In order to evaluate the possibility of using a selective subarachnoid anesthesia in ambulatory surgery we studied its feasibility with a Sprotte needle 24 G in 103 patients (range 15-67 years) undergoing a knee joint arthroscopy. The subarachnoid anesthesia was achieved with Sprotte needle 24 G with introducer with patient on lateral decubitus. Hyperbaric bupivacaine 1% (0.05 mg/cm height) was used in order to obtain a selective homolateral metameric anesthesia between L1 and S3. The numbers of attempts were 1.29 (range 1-4). We involved the homolateral determatomeres in 94% of our attempts, adding a weak anesthetic action on contralateral ones. In the remaining 6% of the patients the anesthesia was bilateral and extending in an overlying manner. Anesthesia was adequate for the time needed to perform all the surgical procedures. Nevertheless in 3 patients the tourniquet was painful. We registered systemic hypotension only in 2% of the patients (a reduced systolic pressure value > 30%). No cases of postdural puncture headache were registered. Non postural atypical headache was seem in 3.9% of the patients. Urinary retention necessitating catheterization was noted only in 3 cases. A weak backache was revealed in 8.8%. When interviewed after surgery, 98% of our patients fulfilled the technique. In spite of technical difficulties that the use of the Sprotte needle may present, the selective subarachnoid anesthesia may be a safe alternative to the other anesthetic procedures for ambulatory surgery of the legs.