Technical Abstract: Federal and state regulations are being promulgated under the Clean Air Act to reduce hazardous air emissions from livestock operations. Although much is known about air emissions from livestock operations in Europe, few data are available on emissions from livestock facilities in the USA and the management practices that may minimize these emissions. The objective of this study was to measure seasonal and diet impacts on ammonia emissions from experimental tie-stall dairy barns located in central Wisconsin. Four experimental chambers each housed four lactating Holstein dairy cows for three 28d trial periods corresponding to spring, early-fall and winter. A 4x4 Latin Square statistical design was used to evaluate four diets [corn silage (CS)- or alfalfa silage (AS)-based diets at low (LP) or high (HP) crude protein] in each chamber for a 4-d ammonia monitoring period. Average ammonia-N emissions during spring (18.8 g cow-1 d-1) were approximately twice the emissions recorded during early-fall (8.4 g cow-1 d-1) and three times greater than emissions during winter (6.7 g cow-1 d-1). Ammonia-N emissions accounted for approximately 1 to 3% of consumed feed N, 2 to 5% of excreted manure N, and 4 to 11% of manure ammonical N. Nighttime ammonia emissions were on average 30% lower than daytime emissions. Forage type did not impact ammonia emissions during winter or early-fall. Only during early-spring were ammonia emissions lower from chambers containing cows fed LP than from cows fed HP. Of the total chamber N inputs (feed and bedding), 93%, 91% and 95% were recovered in N outputs (milk, manure, body weight change, and ammonia N) during spring, early-fall and winter trials, respectively. Confidence in the accuracy of ammonia emission results was gained by the relatively high chamber N balances and favorable comparisons of study data with published relationships among the variables of feed N intake, milk urea N, manure N and urine N excretion, and ammonia emissions.