Each year as an introduction to portraiture using a grid method, I have my Art I students enlarge and draw what appears to be a totally non-objective image. I give them a small square with the image on it, and they draw it onto a larger square.
They don't know what they're drawing until all of the squares are put together to form a picture!
Intro to this exercise involves doing value scales and also doing some preliminary drawings that simply involve copying various shapes.
This has worked so well in previous years that I decided to make it a bit more challenging during the fall of 2004. I gave them 1/2-inch squares this time (instead of inch squares), which resulted in 240 tiny little pieces of a Photoshop-altered picture. Students drew on 2" squares of white paper.
All squares were coded, and when students had completed the 240 2-inch squares, they glued them onto a large sheet of paper. Unknown to them, I had coded the squares so that the image was 90 degrees counter-clockwise from the photo I worked from.