I investigate the conditions upon which atmospheric absorption may participate to the observation of a diffuse
interstellar band (DIB), and the implications it would have. A necessary condition is that the spectrum of reddened stars
comprises a few percent of starlight forward scattered by the interstellar cloud on the line of sight. Reciprocally, this
scattered starlight could, in part, explain the complexity of the DIB spectrum and several observed DIB properties. It will
also affect the interstellar extinction curve and the value of the RV parameter.