Physics Review Letters has a new research article (open access) suggesting, no, make that strongly suggesting that the Deepwater Horizon oil spill this last summer could have easily been stopped by a top kill of cornstarch and water - that magical mix of dilatancy commonly call Oobleck, a term invented by Dr. Suess.
The research is fairly simple: they found that when the mixture is poured into a column of mineral oil, the strand doesn't break up, a critical requirement for achieving a top kill. As long as the fluid stays together, it will eventually fill the well and block flow.
I think the article is a little strange in that it is so oriented towards a top kill and not on what was really discovered. Additionally, there are huge gaps between this initial discovery and the intended application, including temperature (this work was done at RT, the external conditions at the bottom of the Gulf were near freezing, while the exuding petroleum was probably pretty warm although I don't know this for a fact)and also the large differences both chemically and rheologically between crude oil and mineral oil. It's a nice first step and others can certainly begin to study the applicability further for the next time a top kill is needed, which hopefully is not for a very long time.