Early studies found evidence both for and against measurable intelligence using the "g factor" in rats. Part of the difficulty of understanding animal cognition generally, is determining what to measure. One aspect of intelligence is the ability to learn, which can be measured using a maze like the T-maze. Experiments done in the 1920s showed that some rats performed better than others in maze tests, and if these rats were selectively bred, their offspring also performed better, suggesting that in rats an ability to learn was heritable in some way.
Read more, here.