I've always been fascinated by the story of Giotto drawing a freehand circle to demonstrate his prowness. It is an old topos that would have been as much at home in Pliny or an ancient Chinese treatise on art. Did anyone at the time think to measure Giotto's circle against one drawn with a compass? The question continued to worry me. In 1974 I thought I'd have a go myself. In fact, not being Giotto I gave myself rather more chances than one of getting it right. The procedure was quite simple. First I made a broad attempt with a dark colour. Then, over all the canvas, with a light colour, I isolated the putative circle, refining it by eating away at the line from the outside and from within. Over that line I had another try, refining that in its turn. And so on, until I thought that the resultant line was as near a true true circle as I would get. After this, having ascertained where the centre of that circle lay, I drew parallel circles with a compas either side of the freehand line, thus exposing any irregularities it might have. The small group of paintings in which I made several shots might be subtitled The advantages of not being Giotto. Works & Texts to 1992 p.137 (http://www.tomphillips.co.uk/works/paintings/item/5734-attempts-to-draw-a-freehand-circle#)