BACKGROUND Clinicians measure left ventricular dimensions in dogs from both M-mode and two-dimensional images. Little information currently exists as to whether these two methods provide measurements similar enough to be interchangeable. METHODS The animals included in this study are 206 client-owned dogs: 68 healthy, 105 with myxomatous mitral valve disease, 33 with other cardiac or extracardiac disease. METHODS Investigators measured left ventricular diastolic and systolic dimensions from archived M-mode and two-dimensional images obtained from the right parasternal short-axis view. Agreement between the pairs of measurements was examined using limits of agreement (Bland-Altman) plots. RESULTS Left ventricular diastolic dimensions showed no fixed or proportional bias but did show heteroscedasticity. Ninety-five percent limits of agreement for normalized differences approximated ±10%; 95% of the absolute differences for any pair of measurements were <3.9 mm regardless of bodyweight and <2.7 mm for dogs <15 kg. Left ventricular systolic dimensions showed slight proportional bias, with two-dimensional measurements being progressively larger than M-mode measurements as ventricular size increased. Ninety-five percent limits of agreement for normalized differences approximated ±20%; 95% of the absolute differences for any pair of measurements were <4.6 mm regardless of bodyweight and <3.5 mm for dogs <15 kg. Mitral valve disease did not appreciably affect these findings. CONCLUSIONS Left ventricular internal dimensions in dogs with and without cardiac disease measured from two-dimensional right parasternal short-axis images are interchangeable with those measured from M-mode images using the same view.