Anomalous rotation of molecules with applications to the temperature of the upper atmosphere, and a note on maximizing the ratio of "signal to noise" in spectrophotometry of weak light sources