I don't think this has been posted, but, to add to contributions from Gaianne this morning, also from nb41 and from Alvarez, there's a comment from Oil Drum contributor donshan cited by Euan Mearns in his latest article:

The Oil Drum | Fukushima Dai-ichi status and potential outcomes

Like aluminum, zirconium and is alloys (Zircaloy-2) oxidize instantly in air. A thin film of ZrO2 is so impervious to oxygen diffusion that the reaction stops. Even in 300 C (572F) water or steam at over 1000 psi, the oxidation rate is extremely slow and corrosion properties of Zircaloy fuel cladding are outstanding and safe, AS LONG as they are not overheated and cooling water flow is maintained. In fact it is standard practice to autoclave fuel rods in hot-pressured water or steam to precoat these rods with the optimum coating of ZrO2.

But these fuel rods must NEVER be overheated. That is why multiple redundant cooling systems are required. All these backup-cooling systems failed in Japan. Even after reactor shutdown, if the fuel rods are uncovered cladding temperatures can rapidly rise to 800C , or higher, due to fission product decay heat. As in any chemical reaction the rate accelerates rapidly with temperature, but in the case of zirconium, the protective character of a thin ZrO2 film is destroyed by this high temperature and catastrophic oxidation occurs. However this catastrophic oxidation occurs below the melting point, so I object to the media using the common term "meltdown" which is misleading.

This loss of the last battery powered cooling, led to the fuel rods becoming uncovered in a manner similar that also occurred in the Three Mile Island accident (although due to different reasons). When overheated in steam the oxidation reaction above accelerates exponentially. As the zirconium oxidizes the coating thickens, cracks and turns white from internal fractures that increase the diffusion rate of steam to the metal. . It then has the look and mechanical properties of eggshells. Hydrogen from this process is released, but also is absorbed by the underlying metal cladding which causes embrittlement and metal fracture. Soon cracks form in the cladding releasing the trapped fission products inside. This is not "melting', but rather catastrophic disintegration of the cladding structural integrity and containment of fission products. If the process continues the cladding can fracture away exposing the fuel pellets which in the worst-case scenario can drop out and collect on the bottom of the reactor vessel. It is the worse case scenario that I believe is causing the Japanese to inject boric acid. Boron is a neutron absorber and will prevent any possibility of a pile of fuel pellets on the bottom of the vessel from going critical and restarting the chain reaction.