The strong coupling of superconductivity to the orthorhombic distortion in YBa 2 Cu 3 O 7-δ makes possible an analysis of the superconducting fluctuations without the necessity of subtracting any background. The present high-resolution capacitance dilatometry data unambiguously demonstrate the existence of critical, instead of Gaussian, fluctuations over a wide temperature region (± 10 K) around T c . The values of the amplitude ratio A A + − = − / . . 0 9 11 and the leading scaling exponent | | . α ≤ 0 018 , determined via a leastsquares fit of the data, are consistent with the 3D-XY universality class. Small deviations from pure 3D-XY behavior are discussed.PACS numbers: 74.72.-h, 64.60.Fr, 65.70.+y 2 The superconducting transition in classical superconductors is described surprisingly well by mean-field theory. This is due to the large number of Cooper pairs in the coherence volume and is really an exceptional case for second-order transitions. In contrast, the small coherence lengths, high transition temperatures and quasi-2D nature of high-temperature superconductors (HTSC) greatly enlarge the temperature region in which fluctuations of the order parameter are important, and fluctuations in HTSCs have been observed in many types of experiments, e.g. specific heat [1-6], thermal expansion [7,8], resistivity [2,9], penetration depth [10,11] and magnetization [3,6,12] measurements. A large effort has gone into quantifying these effects, because a detailed analysis of the fluctuations can provide important information regarding the dimensionality and the order parameter of the superconducting state. The superconducting transition in the simplest case is expected to belong to the threedimensional XY (3D-XY) universality class with a two-component order parameter, e.g. pure d x y 2 2 − or s-wave. The most plausible order parameter that is most consistent with a large number of experiments on YBa 2 Cu 3 O 7-δ , on the other hand, consists of a real mixture of s + and d x y 2 2 −[13]. Here, we focus on the transition in the absence of an applied magnetic field, but note that many interesting phenomena, e.g. vortex fluctuations leading to first-order vortex melting, occur when a field is applied.The fluctuation signal in most experiments is superimposed upon a large background, and it has proven quite difficult to determine whether the HTSCs are really in the critical regime or whether the fluctuations are better described by first-order corrections to mean-field theory, i.e. Gaussian fluctuations. For example, in specific heat measurements of YBa 2 Cu 3 O 7-δ the fluctuation component is at most 5 % of the large phonon background. By slightly adjusting the unknown phonon background, both critical and 3D-Gaussian type models can be fitted quite well [1,3,[4][5][6]. The fitting residuals of the most recent specific heat measurements are somewhat smaller for the critical than for the Gaussian fits [3,4,6,14], although quite different backgrounds, with and without mean-field contributions, were used for the critical ana...