Levonorgestrel and progesterone plasma levels were measured in women bearing levonorgestrel subdermal implants. Two groups using four or six levonorgestrel covered rods and one group of non-hormonal contraceptors were compared. Blood samples were drawn twice a week for six consecutive weeks at different intervals after treatment administration. The mean levonorgestrel levels (mean +/- S.D.) observed in the four rods group was .49 +/- .13 ng/ml in the first year and decreased to .34 +/- .06 in the fifth year of treatment. The mean values observed in the six rods group were .70 +/- .15 ng/ml in the second year and .43 +/- .11 in the sixth year. The levonorgestrel plasma levels were slightly above those found with Norplantr implants in the four rods group and well above it in the six rod group. The highest plasma progesterone value found in each sampling period was above 9 nmol/l in 7 (14%) out of 50 subjects in the four rods group, in 2 (4.3%) out of 47 subjects in the six rods group and in all the 49 control women. In 4 out of the 7 subjects from the 4 rods group and in the two subjects from the 6 rods group, the highest progesterone value was preceded and followed by values lower than 6 nmol/l. It was concluded that the progesterone levels were seldom compatible with the occurrence of ovulation in women treated with four or six levonorgestrel covered rods which release an estimated daily dose of 70 and 105 micrograms, respectively.