Following my previous post on Gray Scott, I tried to make some variations on the model. I will not go too deep. It could easily take months and years if not lifelong.

some parameters are really sensitive like "feed" and "Kill" and result sometime is not stable. I mean that the pattern can evolve without finding a stable state.

3dxml with Hexagonal pattern at the bottom

1 The grid type, from square to hexagonal

starting from a square of Bs in the middle of the surface, here is the difference between square grid and hexagonal grid. the Laplacian calculation is the same, however, because of the hexagonal grid, all the neighboring pixels are taken into account. If you remember my previous post, I did not take the corners of the square pixels.

​​​​​​​2 The grid resolution

it is also similar to scaling the pattern. a finer resolution makes a downscale of the pattern, a finer pattern.

3 the Feed and Kill parameters

These are the parameters to find Turing patterns. He neither had a laptop nor a powerful and passionate R&D... even less internet...

4 Initial distribution of B substrate

instead of a square, I tried 2 disks of B substrates. there is a slight break of the symmetry at the beginning! because of the pixelization of the disk inside a hexagonal grid.

I think that you know there are many many more possibilities to make good patterns. This is just a superficial test

APPENDIX with the bmp pictures

​​​​​​​

​​​​​​​