Wilmerding chalks up many of his errors to information “which was not available or accessible when I started my work some 40 years ago.” He declined to comment on Craig’s findings, instead referring me to Lisa Peters of New York’s Spanierman Gallery, who was his research assistant for the Lane/Mellen catalogue. (“Fitz Henry Lane & Mary Blood Mellen” moves to the Spanierman starting October 7.) Peters says that Fitz H. Lane “must be read and used with extreme caution due to problems in scholarship, methodology, and issues of accuracy (and innuendo).” Her complaints turned out to be primarily about Craig’s style, clarity, emphasis, and extent of documentation. She offered little actual evidence of his getting his facts wrong, and in at least one case where she questioned his accuracy — the report that Lane himself designed his Gloucester house — it turned out that Craig and Wilmerding were in agreement.

This is not to say that Craig is error-free. Fitz H. Lane states that Mellen lived in Gloucester; Buck’s research into her life for the Lane/Mellen catalogue found that she didn’t. But Wilmerding had that wrong as well.

We all make mistakes. Wilmerding’s strength remains his æsthetic connoisseurship, and, taken one by one, his apparent errors are mostly small and forgivable. But when do they add up to more than just the natural collection of little mistakes that seep into any big research project? When does a scholar become an unreliable source?