Artful Dodger: That is correct, however the
weather effect is derived from this. In any non-linear (
derived system), a small change in a variable can cause a massive shift in the end result. The
global weather system is best represented as a derived equation with as many terms as you wish to consider.
IANAM, but if the pressure exerted by a butterfly's wings is expressed in respect to time/temperature/whatever, the
factor will be multiplied and influence anything after it. This was proven by the
scientist (I forgot his name, but he worked at
Berkeley) and the shift in his graph.
It is trivial whether it looked like a butterfly, smurph or bell curve at the end - what matters is that the graph shifted more and more over time as the 'mistake' was amplified by anything that was calculated with it later on. Hence, the butterfly doesn't exaclty cause a snowstorm in Canada, but it might shift it considerably. That's why, btw, we can't predict the weather for more than a few days at a time - there are too many small factors we can't consider that end up shifting it in the big picture (Airplanes, Cars, butterflies ^_^)