08_2005 Plant Ecol_tropical Dry Forests

Springer 2005

Plant Ecology (2005) 180:117–134 DOI 10.1007/s11258-005-3026-9

-1

Environmental correlates of tree and seedling–sapling distributions in a Mexican tropical dry forest Yalma Luisa Vargas-Rodriguez1,2,*, J. Antonio Va´zquez-Garcı´ a3 and G. Bruce Williamson1 1

Department of Biological Sciences, 107 Life Sciences Building, Louisiana State University, Baton Rouge, LA 70803, USA; 2Current address: Carlos Fuero 543, Colonia Universitaria, Guadalajara, 44840, Jalisco, Me´xico; 3Centro Universitario de Ciencias Biolo´gicas y Agropecuarias, Departamento de Bota´nica y Zoologı´a, Universidad de Guadalajara, Km 15 carretera Guadalajara-Nogales, Las Agujas, Zapopan, 45110, Jalisco, Me´xico; *Author for correspondence (e-mail: [email protected]) Received 28 June 2004; accepted in revised form 28 February 2005

Key words: Bray and Curtis ordination, Disturbance, Importance value, Regeneration, Specialization, Tree diversity, Tropical deciduous forest

Abstract Bray and Curtis ordination was used to explore which environmental variables explained importance values and the presence–absence of tropical tree seedlings, saplings and adults in La Escondida-La Caban˜a, Sierra de Manantla´n, Jalisco, Mexico. The diameters of trees ‡2.5 cm DBH and the presence and height of seedlings and saplings were measured in nine 0.1 ha sites. Four matrices including presence–absence data and importance value indices for trees and seedlings and saplings were analyzed through Bray and Curtis ordination. The matrices were based on density, frequency, and dominance of adult trees as well as seedlings and saplings. The environmental matrix consisted of 18 variables, including elevation, slope, canopy gaps, disturbance, and soil variables. We recorded 63 tree species and 38 seedling and sapling species in the nine sites. The ordination explained 70.9% of the variation in importance value data for trees and 62.6% for seedlings and saplings. The variation explained in presence–absence data for trees was 67.1 and 77.4% for seedlings and saplings. The variance in the ordination axes of seedlings and sapling presence–absence data was poorly explained by the number of gaps in the tree, shrub, or herb layer, suggesting little light specialization by seedlings and saplings. Habitat specialization for soil nutrients appears to be important in explaining the presence–absence of seedlings and saplings. Seedling and sapling specialization along different soil microsites could promote species coexistence in this forest, while heterogeneity in light conditions may instead determine differences in growth and, thus, importance value of trees. We hypothesize that in tropical dry forest in Jalisco, Mexico, a habitat specialization for soil resources is likely more important at early stages in tree life histories than in later life history.

Introduction In Mexico, 60% of the tropical vegetation is tropical dry forest (TDF). Mexican TDF is widely

distributed across the Pacific lowlands on the western side of the country, covering most of Jalisco state and confined to small patches in eastern Mexico (Rzedowski 1978; Trejo and Dirzo

118 2000). The forest is characterized by marked seasonality in rain fall and its occurrence on moderate steep slopes and rocky outcrops. Tropical dry forests in western Mexico, including those in the studied area, are typically deciduous, with short stature trees and high density of small size trees (Trejo 1998). Trees ca. 4 m tall constitute 65%, those with heights between 4 and 8 m are 31% and only 3% are 8–12 m height, some exceptional individuals reach 15 m. Tree mean density is 3610 (800) individuals/ha and basal area is ca. 56.8 m2/ha. Trees with diameter ‡10 cm represent only 20% of individuals and less than 5% are ‡30 cm, thus, the majority of trees have diameters £ 2.5 cm (Rzedowski 1978; Trejo 1998). Trees have extended crowns with bright and peeling barks, as well as compound leaves. Most of the tree species (ca. 85%) are deciduous and few (ca. 15%) are evergreen, such as Ziziphus mexicana Rose and Prosopis laevigata (Willd.) M.C. Johnst. In addition, columnar and Opuntioideae cacti are common in this community (Rzedowski 1978; Trejo 1998). TDF, throughout Mexico, is severely affected by livestock, slash and burn agricultural practices, forest fire, and selective logging. In 1990, only 27% of the original cover of tropical dry forest remained intact and 1.4% continues to be lost annually as a result of deforestation (Trejo and Dirzo 2000). Forest fires and livestock are also common in Sierra de Manantla´n Biosphere Reserve (SMBR). Fire in the area has been mainly associated with agricultural burning, occurring in the dry season (December–May). Livestock activity occurs during the rainy season (June–October). However, forest protection is enforced in core zones through various mechanisms negotiated with agrarian communities (68% of reserve’s area) and private landowners (32% of reserve’s area). In addition, there are management goals in the buffer zone directed to implement sustainable practices in forestry, agriculture, livestock, and other natural resource management activities (INE 2000). TDF in Mexico is characterized by high a and b diversity (Balvanera et al. 2002; Trejo and Dirzo 2002). While species diversity has been shown to be positively correlated with increasing precipitation across wet and dry forest in Puerto Rico (Murphy and Lugo 1986), a comprehensive analysis of Mexican TDF does not show such a relationship (Trejo and Dirzo 2002). Instead, it has

been suggested that Mexican TDF diversity is positively related to potential evapotranspiration (Trejo and Dirzo 2002). In addition, local site factors such as the presence and size of canopy gaps, soil properties, anthropogenic disturbance, and total annual precipitation may have important effects on TDF tree species richness, composition, abundance, and structure (Gonzalez and Zak 1996; Gentry 1988; Oliveira-Filho et al. 1998; Gillespie et al. 2000; Segura et al. 2003). For instance, variation in species diversity in the Neotropics could be explained by total annual precipitation (Gentry 1982, 1988). Associations of species with those factors and specialization to particular microhabitats are hypothesized to contribute to species diversity (Connell 1978; Gentry 1982; Denslow 1987; Welden et al. 1991; Clark et al. 1993). Different tree species are best suited to different habitats, which may lead to habitat specialization. Differential resource utilization might express itself as microhabitat specialization between the species or as differences in geographical distribution. On the other hand, the availability of different resources may be separate in time and may become available during different seasons. For example, canopy gaps provide different soil, light, and moisture conditions than the forest understory and could be exploited by species with specialized competitive abilities (Denslow 1987; Oliveira-Filho et al. 1998). Differences in species’ growth and survival may also occur in the absence of gaps along smaller light gradients, such as between 0.2 and 6.5% available diffuse light (Lieberman et al. 1989; Montgomery and Chazdon 2002). In addition, natural disturbances can cause spatial and temporal variability in the availability of resources leading to habitat differentiation. Consequently, variation in growth and survival of tree species under differing conditions of resource availability (for example micro-topography or light) could result in habitat specialization (Kobe 1999; Pearson et al. 2003). Previous studies have shown that microhabitat specialization with regard to topography and soil characteristics affect the distribution of several plant groups, including tropical trees, melastomate shrubs, herbs, pteridophytes, and palms (Kahn and De Castro 1985; Liberman et al. 1985; Poulsen and Baslev 1991; Basnet 1992;

119 Toumisto and Roukolainen 1993; Clark et al. 1998; Svenning 1999). For instance, distribution and abundance of some deciduous tree families (e.g., Leguminosae), are related to ecological gradients such as edaphic conditions and shade tolerance (Givnish 1999). Cation exchange level affects the distribution of 51 tree species in Ghana (Swaine 1996). Soil organic matter, soil nutrients, texture, and moisture are strongly correlated with main axes of a PCA ordination of tropical trees in a Bornean tropical rain forest (Webb and Peart 2000). In addition, adult trees and seedlings had different strengths of association with those variables (Webb and Peart 2000). Other species can be specialized along moisture gradients, reflected by the greater numbers of species that persist on protected side-slope areas due to their higher dry season moisture levels (Hubbell 1995). Even though previous studies used a quantitative multivariate approach to analyze how tree species relate to environmental variables, only a few studies have considered the relationship between the regeneration of tree species in seasonal tropical dry forest and environmental conditions (Lieberman and Li 1992; Oliveira-Filho et al. 1998), and none have explored the possible contribution of environmental variables at different stages in the life cycle of TDF trees. Here we report the results of a study of tree forest composition and regeneration along a short altitudinal gradient in a tropical dry forest at La Escondida-La Caban˜a, Sierra de Manantla´n Biosphere Reserve, in the Ayuquila watershed, in western Mexico, to generate hypotheses about tropical trees and regeneration patterns along soil and canopy gaps gradients. We addressed the following question: What environmental variables could explain major community gradients for both trees and seedlings and saplings? To answer this question, we employed a multivariate gradient analysis, Bray and Curtis ordination technique (also known as sociological ordination) (Curtis and McIntosh 1951; Beals 1984; McCune and Grace 2002), which produces pure community gradients that can be correlated with measured environmental variables. This approach has provided many insights into the nature, organization and dynamics of ecological communities (Whittaker 1956, 1960; Terborgh 1973; Ludwig and Reynolds 1988).

Methods Study area La Escondida-La Caban˜a gradient is located in the northern boundary of the SMBR, roughly 50 km from the Pacific Ocean, in Jalisco state, Western Mexico. The study area lies between El Aguacate (municipio El Grullo) and Zenzontla (municipio Tuxcacuesco), northwest of ejido Zenzontla, along the Ayuquila-Armerı´ a river, within the tributary watershed La Pasio´n-Cerro Blanco. The elevational gradient extends from Arroyo La Escondida at, 880 m to the top of Cerro la Caban˜a at, 1090 m a.s.l. (1942¢ N, 10407¢ W) (Figure 1). La Escondida-La Caban˜a occupies the steep foothills of the northern Sierra de Manantla´n, in the Sierra Madre del Sur, an area of rough topography, within a hilly and dissected landscape. Relief ranges from 10 to 66% slope inclination, along La Escondida and La Caban˜a hills. Microclimate at La Escondida-La Caban˜a is moist, warm, and highly seasonal with a wet season from June to October and a dry season from December to May (Martı´ nez et al. 1991). Mean annual rainfall is 900 mm (range 600–1000 mm) (Martı´ nez et al. 1991). Mean annual temperature is 22 C (range 22–28 C), free of frost. The temperature lapse rate is 4 C per 1000 m a.s.l. (Martı´ nez et al. 1991). TDF at La Escondida-La Caban˜a corresponds to the Lower Montane Subhumid Tropical Dry Forests of Holdridge (1967) and occupies 12,700 ha (9%) of the SMBR (Cuevas-G. et al. 1998). Tertiary volcanic rocks are prevalent in the study area (Cruz 1989). Acidic intrusive rocks (granites) were found on sites 1, 2, 3, 4, 8 and 9, while intermediate extrusive rocks (rhyolites, andesites, and trachytes) prevailed on sites 5, 6, and 7. Land form is convex with boulders and rocky outcrops. Shallow, well-drained lime soils (Regosol eutrico) were common in sites 5, 6 and 7. Some lime secondary soils such as Feozem ha´plico and some Litosols occured. Litosols (with rocks greater than 7.5 cm in diameter) were frequent at sites 1, 2, 3, 4, 8 and 9. Fluvial sandy soils (Fluvisol eutrico) were found on river banks and at the foothills of La Escondida (CETENAL 1975, 1976). Floristic checklists and descriptions of tropical dry forest species in the SMBR (Va´zquez-Garcı´ a

120

Figure 1. Location of La Escondida-La Caban˜a at Sierra de Manantla´n, Jalisco, Mexico and sampling sites (rectangles).

et al. 1995; Cuevas-G. et al. 1998) contain over 1060 species (38% of the total Sierra de Manantla´n vascular flora). In these checklists, 110 families and 531 genera have been recorded from tropical dry forest of SMBR. The vegetation is largely deciduous for up to 8 months, depending on the length of the dry season. A few species, such as Ziziphus mexicana and Prosopis laevigata are evergreen.

Field sampling The least disturbed forested areas in the Ayuquila watershed were selected through visual interpretation of topographic maps and field reconnaissance. We sampled nine 0.1 ha sites, each one consisting of ten 2 · 50 m transects, using a stratified random sampling design along elevation and exposure to represent variability of the forest.

Tree, shrub and herb gaps quantities represent the averages per site. Total counts of disturbance events and the average of rocks and stones per site are also shown. Soil nutrients units are ppm. Physiographical unit 1 corresponds to lower slope, 2 to middle slope, and 3 to upper slope. Topography (microrelief shape) is considered in two categories, 1 regular, without undulations and 2 corresponds to irregular, with undulations.

6.76 6.76 5.89 6.28 6.28 6.6 6.37 6.41 6.64 250 9.13 250 9.47 250 6.43 250 8.12 180 9.12 120 5.07 250 10.15 250 9.81 60 8.12 50 50 25 25 50 50 50 50 25 12 12 12 12 12 6 12 6 25 12 12 12 12 12 25 12 25 25 53.21 51.74 39.08 35.96 35.04 37.98 37.61 33.76 31.93 2 1 2 1 2 2 1 1 2 4 3 4 3 4 3 2 4 2 5 3 4 3 5 2 2 2 2 1 2 2 3 1 2 2 1 2 65 62 66 58 48 46 24 10 50 20 20 13 10 21 29 33 37 22 62 66 66 50 26 7 10 13 5 34 33 20 25 39 57 59 63 61 17 28 34 23 23 9 11 28 42 880 950 1020 1090 920 990 1060 930 1000 SE SE SE SE W W W E E SITE1 SITE2 SITE3 SITE4 SITE5 SITE6 SITE7 SITE8 SITE9

Tree Shrub Herb Ion exchange capacity Exposure Elevation gaps gaps gaps Disturbance Slope Physiography Rocks Stones Topography meq/100 g P

Table 1. Geographic, edaphic, topographic and disturbance variables used in ordinations.

NO3 Mg K

%Organic matter pH 1:2

121 The total area sampled was 0.9 ha (9 sites · 10 transects/site · 0.01 ha/transect). Sites were spaced at 70 m elevational intervals (Figure 1). This type of sampling allows comparability with the extensive Gentry’s forest data set (Gentry 1982; Phillips and Miller 2002). Four sites (1–4) had southeast exposure, starting near the Ayuquila river, from foot to top of La Escondida Mountain, two sites (8–9) were located at the shoulder of the same mountain, on its eastern slope, three others (5–7) were located on the west-facing slopes of a neighboring mountain, La Caban˜a. The two mountains spaced by a small ravine. Slopes inclination ranged from a 10 to 60% across the nine sites and all sites had a landform of concave shape. At each site, plant and environmental variables were measured. Diameters of all trees ‡2.5 cm diameter at breast height (DBH, at 1.30 m height) were measured. Heights of all seedlings (individuals <130 cm height) and saplings (DBH <2.5 cm and ‡130 cm height) were recorded in each of the nine sites (Olvera et al. 1996). Sampling was conducted from June to October of 1997 (when leaves were present). Species’ identifications were determined by the authors except for difficult taxa that were sent to specialists. Nomenclature generally follows Va´zquez-Garcı´ a et al. (1995). Herbarium specimens were deposited at IBUG herbaria of the University of Guadalajara (Holmgren et al. 1990; Holmgren and Holmgren 1993). Quantitative and qualitative environmental data were recorded for each transect (Table 1). Elevation was taken with an altimeter and slope was measured with a clinometer for each transect. Exposure was measured with a compass. Topography, physiographical unit, and the amount of rocks and stones were registered in each transect. Gap presence was determined by the line intersect method (Battles et al. 1996). Presence of tree, shrub, and herb gaps were recorded at 5-m intervals along each of the 0.01 ha transects, totaling 10 gap measurements per guild per transect. The numbers of tree, shrub, and herb gaps were summarized per site. For gap measurements, we considered as tree strata all woody plants >2.5 m tall with their crowns completely or partially exposed to light; shrub strata all woody plants <2.5 m tall; and herb strata all woody and non woody plants <50 cm tall. While conducting vegetation sampling, two soil samples, 0–30 cm in depth, were

122 taken from random locations at each site. This soil depth is aimed to obtain the best nutrient (N) estimates (Castellanos et al. 2000). Soil pH was measured with potentiometer. Nutrients (NO3, NH4, P, Mn, Mg, K, Ca) were analyzed with the Morgan method. Percent organic matter was measured by the Walkey–Black method, and texture was measured following the Bouyoucos method. Cation exchange capacity was obtained using ammonium acetate (Castellanos et al. 2000). At each site, observations were made on the occurrence of different kinds of natural and human-related disturbance, following Va´zquezGarcı´ a and Givnish (1998). These included the numbers of fallen or damaged trees, tree stumps, browsed or grazed plants, and intensity of grazing (evaluated by the amount of dung, cattle footprints, cattle droppings, and track livestock), as well as the presence absence of signs of past fires, erosion, and forest harvesting. The amount of disturbance was summarized, with a high value signifying high levels of disturbance and low value signifying low levels of disturbance (Va´zquezGarcı´ a and Givnish 1998).

Data analysis Data were summarized using four community matrices and one environmental matrix. Two community matrices consisted of tree importance values and tree presence–absence data for 63 species. Two other community matrices included importance values and presence–absence of seedlings and saplings for 38 species. Matrices with importance values were meant to relate communities in terms of their composition, structure, distribution, and age of the nine forest stands, while presence–absence matrices were meant to relate communities in terms of their composition and distribution, regardless of forest structure and age. Thus, allowing relating environmental variables to different community aspects. Importance values for trees were calculated from relative tree density, frequency, and dominance (basal area) (Curtis and McIntosh 1951; Cottam and Curtis 1956). We adapted importance values for seedlings and saplings by using height, instead of basal area, as a measure of dominance. The environmental matrix included quantitative data for elevation, slope, soil nutrients, soil pH,

soil organic matter, cation exchange capacity, tree gaps, shrub gaps, herb gaps, and disturbance and categorical data for exposure, topography, physiography, stones, and rocks. Bray and Curtis variance-regression ordination was used in connection with the Sørensen coefficient of similarity distance. This type of ordination gives a complete community structure regardless of its relationship to environmental variables and produces clear species patterns that reflect the environmental space the way the biotic community interprets it (Beals 1984; McCune and Grace 2002). Thus, Bray and Curtis, like most indirect (sociological) ordination techniques, has better appeal than direct (environmental) ordination techniques. Bray and Curtis can be applied to a matrix containing any distance measure, including nonEuclidean semi-metrics such as Sørensen (Bray and Curtis) distance. This is important since semimetrics, such as the Sørensen and Jaccard distances, are considered robust measures of ecological distance (Beals 1984; Faith et al. 1987). In contrast, reciprocal averaging (RA) and principal component analysis force a particular distance (Euclidean) measure on the analyst, which produces inadequate results (Beals 1973, 1984) and RA is effective for, but limited to, single dimensional data. Furthermore, papers that have compared Bray and Curtis with other methods provided evidence that Bray and Curtis may perform better than the others (Gauch et al. 1977; Robertson 1978; Gauch and Scrugs 1979; del Moral 1980; McCune 1994) and produces results identical to an ordination based on fuzzy set theory (Roberts 1986). Thus, Bray and Curtis is considered an effective ordination technique and perhaps its only serious rival is Multidimensional Scaling (Beals 1984; Causton 1988; Ludwig and Reynolds 1988; McCune and Beals 1993). Endpoints for ordination were selected by variance-regression, thereby reducing shortcomings of the original technique. The cutoff value used for the ordination biplot was r2=0.444, which results in a r-value that is significantly correlated with ordination axes. The relationship between tree, regeneration, and the environment was evaluated using Pearson correlations between the identified axes of the ordination and the environmental variables. P-values were not assigned because,

123 strictly speaking, the ordination scores are not independent of each other (McCune and Grace 2002). The software used was PCORD v4.0 (multivariate analyses for ecological data) (McCune and Mefford 1999).

Results A total of 63 tree species representing 51 genera and 27 families were recorded in 0.9 ha. Leguminosae had the largest number of species, with 14 species. Acacia was the most speciose genus with five species. A total of 853 individual trees were recorded in the studied area (0.9 ha), providing an estimate of 948 trees/ha. Importance values for trees are shown in Appendix A. All nine sites were considered separate in each analysis and were here summarized in one single Appendix. A total of 38 tree species in 30 genera and 18 families were recorded as seedlings and saplings. The total number of individual seedlings and saplings was 368, thus an estimated 409 individuals/ha. Importance values for seedlings and saplings are shown in Appendix B. All nine sites were considered separate in each analysis and were here summarized in one single Appendix. Bray and Curtis variance-regression ordination axes accounted for a substantial cumulative percentage of variance. For trees, the ordinations explained more of the variation in importance value data (70.9%) than presence–absence data (67.1%). The opposite was true, for seedlings and saplings, for which more variation in presence– absence (77.4%) was explained than for importance value (62.6%). Importance values of trees Sites 3 and 9 were selected as endpoints for axis 1, which extracted 32.3% of the original distance matrix (Figure 2). This axis was explained directly by number of shrub gaps and P concentration and inversely by herb gaps (Table 2). Fifteen of 63 species displayed a positive correlation with this axis and had greater importance values in areas with greater P concentration and more shrub gaps (Table 3). Ceiba aesculifolia (H.B.K.) Britt. and Baker and Phenax hirtus (Sw.) Wedd. displayed a negative correlation with this axis (Table 3).

Figure 2. Ordination diagram for axes 1 and 2 derived from Bray and Curtis ordination using sites (m), importance values of tree species, and environmental variables (vectors). Correlation values for each environmental variable with the two axes are reported in Table 2.

Sites 4 and 8 were selected as endpoints for axis 2, which extracted 20.9% of the original distance matrix. This axis was explained inversely by elevation (Figure 2, Table 2). Ten species displayed a positive correlation with this axis and had a greater importance value with decreasing elevation (Table 3). Sites 2 and 7 were selected as endpoints for axis 3 (not shown), which extracted 17.7% of the original distance matrix. None of the measured environmental variables were correlated with this axis. Heliocarpus terebinthinaceus (DC.) Hochr. and Lysiloma microphyllum Benth. displayed a positive correlation with this axis and decreased with increasing importance value of Phenax hirtus and Thouinia serrata Radlk. (Table 3).

Presence–absence of trees Sites 8 and 3 were selected as endpoints for axis 1, which extracted 28.6% of the original distance matrix (Figure 3). This axis was explained directly by slope and inversely by shrub gaps and disturbance (Table 2). Species that were present in sites with a greater slope and less shrub gaps and disturbance were Acacia riparia H.B.K. (r = 0.841), Celtis iguanea (Jacq.) Sarg. (r = 0.758),

124 Table 2. Pearson correlation (r) of environmental variables and Bray and Curtis ordination axes using importance values and presence–absence data of trees. Variable

Elevation Tree gaps Shrub gaps Herb gaps Disturbance Slope Ion exchange capacity P NO3 Mg K % Organic matter pH

Importance value data

Presence–absence-data

Axis 1

Axis 2

Axis 3

Axis 1

Axis 2

Axis 3

0.082 0.436 0.745 0.689 0.418 0.432 0.485 0.685 0.606 0.132 0.662 0.232 0.375

0.684 0.168 0.379 0.207 0.575 0.556 0.095 0.458 0.396 0.452 0.101 0.288 0.096

0.551 0.491 0.483 0.554 0.469 0.524 0.380 0.117 0.091 0.044 0.018 0.025 0.215

0.523 0.043 0.789 0.655 0.753 0.676 0.213 0.562 0.019 0.575 0.322 0.475 0.536

0.657 0.484 0.331 0.540 0.033 0.244 0.845 0.442 0.523 0.669 0.542 0.188 0.305

0.488 0.568 0.164 0.309 0.310 0.455 0.276 0.329 0.324 0.294 0.283 0.229 0.362

High correlations are in bold.

Table 3. Pearson correlation (r) of tree importance values and Bray and Curtis ordination axes. Axis 1 Acacia cochliacantha Acacia macracantha Acacia pennatula Adelia barbinervis Albizia tomentosa Casearia corymbosa Ceiba aesculifolia Enterolobium cyclocarpum Exostema mexicanum Ficus cotinifolia Ficus insipida Heliocarpus terebinthinaceus Jacaratia mexicana Lasiocarpus ferrugineus Lysiloma microphyllum Malpighia ovata Margaritaria nobilis Nopalea auberi Opuntia fuliginosa Phenax hirtus Pithecellobium acatlense Psidium guajava Senna mollisima Stemmadenia donnell-smithii Stemmadenia tomentosa var. palmeri Tabebuia chrysantha Thouinia serrata Zanthoxylum fagara

Axis 2

Axis 3

0.784 0.854 0.849 0.874 0.923 0.701 0.781 0.849 0.882 0.882 0.882 0.725 0.817 0.801 0.768 0.882 0.892 0.791 0.913 0.709 0.849 0.849 0.880

0.666

0.886 0.821 0.780 0.666 0.849

Only correlations greater than an absolute value of 0.666 are shown.

Conzattia multiflora (B.L. Rob.) Standl. (r = 0.735), Senna atomaria (L.) Irwin and Barneby (r = 0.726) and Spondias purpurea L. (r = 0.758). Species that were absent in sites with greater slope and less shrub gaps and disturbance were Casearia corymbosa H.B.K. (r = 0.695), Nopalea auberi (r= 0.714), Senna mollisima (r= 0.695), and Stemmadenia donnell-smithii (Rose) Woodson (r= 0.701). Sites 9 and 1 were selected as endpoints for axis 2, which extracted 23.2% of the original distance matrix. This axis was explained directly by ion exchange and Mg concentration (Figure 3, Table 2). Acacia macilenta Rose (r = 0.697), Acalypha cincta Muell. Arg. (r = 0.697), Aeschynomene amorphoides (S. Watson) Rose ex B.L. Rob. (r = 0.697), Hamelia jorullensis H.B.K. (r = 0.697), Hintonia latiflora (Sesse´ and Moc. ex DC.) Bullock (r = 0.691), Nopalea karwinskiana (Salm-Dyck) Schumann (r = 0.758), Pachycereus pecten-aboriginum (Engelm.) Britt. and Rose (r = 0.693), Pseudobombax ellipticum (S. Watson) Dugand (r = 0.697), and Triumfetta semitriloba Jacq. (r = 0.697) were present in sites with higher ion exchange and MgO; while Acacia cochliacantha Humb. and Bonpl. ex Willd. (r = 0.733), Heliocarpus terebinthinaceus (r= 0.697), and Iresine cassiniformis Schauer (r= 0.669) were absent in sites with higher ion exchange and concentration Mg. Sites 9 and 7 were selected as endpoints for axis 3, which extracted 15.3% of the original distance

125

Figure 3. Ordination diagram for axes 1 and 2 derived from Bray and Curtis ordination using sites (m), presence–absence data of tree species, and environmental variables (vectors). Correlation values for each environmental variable with the two axes are reported in Table 2.

Figure 4. Ordination diagram for axes 1 and 2 derived from Bray and Curtis ordination using sites (m), importance values of seedlings and saplings species, and environmental variables (vectors). Correlation values for each environmental variable with the two axes are reported in Table 4.

matrix. None of the measured variables explained this axis.

with this axis and its importance value was lower with lower values of Mg and K (Table 5). Sites 5 and 4 were selected as endpoints for axis 3, which extracted 11.1% of the original distance matrix (Figure 4). None of the measured variables explained this axis. Acacia riparia, Bursera simaruba (L.) Sarg., Cordia inermis (Mill.) I.M. Johnst., and Croton fragilis H.B.K. were positively correlated with this axis and their importance values increased with decreasing importance values of Bunchosia palmeri S. Watson, Comocladia engleriana Loes., Senna mollisima, and Tabebuia chrysantha (Jacq.) G. Nicolson (Table 5).

Importance value of seedlings and saplings Sites 8 and 3 were selected as endpoints for axis 1, which extracted 34.3% of the original distance matrix (Figure 4). This axis was explained inversely by shrub gaps (Table 4). Two species (Ceiba aesculifolia and Lysiloma microphyllum) displayed a positive correlation with this axis and had greater importance values in areas with lower number of shrub gaps. Seven species displayed a negative correlation with this axis (Table 5). Sites 7 and 9 were selected as endpoints for axis 2, which extracted 17.2% of the original distance matrix (Figure 4). This axis was explained directly by tree gaps and inversely by Mg and K concentration (Table 4). Six species displayed a positive correlation with this axis and had greater importance values in areas with greater number of tree gaps and lower Mg and K concentration (Table 5). Spondias purpurea displayed a negative correlation

Presence–absence of seedlings and saplings Sites 8 and 4 were selected as endpoints for axis 1, which extracted 36% of the original distance matrix. This axis was explained inversely by percent organic matter (Figure 5). Casearia corymbosa (r= 0.772), Nopalea auberi (r= 0.772), Senna mollisima (r= 0.690), Spondias purpurea (r= 0.721), and Tabebuia chrysantha (r = 0.690) were present in sites with the most organic matter.

126 Table 4. Pearson correlation (r) of environmental variables and Bray and Curtis ordination axes using importance values and presence–absence data for seedlings and saplings. Variable

Elevation Tree gaps Shrub gaps Herb gaps Disturbance Slope Ion exchange capacity P NO3 Mg K % Organic matter pH

Importance value data

Presence–absence-data

Axis 1

Axis 2

Axis 3

Axis 1

Axis 2

Axis 3

0.452 0.247 0.718 0.587 0.629 0.644 0.307 0.502 0.237 0.310 0.288 0.590 0.377

0.131 0.689 0.110 0.248 0.305 0.169 0.527 0.527 0.633 0.704 0.689 0.330 0.006

0.543 0.090 0.251 0.275 0.339 0.210 0.141 0.024 0.058 0.418 0.238 0.192 0.098

0.567 0.201 0.593 0.393 0.665 0.597 0.070 0.257 0.123 0.510 0.064 0.703 0.316

0.256 0.499 0.474 0.642 0.070 0.271 0.807 0.627 0.574 0.527 0.712 0.122 0.065

0.360 0.099 0.294 0.427 0.195 0.077 0.341 0.318 0.016 0.001 0.515 0.549 0.115

High correlations are in bold.

Table 5. Pearson correlation (r) of seedlings and saplings importance values and Bray and Curtis ordination axes. Axis 1 Acacia cochliacantha Acacia riparia Adelia barbinervis Bunchosia palmeri Bursera simaruba Casearia corymbosa Ceiba aesculifolia Comocladia engleriana Cordia inermis Croton fragilis Croton ciliato-glandulifera Lasiocarpus ferrugineus Lysiloma microphyllum Nopalea auberi Opuntia fuliginosa Senna mollisima Spondias purpurea Stemmadenia donnell-smithii Tabebuia chrysantha Zanthoxylum fagara

Axis 2

Axis 3

0.798 0.750 0.762

0.680 0.728 0.750

0.699 0.679 0.728 0.750 0.750 0.798 0.798 0.951 0.821 0.732 0.737

0.726 0.680 0.736

0.897 0.673 0.798

Only correlations greater than an absolute value of 0.666 are shown.

Sites 9 and 2 were selected as endpoints for axis 2, which extracted 27.9% of the original distance matrix. This axis was explained directly by ion exchange and K concentration (Figure 5). Ceiba aesculifolia (r = 0.795) and Lysiloma microphyllum (r = 0.730) were present in sites with high ion exchange and K. Acacia cochliacantha (r = 0.795), Adelia barbinervis Schlecht. and Cham.(r = 0.730),

Croton ciliato-glandulifera Ort. (r = 0.795), Lasiocarpus ferrugineus Gentry (r = 0.795), Opuntia fuliginosa Griff. (r = 0.730), Senna mollisima (r = 0.692), Stemmadenia tomentosa var. palmeri (Rose) Woodson (r = 0.730), Tabebuia chrysantha (r = 0.692), and Zanthoxylum fagara (L.) C. Sargent (r = 0.795) were present in sites with less ion exchange and K concentration.

Importance values of trees

Sites 6 and 1 were selected as endpoints for axis 3, which extracted 13.5% of the original distance matrix. None of the measured variables explained this axis.

Discussion A total of 51 of the 65 species were associated with environmental variables using Bray and Curtis ordination and Pearson correlation. Our data support the hypothesis that the importance values of trees and seedlings and saplings are related to the interaction between soil resources and light gaps because soils in light gaps are found to be nutrient rich, due to increased mineralization of organics under light conditions (Denslow et al. 1998). In addition, presence or absence of trees (i.e. adult stage) might be determined by disturbance and slope, while presence or absence of seedlings and saplings (i.e. early life stage) might be strongly determined by soil variables. Webb and Peart (2000), using an ordination technique also found physiographic and light associations in 21 of 45 species of tropical seedlings and trees.

Availability of P, shrub gaps, and herb gaps explained most of the variation in importance values of the overall tree community and, especially the distribution of 17 tree species. Growth differences in tropical trees are often found by increasing P-supply (Chapin 1980; Burslem et al. 1994). The greater importance value of legume, N-fixing, trees at La Escondida-La Caban˜a (40% of total number of trees) could result from high levels of P and low levels of N, which might enhance symbiosis, survivorship, and growth (Vitousek and Howarth 1991). Legumes trees, such as Acacia cochliacantha, Acacia macracantha, Acacia pennatula (Schlecht. and Cham.) Benth., and Enterolobium cyclocarpum (Jacq.) Griseb., behave as lightdemanding species with a strong response in biomass allocation and growth with increased P availability, as well as a higher phosphorous-use efficiency when P supply is low (Huante et al. 1995). In contrast, late successional (shade tolerant) species, such as Recchia mexicana Moc. and Sesse´, have relative low growth rates and show little or no response in growth to different P concentrations and less dependency on nutrient supply (Huante et al. 1995). The availability of P may not be limiting the overall development of TDF as shown in Chamela (Campo et al. 2001). However, in La Escondida-La Caban˜a, where slash and burn agriculture and pasturing of livestock are common land uses, P availability may be limiting tree species importance value since the effects of fires and other disturbances could increase the rate at which P is lost from the soil (Campo et al. 2001; Louette et al. 2001). Only Ceiba aesculifolia and Phenax hirtus importance values seem to be favored with decreasing P. Disturbance, together with low P, may often limit st49mpo1-559.827598.44mbo96261t001; Louette

128 importance values of 15 species in sites with more shrub gaps and P might be also related to high soil resource availability that is often found in gap sites (Denslow 1980, 1998). The fact that elevation explained secondary axis of importance values of the overall tree community suggests that the effect of short elevation gradients may not be an overriding factor on the organization of community at small scales, where other factors, such as nutrient supply and natural enemies become more important (Lieberman et al. 1985; Vargas-Rodriguez 1998; Va´zquez-Garcı´ a and Givnish 2000). Only rarely has elevation been the major variable explaining community organization along short elevational gradients (Lott et al. 1987). In addition, elevation showed no relationship to species richness along a larger elevational gradient at the El Tecolote ravine, western Sierra de Manantla´n (Cuevas-G. 2002). However, the influence of anthropogenic disturbance to this area, may confound the effects of elevation on the species composition of this community (Louette et al. 2001). In our study, evergreen species such as Ficus cotinifolia H.B.K. Ficus insipida Willd. Albizia tomentosa (Micheli) Standl. Exostema mexicanum A. Gray, and Margaritaria nobilis L. increased in importance value as elevation decreased. This may be attributed more to an increase in soil moisture in wetter sites, close to stream of lower areas. Changes in tree basal area, height, and stem diameter, factors that contribute to our estimation of importance value, were also found to be dependent on (Segura et al. 2003).

Presence–absence of trees Shrub gaps, slope and degree of disturbance were the primary environmental variables that explained the variation in the presence–absence of trees. Competition for light between subcanopy and canopy trees in different successional stages may be occurring, and contribute to species diversity. For example, canopy tree Lysiloma microphyllum formed a dense shade at one site, preventing the establishment of pioneer species. Consequently, gap dynamics and the interaction of canopy and subcanopy gaps, create heterogeneity in light distribution throughout TDF and can therefore influence species composition (Montgomey and Chazdon 2001; Quigley and

Platt 2003). Anthropogenic activity also plays a major role in the composition of La EscondidaLa Caban˜a TDF forest, in which the presences of 14 tree species are related to disturbance. The influence of disturbance on forest composition and structure also has been found in Central and South American TDF forest (Gonzalez and Zak 1996; Gillespie et al. 2000). Species correlations with gaps and disturbance are often found (Denslow 1987; Oliveira-Filho et al. 1998), but in contrast with the present data, soil variables also produced correlations, and explained important variation for axis 2. Patchy availability of nutrients in tropical dry forest confers special patterns in microbiological soil activity (Roy and Singh 1994). Higher amounts of organic C, N, and P are available at fine microsites scales and attract fine roots from surrounding areas to support tree growth (Roy and Singh 1994). Acacia spp. create islands of fertility with larger soil microbial biomass, C and N mineralization, and organic and total N (Reyes-Reyes et al. 2002). Therefore, the patchy availability of ion exchange and Mg may be determining the presence of 12 tree species in La Escondida-La Caban˜a forest, which are probably specialized to microhabitats with these soil characteristics (Burslem et al. 1995; Swaine 1996). In this study, species such as Celtis iguanea, Conzzatia multiflora, and Senna atomaria were associated with slopes. Studies of habitat associations in mesic to wet tropical forests also have found slope-specialists (Clark et al. 1998; Harms et al. 2000; Webb and Peart 2000). The spatial distribution of soil resources may result from disturbance history or differences in slope characteristics, since gaps tend be more abundant on steep slopes (Poorter et al. 1994). In addition, species associated to slopes might be responding to a gradient of soil characteristics such as water availability (Becker et al. 1988), nutrients (Botschek et al. 1996; Gonzalez and Zak 1996), or soil texture (Chauvel et al. 1987).

Importance value of seedlings and saplings Gaps in the shrub and tree layers create heterogeneity of light in the understory. Not surprisingly, importance values of seedlings and saplings were strongly related to the presence of gaps. In high

129 light conditions, tree seedlings achieved higher relative growth rates and net assimilation rates (Rincon and Huante 1993). Pioneer species, in particular, are light demanding and are more negatively affected by low light environments than shade-tolerant species (Rincon and Huante 1993). The pioneers Lysiloma microphyllum and Ceiba aesculifolia had lower importance values in areas with few gaps. These results suggest that the importance values of seedlings and saplings depend on light availability. Light limitation may reduce growth in the sapling stage, but does not contribute to increased mortality because susceptible seedlings will have already died. Seasonal openings due to tree canopy deciduousness provide temporal sites of high light, allowing different growth rates depending on the season (Rincon and Huante 1993). Therefore, seedlings and saplings of TDF species might be able to persist under a range of light environments, with little selection for light resource. In this sense, the classical theory of niche specialization does not apply in seedlings and sapling TDF populations (Welden et al. 1991; Brokaw and Busing 2000). Ordination axes for seedlings and saplings were explained by K and Mg concentrations, nutrients that play important roles in plant physiology. Mg is a key element in chlorophyll structure and K functions mainly as an osmoregulator and can affect cell size. Both elements are especially critical in the TDF community where they help prevent mortality resulting from drought stress. For instance, TDF seedlings respond to drought stress with an increase in chlorophyll concentration (Khurana and Singh 2001). In addition, K and Mg concentrations appear to be correlated with tree species richness in the tropics and are considered as limiting resources in Amazonian forests (Gentry 1988; Burslem et al. 1995; Marschner 1995). A response in growth with increasing Mg has been found in tropical trees (Burslem et al. 1995; Gunatilleke et al. 1997). Seasonal rainfall and its effects on the reduction of microbial activity during the dry season affect the availability of these elements (Campo et al. 1998), and consequently, the distribution and importance value of seedlings and saplings. Pioneer species were positively correlated with this axis which is consistent with the notion that pioneers are more dependent on nutrient availability than shade-tolerant species (Rincon and Huante 1994).

Presence–absence of seedlings and saplings The presence and distribution of tropical deciduous seedlings and saplings may be more influenced by habitat specialization for soil resources than light. The variance in axes from ordinations of our presence and absence data of seedlings and saplings was explained by soil variables and not by gaps. Soil organic matter affects acidity, soil moisture and nitrogen availability resulting in different gradients of those factors, and thereby determining species’ distributions. In addition, soil organic matter contributes to cation exchange capacity, which then controls K retention. Low densities of seedlings between 1–30 cm tall occur in the tropical dry forest of Western Mexico (VargasRodriguez 1998), suggesting that mortality is high in early life stages. This is also consistent with a higher mortality of seedlings in ‘‘suboptimal’’ habitats found in a Bornean rain forest (Webb and Peart 2000). Poor nutrient conditions (low levels of cation exchange capacity), or the inability of seedlings to form symbioses with mycorrhizas could explain limits to establishment rather than differences in light regime resulting from canopy gaps (Givnish 1999; Hubbell et al. 1999; Swaine 1996). Diversity in Mexican TDF may be maintained more by symmetric competition for soil resources and specialization along soil microsite gradients than by predators and pathogens related mortality (Harms et al. 2000).

Concluding remarks Different light conditions created by light gaps and seasonal canopy openings influence species differentiation among canopy and subcanopy trees, and differentiation in growth and importance value of tree populations. In contrast, seedlings and saplings appear mostly to be light generalists and are able to tolerate a wide range of light conditions (Brokaw and Busing 2000; Wright 2002). However, we hypothesize that habitat specialization for soil resources is likely more important in determining species success at early stages in life than in later stages in TDF at Jalisco, Mexico. Also, the pattern of niche differentiation of adult tropical trees observed along soil gradients (Clark et al. 1999; Svenning 1999, 2001; Webb and Peart 2000) appears to occur at the seedling stage.

130 Acknowledgements This project was financed by CONACYT (96-06-002 Project), IDEAWILD and CUCBA-University of Guadalajara. First author thanks E. Fabia´n Vera Torres, Pablo Carrillo, Margarita Ayo´n, Eduardo Herna´ndez, Celso Corte´s, Etelberto Ortiz, and Ana Paula Reyes for their help with field work. Francisco Santana Michel, Luis

Guzma´n and herbaria IBUG staff helped with species identifications. Saara Dewalt, Jennifer Cramer, Heather Passmore, Blanca Figueroa Rangel, and La´zaro Sa´nchez Vela´zquez made important suggestions to this manuscript. Special thanks Saara DeWalt for reviewing drafts of this paper. BIOL 7093 course at Louisiana State University provided important ideas to this research.

Appendix A. Density, frequency, basal area, and importance value for tree species within the 0.9 ha study area in La Escondida-La Caban˜a, Jalisco, Mexico. BA (dm2) % Freq. Density Tree/ha BA dm2/ha Sites found in Leguminosae Leguminosae Leguminosae Leguminosae Leguminosae Euphorbiaceae Euphorbiaceae Leguminosae Leguminosae Malpighiaceae Burseraceae Burseraceae Burseraceae Burseraceae Flacourtiaceae Bombacaceae Ulmaceae Cactaceae Cochlospermaceae Rhamnaceae Anacardiaceae Leguminosae Leguminosae Capparaceae Euphorbiaceae Leguminosae Rubiaceae Moraceae Moraceae Sterculiaceae Rubiaceae Tiliaceae Rubiaceae Amaranthaceae Caricaceae Malpighiaceae Leguminosae Malpighiaceae Euphorbiaceae Cactaceae Cactaceae Cactaceae Cactaceae Urticaceae

Acacia cochliacantha Acacia macilenta Acacia macracantha Acacia pennatula Acacia riparia Acalypha cincta Adelia barbinervis Aeschynomene amorphoides Albizia tomentosa Bunchosia palmeri Bursera fagaroides Bursera grandifolia Bursera kerberi Bursera simaruba Casearia corymbosa Ceiba aesculifolia Celtis iguanaza Cephalocereus alensis Cochlospermum vitifolium Colubrina triflora Comocladia engleriana Conzattia multiflora Coursetia glandulosa Crateva palmeri Croton fragilis Enterolobium cyclocarpum Exostema mexicanum Ficus cotinifolia Ficus insipida Guazuma ulmifolia Hamelia xorullensis Heliocarpus terebinthinaceus Hintonia latiflora Iresine cassiniiformis Jacaratia mexicana Lasiocarpus ferrugineus Lysiloma microphyllum Malpighia ovata Margaritaria nobilis Nopalea Aubert Nopalea karwinskiana Opuntia fuliginosa Pachycereus pecten-aboriginum Phenax hirtus

6.14 3.07 3.97 0.41 5.31 0.13 2.64 0.06 30.06 13.44 6.48 32.94 2.32 19.29 3.92 197.69 1.79 4.42 0.14 1.27 4.95 1.54 6.95 0.08 0.06 10.43 11.89 8.66 22.80 11.38 0.54 40.80 3.04 0.52 46.08 43.87 538.22 0.09 130.38 3.61 0.05 0.38 124.47 144.94

2.22 1.11 4.44 1.11 6.67 1.11 2.22 1.11 15.56 5.56 5.56 14.44 1.11 4.44 7.78 40.00 6.67 2.22 2.22 3.33 3.33 3.33 3.33 1.11 1.11 1.11 2.22 2.22 2.22 4.44 1.11 23.33 2.22 4.44 16.67 13.33 74.44 1.11 3.33 10.00 1.11 1.11 28.89 55.56

5.56 2.22 7.78 1.11 7.78 1.11 2.22 1.11 23.33 7.78 5.56 17.78 1.11 5.56 17.78 51.11 6.67 2.22 2.22 3.33 3.33 3.33 3.33 1.11 1.11 1.11 2.22 3.33 2.22 5.56 1.11 31.11 2.22 4.44 17.78 27.78 341.11 2.22 6.67 10.00 1.11 1.11 43.33 101.11

6.82 3.42 4.41 0.46 5.90 0.15 2.93 0.07 33.40 14.93 7.20 36.61 2.58 21.44 4.35 219.66 1.99 4.91 0.16 1.41 5.50 1.71 7.72 0.09 0.07 11.59 13.22 9.63 25.33 12.64 0.60 45.33 3.38 0.57 51.20 48.75 598.02 0.10 144.87 4.02 0.06 0.43 138.30 161.04

I.V.

4,9 0.45 1 0.20 8,9 0.64 9 0.12 3,4 0.81 1 0.11 8,9 0.27 1 0.11 1,5,6,8 2.37 2,5 0.87 3,4,5,6,9 0.67 1,2,5,7 2.15 5 0.15 4 0.82 5,7,8,9 1.21 1,2,3,4,5,6,7,8,9 7.88 2,3,4 0.71 3,5 0.30 1,2,9 0.23 4,5,8 0.36 5 0.42 1,3,4 0.36 1,3 0.46 5 0.11 4 0.11 9 0.29 8 0.43 8 0.41 8 0.62 4,5,7 0.69 1 0.12 2,3,4,5,6,7,8,9 3.34 1,2,6 0.28 3,4,9 0.46 1,2,4,5,7,8,9 2.53 2,4,9 2.62 1,2,3,4,5,6,7,8,9 26.27 8 0.15 5,8 2.72 5,6,7,8,9 1.07 1,2 0.11 8,9 0.12 1,2,3,5,6,8 5.60 1,2,3,4,5,6,7,8 9.74

131 Appendix A. Continued. BA (dm2) % Freq. Density Tree/ha BA dm2/ha Sites found in Nyctaginaceae Leguminosae Bombacaceae Myrtaceae Simaroubaceae Leguminosae Leguminosae Sapotaceae Anacardiaceae Apocynaceae Apocynaceae Cactaceae Bignoniaceae Burseraceae Sapindaceae Sapindaceae Tiliaceae Rutaceae Rutaceae

Pisonia aculeata var. aculeata Pithecellobium acatlense Pseudobombax ellipticum Psidium guajava Recchia mexicana Senna atomaria Senna mollisima Sideroxylon capiri subsp. tempisque Spondias purpurea Stemmadenia donnell-smithii Stemmadenia tomentosa var. Palmeri Stenocereus queretaroensis Tabebuia chrysantha Terebinthus acuminata Thouinia serrata Thouinidium decandrum Triumfetta semitriloba Zanthoxylum caribaeum Zanthoxylum fagara Totals

2.33 1.06 12.83 0.06 0.56 1.54 0.04 60.30 165.23 4.89 8.53 4.65 10.34 126.25 16.59 5.62 0.38 2.41 0.12

5.56 1.11 1.11 1.11 1.11 1.11 1.11 3.33 43.33 13.33 7.78 3.33 11.11 3.33 4.44 4.44 1.11 6.67 1.11

1914.98

5.56 1.11 1.11 1.11 1.11 1.11 1.11 3.33 66.67 20.00 13.33 3.33 11.11 4.44 5.56 5.56 1.11 8.89 1.11

2.59 1.18 14.26 0.06 0.62 1.71 0.04 67.00 183.59 5.44 9.47 5.16 11.49 140.27 18.43 6.24 0.43 2.68 0.14

947.78

2127.76

I.V.

4,5 9 1 9 5 3 5,6,8,9 2,5 1,2,3,4,5,6,7,8,9 1,5,6,8,9 8,9 2,4,8 1,2,4,5,6,8,9 5 1,2,5,8 2,5,6,8 1 4,5,6,7 9

0.60 0.13 0.34 0.11 0.12 0.14 0.11 1.39 8.08 1.67 1.13 0.42 1.30 2.57 0.78 0.59 0.12 0.79 0.11

All nine sites were considered separate in each analysis and were here summarized in one single Appendix.

Appendix B. Density, frequency, mean height and importance value for seedlings and saplings species within the 0.9 ha in La Escondida-La Caban˜a, Jalisco, Mexico. % Freq. Leguminosae Leguminosae Leguminosae Euphorbiaceae Leguminosae Leguminosae Malpighiaceae Burseraceae Burseraceae Burseraceae Flacourtiaceae Bombacaceae Cactaceae Cochlospermaceae Anacardiaceae Boraginaceae Leguminosae Euphorbiaceae Euphorbiaceae Tiliaceae Rubiaceae Caricaceae Malpighiaceae Leguminosae Euphorbiaceae Cactaceae Cactaceae

Acacia cochliacantha Acacia macracantha Acacia riparia Adelia barbinervis Aeschynomene amorphoides Albizia tomentosa Bunchosia palmeri Bursera fagaroides Bursera grandifolia Bursera simaruba Casearia corymbosa Ceiba aesculifolia Cephalocereus alensis Cochlospermum vitifolium Comocladia engleriana Cordia inermes Coursetia glandulosa Croton fragilis Croton ciliato-glandulifera Heliocarpus terebinthinaceus Hintonia latiflora Jacaratia mexicana Lasiocarpus ferrugineus Lysiloma microphyllum Margaritaria nobilis Nopalea Aubert Nopalea karwinskiana

3.33 1.11 1.11 8.89 1.11 2.22 2.22 4.44 17.78 1.11 21.11 27.78 4.44 2.22 1.11 2.22 2.22 1.11 6.67 6.67 1.11 4.44 6.67 28.89 1.11 11.11 2.22

Density Trees/ha 13.33 1.11 1.11 12.22 5.56 2.22 2.22 4.44 33.33 1.11 41.11 37.78 5.56 2.22 1.11 2.22 5.56 4.44 11.11 6.67 1.11 4.44 8.89 53.33 1.11 28.89 2.22

Mean height

Sites found in

I.V.

0.53 3.00 1.66 0.44 0.10 0.85 1.29 0.45 0.09 0.34 0.22 0.14 0.15 0.42 0.62 1.24 0.30 0.96 0.38 0.54 1.54 0.66 0.51 0.09 2.50 0.19 0.59

9 8 4 8,9 1 1,5 5 4,6,9 1,2,7,8 4 5,7,8,9 1,2,3,4,5,6,7,8 3,5 1,2 5 4 3 4 9 2,6,9 1 1,4,9 9 1,2,3,4,5,6,7 8 5,7,8,9 1,2

2.24 4.28 2.47 2.76 0.73 1.62 2.21 1.55 5.18 0.70 6.42 6.92 1.24 1.04 1.07 2.15 1.15 1.80 2.29 2.15 2.31 1.83 2.28 8.27 3.61 4.08 1.27

132 Appendix B. Continued. % Freq. Cactaceae Cactaceae Urticaceae Leguminosae Anacardiaceae Apocynaceae Apocynaceae Bignoniaceae Sapindaceae Rutaceae Rutaceae

Opuntia fuliginosa Pachycereus pecten-aboriginum Phenax hirtus Senna mollisima Spondias purpurea Stemmadenia donnell-smithii Stemmadenia tomentosa var. Palmeri Tabebuia chrysantha Thouinidium decandrum Zanthoxylum caribaeum Zanthoxylum fagara

3.33 3.33 5.56 7.78 15.56 25.56 2.22 3.33 2.22 7.78 2.22

Totals

Density Trees/ha 3.33 3.33 6.67 10 23.33 45.56 2.22 3.33 3.33 10 3.33

Mean height

Sites found in

I.V.

0.43 0.33 0.51 0.36 0.24 0.18 0.90 0.55 0.49 0.20 0.63

8,9 2,5,8 1,7,8 5,8,9 1,2,3,4,5,7,8 1,5,8 8,9 5,8,9 2,8 5,6,7 9

1.29 1.15 1.96 2.32 4.47 7.32 1.68 1.45 1.22 2.11 1.41

408.89

All nine sites were considered separate in each analysis and were here summarized in one single Appendix.

References Balvanera P., Lott E., Segura G., Siebe C. and Islas A. 2002. Patterns of b-diversity in a Mexican tropical dry forest. J. Veg. Sci. 13: 145–158. Basnet K. 1992. Effect of topography on the pattern of trees in Tabonuco (Dacryodes-Excelsa) dominated rain-forest of Puerto-Rico. Biotropica 24: 31–42. Battles J.J., Dushoff J.G. and Fahey T.J. 1996. Line intersect sampling of forest canopy gaps. Forest Sci. 42: 131–138. Beals E.W. 1973. Ordination: mathematical elegance and ecological naivete. J. Ecol. 61: 23–35. Beals E.W. 1984. Bray-Curtis Ordination: an effective strategy for analysis of multivariate ecological data. Adv. Ecol. Res. 14: 1–55. Becker P., Rabenold P.E., Idol J.R. and Smith A.P. 1988. Water potential gradients for gaps and slopes in a Panamanian tropical moist forest’s dry season. J. Trop. Ecol. 4: 173– 184. Botschek J., Ferraz J., Jahnel M. and Skowronek A. 1996. Soil chemical properties of a toposequence under primary forest in the Itacoatiara vicinity (Amazonas, Brazil). Geoderma 72: 119–132. Brokaw N. and Busing R.T. 2000. Niche versus chance and tree diversity in forest gaps. Tree 15: 183–188. Burslem D.F.R.P., Turner I.M. and Grubb P.J. 1994. Mineral nutrient status of coastal hill dipterocarp forest and adinandra belukar in Singapore: bioassays of nutrient limitation. J. Trop. Ecol. 10: 579–600. Burslem D.F.R.P., Turner I.M. and Grubb P.J. 1995. Responses to nutrient addition among shade-tolerant tree seedlings of lowland rain forest in Singapore. J. Ecol. 83: 113–122. Campo J., Jaramillo V.J. and Mass J.M. 1998. Pulses of soil phosphorus availability in a Mexican tropical dry forest: effects of seasonality and level of wetting. Oecologia 115: 167– 172. Campo J., Mass M., Jaramillo V.J., Martinez-Yrizar A. and Sarukhan J. 2001. Phosphorus cycling in a Mexican tropical dry forest ecosystem. Biogeochemistry 53: 161–179.

Castellanos J.Z., Uvalle-Bueno J.X. and Aguilar-Santelises A. 2000. Memoria del curso sobre interpretacio´n de ana´lisis de suelos, aguas agrı´ colas y plantas. Instituto Nacional de Investigaciones Forestales, Agrı´ colas y Pecuarias, Me´xico. Causton D.R. 1988. An Introduction to Vegetation Analysis: Principles, Practice and Interpretation. London, UK. CETENAL 1975. Carta geolo´gica, escala 1:50 000. Instituto Nacional de Estadı´ stica, Geografı´ a e Informa´tica, Me´xico. CETENAL 1976. Carta edafolo´gica, escala 1:50 000. Instituto Nacional de Estadı´ stica, Geografı´ a e Informa´tica, Me´xico. Chapin F.S.III 1980. The mineral nutrition of wild plants. Annl. Rev. Ecol. Sys. 11: 233–260. Chauvel A., Lucas Y. and Boulet R. 1987. On the genesis of the soil mantle of the region of Manaus, Central Amazonia, Brazil. Experientia 43: 234–241. Clark D.B., Clark D.A. and Rich P.M. 1993. Comparative analysis of mirohabitat utilization by saplings of nine tree species in neotropical rainforest. Biotropica 25: 397–407. Clark D.B., Clark D.A. and Read J.M. 1998. Edaphic variation and the mesoscale distribution of tree species in a neotropical rain forest. J. Ecol. 86: 101–112. Clark D.B., Palmer M.W. and Clark D.A. 1999. Edaphic factors and the landscape-scale distributions of tropical rain forest trees. Ecology 80: 101–112. Connell J. 1978. Diversity in tropical rain forest and coral reefs. Science 299: 1302–1310. Cottam G. and Curtis J.T. 1956. The use of distance measures in phytosociological sampling. Ecology 37: 451–460. Cruz P. 1989. Atlas cartogra´fico de la reserva de la biosfera Sierra de Manantla´n, Jalisco, Me´xico. Laboratorio Natural Las Joyas, Universidad de Guadalajara, Me´xico. Cuevas G.R. 2002. Ana´lisis de gradientes de la vegetacio´n de la can˜ada El Tecolote, en la Sierra de Manantla´n, Jalisco, Me´xico. PhD. Dissertation, Colegio de Postgraduados, Montecillo, Me´xico. Cuevas G.R., Nun˜ez L.N., Guzman H.L. and Santana M.F. 1998. El bosque tropical caducifolio en la reserva de la biosfera Sierra de Manantla´n, Jalisco-Colima, Me´xico. Boletı´ n del Instituto de Bota´nica de la Universidad de Guadalajara 5: 445–491.

Curtis J.T. and McIntosh R.P. 1951. An upland forest continuum in the prairie forest border region of Wisconsin. Ecology 32: 476–496. del Moral R. 1980. On selecting indirect ordination methods. Vegetatio 42: 75–84. Denslow J.S. 1980. Gap partitioning among tropical rainforest trees. Biotropica 12(suppl.): 47–55. Denslow J.S. 1987. Tropical rain forest gaps and tree species diversity. Annl Rev. Ecol. Sys. 18: 431–451. Denslow J.S., Ellison A.M. and Sanford R.E. 1998. Tree fall gap size effects on above- and below-ground processes in a tropical wet forest. J. Ecol. 86: 597–609. Faith D.P., Minchin P.R. and Belbin L. 1987. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69: 57–68. Gauch H.G.Jr. and Scruggs W.M. 1979. Variants of polar ordination. Vegetatio 40: 147–153. Gauch H.G.Jr., Whittaker R.H. and Wentworth T.R. 1977. A comparative study of reciprocal averaging and other ordination techniques. J. Ecol. 65: 157–174. Gentry A.H. 1982. Patterns of neotropical plant species diversity. Evol. Biol. 15: 1–84. Gentry A.H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annl. Missouri Bot. Garden 75: 1–34. Gillespie T.W., Grijalva A. and Farris C.N. 2000. Diversity, composition and structure of tropical dry forest in Central America. Plant Ecol. 147: 37–47. Givnish T.J. 1999. On the causes of gradients in tropical tree diversity. J. Ecol. 87: 193–210. Gonzalez O.J. and Zak D.R. 1996. Tropical dry forest of St. Lucia, West Indies: Vegetation and soil properties. Biotropica 28: 618–626. Gunatilleke C.V.S., Gunatilleke I.A.U.N., Perera G.A.D., Burslem D.F.R.P., Ashton P.M.S. and Ashton P.S. 1997. Responses to nutrient addition among seedlings of eight closely related species of Shorea in Sri Lanka. J. Ecol. 85: 301–311.

134 Murphy P. and Lugo A. 1986. Structure and biomass of a subtropical dry forest in Puerto Rico. Biotropica 2: 89–96. Oliveira-Filho A.T., Curi N., Vilela E.A. and Carvalho D.A. 1998. Effects of canopy gaps, topography, and soils on the distribution of woody species in a central Brazilian deciduous dry forest. Biotropica 30: 362–375. Olvera M., Moreno S. and Figueroa B. 1996. Sitios Permanentes para la investigacio´n silvı´ cola, manual para su establecimiento. Universidad de Guadalajara, Me´xico. Pearson T.R., Burslem D.F.R.P., Goeriz R.E. and Dalling J.W. 2003. Interactions of gap size and herbivory on establishment, growth and survival of three species of neotropical pioneer trees. J. Ecol. 91: 785–796. Phillips O. and Miller J.S. 2002. Global patterns of plant diversity: Alwyn H. Gentry’s forest transect data set. Missouri Botanical Garden Press, Missouri, US, pp. 319. Poorter L., Jans L., Bongers F. and van Rompaey R.S.A.R. 1994. Spatial distribution of gaps along three catenas in the moist forest of Tai National Park, Ivory Coast. J. Trop. Ecol. 10: 385–398. Poulsen A.D. and Baslev H. 1991. Abundance and cover of ground herbs in an Amazonian rainforest. J. Veg. Sci. 2: 315– 322. Quigley M.F. and Platt W.J. 2003. Composition and structure of seasonally deciduous forest in the Americas. Ecol. Monogr. 73: 87–106. Reyes-Reyes G., Baron-Ocampo L., Cuali-Alvarez I., FriasHernandez J.T., Olalde-Portugal V., Fregoso L.V. and Dendooven L. 2002. C and N dynamics in soil from the central highlands of Me´xico as affected by mesquite (Prosopis spp.) and huizache (Acacia tortuoso): a laboratory investigation. Appl. Soil Ecol. 19: 27–34. Rincon E. and Huante P. 1993. Growth responses of tropical deciduous tree seedlings to contrasting light conditions. Trees 7: 202–207. Rincon E. and Huante P. 1994. Influence of mineral nutrient availability on growth of tree seedlings from the tropical deciduous forest. Trees 9: 93–97. Roberts D.W. 1986. Ordination on the basis of fuzzy set theory. Vegetatio 66: 123–131. Robertson P.A. 1978. Comparison of techniques for ordinating and classifying old-growth floodplain forest in southern Illinois. Vegetatio 37: 43–51. Roy S. and Singh J.S. 1994. Consequences of habitat heterogeneity for availability of nutrients in a dry tropical forest. J. Ecol. 82: 503–509. Rzedowski J. 1978. Vegetacio´n de Me´xico. Editorial Limusa, Me´xico. Segura G., Balvanera P., Dura´n E. and Pe´rez A. 2003. Tree community structure and stem mortality along a water availability gradient in a Mexican tropical dry forest. Plant Ecol. 169: 259–271.

Svenning J.C. 1999. Microhabitat specialization in a species-rich palm community in Amazonian Ecuador. J. Ecol. 87: 55–65. Svenning J.C. 2001. Environmental heterogeneity, recruitment limitation and the mesoscale distribution of palms in a tropical montane rain forest (Maquipucuna, Ecuador). J. Trop. Ecol. 17: 97–113. Swaine M.D. 1996. Rainfall and soil fertility as factors limiting forest species distributions in Ghana. J. Ecol. 84: 419–428. Terborgh J. 1973. On the notion of favorableness in plant ecology. Am. Nat. 107: 481–501. Trejo I. 1998. Distribucio´n y diversidad de las selvas bajas de Me´xico: relaciones con el clima y suelo. PhD. Dissertation, Universidad Nacional Auto´noma de Me´xico, Me´xico, DF. Trejo I. and Dirzo R. 2000. Deforestation of seasonally dry tropical forest: a national and local analysis in Mexico. Biol. Conserv. 94: 133–142. Trejo I. and Dirzo R. 2002. Floristic diversity of Mexican seasonally dry tropical forests. Biodiv. Conserv. 11: 2063– 2048. Tuomisto H. and Ruokolainen K. 1993. Distribution of Pteridophyta and Melastomataceae along an edaphic gradient in an Amazonian rain-forest. J. Veg. Sci. 5: 25–34. Vargas-Rodriguez Y.L. 1998.Odenacio´n sociolo´gica de la comunidad arbo´rea del bosque tropical caducifolio en El Aguacate-Zenzontla, Sierra de Manantla´n, Jalisco. Bc.S. thesis, Centro Universitario de Ciencias Biolo´gico Agropecuarias, Universidad de Guadalajara, Guadalajara. Va´zquez-Garcı´ a J.A., Cuevas G. R., Cochrane T.S., Iltis H.H., Santana M.F.J. and Guzman H.L. 1995. Flora de Manantla´n. Sida, Botanical Miscellany 13. Botanical Research Institute of Texas, USA. Va´zquez-Garcı´ a J.A. and Givnish T.J. 1998. Altitudinal gradients in tropical forest composition, structure, and diversity in the Sierra de Manantla´n. J. Ecol. 86: 999–1020. Va´zquez-Garcı´ a J.A. and Givnish T.J. 2000. Vegetation of the cerro Grande massif, Sierra de Manantlan, Mexico: ordination of a long altitudinal gradient with high species turnover. Boletı´ n del Instituto de Bota´nica 6: 227–250. Vitousek P.M and Howarth R.W. 1991. Nitrogen limitation on land and in the sea: how can it occur? Biogeochemistry 13: 87–115. Webb C.O. and Peart D.R. 2000. Habitat associations of trees and seedlings in a Bornean rain forest. J. Ecol. 88: 464–478. Welden C.W., Hewett S.W., Hubbell S.P. and Foster R.B. 1991. Sapling survival, growth, and recruitment: relationship to canopy eight in a neotropical forest. Ecology 72: 35–50. Whittaker R.H. 1956. Vegetation of the Great Smoky Mountains. Ecol. Monogr. 26: 1–80. Whittaker R.H. 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecol. Monogr. 30: 279–338. Wright S.J. 2002. Plant diversity in tropical forest: a review of mechanisms of species coexistence. Oecologia 130: 1–14.