We estimated total arterial compliance (C) in eight anesthetized mongrel dogs with (i) the area method (AM), (ii) the pulse pressure method (PPM), and (iii) the stroke volume-to-pulse pressure ratio (SV/PP). Average compliance was C(AM)=1.1+/-0.73 ml mm Hg(-1) using AM; C(PPM)=0.60+/-0.31 ml mm Hg(-1) using PPM and C(SV/PP)=0.87+/-0.49 ml mm Hg(-1) using SV/PP. Mean aortic pressure was 64+/-23 mm Hg. The overall agreement between C(AM) and C(PPM) was relatively poor (C(AM)=0.15+/-1.61 C(PPM); r2=0.48), with a consistent overestimation of the area method with respect to the pulse pressure method. There was a significant correlation (r=-0.78) between the relative difference between PPM and AM, and the modulus of the first harmonic of the wave reflection coefficient [gamma] which was low in our dog population (0.37+/-0.18). SV/PP overestimated PPM, but both methods were highly correlated (C(SV/PP)=0.06+/-1.60C(PPM); r2=0.97). C(SV/PP) and C(AM) were similar only for [gamma]>0.4. The effect of isolated changes of [gamma] on PPM, AM, and SV/PP was studied using the linear wave separation technique. The area method appeared very sensitive to the wave reflection intensity. For low reflection coefficients, the diastolic wave profile was flattened and compliance was overestimated. PPM and SV/PP were relatively independent of [gamma] and remained even applicable for [gamma]=0. We believe that the pulse pressure method is the most consistent method for the estimation of total arterial compliance in hemodynamic conditions characterized by a low wave reflection intensity.