Virtual species are simplified models of real species that codify the response of those species to the climatic conditions. Virtual species have been used to quantify the response of species to climatic changes, to predict potential shifts in species’ geographic ranges and to test the methods used to predict the geographic ranges of species (Ecological Niche Models). Today, there are different methods used to construct virtual species for biogeographic analysis. All of those methods combine partial suitabilities across variables (temperature, precipitation, etc.) to create one multi-variable habitat suitability index. The normal procedure for combining partial suitabilities is to sum or multiply the individual layers. However, this procedure might yield misleading results. In this paper, we run several analyses that indicate that those methods underestimate the role of the limitation factors (factors with unsuitable conditions that should decrease the habitat suitability index to zero). To solve this problem we programmed an algorithm—the Niche Limitation algorithm (NicheLim). NicheLim uses the same philosophy as the BIOCLIM model: species have independent physiological tolerances to the environmental variables. This means that we must first transform each continuous layer into a presence-absence variable, and then combine them. Here, we discuss the current framework for constructing virtual species in biogeography and its main drawbacks. We then explain the improvements of using NicheLim to create our virtual species and test biogeographic hypotheses.