"How smart are squid?"

I get this question a lot, and I always hedge the answer. The obvious dodge, of course, is to say: what does "smart" really mean and how would you quantify it in a cephalopod? Complexity of puzzles solved? A reading test? Ability to outwit researchers?

But the real problem with squid is that all of the existing smartness metrics come out of laboratory studies. And squid, by and large, are extremely difficult to work with in the laboratory. (That's if you want to keep them alive--if you just need to slice one open and extract a giant axon, then you're good to go.) Squid are nektonic, which means they live in the middle of open water, swimming around from place to place, catching food, schooling with other squid, but never settling down to catch a few winks underneath a rock. To keep an animal like that in the lab, you need a big aquarium with plenty of room to swim, and walls that won't damage the squid if it accidentally bumps into them. That last part is really hard, because squid have delicate skin that tears and abrades very easily (it's not nearly as sturdy as fish scales).

As a result, most of the laboratory work on cephalopod intelligence has focused on the squid's benthic cousin, the octopus, who spends all its time crawling around on the seafloor and hiding under rocks.

However, one group of scientists has managed to conduct some really remarkable behavioral studies on the reef squid, Sepioteuthis sepioidea, by abandoning the laboratory model altogether. They bring the science to the squid by spending hours on snorkel in the field, observing, taking pictures and video and notes. These scientists are confident that the squid are communicating with each other, using some fairly complex signals, and possibly even teaching each other.

So. How smart are squid? I don't know, but I'm pretty sure the answer will come from Sepioteuthis.