ECGs recorded from dogs show characteristic morphology and changes in morphology with various disease states. These changes are determined by comparing individual recordings to reference ranges established from recordings obtained from normal dogs in right lateral (RL) recumbency. Using these reference ranges for ECGs recorded from dogs in other positions may not be valid. We compared ECG complexes from 39 normal dogs obtained in RL, left lateral (LL), and standing (ST) body positions. ECGs from dogs in ST position showed increased Q-wave and R-wave amplitudes in leads I and II, increased R-wave and S-wave amplitudes in leads aVR and aVL, and decreased R-wave and S-wave amplitudes in lead III when compared with recordings obtained in RL position. ECGs from dogs in LL position showed increased R-wave amplitude in leads II, III, and aVF and S-wave amplitude in lead aVL but decreased R-wave amplitude in lead aVR when compared with recordings obtained in RL position. The mean electrical axis (MEA) shifted to the left in ST position but remained within the normal range in LL position. We determined that both a change in the relative position of the recording electrodes with respect to the heart as well as a change in intrathoracic cardiac position contributed to these changes. P-wave amplitude, P-R and S-T intervals, and QRS complex durations remained unaltered by changes in body position. Our findings indicate that ECGs of dogs recorded in RL, LL, and ST positions yield dramatically different results, and investigators should use position-specific reference ranges to minimize potential misinterpretation of ECG results.