Application rates and organic filtering materials for the removal of nitrogen and phosphorus in biofilters in primary house sewage are determined. An assay was prepared with 27 biofilters in an experimental area of the Universidade Federal de Viçosa, Viçosa MG Brazil. Nitrogen and phosphorus concentrations were determined monthly during 153 days of the experiment which followed the sub-subparcel scheme. Whereas parcels consisted of application rates of home sewage (0.5; 1.0; 1.5 m 3 m 2 d -1 ), the subparcels comprised types of organic matter (composted urban organic wastes, sugarcane bagasse and sawdust) and the sub-subparcels were the evaluation periods (August, September, October, November and December 2009), in totally randomized blocks, with three repetitions. Results showed that the filtering organic matters sugarcane bagasse and sawdust were more adequate for N and P removal in home sewage. However, composted urban organic wastes raised their rates in the effluents collected by the biofilters. There was no significant effect of application rates of house sewage on N and P removal of the effluent by biofilters filled with composted urban organic wastes, sugarcane bagasse and sawdust. Biofilter effluents have high N and P quantities for fertilizationirrigation of non-raw consumed cultures.Key words: sustainability; water re-use; treatment; earthworm. RESUMEN

Retention curves describe water dynamics in the soil and represent the relationship between the energy of water retention and moisture content. Several empirical models of water retention curve settings in the soil are analyzed (Van Genuchten-VG, Brooks & Corey BC, Bruce & Luxmore-BL and Driessen-D). Deformed soil samples were collected at 10 sites per plot; a compound sample was established with 36 plots between October and November 2007, at depths 0-20 cm and 20-40 cm in the Irrigation District of Gorutuba in Janaúba MG, Brazil. Tensions used for tests were 10, 30, 100, 500 and 1500 kPa. Soil Water Retention Curve software was employed to obtain the parameters of the empirical equations. The performance of empirical models was obtained by the index suggested by Camargo & Sentelhas, coupled to estimate standard error, standard deviation and coefficient of variation. The results showed that the VG model had the best performance compared to the other models tested for data on water retention in the soil. The BC model had the most unsatisfactory performance.Key words: water use efficiency, performance, irrigation. RESUMO A curva de retenção é utilizada para descrever a dinâmica da água no solo, representando a relação entre a energia de retenção de água e o teor de umidade. O presente trabalho objetivou analisar vários modelos empíricos de ajuste das curvas de retenção de água no solo (Van Genuchten-VG, Brooks & Corey-BC, Bruce & Luxmore-BL e Driessen-D). Para tal, foram coletadas amostras