An island on a former island on a former island: a biogeographer’s paradise

Submitted by editor on 29 July 2014.

Blog post by Robert P. Anderson, Robert A. Boria, and Eliécer E. Gutiérrez


A fascinating system

A tiny habitat island, on a former island connected only by a spit of sand to the mainland, which itself was once an enormous island: what better place to study evolutionary ecology and biogeography? 


This complicated scenario really occurs, in a region of stark climatic contrasts in northern South America.  A fascinating system of naturally fragmented mesic habitats (“sky islands”) exists in northern Venezuela: wet montane forests that transition quickly into drastically different, xerophytic vegetation in the surrounding lowlands.  A peripheral element of the system, the Cerro Santa Ana, holds particular biogeographic interest, catching the eye of an international research team.

Cerro Santa Ana, viewed from the south. Eliécer E. Gutiérrez stands in the foreground for scale.  Photo taken in the dry season by Robert P. Anderson (August 2006).


Cerro Santa Ana lies on the Península de Paraguaná, which is dominated by flat, xeric lowlands that support primarily cacti and dry-adapted legumes.  A low mountain only ca. 4 km long, the Cerro rises abruptly from these lowlands and reaches an elevation of 830 m, with dwarf cloud forest blanketing its upper, wind-swept slopes.  Although broad connections existed during Pleistocene glaciations in periods of lower sea level, the Península was an island during marine introgressions at some intervening interglacials and is linked to the rest of the mainland today only by the narrow, sandy, and sparsely vegetated Istmo de los Médanos (the “Isthmus of the Dunes”).

Profile of Cerro Santa Ana in the background (center left) and the Istmo de los Médanos (dark arc in the center right) viewed from the south, taken on the northern slopes of the Serranía de San Luis.  The pale brown color stretching to the west (left) of the Istmo corresponds to the shallow Golfete de Coro that separates the Península de Paraguaná from the mainland of Venezuela. Photo by Robert P. Anderson (August 2006).


Robert Anderson, José Ochoa-G., and colleagues began a series of field expeditions to the Cerro and other sky islands in northern Venezuela in 2005.  Their interest in the region was sparked by Anderson’s taxonomic studies of spiny pocket mice.  Based on existing specimens in the research collections of natural history museums, that research revealed a new species endemic to mesic areas of the Península (Heteromys oasicus; Anderson 2003).  Combined with previously existing information in natural history museums representing the accumulated past fieldwork of many other researchers, the team’s field inventories for small non-volant mammals allowed for biogeographic characterizations of the system.  Species richness declines neatly with decreasing area in these sky islands, and the species follow strongly nested distributional patterns, with a few exceptions consistent with in situ speciation (Anderson et al. 2012).  Yet, only ecological niche models (reported here) revealed a hidden gem: the distributions of two species of mouse opossums—themselves elements of a radiation of didelphid marsupials emblematic of South America’s former isolation—suggest that interspecific competition greatly restricts the distribution of one of the species, rendering its population on Cerro Santa Ana isolated from all others.


Theory meets empiricism, via insights from ecological niche models

Traditional notions of vicariance (the establishment of a barrier separating populations of a species) do not consider the possibility that geographic isolation — and therefore potentially even allopatric speciation due to a barrier — could be promoted by the emergence and persistence of biotic interactions.  Nevertheless, niche theory and the principle of competitive exclusion lead to clear predictions regarding the distributional patterns of a pair of species undergoing extreme competition.  Here, the authors contribute to the current wave of studies assessing the role of biotic interactions in determining species ranges (Wisz et al. 2013), and propose mechanisms by which such interactions can affect not only the range of a species, but also the connectivity of its populations.

Robinson’s mouse opossum Marmosa robinsoni. Photo by Marcial Quiroga-Carmona (March 2009).


Specifically, Gutiérrez et al. (2014) assess the distributional consequences of competition between a pair of sister species of small neotropical mouse opossums, Marmosa robinsoni (Robinson’s mouse opossum) and M. xerophila (the Guajira mouse opossum).  Clearly, correlative methods such as the ecological niche models used in this paper cannot demonstrate competitive exclusion; however, they can identify areas matching the requirements of both species, allowing comparison of the predictions with observational records from fieldwork.  These approaches also generate specific directional hypotheses for testing in future field and laboratory studies.


In this project, the researchers conducted extensive ecological niche modeling analyses using Maxent.  The occurrence data were based on localities from fieldwork, examination of specimens existing in natural history museums in Venezuela and the United States, and detailed georeferencing using maps, field notes, online resources, and consultation with many of the original collectors (to obtain the most accurate coordinates of latitude and longitude possible).  After conducting analyses to estimate optimal model complexity (“tuning” model settings to select those that led to the best performance), they overlapped the niche models of suitability for each of the two species.  That superimposition highlighted particular areas (in several different regions) that were suitable for both species.  Nevertheless, these sister taxa show complementary parapatric (adjacent but non-overlapping) distributions, with only one, very narrow, contact zone documented to date.  Overall, the niche models showed that the climatically suitable areas for M. robinsoni corresponded to slightly wetter conditions than those for M. xerophila, but that the two species’ tolerances overlapped substantially.  Most interestingly, in areas predicted as suitable for both species near their known contact zone, existing records of each species largely fell in areas more strongly predicted for that same species (and showing a weaker prediction for the congener).  Where did this occur?  The Península de Paraguaná and Cerro Santa Ana. 


Areas of potential sympatry for Marmosa robinsoni and M. xerophila in northern Falcón, Venezuela showing areas more strongly predicted for each species (based on ecological niche models made using Maxent). Top image, binary comparisons: grey shading denotes pixels more strongly predicted for M. xerophila, whereas black shading indicates pixels in which M. robinsoni had higher suitability values. Bottom image, continuous values of suitability: pixels in red represent sites with environmental conditions predicted as more suitable for M. xerophila, whereas pixels in green indicate sites modeled as more suitable for M. robinsoni. Reproduced from Gutiérrez et al. (2014).


Most intriguingly, however, close examination of the predictions on the Península revealed the striking existence of small areas more strongly predicted for Marmosa robinsoni embedded within a matrix holding conditions more suitable for (and occupied by) M. xerophila.  These ‘islands,’ more favorable for M. robinsoni, corresponded to Cerro de Santa Ana and the Fila de Monte Cano, a low ridge with a very small patch of mesic vegetation.  Each of these sites harbors known localities of M. robinsoni.  In fact, the two areas on the peninsula suitable for and occupied by M. robinsoni appear to be disjunct from populations of the same species in suitable areas on the mainland not because intervening regions harbor unsuitable climatic conditions, but rather because they constitute environments even more suitable for, and occupied by, M. xerophila.


Implications for biogeography and evolutionary ecology

The results suggest that competition may maintain (and may even have created) allopatric conditions between populations of Marmosa robinsoni on the Península de Paraguaná and those on the adjacent mainland.  Overall, the study indicates that in addition to affecting species ranges, biotic interactions have the potential to create geographic isolation, and hence promote genetic differentiation.  The researchers conclude the paper by proposing two mechanisms, intrusion or contraction, whereby biotic interactions (including predation/parasitism, mutualism, and commensalism) can lead to allopatry.  Because of the plethora of empirical cases now emerging in which biotic interactions affect species distributions at regional to continental spatial scales (Wisz et al. 2013), the phenomenon whereby such interactions create and/or maintain allopatry may be taxonomically and geographically widespread.


Anderson, Gutiérrez, and colleagues continue studying the biogeography of other small mammals in this system.  In this integrated research program, they are combining the results of fieldwork and examination of museum specimens (the primary data documenting biodiversity across the planet) on the one hand, with the modern cutting-edge tools of ecological niche modeling and molecular phylogeography on the other.  Complementary to the ecological niche models produced here, the first phylogeographic studies of small mammals in this system are ongoing for mouse opossums and spiny pocket mice.  Their work is supported by many sources, including the U.S. National Science Foundation.



Anderson, R. P. 2003. Taxonomy, distribution, and natural history of the genus Heteromys (Rodentia: Heteromyidae) in western Venezuela, with the description of a dwarf species from the Península de Paraguaná. Am. Mus. Novitates 3396:1–43.

Anderson, R. P., E. E. Gutiérrez, J. Ochoa-G., F. J. García, and M. Aguilera. 2012. Faunal nestedness and species–area relationship for small non-volant mammals in “sky islands” of northern Venezuela. - Stud. Neotrop. Fauna Environ. 47:157–170.

Gutiérrez, E. E., R. A. Boria, and R. P. Anderson. 2014. Can biotic interactions cause allopatry? Niche models, competition, and distributions of South American mouse opossums. – Ecography. 37: this issue.

Wisz, M. S. et al. 2012. The role of biotic interactions in shaping distributions and realised assemblages of species: implications for species distribution modelling. – Biol. Rev. 88: 15–30.