Objective: This study aimed to assess the efficiency, in terms of recovered motile spermatozoa with normal morphology, of three sperm selection techniques: migration- sedimentation (SS), swim-up from fresh semen (SF), and swim-up from washed (SL) sperm. Methods: Samples from 20 normozoospermic men were divided into three equal aliquots and processed in parallel. SS was performed in a Jondet tube, using 1 ml of semen and 2.5 ml of Human Tubal Fluid medium (HTF+10% Synthetic Serum Supplement, Irvine, USA). For SF, 1 ml of HTF was layered over 1 ml of fresh semen (SF). For SL, 1 ml of sperm was first centrifuged (300 g, 10 min) and the pellet resuspended in 1 ml of HTF; a second layer of HTF was placed on top. Migration time was 1h (SF and SL) and 1h30' for SS at 37 degrees C. After migration, 200 mu l were removed from the top layer (SF, SL) and from the central cone (SS). Concentration, morphology and motility were determined. Results: Recovery rates were 25% for SS, 10.1% for SF and 4.5% for SL. SS recovery rate was significantly higher (P<0.01) than the two swim-up techniques. Total motility was statistically different (P<0.001), with 93.6% for SS, 91.2% for SF, and 77% for SL. Sperm morphology was similar between the three techniques (P=0.12). Conclusions: SS is an efficient technique for the recovery of motile spermatozoa from native semen preparations and yielded better results than SF and SL. Routine use for assisted reproduction remains to be evaluated.