Tile drainage water from agricultural fields commonly exceeds environmental guidelines for phosphorus (P) in rivers and streams. The loss of P through artificial drainage is spatially and temporally variable, and is related to local factors. This study characterizes variability in total P (TP) and soluble reactive P (SRP) concentrations in weekly drainage samples from 39 agricultural fields in Nova Scotia, Canada, from April 2002 through December 2003. We examined connections between P concentrations and the factors: (i) soil texture; (ii) discharge flow rate; (iii) soil test P (STP); (iv) manure type; and (v) crop cover. Generally, variability between fields and samples was great, and fields with standard deviations exceeding the mean for TP, SRP, and flow rate were 71, 54, and 79%, respectively. It was evident that poultry and swine manure contributed to high STPs, and to constantly high TP concentrations with high proportions of SRP. Concentrations varied from week to week, and particularly in April, May, October, and November when the greatest TP, SRP, and flow rate averages were measured. Mean TP concentrations exceed the USEPA (1994) TP guideline of 0.10 mg L(-1) at 82% of the fields, and periodically concentrations more than 10 times, and occasionally more than 50 times higher than the guideline were found. The proportion of SRP in TP had a tendency to be higher when TP levels were high in coarse textured soils. In Nova Scotia, dairy manure is most often applied on permanent cover crops, which did not show as much P concentration variability as crop rotations. Daily or hourly observation of short-term increases in P concentrations related to the described factors would help to characterize the changes in P concentrations observed during frequent heavy drainage flow events.