Organic matter pools contributing to N mineralization potential have not been adequately defined, We analyzed 61 soil samples (0-15 cm depth) representing all agroecological regions of Saskatchewan, Canada to determine the relationships between mineralization parameters [i,e., potentially mineralizable N (N-0) and the rate coefficient k, measured by aerobic incubation at 35 degrees C for 24 wk] and N fractions defined by a series of chemical and physical extraction procedures, Light-fraction (LF) organic matter, isolated by notation on NaI solution at specific gravity 1.7, soluble organic matter measured in saturated paste extracts, and NH4-N released by two chemical procedures [digestion in 2 M KCl (100 degrees C for 4 h) and steam distillation in phosphate-borate (PB) buffer at pH 11.2] were evaluated as indices of mineralizable N, Nitrogen released by hot KCl represented 2.5 to 16% of N-0, which ranged from 71 to 631 mg kg(-1). Hot KCl N was poorly correlated with N-0 (R-2 = 0.131, P < 0.01), but it was reasonably well related to N mineralized in the first 2 wk of incubation (R-2 = 0.65, P < 0.001), and to dissolved organic matter (R-2 = 0,58, P < 0.001), The value of k, which ranged from 0.025 to 0.18 wk(-1), increased as hot KCl N increased as a proportion of N-0 (KCI N/N-0 accounted for 64% of the variability in k), These results suggest that hot KCI may be selective for the most labile (possibly water-soluble) organic N, Phosphate-borate extracted an average of 3.5 times as much N as hot KCl, but relationships with mineralization parameters were generally similar to those found for hot KCl, Light-fraction N was the largest N pool measured (33-652 mg N kg(-1)), It was well correlated with N-0 (R-2 = 0,83, P < 0,001), but not with N mineralized in the early part of the incubation or with k, indicating that LF N mineralizes slowly compared with chemically extracted N. When LF N was added to PB N, there was roughly enough N in the combined pool to account for N-0.