Six mature Quarter Horse geldings (age, 11.5 ± 4.7 years; body weight [BW], 526 ± 9.2 kg) were used in a replicated 3 × 3 Latin square design to determine the effects of three dietary lysine (Lys) levels on nitrogen (N) retention and plasma amino acid (AA) concentrations. The geldings were fed a basal diet of concentrate and Bermuda grass hay supplemented with synthetic essential amino acids (EAA) to meet estimated requirements for EAA. Geldings were fed one of three dietary treatments: (1) a basal diet deficient in Lys (L−; 0.027 g of Lys•kg−1 BW•d−1); (2) a basal diet supplemented with synthetic Lys to meet National Research Council (NRC; 2007) requirements (L+; 0.036 g of Lys•kg−1 BW•d−1); or (3) a basal diet supplemented at twice the recommended Lys requirement (2 × L; 0.070 g Lys•kg−1 BW•d−1). Horses fed the 2 × L diet had higher N intakes (P = .0056) than horses fed either the L− or L+ diet. However N retention (P = .63) was not different between treatments. Plasma Lys was greater (P < .0001) in 2 × L than L− and L+ diets. Plasma threonine (Thr; P < .01), methionine (Met; P = .03), and total plasma non-EAA (P < .05) concentrations decreased as dietary Lys increased. These results suggest N retention is not a good response criterion for evaluating the AA requirements of mature horses. However, plasma AA data indicated more efficient use of Thr and Met when horses were fed dietary Lys in concentrations greater than the current requirement recommended by the Nutrient Requirements of Horses (National Research Council, 2007).