Vehicular Delay-Tolerant Networks use the delaytolerant architecture and protocols to overcome the disruptions in network connectivity. These concepts help in cases where the network is sparse or with large variations in density or there is no end-to-end connectivity, by providing a communications solution for non real-time applications. This paper presents data fragmentation techniques to optimize the efficiency of data delivery for the case of the short node contacts that characterize vehicle networks. The techniques were tested in a laboratory environment with portable digital assistants and Lego Mindstorm NXT robotic cars. If no fragmentation is used, only small messages are successfully transferred. Proactive fragmentation fragments messages to a predefined size in the source node. Reactive fragmentation adjusts the fragment sizes to the real duration of the contact when it is broken. Reactive fragmentation showed a good efficiency in adapting the fragmentation in real time to the contact duration. Proactive fragmentation can perform slightly better if the fragment sizes are carefully chosen as it requires less processing. As this choice is difficult, reactive fragmentation is more practical to use.