Accelerated evolution may occur when a small population of a species – whose genetic variability is therefore limited – becomes isolated from all other members. This is one explanation for the rise of new species, as in the Galapagos archipelago. Creation of such genetic bottlenecks encourages rapid genetic drift away from the main population. It has been suggested to explain sudden behavioural shifts in anatomically modern humans over the last hundred thousand years or so, partly through rapid and long-distance migrations and partly through a variety of environmental catastrophes, such as the huge Toba eruption around 74 ka. Another example has been proposed for the teemingly diverse flora and fauna of the Amazon Basin, particularly among its ~7500 species of butterflies, which has been ascribed to shrinkage of the Amazonian rain forest to isolated patches that became refuges from dry conditions during the last glacial maximum.
A great deal of evidence suggests that during glacial maxima global climate became considerably drier than that in interglacials, low-latitude deserts and savannah grasslands expanding at the expense of humid forest. Yet the emerging complexity of how climate change proceeds from place to place suggests that evidence such continental drying from one well-documented region, such as tropical Africa, cannot be applied to another without confirming data. Amazonia has been the subject of long-standing controversy about such ecological changes and formation of isolated forest ‘islands’ in the absence of definitive palaeoclimate data from the region itself. A multinational team has now published data on climatic humidity changes over the last 45 ka in what is now an area of dense forest but also receives lower rainfall than most of Amazonia; i.e. where rolling back forest to savannah would have been most likely to occur during the last glacial maximum (Wang, X. et al. 2017. Hydroclimate changes across the Amazon lowlands over the past 45,000 years. Nature, v. 541, p. 204-207; doi:10.1038/nature20787).
Their study area is tropical karst, stalagmites from one of whose caves have yielded detailed oxygen-isotope time series. Using the U/Th dating technique has given the data a time resolution of decades covering the global climatic decline into the last glacial maximum and its recovery to modern times. The relative abundance of oxygen isotopes (expressed by δ18O) in the calcium carbonate layers that make up the stalagmites is proportional to that of the rainwater that carried calcium and carbonate ions dissolved from the limestones. The rainwater δ18O itself depended on the balance between rainfall and evaporation, higher values indicating reduced precipitation. Relative proportions of carbon isotopes in the stalagmites, expressed by δ13C, record the balance of trees and grasses, which have different carbon-isotope signatures. Rainfall in the area did indeed fall during the run-up to the last glacial maximum, to about 60% of that at present, then to rise to ~142% in the mid-Holocene (6 ka). Yet δ13C in the stalagmites remained throughout comparable with those in the Holocene layers, its low values being incompatible with any marked expansion of grasses.
One important factor in converting rain forest to grass-dominated savannah is fire induced by climatic drying. Tree mortality and loss of cover accelerates drying out of the forest floor in a vicious circle towards grassland, expressed today by human influences in much of Amazonia. Fires in Amazonia must therefore have been rare during the last ice age; indeed sediment cores from the Amazon delta do not reveal any significant charcoal ‘spike’.
See also: Bush, M.B 2017. The resilience of Amazonian forests. Nature, v. 541, p. 167-168; doi:10.1038/541167a