Although it is established that the treatment by praziquantel reduces the urinary lesions due to Schistosoma haematobium, the frequency of mass treatment necessary to maintain a low morbidity level remains poorly known. The objective of this work was to study the impact over three years of a single praziquantel mass treatment on schistosomiasis morbidity in two different systems of disease transmission in Niger. The study was performed in 2 villages hyperendemic for schistosomiasis in the South-West of Niger presenting respectively 2 different systems of schistosomiasis transmission: Koutoukalé-Zéno (K Zéno), located close to an irrigated area of the Niger River Valley where the transmission is permanent, and Téguey located along a temporary pond where the transmission is seasonal. After the initial evaluation (1994), we carried out a survey 3 years later (1997) except in K. Zéno where an intermediate evaluation was performed 10 months after the initial survey (1995). Approximately 300 randomised people have been examined as follows: macroscopic examination of urine and reagent sticks for macro- and micro-haematuria, filtration and microscopic examination of urine for Schistosoma eggs, and ultrasound scan of the urinary tract for morbidity. The therapeutic coverage has reached 69.9% in K. Zéno and 78.2% in Téguey. The prevalence of infestation decreased from 74.1% to 56.4 % in K. Zéno (p < 0.001) and from 65.3% to 30.4% in Téguey (p < 0.001) at the end of the 3 years. The prevalence of heavy infestation (eggs > or = 50) went in the same time from 9.9% to 12.8% (p = 0.3) in K. Zéno and from 9.1% to 3.3% in Téguey (p = 0.01). Using ultrasound scan, the prevalence of the bladder lesions reached its previous level in both villages. However the prevalence of hydronephrosis decreased from 21.1% to 3.9% in K. éno (p < 0.001) and from 12.6% to 4.2% in Téguey (p < 0.001). Three years after the single mass treatment, the morbidity did not reach the initial level. The effectiveness of the treatment is better in the pond system where the transmission is seasonal. The lesions of the upper tract decreased more slowly than the bladder lesions, but a long time after the treatment. The re-infestation induced the re-appearance of the bladder lesions sooner than the lesions of the upper tract. The periodicity of the treatment should be variable according to the transmission system. It should occur every 2 years in irrigated areas and could be delayed (3 years) in temporary ponds. The control was beneficial in the pond system and induced a significant reduction of the severe lesions.