IntroductionModern development, consumption and waste have strained the environment to the point of crisis -and music is part of the problem. Consider the poisonous petrochemicals used to manufacture LPs, the non-biodegradable plastics in CDs, the energy-guzzling server farms that power streaming MP3s, and the toxic graveyards of obsolete consumer electronics around the world. But these earthy and potentially ugly material realities typically go unnoticed in musical discourse, probably because they clash with a longstanding but mistaken belief that music is somehow an immaterial phenomenon.Indeed, even though we are accustomed to the accusation that certain types of music are 'disposable', we rarely think about what happens to recordings when they are actually disposed of. And while commonsense tells us that the 'music industry' is in the business of making records, we rarely think about what those recordings are actually made of. Correspondingly, the environmental impact of recorded sound is almost completely 2 unaccounted for in cultural studies of music and media, which treat the musical commodity principally as a product of musical labour, an object of capitalist exchange, or a text for audience consumption. Questions about the actual material composition and decomposition of the musical commodity -about what happens to music before production and after consumption -have been largely ignored. Yet without confronting the relationship between music's cultural and economic value, on the one hand, and its environmental cost, on the other, the field of music studies forfeits the opportunity for a fully interdisciplinary engagement with -and a holistic democratic ethics of -its object of study. One step toward achieving that goal is to formulate a political ecology of the evolving relationship between popular music and sound technology since 1900.This article inscribes a history of recorded music in three main forms of materiality:shellac (1900-1950), plastic (1950-2000) and data (2000-present). These three broad phases in the material history of recorded music correspond to this article's three main sections. They also encompass the five most prevalent recording formats since 1900: 78s (shellac); LPs, cassettes and CDs (plastic); and MP3s (data). I subject these three materials to a double-headed investigation, examining the dynamics of manufacture and obsolescence that mark the shifts between them. As a backdrop to the sections on shellac,