**3. Results**

that must be highlighted for the purposes of conservation. Therefore, in previous works, we

We therefore begin our study for a biogeographical proposal for the island of Hispaniola with the general description by Takhtajan [3] on floristic regions in the world, and the work of Rivas Martínez et al. [4] on North and Central America, which establishes the rank of sector for the island of Hispaniola, and includes it within the Neotropical‐Austro‐American king‐ dom, the Caribbean‐Mesoamerican region and the province of the Antilles. The studies by Borhidi [7] on the island of Cuba [5, 6] all recognise clear differences between the Greater Antilles (Cuba, Hispaniola, Jamaica and Puerto Rico) and the Lesser Antilles, based—among other considerations—fundamentally on the high biodiversity and distribution of species in the *Orchidaceae* family. Together with the existing local studies on the island of Hispaniola [8, 9] and our own recent fieldwork, we can establish a biogeographical typology based on the elemental biogeographical unit or tesela, defined as a variable area—either continuous or discontinuous—with a homogeneous geomorphological and ecological character giving rise

Doninican Republic and Republic of Haiti, it is an island very much studied from the floristic point of view, with very few studies of vegetation, with few studies like the one by Borhidi [7]. The importance of the study is due to the high diversity of endemic species and habitats that are of interest for conservation, with species and habitats subject to high human pressure,

In general, the island is dominated by areas very antropizadas, especially Haiti, where the human pressure is excessive; while in the Dominican Republic, anthropogenic action is some‐ what lower, with areas dedicated to livestock and agriculture. However, there are two large well preserved landscape units: the rainforests of the mountains and the dry sub‐deciduous forests.

We conduct a botanical study from a biogeographical approach; as notwithstanding the numerous botanical and floristic works of investigation by researchers such as Urban, Ekman, Cicero, Donald Dungan, Marcano Fondeur, Jürgen Hoppe, Liogier, Zanoni, Hager, May, Borhidi, Megía, Jiménez, R. García, A. Veloz and others, very few studies have been under‐ taken from the perspective of phytogeography and vegetation science. Recent works have taken a floristic and physiognomic rather than a phytosociological approach [4, 8–18, 20, 21]; studies using a phytosociological methodology include Refs. [22–31]. The authorship of the

The different types of habitats present on the island of Hispaniola are studied. For this pur‐ pose, we have performed over 300 years of phytosociological sampling as per Ref. [32]. At the same time, we do a floristic study on the distribution of 1500 endemisms. In order to explain the distribution of the different plant communities, a bibliographical and field‐based study

is divided between

started our study with a biogeographic proposal for the Spanish island [2].

The island of Hispaniola with a territorial extension of 76,486 km2

species is mentioned only once in the text and is taken from Ref. [36].

to a single type of potential vegetation.

166 Plant Ecology - Traditional Approaches to Recent Trends

despite being a hot spot for the Caribbean.

**2. Methodology**

The island's geological origin, the bioclimatic analysis with thermotypes ranging from the infratropical to the supratropical, with semiarid to hyperhumid ombrotypes, the origin of the flora as a result of migratory routes and the past isolation of the various sierras and moun‐ tains, all account for the large number of endemic species and habitats. The island has 1284 genera, of which 31 are endemic to Hispaniola: *Zombia, Leptogonum, Arcoa, Neobuchia, Fuertesia, Sarcopilea, Salcedoa, Eupatorina, Vegaea, Coeloneurum, Theophrasta, Haitia, Stevensia, Samuelssonia, Hottea, Tortuella* and *Anacaona*, among others. Several of the endemic genera are monotypes and have a restricted area, such as *Vegaea pungens* Urb., *Zephyranthes ciceroana* M. Mejía & R. García*, Gautheria domingensis* Urb., *M. domingensis, Omphalea ekmanii* Alain*, Gonocalyx tet‐ rapterus* A. Liogier*, Goetzea ekmanii, Reinhardtia paiewonskiana* R.W. Read*,* T. Zanoni & M. Mejía, *Pseudophoenix ekmanii* Burret and *Salcedoa mirabaliarum* F. Jiménez & L. Katinas; or else local endemic species such as *Pinguicula casabitoana, Fuertesia domingensis* Urb., *Pereskia quisqueyana, M. jimenezii* Alain and *Salvia montecristina.*

There are a total of 5800 species according to [36], a figure that was subsequently extended by [37] to 6000 vascular species distributed in 1284 genera, with an estimated 2050 endemic spe‐ cies. Our study characterises the various biogeographical territories based on 1582 endemic species distributed in 19 areas (A1–A19), together with their own vegetation catenas, which we in turn break down into two subprovinces: the Central subprovince and the Caribbean‐ Atlantic subprovince, both clearly separate due to differences in their geological, bioclimatic, floristic and vegetation origin.

The province is characterised by a large number of endemic species, of which 114 belong to the family *Melastomataceae* [24]. The presence of a high number of endemic species with a widespread distribution on the island causes the application of Pearson's index to result in a low relation between areas A12 and A16 (*r* = 1.25), due to their different geological and floristic nature; between A16 and A13 (*r* = 1.17); and between A12 and A13 (*r* = 1.23). In this last case, the low relation between the two areas derives from the difference in the number of endemics. Although both zones have calcareous substrates, A13 has suffered greater human impact. A16 and A17 are highly separated, which is unsurprising as the Massif du Nord (A17) is a prolongation of the Central Cordillera range (A16). The frequent presence of calcareous islands and the intense human pressure in A17 is a further reason for their separation, with A17 acquiring a greater similarity with A15 (northwest Haiti).

The Jaccard analysis reveals that areas A12 and A16 have a distance of 0.9, representing a coincidence of only 10% and differences of 90%; this is also the case between A16 and A17, as this analysis corroborates that A17 has greater similarity with A15. The results of Pearson's analysis for areas A12 and A13 are similar to those obtained with the Jaccard analysis [24].
