as well as other columllar cacti populations such as Carnegiea gigantea (Engelm.) Britt. & Rose (Steenbergh and Lowe 1969,
1977) and Pachycereuspringlei (S. Watson) Britton &
Rose (C. Silva-Pereyra, unpublished manuscript).
Within the desertsof south-central Mexico, columnar
cacti are the physiognomic and structural dominant elements of vegetation types called "columnar cactus
forests" (Valiente-Banuet et al. 1996). In these ecosystems, a total of 34 species of bats and 91 species
of birds have been reported (Arizmendi and Espinosa
de los Monteros 1996, Rojas-Martínez and ValienteBanuet 1996), and some ofthese animals maintain tight
biotic interactions with columnar cacti (Valiente-Banuet et al. 1996, 1997a, b). Neobuxbaumia tetetzo is an
8 m tall branched columnar cactus which dominates
the vegetation over ~400 km2 in the Tehuacán-Cuicatlán Valley. Its fruits are consumed by nine species
of bats and birds (Valiente-Banuet et al. 1996). In this
study, we determine the effectiveness of the different
seed dispersal agents of N. tetetzo, and their relative
impact on the population dynamics of this columnar
cactus in the Tehuacán Valley, Mexico. Our primary
goal is to determine the dispersal effectiveness of different frugivorous species, considering both the quantity and the quality components proposed by Schupp
(1993). Finally, as a way to document the effects of
frugivore activity on the cactus population dynamics,
we conduct demographic simulations by incorporating
dispersal effectiveness of different animal vectors into
a population projection matrix model for N. tetetzo. By
doing this, we predict that (1) seed dispersal is a key
process in the population dynamics of this cactus species since seedling establishment and early survivorship only occurs beneath the canopies of perennial
nurse plants, and (2) bats might be playing an outstanding role in the population dynamics becausefruits
ripen during the night and preliminary observations
indicate that a great amount of seeds are removed by
nocturnal frugivores.
ET AL.
Ecology, Vol. 83, No. 9
STUDY SITE AND METHODS
This study was conducted in the semiarid Valley of
Zapotitlán (18°20' N, 97°28' W), a local basin of the
Tehuacán Valley in the state of Puebla, Mexico. This
region owes its aridity to the rain shadow produced by
the Eastern Sierra Madre (Smith 1965). It has an annual
mean rainfall of 380 mm, most of which falls during
the summer months, and an annual mean temperature
of 21°C with rafe frosts (García 1973). The afea is
heterogeneouswith hills and mountains dominating the
landscape. Flat afeas occur only on alluvial deposits
along Río Salado. Different lithologies are present in
the study area inc1uding shale, limestone, and igneous
materials (Brunet 1967). The main vegetation types,
edaphically controlled, are characterized by the dominance of columnar cacti reaching densities of 1800
individuals/ha (Osorio et al. 1996).
The study was conducted in an afea near Cerro Cutá,
within the Helia Bravo Botanical Garden, located -30
km south of Tehuacán city in which Neobuxbaumia
tetetzo dominates with densities of -1200 individualsl
ha (Valiente-Banuet and Ezcurra 1991) (see Plate 1).
Seed germination and seedling establishment occurs
mostly beneath the canopies of the legumes Mimosa
luisana Brandegee, Caesalpinia melanadenia (Rose)
Standley and Senna wi.slizenii (A. Gray) Irwin & Barneby (Valiente-Banuet et al. 1991). Flowering and
fruiting seasonsoccur from mid-April until the end of
June, just before the rainy season.The species is hermaphroditic and its flowers are whitish and mainly nocturnal, opening at dusk (2000) and c1osingin the early
morning (0600). The bats Choeronycteris mexicana
Tschudi and Leptonycteris curasoae Miller are the only
effective pollinators (Valiente-Banuet et al. 1996). The
fruits show characteristics associatedwith dispersal by
bats (van del Pijl 1982): they are green and grow on
the tips of branches measuring 35.5 :!:: 4.3 mm long
(mean :!:: 1 so; n = 39), 25.0 :!:: 1.9 mm wide (n =
39), and weighting
11.5 :!:: 1.8 g (n
=
":;
:?
39). Ripe fruits
dehisce at night, exposing a sugary whitish pulp with
a mean of 933 :!:: 302 (n
= 35)
small black seeds. Fruit
color and odor are very similar to the flowers, corresponding to a bat-dispersal syndrome. Besides the lesser long-nosed bat (L. curasoae), other frugivores have
been observed consuming the fruits of N. tetetzo (Valiente-Banuet et al. 1996). Among these frugivores are
the White-winged Dove (Zenaida asiatica. Linnaeus),
the Gray-breasted Woodpecker (Melanerpes hypopolius Wagler), and the House Finch (Carpodacus mexicanus P. L. S. Muller).
:
Effectiveness of seed dispersal
Dispersal effectiveness was estimated considering
aspectsof the quantity and quality components of seed
dispersal proposed by Schupp (1993). The quantity
component was estimated using data on the relative
abundanceof frugivores and the frequency of visits to
September 2002
SEED DISPERSAL AND POPULATION DYNAMICS
2619
PLATE 1. Colilmnar cactus forest dominated by Neobuxbaumia cerezoin the Tehuacán Valley, Mexico. Photograph by
Alfonso Valiente-Banuet.
the fruits per hour. The quality component was estimated using data on seedgermination after gut passage
and the probability of seeds being deposited under
shrub and tree canopies (i.e., gafe sites according to
Valiente-Banuet ali.d Ezcurra 1991). Effectiveness of
each frugivore species was calculated as the product
of the tour components (Effectiveness = relative abundance x frequency of visits x seedgermination x
deposition probability under shrubs and trees). Because
the fruits of N. tetetzo are eaten during both the day
and the night, we used different methods to estimate
the components of effectiveness for birds and bats. The
methods employed were selected considering the biology of both groups of frugivores. These methodological differences have direct consequenceson the value of effectiveness, and therefore on the comparison
of effectiveness among the different species of birds
and bats. However, we considered that each method
provides useful and real information on which to base
our calculations; thus we used them in order to obtain
estimates of the tour components of dispersal effectiveness for both groups of frugivores.
Quantity component.-Relative abundance of birds
was estimated using data gathered in 1989 and 1998.
During 1989,three mist nets (20 X 3 m) were placed
along a transect within a large cluster of N. tetetzo
separatedby 150 m just before the flowering seasonof
the species in February and March, and during the last
part of the fruiting season in July. Nets were opened
from 0700 to 1200 for Id in February; fram 0700 to
1800 for 6 d in March; and from 0700 to 1500 h for
3 d in July. The total netting effort for this sampling
period was 285 h. During 1998, another census was
carried out in the afea from May 9 to June 7, corresponding to the main fruiting season. In this census,
fivenets (two at 12 X 2.6 m, three at 6 X 2.6 m) were
placed in the same manner detailed above. Nets were
opened fram 0600 to 1800 for a total netting effort of
1500 h (M. Pérez, personal communication). Bird
abundances
wereper
standardized
and calculated
as the
numberof birds
hour per neto
.
The. frequency of visits to N. tetetzo fruits was determined through daily observations of a group ofthree
to five reproductive individuals from a distance of -30
m. Observations were made in the morning (07001400) and in the afternoon (1500-1900) in periods of
15 min to record visitar species, frequency, and du-
~
2620
Ecology, Vol. 83, No. 9
HÉCTOR GODÍNEZ-AL V AREZ ET AL.
r through June
i in cages (60
l. tetetzo fruits
birds and bats
care was taken
ble. After this
eces were col-
ration of visits. 1
May to 15 June
Relative abun<
data gathered ff(
three mist nets (:
a transect within
by 150m from j
of this columnar
plants from 200(
The total netting
h. Bat abundanc
individuals per t
The frequency
was determined
Leptonycteris cu
These observatil
servations condu
es (Nightvoyage
son, Mississippi,
distinguish betwl
visits to the frui
which it is diffil
and seeds.There
fruits was detern
asoae and C. me)
Banuet 2000). In
asoae and three I
May to 7 June 1
placed in parche
Animals were re
m) in which there was a branch 01 N. tetetzo wItn npe
fruits. Bats were observed with dim light to record the
frequency of visits to fruits and the proportion of time
spent feeding and resting. The proportion of time spent
resting was used to estimate the probability of depositio n of seedsby balsas detailed below. One bar was
used in each fria!. A visit was defined as a flight event
in which a portion oí the fruit pulp was removed by
the bat.
Additional observations were made in OUTstudy site
to determine whether there were other nonflying mammal s that red on N. tetetzo fruits. During the fruiting
season, we walked randomly every other day for 30
min in OUTstudy gire looking for reces that contain N.
tetetzo.seeds. Feces found in the field were identified
and collected in paper bags. We did not determine the
density and the number of seeds per reces, since we
considered that consumption of N. tetetzo fruits by terrestrial mammals was minar in compa!ison to fruit consumption by birds and bats.
Quality component.1. Seed germination.- The effect of gut passageon
seedswas estimated through germination experiments
to determine the proportion of germinated seeds and
their germination rateoIn 'theseexperiments seedsdefecated by different speciesof birds and bats were used.
The animal s were captured with three mist nets (20 X
3 m) placed in a N. tetetzo patch during day (0700-
.
;uch as coyote
yon cinereoartly from feces
cted and stored
eriments were
mi mal species
per (Whatman
. USA) moisttook place at
ht conditions.
,he number of
package version 3
Cra
considered an equal proportion
between the treatments and the
:'1Jl
y -
ea+bt
1+
k
ea+bt
where y is the number of germinated seeds at time t,
a is intercept at t = O, b is the germination rate, and
k is the maximum number of germinated seeds.
2. Location oi seed deposition.- The probability of
depositing seeds under shrub and tree canopies was
estimated by focal observations to determine the number of visits to different perching plants used by birds
after fruit consumption. A group of three to five N.
tetetzo plants with ripe fruits was selected and observations were made daily through binoculars from a
distance of -50-100 m. The number of visits to each
~
SEED DISPERSAL AND POPULATION DYNAMICS
September2002
bat behavior during feeding periods, the proportion of
time spent resting could be taken as a rough estimate
of the probability that bats deposit seedsin gafe sites.
Seed removal by birds and bats
en
Q)
"(3
Q)
.
o..
en
"E
15
-en
1:
(ti
15.
C>
c: .
:c
o
(¡¡
o..
o
z
Total observation time (h)
FIG. 1. Cumulative number of perching plants observed,
after N. tetetzo fruit consumption, for different bird species
in the Tehuacán Valley, Mexico.
species oí perching plant per bird was recorded and
grouped into three categories: (1) visits to N. tetetzo
where birds spent most of the time on the tips of the
columnár cactus branches, (2) visits to creesand shrubs
where birds spent most of the time among branches of
different woody plants, and (3) visits to other nonshruby plants where birds spent most of the time in
plants such as agaves and prickly pears. Focal observations ended when animals could no longer be followed. Observations were made for 20 consecutive
days, starting at 0700 and ending at 1900. Total observation time was 50 h and, after this time, presumably
all potential perching plant species were detected (Fig.
1). A contingency table was used to analyze if there
were significant differences in the number of visits to
each category of perching plants among the different
bird species recorded. The null hypothesis considered
an equal number of visits between categories.The probability that seedswould be deposited in gafe sites (Le.,
trees and shrubs) was determined per bird species estimating the proportion of visits to a particular perching
plant category with respect to the total number of visits
recorded.
2621
.
The probability of seed deposition by the bats Leptonycteris curasoae and Choeronycteris mexicana was
estimated using the methodology previously described
to estimate the frequency of visits to N. tetetzo fruits.
This deposition probability was calculated as the proportion of time that bats spent resting, because focal
observations of N. tetetzo fruits conducted with night
vision lenses(Nightvoyager modell/l) showed that bats
perch in the dominant trees and shrubs immediately
after removing a piece of fruit. Once in the perch, they
handle and consume the piece of fruit and rest until
the next feeding flight (Godínez-Alvarez and ValienteBanuet 2000). Considering the previous description of
A fruit-removal experiment was conducted in the
field from May through June 1997 to estímate the IDean
number of seedsremoved by birds and bats. Each day,
all of the available ripe dehiscent fruits were tagged
and assigned randomly to one of the following treatments: (1) bird removal, where fruits were bagged with
nylon mesh during the night and exposed to birds from
0700 to 1800, and (2) bat removal, where fruits were
covered during the day and exposed to bats during the
night from 2000 to 0600. The fruits used in this experiment were selectedfrom at least 10 different plants.
.Because N. tetetzo fruits ripen gradually during the
reproductive season,the number.of fruits usedper treatment varied according to the number of fruits available
each night. For each treatment, the proportion of pulp
and seedsremoved per fruit was estimated and placed
in one of four categories: (1) consumption of <25%
of the fruit, (2) consumption of 25-50% of fruit, (3)
consumption of 50-75% of fruit, or (4) consumption
of >75% ofthe fruit. A fruit was consideredeatenwhen
a portion of its pulp and seeds was absent and no evidence could be found that it had fallen to the ground.
All fruits used each night were removed from the
plants..Data were analyzed with a contingency table to
determine if the proportion of pulp and seedsremoved
differed significantly between the bird and the bat
groups. The null hypothesis stated that there were no
significant differences between birds and bats in the
proportion of pulp and seeds they removed. The proportion of seedsremoved by each group of frugivores
was multiplied by the IDeannumber of seedsper fruit
to estímatethe number of seedsremoved by eachgroup.
Effect oi seed dispersal on N. tetetzo
population dynamics
The effect of seed dispersal on the finite cate of increase of N. tetetzo populations was estimated using a
Lefkovitch matrix model previously built for this population at this study site (Table 1; Godínez-Alvarez et
al. 1999). In this demographic model, individuals were
grouped in size categories, according to the height of
theprincipal trunk, and a l-yr projection matrix was
constructed considering the survival, reproduction, and
growth probabilities of the individuals in each category. The size categories used in this matrix were as
follows: <2 cm, seedling 1; 2-8 cm, seedling 2; 8-15
cm, seedling 3; 15-45 cm, sapling; 45-100 cm, juvenile; 100-150 cm, immature; 150-250 cm, matuTe 1;
250-350 cm, matuTe2;350-450 cm, matuTe3; 450550 cm, matuTe4; 550-650 cm, mature 5; >650 cm,
matuTe6. This matrix consists of three main parts: (1)
The first row includes the fecundity values for all the
reproductive classes. These fecundity values were es-
2622
HÉCTOR GODÍNEZ-ALV AREZ ET AL.
Ecology,Vol. 83, No. 9
T ABLE 1. Matrix model of Neobuxbaumia tetetzo used to simulate the effect of seed dispersal by different species of
frugivores
Size
on the finite late of increase.
Size categories (n,)
m2
m4
m3
m5
m6
..
0.925
0.057
0.93
0.059
0.952
0.048
0.822
0.107
0.949
Note: The size categories are as follows: se1, <2 ~m; se2, 2-8 cm; se3, 8-15 cm; sa, 15-45 cm; j, 45-100 cm; im, 100150 cm; mI, 150-250 cm; m2, 250-350 cm; m3, 350-450 cm; m4, 450-550 cm; m5, 550-650 ém; m6, >650 cm.
timated as the product of the probability of reproduction, the IDeannumber of seedsproduced by an average
individual of each reproductive class, and the probability of passing from seed to seedling. (2) The main
diagonal includes the probabilities that individual s remain in the same size class after one year. (3) The first
lower subdiagonal incorporates the probabilities that
individuals grow to the next size class after one year.
We incorporated the effect of seeddispersal by each
animal vector into this matrix model by modifying the
fecundity values. This modification was a product of
(1) the probability of seed removal by each disperser
species, (2). the proportion of germinated seeds after
gut passage,(3) the probability that the disperser would
deposit the seeds beneath the canopy of trees and
shrubs; and (4) the survi val probability of seeds beneath.tre.esand shrubs (0.00139) which is assumedto
be constant for all dispersers. In turn, the calculation
of the probability of seed removal was a function of
several measures:We first multiplied the relative abundance of each disperser species within either birds or
bats by its frequency of visits to N. tetetzo fruits. We
then multiplied the result by the IDeannumber of seeds
removed by the relevant group (Le., 233 seeds removed, on average, by birds, and 700 by bats; see
Results: Seedremoval by birds and bats). Finally, these
data were added up across all disperser species.Of this
total, the relative proportion corresponding to each disperser specieswas interpreted as the probability of seed
removal per species. Once fue modification of fecundity values was calculated for each disperser species,
it was incorporated in the matrix modeI by substituting
it as the probability of passing from seed to seedling
in the calculation of fecundity values.
To test the rol e of seed dispersal in the population
dynamics of N. tetetzo, we performed two matrix simulations considering (1) the effect of all dispersers acting together, and (2) the absenceof seed dispersers by
assuming that N. tetetzo only interact with the seed
predator Carpodacus mexicanus. Additionally, other
matrix simulations were performed to analyze the individual effect of particular dispersers.To simulate the
effects of all disperser species acting together, wecalculated the probability of seed removal as the sum of
all the probabilities estimated for each specieswhereas
the germination proportion after gut passage and the
deposition probability were estimated as the IDean
weighted by the seed removal probability.
For each simulation, we estimated the finite Tale of
increase using the program STAGECOACH, version
2.3 (Cochran and Ellner 1992), and its 90% confidence
limits to determine whether it differed from unity (Le.,
whether the population may be considered in a numerical equilibrium). Confidence limits for X.were obtained using Monte Carlo simulation models in which
the variances for matrix entries were estimated assuming a lag-normal distribution of errors for fecundity
values and a multinomial distribution of errors for transitian probabilities (Alvarez-Buylla and Slatkin 1993).
With these assumptions, 1000 simulations were conducted and the sampling distribution of lambda was
constructed for each matrix. Confidence limits were
determined considering the values of the 5 and 95 percentiles.
RESULTS
Ejfect 01 seed dispersal
Quantity component.-Seven species of birds belonging to fue families Emberizidae, Fringillidae, Mimidae, Picidae, and Troglodytidae visited the fruits of N.
tetetzo (Table 2). All observed speciesred only on fruits
whereas Carpodacus mexicanus and Melanerpes hypopolius algo red on flowers. The Lesser Goldfinch
(Carduelis psaltria Say) and Campylorhynchus brunneicapillus were the species with the highest relative
abundancefollowed by the Varied Bunting (Passerina
versicolor Bonaparte; Table 2).
Zenaida asiatica and C. mexicanus were the most
frequent visitors to N. tetetzo fruits. Their visits added
September 2002
SEED DISPERSAL AND POPULATION DYNAMICS
TABLE 2. Relative abundance(individuals.h-I.net-I) fordifferent birds and bats inpatches oí Neobuxbaumia tetetzo
in the Tehuacán Valley, Mexico.
2623
glottos and Aimophyla mystacalis were the specieswith
the highest and lowest rates, respectively (Fig. 2). In
spite of these differences, the 95% confidence intervals
showed that there were no statistically significant difRelative
abundance ferences among them. Therefore, germination rates
Family
Species
were not included in the estimation of dispersal effecBirds
tiveness. Seedgermination curves after gut passagefor
Passerina versicolor
0.007
Emberizidae
0.011
Carduelis psaltria
Fringillidae
different mammal and bird species showed that gerCarpodacus mexicanus
0.004
mination
was a rapid process that occurred within the
Mimidae
Toxostoma curvirostre
0.004
first 3-5 d after sowing, depending on the frugivore
0.004
Picidae
Melanerpes hypopolius
0.004
Picoides scalaris
species (Fig. 2).
0.011
Troglodytidae
Campylorhynchus
2. Location 01 seed deposition.-After fruit conbrunneicapillus
sumption, birds perched mostly ón various shrubs,
Bats
trees, and columnar cacti (Fig. 3). N. tetetzo was strong0.002
Artibeus jamaicensis
Phyllostomidae
ly preferred by Z. asiatica and M. hypopolius, while
Choeronycteris mexicana
0.106
shrubs were more frequently preferred by M. polyLeptonycteris curasoae
,0.161
glottos (G = 257.2, df = 12, P < 0.0001). For the Test
Leptonycteris nivalis
0.016
of the species, we did not find significant differences
in their preference for N. tetetzo and shrubs compared
up to >70% of the total number of visits recorded for to expected values (Fig. 3).
birds 'and their frequency of visits were over one visit
The bats L. curasoae and C. mexicana consumed the
per honroIn contrast, the visits of the rest of the species pulp and the seedsof the fruits beneath their perches.
were < 10% of the total number of visits and their fre- The proportion of time spent resting for both species
quency .of visits was lower tl1an one visit per hour was 71.7 :t 4.8% (mean :t 1 SE; n = 5) and 73.8 :t
(Table 3). The number of visits to N. tetetzo fruits var- 5.3% (n = 3), respectively, thus the probability of seed
depositionwas 0.72 for L. curasoaeand 0.74 for C.
ied throughout the dar, showing two peaks of activity:
the first peak was observed in the morning around 0930, mexicana.
Of all the species.of birds and mammals studied, we
the second occurred in the afternoon at ~1730. Bird
obtained the complete data necessary to estimate the
activity decreasedat noon between 1200 and 1600.
The main bats that consumed N. tetetzo fruits were effectiveness of dispersal only for four speciesof birds
the phyllostomids Leptonycteris curasoae and Cho- and one species of bat (Table 4). The dispersal effeceronycteris mexicana, which had the highest relative tiveness estimated for these frugivores showed that the
abundance(Table 2). The IDeanfrequency of visits per lesser long-no sed bat (Leptonycteris curasoae) had the
hour (:!::l SE) for L. curasoae was 17.9 :!::4.0 (n = 5), highest value given by both a high frequency of visits
whereas for C. mexicana it was 20.7 :!:: 6.4 (n = 3). to fruits and a high probability of seed deposition in
Besides birds and bats, the coyote (Canis latrans) and gafe sites. Other speciessuch as the Cactus Wren (Camthe gray fox (Urocyon cinereoargenteus) algo fed on pylorhynchus brunneicapillus), Gray-breasted Woodthe fruits' of N. tetetzo once they fell to the ground. pecker (Melanerpes hypopolius), and Curve-billed
Fecesof these animals had numerous seedsand always Thrasher (Toxostoma curvirostre) presentedhigher germination percentagesthan bats but intermediate values
were found in open spaces.
of dispersal effectiveness. The House Finch (CarpoQuality component.1. Seed germination.-The effect of gut passageon
the proportion of germinated seeds and their germination cate was estimated only for five speciesof birds TABLE 3. Data on visits made to N. tetetzo fruits by different
frugivorous birds in the Tehuacán Valley, Mexico.
and three speciesof mammals (Fig. 2). Seedsdefecated
by the birds C. brunneicapillus, Toxostoma curvirostre
Mean
Swainson, Mimus polyglottos, and Melanerpes hypoduration
Frequency ::': 1 so
polius presented the highest germination percentages
Number (visitslh)
Bird species
(mjn)
(>95%), not differing from control seeds (99%). In
2;1
:t 2.9
1.84
Carpodaeus
mexieanus
116
contrast, germination of seeds defecated by bats (L.
66
1.04
2.3 :t. 3.6
curasoae) and gray foxes (U. cinereoargenteus) varied Zenaida asiatiea
0:9 :t 0.7
27
0.43
Melanerpes hypopolius
between 80% and 86%, differing significantly from the Cyanoeompsaeyanoides
1.4 :t 1.2
11
0.17
2.2 :t 1.8
9
0.14
control. Seedsdefecated by Canis latrans and A. mys- Toxostoma eurvirostre
0.6 :t 0.5
7
0.11
tacalis had a germination lower than 50%. The. birds Campylorhynehus
brunneieapillus
Z. asiatica, C. psaltria, and C. mexicanus destroyed all Mimus
4.0
:t 4.0
7
0.11
polyglottos
0.8
:t 0:6
'7
0.11
seeds and no germination was observed.
Picoides sealaris
The estimated germination catesfor all speciesvaried
Notes: Total observation time, 63.2 h; global frequency of
from 1.0 to 5.7 germinated seeds per dar. M. poly- visits, 4.0 visits/h.
HÉCTOR GODÍNEZ-AL V AREZ ET AL.
2624
Ecology, Vol. 83, No. 9
.
.
.
.
~t::
.Q
1i:i
t::
.E
.....
Q)
CJ
....
80
60
,'"
~
:cf
.¡jl
te;¡.,
.
40
20
o
Time (d)
Fro. 2. Gerrnination of N. tetetzo seeds after gut passage for different species of birds and rnarnrnals: (a) Urocyon
cinereoargenteus (filled squares), Canis latrans (open squares), control (open circles); (b) Leptonycteris curasoae (filled
squares), Campylorhynchus brunneicapillus (open square.s),control (open circles); (c) Toxostoma curvirostre (filled squares),
Aimophyla mystacalis (open squares), control (open circles); (d) Melanerpes hypopolius (filled squares), Mimus polyglottos
(open squares), control (open circles). Syrnbols and lines refer to data and fitted logistic regressiuns, respectively.
dacus mexicanus) destroys seeds either when feeding
on fruits or during gut passage, and thus shows null
dispersal effectiveness;this speciesacts rather as a seed
predator (Table 4).
Seed removal by birds and bats
Bats removed a greater proportion of pulp and seeds
from fruits than birds dictoThe number of fruits with
>75% removal by bats were higher than expected by
chanceoIn contrast, the number of fruits with <25%
removal by birds were higher than expected by chance
(G = 38.6, df = 3, P < 0.0001;Fig. 4). None of the
birdspecies that visited N. tetetzo fruits have nocturnal
activity, thus they do not remove fruits at night. Considering that the IDeannumber of seedsper fruit is 933
:t 302 (mean :t 1 SD; n = 35), these results suggest
that bats could remove a IDean of 700 seeds per fruit
each night whereas birds may remove only a IDeanof
233 seedsper fruit each day.
Effect 01 seed dispersal on N. tetetzo
population dynamics
Table 4 shows the estimated finite rates of increase
aimed at evaluating the effects of seed dispersal by
each animal vector, estimated by the modification of
the fecundity values. This modification was a product
of the probability of seed removal by each disperser
species, the proportion of germinated seeds after gut
passage, the probability that the disperser would deposit the seedsbeneath the canopy of trees and shrubs,
and the survival probability of seedsbeneath trees and
shrubs. The late of increase considering the effects for
all dispersersacting together showed that the N. tetetzo
population is in a numerical equilibrium. This result
contrasts with the finite late of increase obtained when
it is assumed an absence of seed dispersers (Le., N.
tetetzo only interacts with the seed predator Carpodacus mexicanus) since the population of N. tetetzo
decreased(Table 4).
When we considered the effects of each disperser
individually, the estimated population growth rates varied among species. For instance, if only the effect of
C. brunneicapillus was considered, the population of
N. tetetzo showed a growth late lower than unity,
whereas if only the effect of the most effective disperser, L. curasoae, was inc1uded, a growth late equal
\
2625
SEED DISPERSAL AND POPULATION DYNAMICS
September 2002
. N. tetetzo OShrubs
Ii§\jOthers
Cyanocompsa cyanoides
Campyromynchusbrunne~apmus
Carpodacus mexicanus
..
Toxostoma curvirostre
Zenaida asiatica
**
',',',
**
**
***
***
-8
Mimuspolyglottos
',',',','
"""',***
Melanerpes
***
',',',',',',',',',',""',',',',
-6
-4
-2
O
2
4
hypopolius
6
8
10
Residuals
FIG. 3. Residuals of a contingency táble analysis including bird species and perching plants. Bars depict residual frequencies and their significance, indicating preference (positive residuals) or avoidance (negative residuals) by each bird
species.
* P < 0.05; **.p < 0.01; *** P <0.001.
to unity Was obtained. The 90% confidence intervals
for these values showed that the calculated finite rates
of increase for the N. tetetzo population incorporating
the individu~l effectof each disp~rser di:ffered marginally from unity except for the lesser long-nosed bat
(L. curasoae).
Considering the results of this study, in Fig. 5 we
present some of the possiblt; paths followed by N. tec
tetzo seedsduring the dispersal phase, along with their
potential effects on the finite late of increase. According to this diagram, seedscould be removed from fruits
by birds and/or bats, or el se they could fall down to
T ABLE 4.
the ground. The probability of this last event (0.03)
was calculated, independently of the probability of seed
removal, as the proportion of the IDeannumber of fruits
found in a l-m2 afea (1.7 :!::0.2, IDean :!:: 1 SE,n = 50)
at the base of N. tetetza plants with respect to the IDean
number of fruits per plant (49.2 :!::7.7, IDean :!:: 1 SE,
n = 30). Once on the ground, seedscould be exposed
to seed dispersers such as gray foxes (u. cinereoargenteus) and coyotes (c latrans) as well as to seed
predators such as harvester anís (Pogonomyrmex barbatus F. R. Smith). The seedsof the fruits that remain
attached to the plants could be eaten by House Finches
Dispersal effectiveness for different frugivores and their effects on N. tetetzo population growth Tale in the
Tehuacán Valley, Mexico.
Frugivores
Leptonycteris curasoae
Relative Frequency
abundance
of
(individuals.
visits
h-l.net-1)
(visits/h)
0.161
17.9
Seed
germinationt
-
Deposition
probability:j:
-
0.86
0.72
Estimated rate
of increase
Effectiveness§
Campylorhynchus brunneicapillus
0.011
0.11
0.975
0.43
5.1 X 10-4
Melanerpes hypopolius
0.004
0.43
0.955
0.03
.4.9 x 10-5
Toxostoma curvirostre
0.004
0.14
0.975
0.12
6.6 X 10-5
Carpodacus mexicanus
0.004
1.84
o
0.18
o
All dispersers
t
Proportion of seeds that germinate after gut passage.
:j:Probability of seeds being deposited in safe sites under shrub and tree canopies.
§ Effectiveness = relative abundance X frequency of visits X seed germination X deposition probability.
(90% límits)
1.003
(0.94-1.01)
0.962
(0.94-0.99)
0.959
(0.94-0.99)
0.959
(0.94-0.99)
0.959
(0.94-0.99)
1.003
(0.94-1.01)
2626
HÉCTOR GO])ÍNEZ-ALVAREZ
ET AL.
Ecology, Vol. 83, No. 9
as well as in captivity for different species. For example, C. brunneicapillus spent most of the time on
the ground or on small perennial shrubs where they
forage insects and construct their nests, simultaneously
avoiding potential predators (Simons and Simons
1990). This behavior increases the probability of dispersing seeds under shrubs (Milton et al. 1998). In
contrast, C. mexicanus and M. hypopolius showed a
high frequency of visits to N. tetetzo fruits. However,
these birds spent ver y liUle time on shrubs and, in the
case of C. mexicanus, had a negative effect on seed
germination. On the other hand, Melanerpes hypopolius did not affect seedgermination. However, Ibis bird
spent all of its time on columnar cacti (Hendricks et
al. 1990), suggesting that Ibis species is not a good
disperser.
For all frugivores, seed germination and probability
of seed deposition have higher values Iban those reported for their relative abundance.However, for some
species such as the bar L. curasoae, and the birds M.
hypopolius and T. curvirostre, the frequencyof visits
to fruits may algo have an important contribution to
dispersal effectiveness. The quality component is important for the successful recruitment of new N. tetetzo
individuals, since it occurs only beneath the canopies
of perennial trees and shrubs (Valiente-Banuet and
Ezcurra 1991). Therefore, the dispersal of intact and
viable seedsto these gires is an important step for the
FIG. 4. Proportion of pulp and seeds removed from N.
tetetzo fruits by birds and bats in the Tehuacán Valley, Mexpopulation dynamics of Ibis columnar cactus. Godínezico. apeo and closed bars refer to expected and observed Alvarez and'Valiente-Banuet (2000) found that L. curvalues, respectively. Asterisks denote significant differences
asoae feeds on N. tetetzo fruits while perched, con(G = 38.6, df = 3, P < 0.00001).
suming the pulp and the seeds. These same authors
reported that Ibis behavior favors the dispersal of seeds
(c. mexicanus). Gn the other hand, bat- and bird-dis- to a great variety of habitats including plants, rocks,
persed seedscould be transported to gafe sites beneath and rock crevices. In the TehuacánValley, night roosts
the canopies of trees and shrubs. This last path is of used by L. curasoae and C. mexicana include the domcrucial importance for the maintenance of N. tetetzo inant shrubs Mimosa luisana, Caesalpinia melanadenpopulations under natural conditions since, as the ma- ia, Cercidium praecox (Ruíz & Pavón) Harms and the
trix simulations showed, the estimated population columnar cacti Neobuxbaumia tetetzo and Stenocereus
growth rate, considering the effects for all dispersers stellatus (Pfeiffer) Riccob. (A. Rojas-Martínez, personal communication). Some of these plants, at our
acting together, did not differ from unity (Table 4).
study gire, have significantly higher numbers of young
DISCUSSION
N. tetetzo individual s growing beneath their canopies
Seeddispersal systemsin the columnar cactus forests Iban expected by chance (Valiente-Banuet et al. 1991).
Matrix simulations indicate that seed dispersal is a
dominated by N. tetetzo are complex. Different species
of birds and mammals eat N. tetetzo fruits dispersing key process since it maintains the population of N.
seeds in varyingquantities and with varying degrees tetetzo at equilibrium. This result is supported algo by
of dispersal quality, affecting the recruitment of new the estimated rates of increase in absenceof seed disindividuals. Therefore, these fruit-eating animal s have persers (Le., interaction only with the seed predator
'a direct effect on the population dynamics of this co- Carpodacus mexicanus) in which the N. tetetzo population decreased.When we considered only the effects
lumnar cactus.
The analysis of dispersal effectiveness showed that of L. curasoae, Ibis species maximize the value of }..
the wide array of fruit-eating animals may act both as suggesting that Ibis bar plays an outstanding role in the
seed predators (e.g., C. mexicanus) or as seed dispers- maintenance of N. tetetzo population. However, the relers with varying degrees of effectiveness (e.g., L. cur- ative importance of the different species of frugivores
asoae, C. brunneicapillus, and T. curvirostre). These may vary spatially and temporally, in our case more
results are supported by observations made in the field intense field work is necessary in order to obtain a
SEED DISPERSAL AND POPULATION DYNAMICS
September2002
...
1..<1
L curasoae
t
"
~
,
2627
C. brunneicapillus
M. hypopolius
T. curvirostre
C. mexicanus
(Seed predation)
(Seed dispersal)
0.99
0.72
0.0013
.
\
t
P. barba tus
0.86
Effect of fruit consumers
on seed germination
(Seed predation)
~
Possible paths for N. tetetzo seeds and their effects on fue finite Tale of increase in the Tehuacán Valley, Mexico.
better understaQdingof the relative impact of the different species of birds and bats.
With respect to the effects of other species of frugivores, coyotes (c latrans), and gray foxes (U. cinereoargenteus) did not affect seed germination but,
considering that these species defecate in open spaces
ami/or in caves where seed germination and seedling
establishment are unlikely, and also that secondary dispersers were not detected, it is possible to assume that
they have a negative effect on population dynamics,
acting mainly as seed predators. On the other hand,
field observations and Valiente-Banuet and Ezcurra
(1991) indicated that harvester ants (Pogonomyrmex
barbatus) are predators that consumea great proportion
of the seeds that fall down to the ground.
One of the reasons why the assessmentof the ecological role of species in seed dispersal has inhe.rent
limitations is related to the fact that seeddispersal systems in the tropics are complex. Pruits are consumed
by a wide array of species, and consequently, different
methods are neededto determine the quantity and quality components of seed dispersal for diurnal and nocturnal vectors (i.e., birds and bats). This aspect may
produce over- or underestimation of effectiveness.
However, by comparing OUTestimations of effectiveness for birds and bats we found similar patterns to
those found in other studies. Por example, Poulin et al.
(1993) studied the bird community of a tropical semi-
arid habitat in northeastern Venezuela, using mist nets
to evaluate the abundances of different species. This
study was conducted during one year and the total netting effort was 1152 h. The abundances reported by
these authors for birds of the same families (i.e., Fringillidae, Mimidae, Picidae, and Troglodytidae) analyzed in this study varied from 0.0009 to 0.13
birds.net-1.h-l. Some of these abundances (0.13
birds'net-1.h-l) are higher than those reported in this
study. However, these estimates do not increase the
effectiveness of birds in relation to bats. Different studies have reported that bird abundancesin arid and semiarid habitats vary temporally and spatially (Raitt and
Pimm 1976, Poulin et al. 1993). These fluctuations in
abundanceshould be ctmsidered in arder to understand
the effects of birds on the demography of N. tetetzo in
the long termo
In the same way, our results showed that the bat L.
curasoae presented the highest effectiveness because
the frequency of visits and the seed deposition probability under trees and shrubs were higher than those
found for the regí of the species. These values were
obtained in captivity and it is possible that they could
be overestimated. However, the consideration of the
results reported by other studies and the incorporation
of them in the estimation of the effectiveness of dispersal gave similar results. For instance, Bizerril and
Raw (1998) studied the feeding behavior of two other
2628
HÉCTOR GODÍNEZ-ALVAREZ
species of phyllostomids, Carollia perspicillata Linnaeus and Glossophaga sorlcina Pallas, in a gallery
forest, and found that the mean frequency of visits to
Piper fruits for both species was 14.2 :1::8.4 visits/h.
This figure is lower than the frequency of visits r,eported here for L. curasoae (17.9). However, if we
consider the estimation reported by these authors in
our calculations, bats are still the best dispersers of N.
tetetzo seeds. Howell (1979) studied the foraging behavior of L. curasoae in the Sonoran Desert, and found
that bats spent 3 h fiying every night and at around
midnight they roosted for 2-3 h. This author also reported that while fiying, bats red in a series of discontinuous bouts of :1::20min alternating with resting periods of :1::20mino Considering this information, we
estimated that the proportion of time that Howell's bats
spent resting eaclÍ night varied between 70 and 75%,
which is similar to our estimation (72%); In contrast
Horner et al. (1998), using radio transmittersto study
the foraging behavior of L. curasoae reported that this
species spent only 17~5%of its time resting in night
roosts. When these two figures (70-75% and 17.5%)
were considered in our effectiveness estimations, bats
were still the best seed dispersers for N. tetetzo seeds.
At the same time, our results indicate that bats removed a higher proportion of seedsand pulp from fruits
than birds. A similar situation was reported in the Sonoran Desert by Fleming and Sosa (1994), in which L.
curasoae removed 10-80% of the available seeds of
the columnar cactus Pachycereus pringlei each night.
However, germination peicentage and germination cate
of bat-ingested seedsdecreasedto 86% and 2.9 seeds
germinated per day, respectively. In contrast to our
results on the germination of bat-ingested seeds,GodÍnez-Alvarez and Valiente-Banuet (1998) found that N.
tetetzo seedsunder experimental treatments simulating
gut passageshowed the same germinability as the control treatment.
.
Considering its dispersal effectiveness and its impacts on the finite cate of increase of the N. tetetzo
population, it is possible to hypothesize that L. curasoae could represent a key species in the ecology of
this columnar cactus because this bat plays an important ecological role not only as seed disperser but also
as an effective pollinator. Studying N. tetetzo pollination in the Tehuacán Valley, Valiente-Banuet et al.
(1996) found that successful pollination dependson the
presenceof some bats such as L. curasoae and C. mexicana, among others. Moreover, studies conducted with
other columnar cacti (Neobuxbaumia mezcalaensis
(Bravo) Backeberg, Neobuxbaumia macrocephala
(Weber) Dawson, Valiente-Banuet et al. 1997a; Pachycereus weberi (Coult.). Buxb., Pilosocereus chrysacanthus (R A. C. Weber) Byles & Rowley, ValienteBanuet et al. 1997b; Stenocereus stellatus (Pfeiffer)
Riccobono, Casaset al. 1999) in the same region have
indicated that bats are algo necessl¡lryfor the reproduction of these plants. Indeed, L. curaSoaemaintains
ET AL.
Ecology, Vol. 83, No. 9
resident populations in the Tehuacán Valley and field
observations on the consumption of columnar cacti
fruits by bats have shown that L. curasoae also feeds
on the fruits of other 17 species, suggesting that seed
dispersal by bats should be important in population
dynami«s and community diversity of columnar cact\ls
forests (Rojas-Martínez et al. 1999).
In conclusion, our results indicate that N. tetetzo
fruits are eaten by several fruit-eating birds and mammals which differ in their effects on seed dispersal
successand hence on the population dynamics of this
co1umnar cactus. At the same time, this study support
the hypothesis that seed dispersal is a key process in
the popu1ationdynamics of N. tetetzo, and that from
all frugivores, the lesser long-nosed bat (L. curasoae)
could be considered as a key species in the population
dynamics of cacti in central Mexico.
ACKNOWLEDGMENTS
We thank Ariel Alcántara, Oscar Osorio, Adolfo Vital, Antonio Soriano, Leticia Ríos, Verónica Alvarado, and Rocío
José for field assistance. To Coro Arizmendi and Mónica
Pérez for providing information on bird abundance in the
study afea. Alejandro Casas provided critical reviews on the
first draft of the manuscript. The authors particularly want to
thank to Pedro Jordano and TeresaVal verde for their helpful
comments and suggestions to improve the final version. This
study was supported by Fondo Mexicano para la Conservación de la Naturaleza project AI-97/036, Dirección General
de Asuntos del Personal Académico, UNAM project IN207798, and a grant provided by Programa de Apoyo a las
Divisiones de Estudio de Posgrado project 030705 to H. Godínez-Alvarez. This studyis
research.
' part of the first author's doctoral
LITERATURE
CITED
Alvarez-Buylla,
E., and M. Slatkin. 1993. Finding- confidence
.
limits
on population
growth
rates: Monte Caria test of a
simole analvtic methad. Oikos 68:273-282.
Arizmendi, M: del C., and A. Espinosa de los Monteros. 1996.
A vifauna de los bosques de cactáceascolumnares del Valle
de Tehuacán, Puebla. Acta Zoológica Mexicana 67:25-46.
Bizerril, M. X. A, and A. Raw. 1998. Feedirig behaviour of
bats and the dispersal of Piper arboreum seeds in Brazil.
Journal of Trouical
Ecolol!v
.
-- 14:109-114.
Bohning-Gaese, K., B. H. Gaese, and S. B. Rabemanantsoa.
1999. Importance ofprimary and secondary seeddispersal
in the Malagasy tree Commiphora guilla'umi. Ecology 80:
R?1-R1?
Bronstein, J. L. 1994. Our current understanding of mutua1ism. Quarterly Review of Bio1ogy 69:31-5l.
Brunet. J. 1967. Geo1o!!ic
studies. Pa!!es
66-90 in D. S.
Byers, editor. The prehistory of the Tehuacan Valley. Volume 1. Environment and subsistence. University of Texas
Press, Austin, Texas, USA.
Casas,A., A. Valiente-Banuet, A. Rojas-Martínez, and p, Dávila. 1999. Reproductive biology and the process of domestication of the columnar cactus Stenocereusstellatus in
Central Mexico. American Journal of Botany 86:534-542.
Cochran, M. E., and S. Ellner. 1992. Simple methods for
calculating age-based life history parameters for stagestructured populations. Ecological Monographs 62:345364.
Crawley, M. J. 1993. GLIM for ecologists. Blackwell, Cambridge, UK.
Fleming, T. H., and V. J. Sosa. 1994. Effects ofnectarivorous
September 2002
SEED DISPERSAL AND POPULA nON DYNAMICS
and frugivorous mammals on reproductive success of
plants. Journal of Mammalogy 75:845-851.
García, E. 1973. Modificaciones al sistema de clasificación
climática de Koppen. Instituto de Geografía, Universidad
Nacional Autónoma de México, México.
Godínez-Alvarez, H., and A. Valiente-Banuet. 1998. Germination and early seedling growth of Tehuacán Valley
cacti species: the role of seed ingestion by dispersers and
soils on seedling growth. Journal of Arid Environments 39:
21-31.
Godínez-Alvarez, H., and A. Valiente-Banuet. 2000. Fruitfeeding behavior of Leptonycteris curasoae and Choeronycteris mexicana in flight cage experiments: consequences
for dispersal of columnar cactus seeds.Biotropica 32:552556.
Godínez-Alvarez, H., A. Valiente-Banuet, and L. Valiente
Banuet. 1999. Biotic interactions and the population dynamics of the long-lived columnar cactus Neobuxbaumia
tetetzo in the Tehuacán Valley, Mexico. Canadian Journal
of Botany 77:203-208.
Hendricks, P.,J. R. McAuliffe, and A. Valiente-Banuet. 1990.
On communal roosting and associated winter social behavior of gray-breasted woodpeckers. Condor 92:254-255.
HeITera, C. M., P. Jordano, L. López-Soria, and J. A. Amat.
1994. Recruitmentof a mast-fruiting,bird dispersedtree:
bridging frugivore activity and seedling establishment.
Ecological Monographs 64:315-344.
Horner, M. A., T. H. Fleming, and C. T. Sahley. 1998. Foraging behaviour and energetics of a nectar-feeding bat, Leptonycteris curasoae (Chiroptera: Phyllostomidae). Journal
of Zoology, London 244:575-586.
Howell, D. J. 1979. Flock foraging in nectar-feedingbats:
advantages to the bats and to the host plants. American
Naturalist 114:23-49.
Jordano, P. 1992. Fruits and frugivory. Pages 105-156 in M.
Fenner, editor. Seeds: the ecology of regeneration in plant
communities..CAB International, Wallingford, UK.
Jordano, P., and C. M. Herrera. 1995. Shuffling the offspring:
uncoupling and spatial discordance of multiple stages in
vertebrate seed dispersa!. Écoscience 2:230-237.
Jordano, P., and E. W. Schupp. 2000. Deterrninants of seed
dispersal effectiveness: the quantity component in the Prunus mahaleb-frugivorous bird interaction. Ecological
Monographs 70:591-615.
Loiselle, B. A., andJ. G. Blake. 1999. Dispersalofmelastome
seeds by fruit-eating birds of tropical forest understory.
Ecology 80:330-336.
Milton, S. J., W. R. J. Dean, G. 1. H. Kerley, M. T. Hoffman,
and W. G. Whitford. 1998. Dispersal of seeds as nest material by the cactus wren. Southwestern Naturalist 43:449452.
Osorio, B. O., A. Valiente-Banuet, P. Dávila. and R. Medina.
1996. Tipos de vegetacióny diversidad~ en el Valle de
Zapotitlán de las Salinas, Puebla, México. Boletín de la
Sociedad Botánica de México 59:35-36.
Poulin, B., G. Lefebvre, and R. McNeil. 1993. Variations in
bird abundance in tropical arid and semi-arid habitats. Ibis
135:432-441.
Raitt, R. J., and S. L. Pimm. 1976. Dynamicsof bird communities in the Chihuahuan Desert, New Mexico. Condor
78:427-442.
Reid, N. 1989. Dispersal oí mistletoes by honeyeaters and
2629
flowerpeckers: components of seed dispersal quality. Ecology 70: 137-145.
Rojas-Martínez, A., and A. Valiente-Banuet. 1996. Análisis
comparativo de la quiropterofauna del Valle de TehuacánCuicatlán, Puebla-Oaxaca. Acta Zoológica Mexicana 67:
1-23.
.
Rojas-Martínez, A., Valiente-Banuet, M. C. Arizmendi, A.
A1cántara-Eguren,
and H. Arita. 1999. Seasonaldistribution ofthe long-nosed bat (Leptonycteris curasoae) in North
America: does a generalized migration pattern really exist?
Journal of Biogeography 26: 1065-1077.
Schupp, E. W. 1993. Quantity, quality and the effectiveness
of seed dispersal by animals. Vegetatio 1071108:15-29.
Schupp, E. W., and M. Fuentes. 1995. Spatial patterns of
seed dispersal and the unification of plant population ecology. Écoscience 2:267-275.
.
Shreve, E 1931. Physical conditions in SUDand shade.Ecology 12:96-104.
Simons, L. S., and L. H. Simons. 1990. Experimental studies
of nest-destroying behavior by cactus wrens. Candor 92:
855-860.
Smith, C. E. 1965. Flora of Tehuacan Valley. Fieldiana Botany 31:107-143.
Steenbergh, W. H., and C. H. Lowe. 1969. Critical factors
during the first years of life of the saguaro (Cereus giganteus) at the Saguaro National Monument, Arizona. Ecology
50:825-834.
Steenbergh, W. H., and C. H. Lowe. 1977. Ecology of the
Saguaro. n. Reproduction, germination, establishment,
growth and survival of the young planto National Park Service Scientific Monograph Series 8. U.S. Government
Printing Office, Washington, D.C., USA.
Tumer, R. M., S. M. A1com, G. Olin, and J. A. Booth. 1966.
The influence of shade, soil and water on saguaro seedling
establishment. Botanical Gazette 127:95-102.
Valiente-Banuet, A., M. del C. Arizmendi, A. Rojas-Martínez,
and L. Domínguez-Canseco. 1996. Ecological relationships between columnar cacti and nectar feeding bats in
Mexico. Journal of Tropical Ecology 12:103-119.
Valiente-Banuet, A., O. Briones, A. Bolongaro-Crevenna, E.
Ezcurra, M. Rosas, H. Núñez, G. Barnard, and E. Vázquez.
1991. Spatial relationships between cacti and nurse shrubs
in a semi-arid environment in central Mexico. Journal of
Vegetation Science 2: 15-20.
Valiente-Banuet, A., and E. Ezcurra. 1991. Shade as a cause
oí the association between the cactusNeobuxbaumia tetetzo
and the nurse-p1antMimosa luisana in the TehuacánValley,
Mexico. Joumal oí Eco1ogy 79:961-971.
Valiente-Banuet, A., A. Rojas-Martínez, M. del C. Arizmendi,
and P. Dávila. 1997a. Pollination biology oí two co1umnar
cacti (Neobuxbaumia mezcalaensisand Neobuxbaumia macrocephala) in the TehuacánValley, central Mexico. American Joumal oí Botany 84:452-455.
Valiente-Banuet, A., A. Rojas-Martínez, A. Casas, M. del C.
Arizmendi, and P. Dávila. 1997b. Pollination bio1ogy oí
two winter-blooming giant columnar cacti in the Tehuacán
Valley, central Mexico. Joumal oí Arid Environments 37:
331-341.
van der Pijl, L. 1982. PrincipIes oí dispersal in higher plants.
Springer-Verlag, Berlin, Germany.
Wenny, D. G. 2000. Seeddispersal, seedpredation, and seedling recruitment oí a neotropical montane cree.Ecological
Monographs 70:331-351.
Wilkinson, L. 1993. SYSTAT 5.02 Cor MicrosoCtWindows.
SYSTAT, Evanston, Illinois, USA.