Evidence suggests that, from birth, fathers treat sons differently than daughters in the U.S., as well as in developing countries. Fathers' time investments in children are one channel through which differential treatment by gender may affect children's outcomes. This paper uses data from the 2003 American Time Use Survey to explore three questions about paternal time in married two-parent families: Does the gender composition of his children affect the amount of time a father spends with them? If so, does the gender of the individual child have an additional effect? And is a girl advantaged or disadvantaged by the presence of brothers in spending time with her father? Father-level and child-level regressions examine the effects of gender composition and gender. Father fixed effects specifications show that gender is important within families as well as across families. The results show that fathers of boys invest more of their time in children than those with only daughters, and reduce their leisure time without children to do so. Boys get more of this time than girls in all-girl families and compared to their own sisters. To the extent that fathers' time affects children's outcomes, girls are at a disadvantage, especially girls in all-girl families. Girls with brothers do receive more of fathers' time than girls with only sisters, but this is primarily in television watching, so whether this advantages them is open to question.