I found an interesting article on bivalve predation by the Giant Pacific Octopus. Of course, I only have access to the abstract...
The packaging problem: Bivalve prey selection and prey entry techniques of the octopus Enteroctopus dofleini.
Anderson, Roland C.; Mather, Jennifer A. , 2007. Journal of Comparative Psychology. Vol 121(3) 300-305.
Abstract
Many predators face a complex step of prey preparation before consumption. Octopuses faced with bivalve prey use several techniques to penetrate the shells to gain access to the meat inside. When given prey of mussels Mytilus trossulus, Manila clams Venerupis philippinarum, and littleneck clams Protothaca staminea, Enteroctopus dofleini solved the problem differently. They pulled apart V. philippinarum and M. trossulus, which had the thinnest shells and the least pulling resistance. P. staminea were eaten after the shells had been chipped or had been penetrated by drilling, presumably to inject a toxin. Likely because of these differences, octopuses consumed more V. philippinarum and M. trossulus than P. staminea when the mollusks were given to them either 1 species at a time or all together. However, when the shells were separated and the penetration problem removed, the octopuses predominantly chose P. staminea and nearly ignored M. trossulus. When V. philippinarum were wired shut, octopuses switched techniques. These results emphasize that octopuses can learn on the basis of nonvisual information and monitor their body position to carry out feeding actions.