Mick, you've got a very interesting point there: I can't find any independent corroboration. Steve Roud didn't mention Clifton in his 2000 liner notes for the Harry Cox recording, only the undated and unattributing broadsides published by Such and Sanderson, "later 19th century." If the liner notes for George Dunn's Chainmaker album reflect what Palmer wrote in his Everyman's Book, then Palmer mistook "Water Cresses" and "The Watercress Girl" for the same song. The 1863 date corresponds to the former, which is directly associated with Clifton; consequently, anyone accepting Palmer's equation would attribute the latter song to Clifton as well ... and thus is folklore made.

Steve, have you seen an announcement or H&C catalog listing for "The Watercress Girl" on the back of any sheet music you have?