Particulate nitrogen (PN) and chlorophyll a (Chla) were measured in the northern reach of San Francisco Bay throughout 1980. The PN values were calculated as the differences between unfiltered and filtered (0·4 μm) samples analyzed using the UV-catalyzed peroxide digestion method. The Chla values were measured spectrophotometrically, with corrections made for phaeopigments. The plot of all PNChla data was found to be non-linear, and the concentration of suspended particulate matter (SPM) was found to be the best selector for linear subsets of the data. The best-fit slopes of PNChla plots, as determined by linear regression (model II), were interpreted to be the N: Chla ratios of phytoplankton. The Y-intercepts of the regression lines were considered to represent easily-oxidizable detrital nitrogen (EDN). In clear water ( < 10 mg l⁻¹ SPM), the N: Chla ratio was 1·07 μg-at N per μg Chla. It decreased to 0·60 in the 10–18 mg l⁻¹ range and averaged 0·31 in the remaining four ranges (18–35, 35–65, 65–155, and 155–470 mg l⁻¹). The EDN values were less than 1 μg-at N l⁻¹ in the clear water and increased monotonically to almost 12 μg-at N l⁻¹ in the highest SPM range. The N: Chla ratios for the four highest SPM ranges agree well with data for phytoplankton in light-limited cultures. In these ranges, phytoplankton-N averaged only 20% of the PN, while EDN averaged 39% and refractory-N 41%.