This study establishes empirically a nonlinear relationship between hours worked per week and hourly wage growth: for workers who put in 48 hours per week or more, working 5 extra hours per week increases annual wage growth by about 1 percent. The average effect is zero when hours are below 48. This relationship is especially strong for young professional workers. I provide evidence in support of a model of promotions that combines higher skill-sensitivity of output in upper levels of the job ladder with worker heterogeneity. The results can be used to account for part of the gender wage gap.