Dogs differ greatly in size, heart (HR) and breathing rates (BR). In addition, they have a clear Respiratory Sinus Arrhythmia (RSA) at rest. Therefore, better than any other mammalian species, dogs offer an opportunity to test whether resting RSA varies with body weight, HR or BR. Sequences of inter-beat-intervals (IBI, ms) a few-minutes long were collected in twenty-three resting dogs of different sizes, together with pneumograms. IBI variability was quantified by standard time-domain criteria. From beat-to-beat instantaneous heart rate (hR, beats/min), RSA was the difference between inspiratory peak (hR-peak) and expiratory trough (hR-trough), in percent of mean HR. RSA averaged 40.1 % ±4.5, or more than three times that of humans, with large inter-animal variability. On average, RSA contributed 38 % of the total IBI variability. RSA did not differ between sexes and did not correlate with body weight. It had modest negative correlations with HR (P < 0.05) and BR (P < 0.05), and a very strong negative correlation with hR-trough (P < 0.001). In two separate dogs, during panting, RSA was absent. In the transition from resting to panting, RSA continued like at rest for several breaths, despite the tachypnea, underlying the importance of central mechanisms in the origin of RSA. In conclusion, RSA in dogs is very large and explains less than half of their sinus arrhythmia. Rather than HR, BR or hR-peak, changes in the vago-sympathetic control, represented by hR-trough, are the most likely source of variability of RSA among subjects.