Thank you, Jerry, for the prod for comments on the question of whether METS ought or ought not have a processMD. My only elaboration on your summary is to clarify our sourceMD does not contain "descriptive metadata" in the conventional bibliographic sense, although you could think of it in terms of a much expanded MARC 300 field. The processMD slot, of course, is a place to record what we did to get from the source to the file we have before us now. Thanks for the continuing effort! Carl/Morgan/Dick/Neil - - - - - - Here are some of the attributes we have in sourceMD, just to give the flavor: CONDITION DIMENSIONS DISPOSITION NOTE BASE_MATERIAL BINDER CALIBRATION_TYPE CALIBRATION_INFO CALIBRATION_LOCATION CHANNEL_TRACK_TYPE CHANNEL_TRACK_MAPPING DATA_RATE_FIXED DATA_RATE_VARIABLE DISK_SURFACE DIGITAL EQUALIZATION FORMAT_SPECIFICATION_NAME FORMAT_SPECIFICATION_INFO DURATION GROOVE GUAGE LENGTH NOISE_REDUCTION OXIDE SAMPLING_FREQUENCY SOUND_FIELD SPEED SPEED_ADJUSTMENT STOCK_BRAND TRACK_FORMAT