A recent paper not yet mentioned:
====
At the beginning of the 20th century, the fossil collector and trader Charles Hazelius Sternberg and his sons discovered two Edmontosaurus mummies in fluvial sandstones of the Late Cretaceous, Maastrichtian, Lance Formation, Wyoming, U.S.A. The first mummy was sold to the American Museum of Natural History in New York (U.S.A.) and the second to the Senckenberg Naturmuseum in Frankfurt/Main (Germany). The putative stomach contents of this second specimen, publicished almost a century ago in form of a short note, has been used by various authors as a kind of "crown witness" for inferences about the diet of Edmontosaurus and hadrosaurs, in general. However, it has repeatedly been questioned whether this material represents stomach contents or not. In the course of an ongoing Edmontosaurus project conducted at the Senckenberg Forschungsinstitut und Naturmuseum Frankfurt, microscopic plant remains from the reported stomach contents, as well as leaf impressions and sediment samples bearing abundant plant debris from the sandstone infill of the body cavity, were analysed based on published data from this and other dinosaur mummies to evaluate whether it is likely that this material could represent stomach contents. Based on the sedimentological and taphonomical context, as well as considerations about hadrosaur food and food processing, it seems very unlikely that the remains represent stomach or gut contents of the Edmontosaurus mummy on exhibit in Frankfurt.
For this 1922 article:
KrÃusel, R. (1922) Die Nahrung von Trachodon. PalÃontologische Zeitschrift 4: 80-87.
(no online link that I can find... It's not in the Springer archive for the journal.)
====
News (in German)