Fish are smarter than what we thought.

In the Thanksgiving, I helped my girlfriend fishing - she is working on mating preference behaviors with mosquito fish. There are three continuous ponds on her campus, upstream-pond, middle-pond and downstream-pond. The upstream pond was occupied by a kind of large fish (I cannot remember the name) which feed on other small fish so that fish is the only species in upstream pond. In the middle pond, there are mosquito fish which are of large size and they are mature fish. Downstream is dominated by sailfin molly together with small mosquito fish. She wants only large mosquito fish which can only be caught in the middle pond. We used traps with large entrance but small exit to catch mosquito fish. For the first time, we put traps at the bottom of the middle pond and made some dog food as bait inside. 20 mins later, it was empty. As mosquito fish swim in the top layer of, we turned to trap them on the top layer of water. This time, over 30 fish got trapped and she was so satisfied. It was a pity that when we tried to replicate this success, the second trial gave us only several fish and the third trial gave us zero. It is hard to imagine that fish can learn so quickly because those fish which got trapped cannot communicate with other fish about the experience being trapped. Fish can actually learn by only watching and twice is many enough for them to adjust their behaviors, somehow faster than some humans.