> Obviously if we came in one day to see an octopus in the flounder tank the mystery would be solved day 1.

It's hard not to speculate that the octopus had a sufficiently sophisticated model of its keepers in relation to itself (which implies self-consciousness) that it knew this.

I can come up with other theories (some territorial affinity, perhaps), and I'm sure people have. I wonder which have been tested/explored...