This paper explores whether and why a maternal "child penalty" to earnings would emerge even without changes in employment and hours worked. Using a matched event study design, we trace monthly changes in determinants of wages (job performance, human capital accumulation, and promotions). Data come from a usefully unusual setting with required multiyear employment and detailed personnel data: the United States Marine Corps. Mothers' job performance initially declines, and gaps in promotion grow through 24 months postbirth. Fathers' physical fitness performance drops somewhat but recovers. These patterns lead mothers to earn relatively lower wages, even absent changes in employment postbirth.