The sources for ancient masks are, of course, regrettably few. We used what little text evidence there is, particularly the Suda entry on Thespis, which says that the playwright introduced masks constructed of linen, and looked closely at extant imagery. Many mask pictures are from much later than the original productions and we confined ourselves to the vases that are most contemporaneous with extant plays and that show masks as masks, rather than simply as the integrated head of a character, in particular the Pronomos vase from the end of the fifth century (A.R.V.2 1336.1). We also followed the same logic that many of the scholars and practitioners mentioned below have used: the masks must have been constructed of something not capable of surviving in the archaeological record; they must have been light enough to be worn without too much discomfort; their materials must have been able to have been carved or moulded into a permanent shape that could also be decorated.