I'm curious about the old myth that "every snowflake is unique". My fellow CS profs Glenn Chappell and Chris Hartman point out that *if* snowflakes really have 6-sided symmetry, that means two points on opposite sides of a flake experience the same growth conditions. This implies nearby flakes should actually have quite *similar* shapes.
So this is a simple experiment--I photographed a piece of aluminum foil without any snowflakes, took it outside into the gently drifting Fairbanks Alaska snow, and took pictures of the flakes.
Conclusions are difficult to reach--it seems to be common in this snowfall for adjacent flakes to glom together into big clumps, hiding any symmetry that might exist. More resolution would improve matters, but hexagonal symmetry is indeed evident on almost all the isolated flakes. They appear to be mostly long, pointy dendrites, not the cool branching patterns I've seen elsewhere.
Ken Libbrecht has some way cooler pictures: http://www.its.caltech.edu/~atomic/snowcrystals/photos/photos.htm