Many scholars of the arts of Africa and the African diaspora have noted the close connection of dance and music. African and diasporic musicians play music as movement, and dancers feel movement as music. In some cases, the mutuality of dance and music is tightly choreographed; in others, it is improvised yet still somehow achieved. However, there has been little published that analyzes how such mutuality is accomplished. The following article analyzes dancer-drummer interaction in bele, a performance art from Martinique in the French West Indies. It examines both interactive practices that are made explicit by performers and some underlying conditions of those practices that remain tacit. It may be considered a study in interactive and improvisational competence. American writers, both black and white, have often looked to the dance-music link to validate African-American expression. Among scholars, attention to the dance-music link began at least as early as Herskovits' provocative, although brief, remarks on the cultural imponderables of musical style and motor habits (Herskovits 1966, 59; Herskovits 1990, 265). Herskovits' attention to these subliminal, tenacious (1966, 13), resilient (23), and resistant (60) elements of culture led diasporic studies away from the narrow focus on static retentions of strictly delimited Africanisms, as demanded by more conservative scholars such as M. G. Smith (1957), and toward a broader appreciation of the syncretisms and reinterpretations underlying even apparently acculturated phenomena (Herskovits 1966, 57, 60). Increasingly, diasporic studies has emphasized the transformational resilience of black identity and culture. Current writers recognize the stereotyping, romanticization, commercialization, and exploitation of black culture and yet celebrate it as a changing same (Gilroy 1991). A number of works have continued the project of validating black expression by insisting on the moral, symbolic, and spiritual importance of African dance (e.g., Asante 1990; Nketia 1974; Thompson 1966; Thompson 1974; Wilson 1985), by using African-American aesthetics to criticize orthodox musicological assumptions about musical meaning (Keil 1966), or by recognizing the tremendous role that dance and the body have played in the construction, stereotyping, repression, and resilience of black identity (e.g., Hazzard-Gordon 1983; McClary and Walser 1994). However, all of these writings treat the dance-music link more or less as a given and describe it only in general terms. An initial attempt to define the dance-music connection in greater detail may be found in Waterman's concept of a sense: the kinesthetic perception of regular beats by both dancers and musicians, whether or not a given beat is actually expressed in movement or sound (Waterman 1952, 211). Waterman introduced four aesthetic concepts (dominance of percussion, polymeter, off-beat phrasing of melodic accents, and overlapping call-and-response patterns) that have proved remarkably durable, being cited or paraphrased by numerous other authors. Waterman's ideas have appeal because they describe processes of music making rather than the static retention of specific instruments, texts, rhythms, dance steps, and so forth. Thus, they help not only to describe black music but to explain its transformations and resilience. But Waterman's concept of a metronome sense remains much less widely used than his four aesthetic principles. How does one demonstrate that such a thing exists or investigate its workings? If a metronome sense exists, it would be an underlying competence, a mostly automatic and tacit aspect of perception. It would help explain the perceptual foundation of polyrhythm; it would enable performers and listeners to make sense of varied, complex, asymmetrical patterns of notes and movements by relating them to one another via the perception of an underlying layer of regularity to which they also relate. …