Contribution of Nontrees to Species Richness of a Tropical Rain Forest

Abstract

We report the results of the first complete samples of all plant species and individuals for any lowland tropical forest in the world. The three forests sampled are in western Ecuador; Rio Palenque, Jauneche, and Capeira are, respectively, wet, moist, and dry forests. In each forest we sampled all vascular plants in a 0.1-ha area. At wet forest Rio Palenque, nontree habit groups make up most of the sampled species and individuals. Over a third of the species and almost half the individual plants are epiphytes, 13 percent of the species are terrestrial herbs, 10 percent are shrubs, and 9 percent nonepiphytic climbers. The moist and dry forest samples have many fewer species, largely due to many fewer epiphytes. The new data are compared with the most diverse 0.1-ha samples from elsewhere in the world. Our wet forest sample is by far the most species-rich such sample yet recorded and would remain so even if all tree species were excluded from the data. ALTHOUGH IT IS WIDELY ACCEPTED that tropical rain forests are the world's most species-rich plant communities (e.g., Cain & Castro 1959, Walter 1971), the data base for this assertion is weak and almost entirely restricted to trees. Recently it has been suggested that certain nontropical plant communities, rather than tropical rain forests, may be the most species-rich vegetation types in the world (Richards 1969, Naveh & Whittaker 1979), at least for small sample plots (Parsons & Cameron 1974). Others have suggested that various types of nonforest vegetation are the most diverse in the world, at least for plants of particular habits, e.g., shrubs or herbs. Such claims have been made for the fynbos of the South African Cape region (Richards 1969, Taylor 1978), the Australian sclerophyllous forests (Richards 1969, Milewski 1983; see also Rice & Westoby 1983, Lamont et al. 1977), and the Mediterranean shrublands and maquis (Naveh & Whittaker 1979). The cerrado of Brazil, also sclerophyllous, typically has more species per 0.1 ha (or 1 ha) than any vegetation of Mediterranean-type climates (Eiten 1978, pers. comm.), but the species-rich types of cerrado include many trees (usually 30-60 species/ha) and may be better considered as a distinctive type of open-canopied tropical woodland (Eiten 1972, 1978). The highest recorded plant species diversities for samples of ?0.1 ha in each of these vegetation types are recorded in Table 1 and Figure 1. Direct comparisons of the species richness of different plant communities are greatly complicated by problems of scale (Rice & Westoby 1983). One reason for the lack of comparisons of plant diversity between tropical and other vegetations is that nearly all the available data on tropical forest diversity are restricted to trees. Sample areas Received 23 September 1985, revision accepted 30 December 1985. of 1 ha are needed to adequately measure tree species diversity. Since few individual trees are included in small plots, the great richness of tropical forest tree species is not apparent in samples comparable to the 0.1-ha ones commonly used to census other growth forms (e.g., Naveh & Whittaker 1979; Cowling 1983; Rice & Westoby 1983; Shmida, in press and included references). Even when only trees are considered, 1-ha samples of tropical forest may have as many as 300 species of 10 cm DBH in Amazonia (Gentry 1987), 227 in Malesia (Cousens 1951, Whitmore 1975), and 223 in Borneo (Proctor et ad. 1983). Other tropical forests on each continent have many fewer tree species, often only 50 or 100/ha. We may conclude from these figures, and from the very different shapes of the tropical and nontropical species area curves in Figure 1, that for relatively large areas of 1 ha, the richest closedcanopy tropical forests have more tree species than the richest nonforest communities have species of all plants. However, for smaller sample areas of <0. 1 ha, the richest sclerophyllous vegetations are apparently as species rich as are the majority of tropical forest communities when only the latter's trees are considered. Whether tropical forests, as a general rule and even for sample plots of < 1 ha, have more plant species than these rich nonforest communities depends on how rich tropical forests are in the rarely censused nontree habit groups. Lianas are one nontree component of tropical forests that contribute significantly to their diversity. One of us has recently published data that focus on the species richness of tropical forest lianas (Gentry 1982b, 1983, 1985a). In a series of 0.1-ha samples of plants -2.5 cm DBH in a broad series of lowland Neotropical forests, lianas (including some hemiepiphytic climbers; see definitions in Gentry 1985b) constituted 19 percent of the species in dry and pluvial forests, 22 percent in wet forests, and 23 BIOTROPICA 19(2): 149-156 1987 149 This content downloaded from 207.46.13.122 on Thu, 19 May 2016 04:17:50 UTC All use subject to http://about.jstor.org/terms TABLE 1. Plant species richness of 0. 1-ha samples in different vegetation types. Only the most species-rich sites (and some ranges of normal variation) for each vegetation type are included. Range (ave.) for Site or vegetation Reference 0.1 ha 0.1-ha plots Herbsa Woodya Mediterranean climate South African fynboschb Tussen die Riviere Werger 1972 64 Cape Province fynbos Naveh & Whittaker 1979 52-128 (ave. 75) 39.8 35.3 Tygerberg Cowling 1983 103 68 30 Cape Town Cowling 1983 99 71 28 Subtropical transitionalthicket Cowling 1983 98 48 50