Among the Galenic texts attracting special attention at the end of the thirteenth century was De complexionibus, one of whose features was a crude protocol for determining experimentally the qualitative character and intensity of any given medicine. This protocol caught the attention of the medical faculty at Montpellier, especially in the somewhat more structured form given it in Avicenna’s Canon. Three generations of writers at that school (1300–1340) successively articulated this initial model and made it into an elaborate, carefully structured test-procedure for identifying the nature and extent of a drug’s effect on healthful function, a procedure they referred to as a via experimenti: they introduced a null point as the referent for their measurements (something not present in Galen or Avicenna), identified an increasing range of contingent factors that might affect the results and that the tester had therefore to control for, and devised ways to standardize the purity and weight of the sample to be tested. Their increasingly precise protocol was certainly designed to be used, yet there is no sign that it ever was systematically employed; in practice, the masters seemed to have preferred an alternative via rationis that inferred the nature and strength of a medicine from sensory attributes, such as taste and color, that were understood to be the necessary expressions of qualitative nature. They freely acknowledged that a taste test’s results were coarser and less certain than a structured experimental procedure, but it was far easier and quicker to perform than the elaborate alternative, and when it went wrong their mistakes would have no serious medical consequences for their patients as long as they ensured that the doses they prescribed were initially very small—yet another method of testing, of course, though not, for them, of the same sort.