Here is what Rex had to say:
As you can see one of the primary boundries is the mountain range that divides the pacific side from the atlantic side. Only a few species are found on both slopes and this makes it an obvious choice. The area where 4 regions meet in south Mexico is based on drainage basins that provides natural barriers.
There are several changes to be made though. A portion of texas needs to be included with the Atlantic slope of North Mexico. Panama's division is most likely temporary until I can find more accurate maps and cichlid distribution information. A map containing all of Panama's rivers and labeled is something I have not found.
On the pacific slope, I included the Honduras with Nicaragua, Costa Rica and Panama, although I actually feel like the boundry should be a river (possibly Rio Choluteca.) Likewise on the atlantic side I have the boundary set between Honduras and Nicaragua while in my opinion this might be improved upon if it can be set on river basin.
The attached map is his rough draft, if there is any trouble viewing it let me know.