The physicists scanned a series of Pollock's artworks into a computer and masked each painting with a series of grids. They then counted the number of squares that contained part of the painted pattern (N) - using the well-know 'box-counting' method of fractal geometry - and reduced the size of the squares (L). The largest square was the size of the painting while the smallest was 1mm. The fractal behaviour of the painting could then be pinpointed by plotting a log graph between the two values N and L. They discovered that Pollock's paintings consist of two distinct fractal patterns. One was determined by the paint dripping on the canvas and gradually becomes finer over time. The other pattern was shaped by his motions around the canvas.