Glasses and other noncrystalline solids exhibit thermal and acoustic properties at low temperatures anomalously different from those found in crystalline solids, and with a remarkable degree of universality. Below a few kelvin, these universal properties have been successfully interpreted using the tunneling model, which has enjoyed (almost) unanimous recognition for decades. Here we present low-temperature specific-heat measurements of ultrastable glasses of indomethacin that clearly show the disappearance of the ubiquitous linear contribution traditionally ascribed to the existence of tunneling twolevel systems (TLS). When the ultrastable thin-film sample is thermally converted into a conventional glass, the material recovers a typical amount of TLS. This remarkable suppression of the TLS found in ultrastable glasses of indomethacin is argued to be due to their particular anisotropic and layered character, which strongly influences the dynamical network and may hinder isotropic interactions among low-energy defects, rather than to the thermodynamic stabilization itself. G lasses or amorphous solids are well known (1, 2) to exhibit thermal and acoustic properties very different from those of their crystalline counterparts. Even more strikingly, many of these properties are very similar for any glass, irrespective of the type of material, chemical bonding, etc. Hence the low-temperature properties of noncrystalline solids are said to exhibit a universal "glassy behavior." In particular, below 1−2 K the specific heat of glasses depends approximately linearly on temperature, C p ∝T, and the thermal conductivity almost quadratically, κ ∝ T 2 , in clear contrast with the cubic dependences successfully predicted by Debye theory for crystals. In addition, a broad maximum in C p /T 3 [originated from the so-called "boson peak" in the reduced vibrational density of states g(ω)/ω 2 ] is also typically observed in glasses around 3−10 K, as well as a universal plateau in the thermal conductivity κ(T) in the same temperature range (1, 2).Very soon after the seminal paper by Zeller and Pohl (1) in 1971, Phillips (3) and Anderson et al. (4) independently introduced the well-known standard tunneling model (TM). The fundamental idea of the TM is the ubiquitous existence of atoms or small groups of atoms in amorphous solids due to the intrinsic atomic disorder, which can perform quantum tunneling between two configurations of very similar energy, usually named twolevel systems (TLS). This simple model was able to account for the abovementioned thermal and acoustic anomalies of glasses below 1−2 K, and soon acquired unanimous recognition. Only very few authors (5) posed then criticisms against the standard TM, pointing out how improbable it was that a random ensemble of independent tunneling states would produce essentially the same universal constant for the thermal conductivity or the acoustic attenuation in any substance. Indeed, significant discrepancies with the TM below ∼100 mK have also been reported (6-9), in particular ...