Re: Ellen's elastomeric coatings and "all the crevices of metal shingles" I've seen satisfactory performance of applied coatings on copper roofs which had developed "pinholes" in undocumentable locations. But I have trouble imagining how thin elastomerics (thin enough to not obscure the roofing pattern very much) effectively span seams or gaps, especially when even small movements of the substrate could tear open the membrane. That is to say that if the elastomeric is well bonded to each of two metal pieces, as it should be, then movement between those two pieces stresses the inner face of the membrane with zero dimension of tolerance. The outer face of the membrane might stretch a bit, but doesn't the inside face of the membrane propigate into a tear?! Therefore, I wouldn't expect much from thin coating seamed material subject to expansion, contraction or other movement. Thick sloppy bituminous coatings are another thing, and that would considerably obscure the pattern considerably. It's the same principle as for joint sealants that are bonded to all three concealed faces of the joint (instead of only the two opposed faces). Movement or opening of the joint would tend to tear open from the inside out, causing the joint to fail. I assume coating a shingle roof is a stopgap measure to keep from a more thorough repair or treatment. --Jim