We compared the sensitivity of two juvenile unionid mussels (Lampsilis cardium and Lampsilis higginsii) to ammonia in 96-h water-only and sediment tests by use of mortality and growth measurements. Twenty mussels were placed in chambers buried 2.5 cm into reference sediments to approximate pore-water exposure (sediment tests) or elevated above the bottom of the experimental units (water-only tests). In the sediment tests, a pH gradient existed between the overlying water (mean 8.0), sediment-water interface (mean 7.7), and 2.5 cm depth (mean 7.4). We assumed that mussels were exposed to ammonia in pore water and report effect concentrations in pore water, but if they were exposed to the higher pH water, more of the ammonia would be in the toxic un-ionized (NH 3) form. The only differences in toxicity and growth between mussel species occurred in some of the water-only tests. In sediment tests, median lethal concentrations (LC50s) ranged from 124 to 125 ??g NH3-N/L. In water-only tests, LC50s ranged from 157 to 372 ??g NH3-N/L. In sediment tests, median effective concentrations (EC50s based on growth) ranged from 30 to 32 ??g NH3-N/L. Juvenile mussels in the water-only tests grew poorly and did not exhibit a dose-response relation. These data demonstrate that growth is a sensitive and valuable endpoint for studies on ammonia toxicity with juvenile freshwater mussels and that growth should be measured via sediment tests. ?? 2007 SETAC.