Take a piece of graph paper. Now take a large population of people for whom you know their average daily alcohol consumption, and their mortality (when they die). Plot the mortality on the upright axis, and on the horizontal axis plot alcohol average daily consumption. Now fit a curve to the data points. If you have a large enough population, you'll see a nice curve that has the shape of a J (see figure, right). That is, those who drink nothing have what is taken to be baseline mortality -- the risk of death we use as a reference point. As daily consumption increases, what we find is that risk of death drops slightly. Then, as alcohol consumption increases, the risk of death increases until it reaches the baseline (these people have the same risk of death as those who drink nothing at all), and as drinking gets heavier, the risk of death continues to increase steadily. This is called the J-shaped curve, and is a remarkably consistent finding from many large studies. It is the solid body of data that has led to the conclusion that moderate drinking is indeed beneficial to health, although this throws up a whole set of new questions which are explored further in other articles here.