Arnold's cat mapping is the self-mapping of the 2-torus $\mathbb{R}^2/\mathbb{Z}^2$ given by $\ F(x,y)\equiv(2x+y, x+y)\; \mod 1$. If one draws the image of a cat on the 2-torus, discretizes it and applies the mapping F iterated, the cat becomes unrecognizable at first. After finitely many steps, however, one gets back the original image of the cat. The goal of this project is to find an image of a cat for which an image of a dog is generated at least once during the iteration.