Since the mid-1970's it has been apparent that giant stars of similar V and B-V within a “normal” globular cluster [i.e., one with a narrow giant branch in its color-magnitude diagram (CMD)] exhibit a perplexing range of strengths for such spectral features as CH, CN and NH. This complex subject has been reviewed by Kraft (1979), McClure (1979) and Freeman and Norris (1981). DDO photometry first revealed star-to-star differences of CN strengths at MV>+1, where observational confusion between asymptotic and first-ascent giant stars is removed (Hesser, Hartwick and McClure 1976, 1977; Hesser 1978). Subsequently, we have sought to place observational constraints on possible mechanisms by studying such questions as: At what MV's do spectral differences first become observable? Do spectral features other than those from CNO-based molecules vary from star-to-star? Can small temperature or gravity differences produce the observed ranges?