Maintenance of urinary continence is multifactorial and depends mainly on detrusor control and urethral closure function. The closure forces can be categorized as permanent closure forces active at rest, and adjunctive closure forces active during physical activities. The efficiency of these forces depends on the structural components in the urethral wall, the position of the bladder neck and proximal urethra, the periurethral striated muscles, and the pelvic floor muscles. By means of pudendal blockade and simultaneous recordings of pressure and cross-sectional area in the urethra, it has been demonstrated that the striated periurethral muscles and the pelvic floor muscles are of paramount importance for the closure function. This emphasizes the importance of well-functioning pelvic floor muscles to obtain continence, and probably explains the rationale for the effect of pelvic floor training in treating urinary incontinence. This study presents a review of the literature on female urinary incontinence, continence mechanisms, pelvic floor muscles, and pelvic floor training. Furthermore, a review of the literature on estrogen receptors in the pelvic floor muscles is given. Perineal ultrasonography, a method for visualization and measurement of thickness of the pelvic floor muscle, was developed and evaluated. This method was used to gain information on the thickness of the pelvic floor muscles in younger physiotherapists, healthy women, and women suffering from urinary incontinence, and to evaluate the effect of pelvic floor training. Additionally, a study of the Pelvic floor muscles was performed to assess the presence of estrogen receptors. Muscle thickness seems to decrease with age. In women over age 60 years, a significantly thinner pelvic floor muscle was found compared to younger women. The muscle increment during contraction decreased significantly with age, probably reflecting a stronger pelvic floor or a better awareness of pelvic floor function in the younger women. Incontinent women had a thinner pelvic floor muscle compared to healthy women. Hypertrophy of the muscles was demonstrated in urinary-incontinent women after pelvic floor training, and the difference in thickness of the muscles in these women before training compared to healthy women was eliminated by training. pelvic floor training reduced the use of incontinence appliances and urinary leakage both in stress and urge-incontinent women. Subjectively, 60% of the women gained a positive effect of the training. In spite of the fact that training increased muscle thickness and the increment of muscle thickness during contraction, no correlation between these parameters and subjective improvement or reduced urine loss in the pad weighing test could be demonstrated. Training may strengthen the pelvic floor without effect on the multifactorial continence mechanism in cases where urinary incontinence is caused by destruction of the urethral attachment to the surrounding tissue. No estrogen receptors were found in the nuclei of striated muscle cells in biopsies from levator ani muscles, using an immunohistochemical technique. Thus, the effect of estrogen treatment on the striated pelvic floor muscles is doubtful. A possible effect of estrogen treatment of urinary incontinence must be mediated via other structures than the pelvic floor muscles.