The model drug, 17-beta-Estradiol (E2) was dissolved in an acrylic pressure-sensitive adhesive (PSA) DURO-TAK((R)) 87-2516 to manufacture the DURO-TAK matrix (DTM). Eight loading doses of E2 from 0.25% to 12% (w/w) were prepared. The in vitro penetration flux of E2 across hairless mouse skin from the DTM increased proportionally to E2 concentrations of up to 6% (w/w) and decreased thereafter. When the DTM with more than 2% E2 (w/w) was stored for a period of 140 and 360 days, the in vitro penetration flux of E2 decreased gradually. CCD microscope observations revealed that this phenomenon was attributed to induction and growth of E2 crystal in the DTM. By segmenting the CCD micrograph, the volume ratio of crystal occupied in the DTM was determined and the concentration of E2 dissolved in the DTM was estimated. The E2 flux calculated from the estimated concentration agreed with the value gathered from the in vitro penetration experiments. These results showed that the in vitro flux at steady-state (dQ/dt)(ss) of E2 was mainly dependent on the concentration of E2 dissolved in the supersaturated DTM rather than loading dose. Furthermore, we obtained a simple method for predicting the effect of crystallization on the in vitro flux (dQ/dt)(ss) by using a differential scanning calorimeter (DSC).