OBJECTIVE: To determine correlations between time-domain and frequency-domain variables of heart rate variability (HRV) derived from 24-hour recordings obtained by use of an ambulatory electrocardiographic recorder (Holter monitor). ANIMALS: 59 overtly healthy Doberman Pinschers (41 without echocardiographic evidence of cardiomyopathy and 18 with precongestive heart failure attributable to cardiomyopathy). PROCEDURE: The HRV was analyzed from 24-hour recordings. Variables were calculated from the entire 24-hour recording as well as 4 user-selected time epochs. Comparisons were made for total power to SD of normal beat-to-normal-beat (NN) intervals (SDNN), ultra-low frequency power to SD of the means of NN intervals, low-frequency power and very-low-frequency power to mean of the SD of NN intervals, and high-frequency (HF) power to the root mean square successive difference of NN intervals (RMSSD) and percentage of NN intervals that varied from the previous NN interval by > 50 milliseconds (PNN50). RESULTS: 58 of 66 (88%) comparisons revealed significant values, indicating that relationships between variables were not random (r > 0.7 in 41 of 66 [62%)) comparisons). Strong correlations (r > 0.8) were found between the square root of total power and SDNN and between HF power and RMSSD. CONCLUSIONS AND CLINICAL RELEVANCE: Time-domain surrogates for variables of frequency-domain analysis variables that correlated in the dogs reported here are the same ones that reportedly correlate in humans. When 24-hour recordings obtained by use of a Holter monitor are used to calculate HRV, SDNN and total power as well as RMSSD and HF power are interchangeable.