For quite some time it was believed that geographical isolation was the key to speciation, even when other models were gaining ground it was still considered to be the most prevalent occurrence. A recent study set out to test the importance of geographical isolation in an ideal situation. In Martinique scientists have genetically tested lizards separated by around 8 Ma of isolation to find that they have not evolved into separate species. The study also provided the opportunity to compare populations which are not separated geographically but inhabit separate ecological conditions. There was more genetic isolation between the ecologically separate lizards than the geographically separated lizards.
This is being claimed as a blow against allopatric speciation, however, it seems more to be a blow against the role of genetic drift in speciation. With geographical isolation selection pressures may differ or may remain the same. If the habitat is similar and doesn't change then only drift may drive speciation, which it seems does not occur. The study shows clearly that ecological factors are the most effective, which is what is found when natural selection drives speciation. Mere geographical isolation alone is not enough, but when it means a change in habitat it seems as though speciation is almost inevitable.