Mixed commercial and residential land use is observed in most cities around the world. This is in contrast to a myriad of bid rent models, which predict that mixed land use does not occur. The main exception are the models by Fujita and Ogawa (1982) and Lucas and Rossi-Hansberg (2002) that predict the presence of a very restrictive type of mixed land use. The latter study derives the equilibrium distribution of residential and commercial land uses while allowing for endogenous agglomeration externalities. We extend this model by introducing a traffic congestion externality. We show that congestion induces a general type of mixed land use zone, which is comparable to the type of zone assumed in the model of Wheaton (2004). The interplay between these externalities is then demonstrated, as reduced congestion leads to commercial concentration and agglomeration gains.