This study investigates how principals in a large US urban school district responded to two different superintendents who employed contrasting leadership styles and utilised divergent organisational schemes. We originally conducted interviews with principals in 2007, when the district's superintendent asserted fierce performance demands and limited principals' site-based discretion in favour of protecting and exerting central office power. We conducted interviews again in 2013 after a new superintendent had relaxed school test score expectations and distributed the central office's previously tight, centralised control into largely self-directing sub-regions. Our findings demonstrate that superintendent change noticeably affected how principals understood and encountered accountability, autonomy and stress. To help make sense of our findings, we employ a three-part conceptual framework drawn from the study of educational leadership. We conclude by considering implications, including the notion that unrelenting stress has become a permanent part of the modern urban US principalship.This study situates at the nexus of district and building leadership in order to provide particular attention to the lived experiences of school-based leaders. In doing so, we build on previous research that has examined the work lives and attitudes of principals (Leithwood et al. 2004;Mintrop 2012;Shipps 2012;Spillane and Hunt 2010). Specifically, we investigate how principals responded to two different superintendents (chief school system leaders) in a large urban school district in the Southeastern United States. In 2007, the superintendent, Dr Morgan (a pseudonym), exhibited an aggressive, top-down management style in which he demanded immediate results and exerted constant accountability performance pressure regarding school aggregate test scores. In 2008, a new superintendent, Mr Davis (a pseudonym), took over the district. By 2013, he had established greater system-wide autonomy and deemphasised using aggregate school test scores as the predominant determinant of principal performance. Symbolic of the differences between the two superintendents' approaches to principals, Dr Morgan was well known for either demoting principals or transferring them to a different school two or three years after they took their positions. Under Mr Davis, conversely, ensuring principal stability became a stated systemic goal, and the district introduced a modest financial incentive plan intended to reward principals who stayed at their schools for at least five years.