Revenge of Flexible Polyominoes

Several months ago, I felt myself at a bit of a creative ebb. I wasn’t coming up with any bold new polyform ideas, so the best I could do would be to tinker around in a space that was already well-trodden. In this state of mind, I asked myself: did Abaroth miss anything good?

When I came up with the idea of letting the cells in polyominoes be flexible rhombi, Abaroth ran with it, and made an entire gallery of tilable shapes with solutions. Some of Abaroth’s discoveries used rows of squares as seams between the “leaves” of a target shape, but he missed this nice pentomino star:

An obvious thing to want after seeing a tiling like this with five-fold dihedral symmetry is one with sixfold dihedral symmetry. So far, the attempts have run into some problems.

The tiling on the left is Abaroth’s. It contains a couple of ambiguous pentominoes in the upper right. Where the green one wraps around a degree-3 vertex, it could be “unglued” to form either an X or an F pentomino. The red one could be an L, an N, or a Y.

The tiling on the right is mine, and has a different problem. The P pentomino in the upper left is not ambiguous, but it is split. This type of flaw can only occur in a polyomino that contains a square tetromino; P is the only pentomino that does.

Problem #53: Find a flexible pentomino tiling with sixfold dihedral symmetry without ambiguous or split pentominoes.

One-sided flexible polyominoes were another area that had been missed. It turns out that there are some nice tilings here:

George Sicherman, Abaroth, and Edo Timmermans all found one-sided pentomino tilings for the above double star. This double balanced three-coloring found by Edo Timmermans is particularly nice. Remarkably, the one-sided hexominoes also admit a double star:

(Solution again by Edo Timmermans.)

It might not be clear at first that other symmetry variations on polyominoes will survive in this weird flexible world, but in fact some can. If squares can flex into rhombi, then rectangles can flex into parallelograms, and we can get tilings like the following, using the 3-, 4-, and 5-rects:

For the second post in a row, I’m going to stop with at least another post’s worth of material left to share. If I leave you in suspense, you’ll have to keep coming back, right?

Leave a Reply

Your email address will not be published. Required fields are marked *