1. The reducing method of Talbot, Saltzman, Wixom & Wolfe [1945], and the formaldehyde method of Corcoran & Page [1948], as well as a modification of that of Daughaday, Jaffe & Williams [1948] for the estimation of urinary corticosteroids, have been critically examined. 2. A study has been made of the distribution of reducing material, of formaldehydogenic material and of reducing and Δ4-3-ketosteroids, in the various fractions obtained during the various purification procedures applied to the crude CHCl3 extract. 3. Caustic soda washing does not remove significant quantities of the steroids investigated, although non-steroid reducing and uncharacterized formaldehydogenic material is removed. 4. The water fraction of the benzene-water partition contains all the reducing and Δ4-3-ketosteroids so far detected in urine. 5. The formaldehydogenic method of Corcoran & Page gives slightly higher results than the modified method of Daughaday et al. 6. The acid hydrolysis curves obtained with the Talbot method resembled those described by Paterson, Cox & Marrian [1950], but no regular form of hydrolysis curve was obtained when either formaldehydogenic method was used. 7. A comparison of normal values obtained by the authors using the chemical and paper chromatography techniques revealed that the formaldehydogenic methods estimate a very high proportion of uncharacterized material, whereas results by the reducing method of Talbot et al. approximate relatively closely to those obtained by semiquantitative paper chromatography.