We question recent studies invoking the existence of a traditional ‘logarithmic surface layer’, or log layer, in the boundary layer of the rapidly rotating core of a hurricane. One such study argues that boundary-layer parametrization schemes that do not include a log layer are ‘badly flawed’. Another study assumes the existence of a log layer to infer drag coefficients at hurricane wind speeds. We provide theoretical reasoning supported by observational evidence as to why significant departures from the normally assumed logarithmic layer might be expected, questioning its use in the inference of the drag coefficient at high wind speeds and laying bare suggestions that hurricane models using boundary-layer schemes that do not represent the log layer should not be used. The ramifications of these findings for hurricane modelling are discussed. Finally, we draw attention to a study examining a range of boundary-layer schemes demonstrating that a recently articulated boundary-layer spin-up mechanism transcends the presence of a log layer.