Establishment of Biogeographic Areas by Distributing Endemic Flora and Habitats (Dominican Republic, Haiti R.)

Despite the large number of botanical studies conducted on the flora of the Island of Hispaniola, some of which adopted a floristic or physiognomical approach (e.g., Zanoni et al., 1990; Höner and Jiménez 1994; Guerrero et al., 1997; May, 1997, 2000, 2001; Mejía and Jiménez 1998; Rivas-Martínez et al. 1999; May and Peguero, 2000; Mejía et al., 2000; Slocum et al., 2000; García and Clase 2002; García et al. 2002; Peguero and Salazar 2002; Veloz and Pequero 2002); and Cano et al., 2009a, 2009b, 2010a, 2010b, 2011, 2012 with a phytosociological methodology, only few have adopted phytogeographical or phytosociological approaches. Here we present a biogeographic profile on the flora of the Island of Hispanola.


Study area
Our study area is the Island of Hispaniola (Dominican Republic, Republic of Haiti); (Fig. 1). The island belongs to the so-called Greater Antilles (Cuba, Hispaniola, Jamaica and Puerto Rico), which are located approximately in the centre of the Antillean Arc. With an area of 76,484 km 2 Hispaniola is the second largest island of the group, second in size only to Cuba (110,861 km 2 ). By comparison, the Bahamas (Grand Bahama, Andros, Mayaguana, Great Inagua and Grand Caicos) are a group of islands open to the Atlantic and extend from SE to NW. The arc of the Lesser Antilles, with the Virgin Islands, St Kitts Nevis, Antigua and Barbuda, Dominica, St Lucia, St Vincent and the Grenadines, Barbados, Grenada, Trinidad and Tobago, is located to the SE of Puerto Rico.

Methods
Our biogeographical approach for the vegetation of the Island of Hispaniola relies both on previous geological studies compiled by Liogier (2000), Mollat et al. (2004), Cano et al. (2009aCano et al. ( , 2010a and also on some geomorphological studies (A.R.N. 2004). We carried out a total of 300 vegetation relevếs on the island These samples, together with the numerous floristic studies used references, allowed us to plot the different biogeographical areas of the island. For an accurate mapping of the floristic differences in the territory, we first studied the endemic plants growing not only all over the island but also on restricted sites of it. We  (Cano et al., 2009a) provide here the number of endemic species per sampling unit. The area of these sample units ranges from 500 to 2,000 m 2 , depending on the vegetation unit involved in each case, either of herbaceous, scrub or forest species. To check that the units defined in this manner were suitable as far as the differences recorded in the flora and vegetation are concerned, we simultaneously applied our own Jaccard's and Pearson's numerical analyses already published. Our approach relies heavily on the bioclimatic and biogeographical analyses up to the rank of biogeographical sector, as conducted by us in previous studies (Cano et al. 2009a(Cano et al. , 2010a. With the presentation of some bioclimatic charts representing the study area in this paper we further extend this approach. For that purpose we also made use of the globalbioclimatic.org website Rivas-Martínez (2009). Our biogeographical mapping is based on edaphological, bioclimatic, floristic, vegetational and historical criteria. Our analysis of the flora makes use of 1,582 endemic species recorded either in the references or in our field samples. We followed the taxonomy of Liogier (1996Liogier ( -2000 and Martín & Cremers (2007).
We determined the degree of kinship between each pair of 19 biogeographic areas (Fig. 2), and we consulted previous studies (Cano et al., 2010a). These references were further extended through application of the Jaccard's coefficient and the Pearson's index between the 19 areas according to the presence/absence of the species.

Analysis of vegetation
Island of Hispaniola exhibits a wide altitudinal range, from 0 at sea level to 3,175 masl on Pico Duarte (Cordillera Central), a great variety of soils, and a rainfall gradient ranging from 400 to 4,600 mm. These three parameters, together with the isolation of the territories involved have been crucial for the emergence of the current vegetation. To study the vegetation we defined some large areas according to rainfall and temperature records. These areas include dry areas, subhumid areas, humid-hyperhumid areas and high mountain areas, as defined in Cano et al. (2009a, b).
Dry areas exhibit a tropical xeric macrobioclimate dominated by an infratropical semiarid and dry thermotype ( Fig. 3 and 4). These areas also support a high richness of endemic species and correspond with our study areas A3, A9 and A12.    Most of the territory of Hispaniola presents a tropical pluviseasonal macrobioclimate (Figs. 5, 6) and dominance of the subhumid ombrotype, with rainfall rates ranging from 1,000 to 2,000 mm and Oi values ranging from 3.7-4.3 (Parque Nacional del Este), Oi = 4 (El Seibo), Oi = 6.2 (Miches), Oi = 5.9 (Mayaguana), to Oi = 9.3 (Jarabacoa). The dominant vegetation in these areas is a subhumid, broad-leaved forest undergoing a dry season from December to April. As a result of the water stress, this flora includes tree-like, deciduous species. This is the case of Bursera simaruba (L.) Sarg., Swietenia mahagoni (L.) Jacq. and other species, such as Metopium toxiferum (L.) Krug & Urb., Krugidendron ferreum (Vahl) Urb., Acacia macracantha H. & B. ex Willd., Coccoloba diversifolia Jacq. and Bucida buceras L. These formations include important endemic taxa, such as the climbing plant Aristolochia bilobata L., the tree-like Melicoccus jimenezii (Alain) Acev. Rodr. and scrub-like plants, such as Lonchocarpus neurophyllus Benth. There also are other dominant scrub formations playing the role of dynamic substitution stages, such as Zamia debilis L., which occus with the endemic taxa Pereskia quisqueyana Alain and Goetzea ekmanii O.E. Schulz.
If these subhumid forests are located on reef-perforated limestones, the territory adopts a dry profile as a result of the heavy water loss through the soil. Such settings include floristic elements such as P. polygonus, P. unguis-cati, L. weingartianus and Hylocereus undatus (Haw.) Britt. & Rose. The plant formations peculiar to A7 are associated with the dry forests of the southwest region of the island, with which they also share some physiognomical features. A similar situation occurs in in the rocky escarpments of Samaná, where B. simaruba, Coccotrinax gracilis Burret, A. antillarum, L. weingartianum and O. dilleni occur frequently. As a result of water stress, these habitats tend to exhibit deciduous species related to the    In humid areas the macrobioclimate is tropical pluvial (Figs. 7,8) with no dry season. Actually, rainfall rates are higher than 2,000 mm. These humid areas tend to be located on mountain ranges, such as the Cordillera Septentrional, Cordillera Central, Sierra de Photo 2. Dry, edaphoxerophilous forest (Parque Nacional del Este, Dominican Republic) in Area A7.

Establishment of Biogeographic Areas by Distributing Endemic Flora and Habitats (Dominican Republic, Haiti R.) 109
Our study of high mountain areas was carried out in the Sierra de Bahoruco (A12) and the Cordillera Central (A16), which we crossed from Constanza to Sán José de Ocoa. The high mountain macrobioclimate is tropical pluviseasonal and mesophytic. From a physiognomical point of view, the plant formations sampled between 1,203 m (Sierra Bahoruco) and 2,383 m (Cordillera Central) are similar and are a pine forest of Pinus occidentalis Sw. In these territories precipitation rates are lower, as the sea of clouds carried by trade winds that supports broad-leaved forest at lower elevations never reaches these higher altitudes. Winter time temperatures can fall below 0 °C. Cold, xeric conditions in the high mountain support the P. occidentalis forest, which in the Cordillera Central is accompanied 8-10 endemic species per sampling relevế.. A similar scenario has been observed in Sierra de Bahoruco (Photograph 5), where pine forest support an average record of 20 endemic plants per relevế. The high level of endemism in these two mountain ranges is likely attributable to a lengthy period of isolation. By contrast, the pine forest of P. occidentalis develops on limestone soils in Sierra de Bahoruco and includes a different floristic composition. As particularly interesting plants, special mention must be made of the endemic species Coccotrinax scoparia Becc. Agave intermixta Trel., Senecio barahonensis Urb., Cestrum brevifolium Urb., Eupatorium gabbii Urb., Lyonia truncatula Urb., Sideroxylon repens (Urb. & Ekm.) T. Pennigton, Cordia selleana Urb., Narvalina domingensis Cass., and Galactia rudolphiodes (Griseb.) Benth. & Hook. var. haitiensis Urb., together with other endemic grasses, including Pilea spathulifolia Groult, Tetramicra ekmanii Mansf., Artemisia domingensis Urb., Gnaphalium eggersii Urban, and Polygala crucianelloides DC. In our opinion, high mountain pine forest habitats endemic to Hispaniola (Cano et al., 2011a) include Dendropemom phycnophylli-Pinetum occidentalis Cano, Velóz &Cano-Ortiz 2011 andCocotrino scopari-Pinetum occidentalis Cano, Velóz &Cano-Ortiz 2011 (Photograph 6).

Distribution analysis of endemic species
Endemic plant species are found in many of the 19 floristic areas of Hispaniola. The total number of endemic species of all 19 floristic areas combined is 2,094. Meanwhile, the total number of endemic taxa is 1,162. The difference of these two figures, 932 taxa, indicates that a large number of endemic species is widely distributed, occurring widely across the island. However, there are three areas that should be considered as hot spots of endemism, including areas A12, A16, A13 (Fig. 2). Second in levels of endemism, A4 and A9 exhibit endemicity levels far above average, and are particularly interesting because of their endemic species exclusive to the territory. Also, in A18 and A19 are found endemic plant genera.

Biogeographical analysis
The geological background of the island, the wide spectrum of bioclimatic thermotypes (which range from infratropical to supratropical standards) and ombrotypes (which range from semiarid to hyperhumid standards), the origin of the flora as a result of dispersal routes, and the intense isolation of sierras and mountains have generated a high level of endemic habitats and species.. The island supports 1,284 plant genera, of which 31 are endemic, including: Zombia, Leptogonum, Arcoa, Neobuchia, Fuertesia, Sarcopilea, Salcedoa, Eupatorina, Vegaea, Coeloneurum, Theophrasta, Haitia, Stevensia, Samuelssonia, Hottea, Tortuella, and Anacaona, among others. Some endemic genera are monotypic and have a fairly restricted distributional area. This is the case of Vegaea pungens Urb., According to Liogier (1996), the total richness of Hispaniola plants is 5,800 taxa, Including islands Beata, Saona, Gonave and Tortuga.. Subsequently, Mejía (2006) increased the figure to 6,000 vascular species encompassing 1,284 genera, including 2,050 endemic species (34.1% of all Hispaniola plant species). This high level of endemism makes Hispaniola one of the world's hot spots for the conservation of the flora, and is consistent with the high levels of endemism on other Carribean islands. For example, Cuba supports a total of 6,500 species, approximately 50% of the species of which are endemic, and including 66 endemic genera. In comparison, Madagascar, in the Palaeotropical Kingdom, African Subkingdom supports a total of 12,000 plant species, 80% of which are endemic, and distributed among 12 families and 350 genera (Costa, 1997).
Our biogeographical description of island of Hispaniola includes 1,582 endemic species distributed in 19 floristic areas (Cano et al., 2010a). This high proportion of endemic taxa, together with the existence of peculiar vegetation catenae supports ascription of the rank of biogeographical province to Hispaniola. The Hispaniola Province supports 154 endemic species that are broadly distributed across the island (Table 1; Fig. 9). Of these widely distributed species, 114 are in the family Melastomataceae (Cano et al., 2010a).
Despite the presence of a relatively large number of widely distributed endemic species, Pearson analysis produced low pairwise correlations between floristic areas A12 and A16 (r = 1.25), A16 and A13 (r = 1.17), and A12 and A13 (r = 1.23). We attribute these low scores todifferences in geological, edaphic, climatological, and land use factors. Low pairwise correlation. For example, in the latter case, both areas A12 and A13 have calcareous substrates, but A13 has sustainedmore intense human land use impacts. A16 and A17 are distinctively different floristically because the Massif du Nord (A17) is an extension of the Cordillera Central (A16). The common occurrence of calcareous outcrops in A17, and the high intensity of human use there increased the difference between these two areas, making A17 more similar to A15 (northwest of Haiti; Jaccard analysis revealed a dissimilarlity distance of 0.9 between areas A12 and A16 a 10% match and a 90% difference; (Fig. 10). We obtained similar results in the comparison of A16 and A17, and this analysis confirmed that A17 was more similar to A15. Also, Jaccard analysis for areas A12 and A13 corroborated the Pearson analysis (Cano et al., 2010 a).