WiFi-enabled buses and stops may form the backbone of a metropolitan delay tolerant network, that exploits nearby communications, temporary storage at stops, and predictable bus mobility to deliver non-real time information. This paper studies the routing problem in such a network. Assuming the bus schedule is known, we maximize the delivery probability by a given deadline for each packet. Our approach takes the randomness into account, which stems from road traffic conditions, passengers boarding and alighting, and other factors that affect the bus mobility. In this sense, this paper is one of the first to tackle quasi-deterministic mobility scenarios. We propose a simple stochastic model for bus arrivals at stops, supported by a study of real-life traces collected in a large urban network. A succinct graph representation of this model allows us to devise an optimal (under our model) single-copy routing algorithm and then extend it to cases where several copies of the same data are permitted. Through an extensive simulation study, we compare the optimal routing algorithm with three other approaches: minimizing the expected traversal time over our graph, maximizing the delivery probability over an infinite time-horizon, and a recently proposed heuristic based on bus frequencies. We show that our optimal algorithm shows the best performance, but it essentially reduces to minimizing the expected traversal time. When transmissions fail frequently (more than half of the times), the algorithm behaves similarly to a heuristic that maximizes the delivery probability over an infinite time-horizon. For reliable transmissions and values of deadlines close to the expected delivery time, the multi-copy extension requires only 10 copies to almost reach the performance of the costly flooding approach. I. INTRODUCTION We consider an opportunistic data network formed by (some) buses and bus stops in a town equipped with wireless devices, e.g. based on WiFi technologies, like in DieselNet [1]. Most of the stops act as disconnected relay nodes (the throwboxes in [2]), and a few of them are also connected to the Internet. Data are delivered across the town following the store-carry-forward network paradigm [3], based on multihop communication in which two nodes may exchange data