Four sources (A, B, C, D) of magnesium oxide were assessed for both in vivo and in vitro estimates of ruminal availability. Samples were selected from a population of 41 samples based on laboratory screening. In vivo assessment consisted of 4 lactating cows (250 ± 48 DIM, producing 34 ± 3 kg/day of milk) in a 4 × 4 Latin square design of cows and periods. Cows were individually fed a common diet restricted to 90% of pre-experiment ad libitum intake to maintain consistent intakes throughout the experiment (overall average DM intake 26.1 kg/cow/day). Cows were dosed intraruminally with 40 g Mg per day delivered in 2 equal doses at a.m. and p.m. feedings for 3 d before sampling blood and ruminal fluid. Serum Mg concentrations 3 h post-dosing (overall average 1.35 mM) were unaffected (P = 0.35) by source. Blank-corrected weighted mean ruminal fluid Mg concentrations (sampled at various times up to 12 h post-dosing) differed (P = 0.029) among magnesium oxide sources, averaging 2.29 mM (A), 3.04 mM (B), 2.37 mM (C), and 4.19 mM (D). In vitro assessment consisted of a pH-stat titration procedure whereby the acid (HCl) consumed due to magnesium oxide solubilization to maintain pH 6.5 in a suspension of magnesium oxide in distilled water was measured with continuous stirring over 12 h. Time-course titration curves (mmol acid consumed/g of magnesium oxide) were described using a generalized Michaelis-Menten equation, Y_t = Y_max × t^c/(K^c + t^c), where Y_t is the acid consumed (mmol acid/g magnesium oxide) at any time (t, hours), Y_max is the asymptotic value of Y at infinite time, K is the time for one-half maximal solubilization and c is a dimensionless shape parameter determining the time to reach near asymptotic acid consumption. Both K (P = 0.006) and c (P = 0.035) but not Y_max (average = 105.3, P = 0.22) differed due to source. Values of K and c were 17.2 and 1.10 (A), 6.51 and 1.05 (B), 29.2 and 1.26 (C), and 0.83 and 1.44 (D), respectively. Results indicated that the kinetics of in vitro magnesium oxide solubilization may be reflective of in vivo ruminal magnesium oxide availability.