Near-infrared diffuse reflectance sensing (NIRS) of soils has been the object of considerable interest and research in the last few years. This has been motivated by the prospect that this method seems to provide a cheap, convenient alternative to conventional, time-consuming methods for the measurement of a wide range of soil parameters. In particular, various authors have advocated that NIRS could be used to measure rapidly and non-destructively the concentration of trace metals in surface soils. Correlation analyses between NIRS spectra and trace metal concentration have yielded inconclusive results to date, suggesting that trace metal concentration may belong to a class of “tertiary” soil parameters, linked to NIRS spectra through “surrogate”, or indirect, correlations, involving some other primary or secondary parameter like clay or organic matter content, to which NIRS spectra are very sensitive. To assess the validity of this surrogate correlation hypothesis in the case of trace metals, experiments were carried out with soil samples varying only in the amount of trace metals they contain. Field-aged Hudson and Arkport soil pots spiked with Cu and Zn, freshly spiked samples of the same soils, and samples of a metalliferous peat soil from Western New York naturally rich in Cd and Zn were subjected to NIRS under laboratory conditions. Detailed analysis indicates that the NIR spectrum is sensitive to sample handling, including the orientation of the samples in the NIRS instrument, but that, at the same time, there is no discernable effect of the presence of trace metals on any part of the NIR spectrum. These results provide strong experimental support to the hypothesis of “surrogate” correlation for trace metals, and indicate that trace metals, even in severely contaminated soils, should not interfere with the NIR sensing of primary or secondary parameters, like organic matter content. Further work is needed to determine if this feature of NIR spectra extends to other soil chemical parameters.