> If you assume a solid, you assume a certain constant thermal conductivity, as Kelvin did, even though with a liquid, that's not true I think the liquid is a

> If you assume a solid, you assume a certain constant thermal conductivity, as Kelvin did, even though with a liquid, that's not true I think the liquid is a red herring here. The important thing is a that there are two layers with different properties. Would work just as well with an inner highly conductive solid.