This ethnographic study of teaching and learning in urban high school science classes investigates the ways in which teachers and students talk, gesture, and use space and time in interaction rituals. In situations where teachers coteach as a means of learning to teach in inner-city schools, successful teacher-teacher collaborations are characterized by prosodic expressions that converge over time and adapt to match the prosodic parameters of students' talk. In these situations our ethnographic data provide evidence of solidarity and positive emotions among the teachers and also between students and teachers. Unsuccessful collaborations are associated with considerable differences in pitch between consecutive speakers participating in turns-at-talk, these being related to the production of negative emotions and conflicts at longer time scales. Situational conflicts are co-expressed by increases in pitch levels, speech intensities, and speech rates; and conflict resolution is accelerated by the coordination of pitch levels. Our study therefore suggests that prosodic alignment and misalignment are resources that are pragmatically deployed to manage face-to-face interactions that have solidarity and conflict as their longer-term outcomes. (Contains 2 footnotes and 9 figures)