The particulate backscattering ratio ( $b_{bp}/b_p$ ) is a useful indicator of the angular scattering characteristics of natural waters. Recent studies have shown evidence both for and against significant spectral variability in $b_{bp}/b_p$ in the visible domain, but most show significant variability in its magnitude. We present results from a case study in which both backscattering and scattering coefficients were measured at nine wavelengths in a region of UK coastal waters where optical scattering is strongly influenced by inorganic particles and where a wide range of turbidities is found in a small geographic area. Using a new approach based on regression analysis of in situ signals, it is shown that, for this study site, most of the apparent variability in the magnitude of the backscattering ratio can be attributed to measurement uncertainties. Regression analysis suggests that $b_{bp}/b_p$ is wavelength dependent for these mineral-rich waters. This conclusion can only be avoided by positing the existence of undocumented, systematic, wavelength- dependent errors in backscattering measurements made by two independently calibrated sensors. These results are important for radiative transfer simulations in mineral-dominated waters where the backscattering ratio has often been assumed to be spectrally flat. Furthermore, spectral dependence also has profound implications for our understanding of the relationship between $b_{bp}/b_p$ and particle size distributions in coastal waters since the commonly assumed power-law distribution is associated with a spectrally flat particulate backscattering ratio for nonabsorbing particles.