In the last years the determination of serum digoxin levels has become absolutely necessary not only in specialized centers but also in all medical departments for its pharmacological and clinical implications. In order to verify the reproducibility, accurateness and sensibility of the techniques currently employed we performed a comparative analysis of the serum digoxin levels obtained by using one immuno-enzymatic and four different radio-immuno-assay kits. The first in vitro trial was performed by evaluating serum digitalis on blood specimens from eight healthy human donors to which known doses of digoxin and beta-methyl-digoxin had been added. The second in vivo trial was performed by giving for three days to sixteen cardiac patients, previously off any digitalis treatment, only one oral morning dose of 0.125 mg of digoxin (8 patients), and 0.135 mg of beta-methyl-digoxin (8 patients). A monoexponential decay constant was calculated on the curves of serum digitalis levels, measured in the third day of treatment by means of the five different techniques. The statistical analysis, performed on the results we obtained, demonstrates a satisfying degree of reproducibility and accurateness only for some of the kits used. This is due to a poor sensibility of some kits in the presence of low digitalis levels, which can produce significantly different value for the decay constant (p < 0.02). Such a behaviour possibly due to different manufacturing processes of the kits raises some questions on the validity in comparing results from various laboratories when different testing procedures are employed.