Note the slight distortion of the shells and how they all distort in the same direction. This is evidence that the rocks stretched slightly when they were folded during the formation of the Appalachian mountains. This makes sense since they are located on the bending edge of a syncline (a "U" shaped fold of rocks). Below are impressions of the pedicle valves of
These are whole shells still embedded within the shale. Note that there is one oriented with the pedicle valve down (with relation to the bedding plane of the rock) and one with the brachial valve oriented down. This tells me that the shells are not likely in life position but likely were tumbled from their original location and buried during a storm event. The fact that the shells are articulated (meaning they have both valves attached together still) is yet more evidence for that scenario.
A profile view of one of the shells within the rock. The pedicle valve is pointed down thus the brachial valve is up. Note then that the brachial valve is convex, as is typical in Emanuella praeumbonata, instead of concave to flat as it would be in Ambocoelia umbonata.
These specimens were (probably) collected from the Mahantango formation, Tully member which are middle Devonian (Givetian stage) in age. This is my first known example of Emanuella praeumbonata from the Mahantango formation.
I've found Ambocoelia umbonata before from the Widder formation (Hungry Hollow member) at Arkona, Ontario and the Ludlowville formation (Wanakah shale) at Lake View, NY and the general shape and size of those specimens are what lead me to jump to my earlier conclusion that the specimens I'd found at the Roseburg pit were the same genera.