An eco-logical simulation of n burgeoning qabbits and m tasty qarrots (median width ???) at some fps, some hit tests per frame.
Initially there is a lot of variation among the qabbits and none among the qarrots; over time some groups of qabbits die out, while others prosper, largely as a result of the qarrot environment they find themselves in. Larger qabbits can more easily find qarrots and other qabbits, so over time the qabbits evolve to be larger, which also decreases the number of separate qabbit groups. At first the qarrots evolve to be smaller to reduce the risk of qabbit predation, but larger qarrots can propagate further, so they have a competitive advantage repopulating deserts where the qabbits have been extirpated by starvation, so over time the qarrots also evolve to be larger.
The frequent genetic bottlenecks that result from local Malthusian ecosystem collapse foster regional phenotypic variation in both qabbits and qarrots. Over time the qabbits become more uniform, both because they have sex and because their enormous size becomes larger and larger compared to their mutations.
Near the edges of the world, qarrots of large dimension perpendicular to the edge tend to waste more of their seedlings by casting them into the abyss beyond the edge, so there’s a countervailing evolutionary pressure there that favors qarrots that are not too large in the relevant dimension.
I suspect that eventually the progressively larger qabbits will develop progressively more unstable populations as the number of separate groups diminishes, and ultimately a random fluctuation will wipe out the single remaining group, leaving only qarrots.