Fig. 5. Tile drain E. coli and C. perfringens, in cfu per 15 min, over the study period (JD > 295). Data were averaged by treatment and by tile hydrograph event (or baseflow sampling time). First-order exponential decay models (Y = predicted mass load, x = day) were fit to the E. coli data (model significant at p < 0.1); but models for C. perfringens data were not significant (p > 0.1). Error bars represent standard error.