> ""natural selection" - as I understand it - narrows down variations, it does not expand them." The simple answer is that natural selection often reduces vari

> ""natural selection" - as I understand it - narrows down variations, it does not expand them." The simple answer is that natural selection often reduces variability, but that does not mean it will reduce a species all the way to a single variation. Sometimes natural selection actually increases variability by favoring outliers (specialists) over the average (generalists). Consider a population with two food sources that require slightly different characteristics to access, for example, food sources at different water depths. They'll be most easily accessed by creatures whose neutral buoyancy puts them at that depth of water. Natural selection may favor individuals who are most able to generalize (that is, those who float right in between) or those who are most specialized (those who float right at the level of one food source or the other), depending on various details of the system. If you have the relevant mathematical background, you may want to look up some of the recent models of "adaptive speciation" (Ulf Dieckmann and Michael Doebeli have done some great work in this area; their book [0] has some great examples.) [0] http://books.google.com/books?id=Z-gCf7TD9-MC&lpg=PP1&ots=6XAgn7RZPJ&dq=adaptive%20speciation&pg=PP1#v=onepage&q&f=false http://books.google.com/books?id=Z-gCf7TD9-MC&lpg=PP1&#3...