I am currently between conferences (Barcelona a week ago, and Warsaw next week), and so I am working on a paper. It is on a model of crystallisation via a process that has many steps. I have a general model of this sort of process and am making predictions for how this differs from when crystallisation occurs via one step. The main prediction is for a long time at the beginning of an experiment, nothing happens, then crystallisation rapidly occurs.

Many other processes also occur via a number of steps, and show the same behaviour of nothing much apparently occurring early on. One example is lung cancer. The fact that that there are a number of individual steps from a completely healthy lung to a cancerous lung has consequences. You can see these consequences here, which is the figure in an article on how stopping smoking affects the risk of lung cancer by Peto.

Most smokers start smoking in their teens, but even in those that puff like chimneys, lung cancer is very rare before they are in their 40s. But then as the figure shows, deaths from lung cancer rapidly increase. The figure also shows that if a smoker quits at say 40 or 50 years old, then their risk of getting lung cancer is lower than if they had continued smoking. But it still increases, and is much higher than if they had never smoked. Smoking for 20 or 30 years has permanently damaged their lungs, and so permanently and greatly increased their chances of getting lung cancer.

This is all due to lung cancer being the result of not one but many steps. If it was just one step, then smoking would either give you cancer or it wouldn’t, and as soon as you stopped, then your risk would drop to the level of someone who had never smoked. If for example, we assume 5 mutations in DNA are required, then if smoking causes 4 of them, then if you quit, some other cause can produce the 5th mutation. By this way, smoking can increase the risk even after you have stopped.

I guess the moral of this, is that if you smoke, you should stop smoking ASAP.