Despite their use in cardiac risk stratification, the physiological meaning of nonlinear heart rate variability (HRV) measures is not well understood. The aim of this study was to elucidate effects of breathing frequency, tidal volume, and light exercise on nonlinear HRV and to determine associations with traditional HRV indices. R-R intervals, blood pressure, minute ventilation, breathing frequency, and respiratory gas concentrations were measured in 24 healthy male volunteers during 7 conditions: voluntary breathing at rest, and metronome guided breathing (0.1, 0.2 and 0.4 Hz) during rest, and cycling, respectively. The effect of physical load was significant for heart rate (HR; p < 0.001) and traditional HRV indices SDNN, RMSSD, lnLFP, and lnHFP (p < 0.01 for all). It approached significance for sample entropy (SampEn) and correlation dimension (D2) (p < 0.1 for both), while HRV detrended fluctuation analysis (DFA) measures DFAalpha1 and DFAalpha2 were not affected by load condition. Breathing did not affect HR but affected all traditional HRV measures. D2 was not affected by breathing; DFAalpha1 was moderately affected by breathing; and DFAalpha2, approximate entropy (ApEn), and SampEn were strongly affected by breathing. DFAalpha1 was strongly increased, whereas DFAalpha2, ApEn, and SampEn were decreased by slow breathing. No interaction effect of load and breathing pattern was evident. Correlations to traditional HRV indices were modest (r from -0.14 to -0.67, p < 0.05 to <0.01). In conclusion, while light exercise does not significantly affect short-time HRV nonlinear indices, respiratory activity has to be considered as a potential contributor at rest and during light dynamic exercise.