I wrote:

Just to be clear: my question for John Conway is whether he learned about the problem from Gardner, and if not, where he learned about the problem from.

but I wasn't clear about what problem I was talking about! Here's the quote from Gardner's letter, sent to David Klarner on April 16, 1990: You have a triangular array of points, the number of points a multiple of 3. You wish to tile the lattice with lines that cover three points, like so: o / o o / \ o o o \ o---o---o o o o o---o---o It is almost certain that such tiling is never possible, but how to prove it? All my efforts to find a coloring proof have failed. Proving similar impossibility for triangles of three points seems equally hard; perhaps a proof for one would apply to the other. o / \ o---o o o---o / \ \ / o---o o o o o---o o o / \ \ / / \ o---o o o---o o Jim Propp