There is a style of thinking and approaching a problem to be learned from every good teacher. In this presentation, an attempt is made at giving an objective account of two subjects that I worked on as W. Brutsaert’s graduate student (radiation and scalar similarity), but highlighting the personal influence and style that I came to know and that helped to illuminate the way. Longwave radiation interacts with temperature fluctuations, by dissipating them in a process which has similarities to molecular effects, but which has a very different spectral behavior. The pioneering work of Brutsaert on the subject served as a basis to extend the original ideas to the spectral domain and, by doing this, identifying some new governing dimensionless parameters that can help to determine the extent to which radiative effects are important in the stable surface layer. A full comparison of the spectral approach with Brutsaert’s early formulation for the radiative dissipation had not been made, however, and some simple spectral shapes are used to show how this can be done. Moving to scalar similarity, the motivations for questioning the validity of Reynolds’ analogy between any two scalars are briefly reviewed. A lot of the physical basis to assume scalar similarity rests on the validity of a simple equilibrium between gradient production and molecular dissipation of scalar variance, and on the ratio between their molecular diffusivities being of order 1. As the essence of these results is revisited, they are recast in a way that gives them more generality, away from the original emphasis on temperature-humidity similarity and towards (hopefully) applicability to such important atmospheric components as carbon dioxide and methane. As in the case of radiation, some further results are also commented: the subject continued to attract the attention of other researchers from the Brutsaert School, and such topics as chessboard variability and its prediction of the importance of the active/passive role of scalars; experimental evidence on differences between the Monin-Obukhov Similarity functions for temperature and humidity; and the similarity hypothesis application to the bandpass eddy-covariance method are briefly reviewed. Finally, some on-going research is reported on scalar similarity that allows diagnosing the validity of the aforementioned equilibrium between gradient production and molecular dissipation with dimensionless parameters, and to further the understanding of the effects of molecular diffusivities and the active/passive role of scalars by means of numerical simulations of a chaotic system.