Angiosperms, Los Tuxtlas Biosphere Reserve, Veracruz, Mexico

: The Los Tuxtlas Reserve has been heavily deforested and fragmented since the 1970’s. Although the flora of Los Tuxtlas has been described previously, most floristic lists come from the large forest reserve of the Los Tuxtlas field station. Here we present a check list of Angiosperms recorded in 45 rainforest fragments (< 1 to 266 ha) located in three landscapes with different levels of deforestation. We sampled all trees, shrubs, lianas, palms and herbs with diameter at breast height (dbh) Ā 2.5 cm within ten 50 m x 2 m plots per fragment. We recorded 9,435 plants belonging to 73 families and 372 species. Fabaceae, Rubiaceae, and Moraceae were best represented. Eight species are classified as Endangered by the Mexican government, and five are human-introduced species. We conclude that the conservation and restoration of all the remaining rainforest fragments are necessary to effectively preserve the plant diversity of this region.


Introduction
Accelerated deforestation and fragmentation of primary forest in tropical regions (Achard et al. 2002) are threatening global biodiversity at an alarming rate (FAO 2006). These two processes modify the spatial pattern of the remaining forest (e.g. reduced patch area, increased patch isolation, and increased proportion of forest edges; Andrén 1994), potentially affecting ecological processes such as pollination, seed dispersal, recruitment, competition, migration, herbivory and extinction (Turner et al. 1996;Benítez-Malvido 1998;Benítez-Malvido et al. 1999;Cordeiro and Howe 2001;Wright and Duber 2001;Chacoff et al. 2004;Aguirre and Dirzo 2008). As a consequence, habitat fragmentation can not only decrease the number of plant species, but also lead to significant changes in composition and vegetation structure (Turner et al. 1996;Laurance et al. 1998;Hill and Curran 2003;Arroyo-Rodríguez and Mandujano 2006;Chazdon et al. 2007;Santos et al. 2008;Arroyo-Rodríguez et al. 2009;Dirzo et al. 2009).
In Mexico deforestation has led to the loss of approximately 90 % of the tropical rainforest (Flores-Villela and Gerez 1994), particularly affecting the Los Tuxtlas Biosphere Reserve (Dirzo and García 1992;Guevara et al. 2004). This region represents the northern limit of tropical rainforest distribution in the Neotropics (Dirzo and Miranda 1991), and 95 % of the original rainforest here has already disappeared (Castillo-Campos and Laborde 2004; Guevara et al. 2004). While several studies have analyzed the plant communities in this region (e.g. Bongers et al. 1988;Ibarra-Manríquez et al. 1995;1996a;b;1997a;b), most were carried out in the large forest reserve (700 ha) of the Los Tuxtlas biological field station of the National Autonomous University of Mexico (UNAM), with little attention being paid to the changes in plant communities that arise from deforestation and forest fragmentation (but see Arroyo-Rodríguez and Mandujano 2006;Arroyo-Rodríguez et al. 2009;Dirzo et al. 2009).
To contribute to the understanding of how the loss and fragmentation of the rainforest affect the vegetation in Los Tuxtlas, and to assess the potential conservation value of forest fragments, we sampled vegetation in 45 rainforest fragments (< 1 to 266 ha) located in three landscapes with different levels of deforestation (24 %, 11 % and 4 % of remaining forest cover; Arroyo-Rodríguez et al. 2007;2009). Here we present a species list of all the plants recorded in the study landscapes and suggest some conservation priorities.

Study Site
The Los Tuxtlas region is located in the southeast of the state of Veracruz, Mexico (18°8' -18°45' N, 94°37' -95°22' W; Figure 1). The climate is warm and humid, with a mean annual temperature of 25 ºC, and annual rainfall between 3,000 and 4,600 mm. This region covers an area of 155,122 ha, with elevation ranging from 0 to 1,780 m above sea level (a.s.l.). Los Tuxtlas was decreed a Biosphere Reserve in 1998 owing to its exceptional biodiversity (CONABIO 2000). The original dominant vegetation type (below 700 m a.s.l.) was tropical rainforest, but the reserve was heavily deforested and fragmented between 1972 and 1993, and the remaining rainforest is surrounded by a matrix of pastures and croplands (Castillo-Campos and Laborde 2004; Guevara et al. 2004; Figure 2).
We selected three landscape fragmentation units ( Figure 1) considering that: (1) they represent a gradient of rainforest deforestation; (2) they are all situated between 0 and 400 m a.s.l. (to avoid changes in vegetation associated with altitude; see Castillo-Campos and Laborde 2004); and (3) they each occupy a similar area (ca. 5,000 ha). Elsewhere we have presented a full description of the methods used to digitize the landscapes, and detailed the differences in spatial attributes between landscapes (Arroyo-Rodríguez et al. 2007;2009). Only a brief overview is given here. The three landscapes have been highly deforested, but there were notable differences in the degree of deforestation. The landscape with the lowest deforestation level (LDL) covered 5,356 ha, 24 % of which was rainforest distributed among 75 patches ranging from 0.5 to 700 ha; the landscape with intermediate deforestation level (IDL) covered 4,965 ha, 11 % of which was rainforest distributed among 88 patches ranging from 0.5 to 76 ha; and the landscape with the highest deforestation level (HDL) covered 5,046 ha, 4 % of which was rainforest distributed among 46 patches ranging from 0.5 to 68 ha.  . Vegetation within forest fragments is highly variable in composition and structure (E-G), with larger fragments dominated by large trees, with a closed canopy (E-F), and smaller fragments dominated by a higher density of smaller trees (G).

Data Collection
Vegetation was sampled in 45 randomly selected rainforest fragments (15 per landscape) using the Gentry (1982) protocol. Within each fragment we randomly located ten 50 m x 2 m plots. All trees, shrubs, lianas, palms and herbs with dbh Ā 2.5 cm were recorded. Lianas were measured at the base, not at dbh (Gentry 1982). Species not identified in the field were collected for identification in the MEXU (Institute of Biology, UNAM, Mexico City) and XAL (Institute of Ecology A.C., Xalapa, Veracruz) herbaria (see further details in Arroyo-Rodríguez and Mandujano 2006;Arroyo-Rodríguez et al. 2007;2009).

Results and Discussion
In total, we recorded 9,435 plants belonging to 73 families and 372 species (Table 1). We identified 88.4 % (n = 329) of the species and 98.8 % of the stems sampled. Of the 329 identified species, 320 were dicotyledonous and 9 monocotyledonous (i.e. families Arecaceae, Heliconiaceae, and Smilacaceae) ( Table 1). The 320 identified dicotyledonous species represent ca. 51 % of the dicotyledonous species reported for the Los Tuxtlas biology field station (627 spp.; Ibarra-Manríquez et al. 1995;1996a;b), and ca. 17 % of all the plant species (including epiphytes) reported for the Los Tuxtlas Biosphere Reserve (1,873 spp.; Castillo-Campos and Laborde 2004). As two well-known ecological theories predict (island biogeography theory and metapopulation theory: MacArthur and Wilson 1967;Hanski 1999) we found that the number of species was higher in the landscape with lowest deforestation level (LDL = 253 species), than in the other two landscapes (IDL = 160 species; HDL = 180 species) ( Table  1).
The families with the highest number of species were Fabaceae (31 species), Rubiaceae (19), and Moraceae (19), together representing 21 % of all the identified species (Table 1). These are also the best represented families in the Los Tuxtlas biological field station (Bongers et al. 1988;Ibarra-Manríquez et al. 1997a). In general, the most common species were Astrocaryum mexicanum (Arecaceae), Siparuna andina (Monimiaceae), Croton schiedeanus (Euphorbiaceae), Vochysia guatemalensis (Vochysiaceae) and Stemmadenia donnell-smithii (Apocynaceae) (together representing ca. 19 % of all stems) ( Table 1). All these species (except A. mexicanum) are light-demanding; a functional group that is common in forest gaps and close to forest edges (Benítez-Malvido 1998). As the majority of the study fragments were small (60 % < 5 ha), and therefore, highly affected by edge effects (e.g. increases in light, temperature and wind intensity; Saunders et al. 1991), we expected light-demanding species to be relatively dominant in terms of abundance of stems. However, as we have previously reported, most of the identified species are old-growth forest species (Arroyo-Rodríguez et al. 2009 SEMARNAT 2002). In the LDL, we found 82 stems of 7 of these species, in the IDL 119 stems of 5 species, and in the HDL only 35 stems of 5 species. Furthermore, of the 148 plant species reported as useful for commerce (i.e. timber, fuelwood, ornamental, artwork, and others) in the Los Tuxtlas rainforest (not including epiphytes; Ibarra-Manríquez et al. 1997b), 113 species (76 %) were sampled in our study fragments (Table  1). These species made up 34 % of all identified species, and 52 % of all identified stems. Of the 249 species identified in the LDL, 95 (38 %) were useful species. Of the 145 species identified in the IDL, 63 (44 %) were useful species, and of the 154 species identified in the HDL, 65 (42 %) were useful species.
Although evidence indicates that fragmentation may favor the invasion of exotic plant species in forest fragments (Turner et al. 1996;Dislich and Pivello 2002), in our sample most of the species (369 species, 99 %) were native to the region, and only five (1 %) were human-introduced species (Citrus sp., Coffea arabica, Psidium guajava, Theobroma cacao, and Manguifera indica) representing only 0.4 % of the stems sampled (Table 1). This finding is similar to that reported by Dirzo et al. (2009) in the same region, and could be caused by, on the one hand, a relatively short amount of isolation time for the fragments (Turner et al. 1996;Santos et al. 2008), and, on the other hand, the lack of environmental conditions for the natural dispersion, establishment and development of these cultivated species.
In conclusion, our results demonstrate that forest fragments may serve as reservoirs of diverse native plant communities, including endangered and economically important plant species. In spite of the small size of most of the fragments, in a previous paper we demonstrated that the smallest fragments present a similar species density to the biggest fragments, and also that the species turnover (beta diversity) among fragments and landscapes is very high (Arroyo-Rodríguez et al. 2009). Therefore, and in accordance with Dirzo et al. (2009), we believe that the conservation and restoration of all of the remaining forest fragments is necessary in order to effectively preserve the plant biodiversity in Los Tuxtlas region.