Two kinds of vesicles with and without the presence of sodium cholate (flexible vesicles and conventional vesicles) were prepared, using cyclosporin A as model drug. When applied onto the excised abdominal skin of mice non-occlusively, the enhancing effects of vesicles on the penetration of cyclosporin A were assessed by an in vitro permeation technique. The effect of sodium cholate micelles was also studied. In vivo study was carried out by topical application of vesicles onto the mice skin and drug serum concentration was detected. Results showed that after 8 h of administration, flexible vesicles transported 1.16 microg of cyclosporin A through per cm(2) mice skin and amounted to 1.88 microg 24 h later. The residual amount in the skin was 1.78+/-0.51 microg/cm(2). However, flexible vesicles failed to transport measurable amount of drug through pre-hydrated skin while deposited 2.39+/-0.26 microg/cm(2) into the skin. Conventional vesicles failed to transfer cyclosporin A into the receiver while accumulated 0. 72+/-0.19 microg/cm(2) of drug in the skin. Furthermore, 1 and 40% sodium cholate micelles precluded the transport of cyclosporin A. In vivo studies indicated that with the application of flexible vesicles, serum drug concentration of 53.43+/-9.24 ng/ml was detected 2 h later. After the stratum corneum of mouse skin has been destroyed by shaving, flexible vesicles transferred large amount of drug into blood, up to 187.32+/-53.21 ng/ml after 1 h of application. Conventional vesicles failed to deliver measurable amount of drug into the blood under normal skin condition. In conclusion, flexible vesicle is better than conventional vesicle as the carrier for transdermal delivery of cyclosporin A. Penetration and fusion have been suggested to be two major functional mechanisms. Hydration is detrimental to the enhancement effect. Stratum corneum constitutes main barrier to the transport of lipophilic cyclosporin A.