I'm trying to finish up my Jefferson pump house. Dimmler tells us the roof was shingled, and that the stack was 16" x 45', riveted and guyed.
My first inclination was that there would be a 2' to 3' square sheet-iron flashing separating the shingles from the stack. But the only picture I can find (the Dome Rock pump house) shows nothing that substantial (although I can't tell exactly what it does show):
Anyone know of any better pictures or any other guidance?
Speaking more to residential and commercial building practices of the period,
not much consideration was given to flashing like it is today. I do a LOT of historical
work, and am often faced with the dilemma of installing a practical flashing, where
none was ever present, creating a very different appearance where a roofline will
intersect a vertical wall surface. Code calls for a 1.5" gap from siding surfaces to
bare roof sheathing. This is quite a shadow line that historically was just where the
siding and roof shingles met.
In the case of flues and stove pipe, one must suspend the hot pipe from nearby
materials that might catch fire. Code today is 18" of clearance. Back then, there
was no code. It was just built to a distance that common sense dictated. No one
wanted a fire.
The flashing was there to keep the water out. The roofing material would probably overlap the flashing on the top and possibly sides, while the flashing would be on top below (although it could overlap a base layer, with a top layer covering most of the flashing), much like how shingles and other roof materials are applied from the bottom us so water doesn't leak under it. Where the pipe could get hot, like the flue, there needs to be some space from combustible materials. This can be managed with double pipe with an air gap to insulate the outer pipe, as was common on rolling stock:
or at the time, it may have often just been a larger hole in the roof sheathing, which would need more flashing to cover (which could be hidden by roofing, but the roofing shouldn't get too close if combustible). Here's a photo of a mill at Ohio City showing a bit more flashing around a flue penetration:
Flashing bottom edge sits atop a row of shingles, other three edges covered by shingles, nestled around flue pipe in circular arrangement.
To my eye, the support wires (3) attach to top of flue and fasten to roof ridge and lower roof edge. Note foreground wire faintly visible against darker flue pipe and its shadow attaching to flue shadow at angle to the left. Both terminate at lower edge of roof.
While Pando is about what I would have expected, I think I'll copy Baileys, as that seems to me a more common pattern on the C&S (you can imagine the Dome Rock photo showing the same thing, whereas it's harder to imagine it showing some of the others).