We opted to use a Dentiform model (above) as our “patient” so that we could have nine crowns in our experiment without having to put a live patient through nine impressions. Our dentist-consultant modified the Dentiform so that it appeared more lifelike; he then prepared tooth #14 with a shoulder preparation, took nine full-arch impressions of the Dentiform and sent each one to a different laboratory with an identical prescription.

Crown S lost points during evaluation because it was returned on a broken model and because there are cracks in the porcelain on the buccal and lingual surfaces (below).

What concerned our evaluators most was that none of the crowns in our experiment has an acceptable marginal fit when evaluated on the die. Shown here are three crowns—(from l. to r.): Crown Z, Crown T and Crown R—that were among the worst of the lot.