Our abstractions---as with most high-level mobile ad hoc network application specifications---make strong assumptions about the behavior of the communication medium. Accordingly, we are investigating low-level algorithms that can provide behavior that matches these assumptions. Specifically, we have studied the problem of solving single-hop consensus, using receiver-side collision detection, in an environment with non-uniform message loss and a dynamic set of participants. We have also begun work on developing communication and coordination primitives that are resistant to malicious interference.