This paper uses a version of theBergstrom/Goodman median voter model toexamine whether there are aspects ofpublicness to public school music. It alsoexamines whether the provision of publicschool music is an example of the abilityof small groups to seek benefits forthemselves at the expense of larger groups. School music programs provide a uniquesetting in which it is possible to separatethe demand for school music programs by theconsumers of school music (the privatebeneficiaries) from the demand for schoolmusic programs by those who do not directlyconsume the product. If those who are notdirect consumers of music programs appearto value school music programs as a publicgood, their valuation cannot be confusedwith a jointly consumed private good, orconsumption of a private good with scaleeconomies, because they are not themselvesconsuming it as a private good. Using datafrom the 1987 Schools and Staffing Survey,the results from a logarithmic, normative,demand model show that (parents of)nonmusic students appear to value musicprograms as a public good. Once it isprovided for music students, additionalproportions of nonmusic students do notdemand additional music classes. In thatmodel, music students appear to value musicclasses as a private good. In thepositive, additive, demand model, moremusic students mean that more music classeswill be provided. Further, since music isvalued as a non-subtractable public goodwhose costs can easily be spread,increasing numbers of non-music studentsalso result in (slightly) more musicclasses.