Melts of linear polyethylenes (HDPE) have broad retardation and relaxation spectra that can be determined by combining oscillatory and long duration creep/creep recovery tests. For a commercial HDPE the necessary sample preparation (stabilization), testing (at 170 degrees C), and data conversion are described. For long time testing incomplete creep/creep recovery is preferred because the strain is kept within the linear range and reaches a final, constant value. The measured recoverable compliance, J(e)(o)(t), is converted into a discrete retardation spectrum {L-i, tau(i)＇}. For the short time range the dynamic moduli are measured and converted into the dynamic compliances and finally the retardation spectrum L(tau＇). Both spectra overlap and form a combined spectrum over 7.5 decades of retardation time tau＇. From this the dynamic compliances are recalculated and converted into the dynamic moduli. These are used for computing the weighted relaxation spectrum, tau H(tau), ranging from tau = 0.01-2.5 x 10(5) s. tau H(tau) has a maximum at tau = 10(4) s, explaining the difficulties in characterizing HDPE melts. The viscosity functions eta(t) and the first normal stress functions Psi(1)(t) are measured at shear rates between 10(-3) and 10 s(-1) and compared with their limits for zero shear rate, eta(0)(t) and Psi(1)(0)(t), respectively, predicted from H(tau). The measured results are close to these predictions, but only at short times, such that the equilibrium values for t --> infinity, eta(0) and Psi(1,0), are never reached in such tests. In contrast, creep/creep recovery reveals eta(0) = 113.5 kPas, the thigh!) linear equilibrium compliance J(e) = 0.05 Pa-1, and 2 eta(0)(2)J(e) = Psi(1,0) = 1.3 x 10(9) Pa s(2). The latter value coincides well with the one calculated from the second moment of H(tau).