We reinterpret the scanning-probe Raman spectra shown in the above paper [C.C. Neacsu et al., Phys. Rev. B 73, 193406 (2006)] and compare it to a variety of single-molecule surface-enhanced Raman (SER) studies. The observed blinking behaviour and spectral features must be attributed to carbon contaminations rather than to Malachite Green (MG) single molecules, because, under the given experimental conditions, the extremely high field enhancement of 5 x 10 9 will inevitably lead to a quick (photo)decomposition of the adsorbate.PACS numbers: 68.37. Uv, 33.20.Fb In a recent paper, 1 Neacsu et al. present an AFMtip-enhanced Raman (TER) study on Malachite Green (MG) at gold films. They analyze near-field (tip approx. 5 nm above sample surface) and far-field (tip several 100 nm above sample surface) spectra of a Au surface covered with MG at low adsorbate concentration in correlation with a DFT calculation as well as of the clean Au surface. In a time series of 100 AFM-TER spectra (1 s integration time per spectrum) for a submonolayer MG surface coverage, spectral diffusion is observed and interpreted as characteristic single-molecule (SM) behaviour due to "random surface diffusion of MG in and out of the near-field-confined surface area under the tip, facilitated by the thin most likely liquid water layer on the gold surface". A Raman enhancement of up to 5 x 10 9 is derived from comparison of tip-enhanced versus far-field response of the same surface monolayer.Figures 2b and 2c in Ref. 1 show a near-field (tip approached) and a far-field (no tip) spectrum of MG on Au, respectively. The far-field spectrum exhibits the typical Raman features of MG which can be ssigned according to Lueck et al. 2 ) and a DFT calculation (Fig. 2c in Ref. 1), but the near-field spectrum does not resemble neither the far-field nor the DFT spectrum.The authors state that "the pronounced spectral difference between the tip-enhanced and the far-field response resembles the observation frequently made in SERS" (surface-enhanced Raman spectroscopy) and claim that "this characteristical difference is the result of the strong optical field localization, and related to different selection rules for the tip-scattered response, akin to SERS". Different selection rules cannot be held responsible for the so remarkable differences between near-field and farfield spectra presented by Neacsu et al. Note in this context that SER as well as TER spectra usually strongly resemble the far-field spectra of the investigated species in the band positions, in a way that the unambiguous identification of the molecule is always possible. In fact, a comparison of the Raman bands found in literature for MG in water (far-field, Ref. 2) and MG on silver colloids (near-field, Ref.3) does not reveal any band displacements larger than ± 3 cm −1 , and also the relative band intensities are similar. Variations in band intensities may occur according to surface selection rules (changes in the polarizability perpendicularly to the surface, i.e. parallel to the incident...