We study end-to-end multihop transmission of packets in a Poisson field of interferers
with guaranteed packet delivery.  In contrast to existing information theoretic work on
multihop throughput and capacity, we explicitly consider both interference and noise.
Compared to the large body of work on end-to-end transport capacity scaling laws, we are
able to compute several transport capacity results in closed-form.  The main simplifying
assumptions that allow this are (i) randomly placed interfering nodes and (ii)
deterministically placed relay nodes.  Each packet is retransmitted over $n_h$ hops until
it is successfully received, hence each hop has a geometric distribution on the number of
transmissions.  By finding the optimal number of hops $n_h^*$ for relevant integer path
loss exponents and the success probability of each hop, the transport capacity for these
cases is obtainable in closed-form and we show it follows a $\Theta(\sqrt{n})$ scaling law.
That is, uncoordinated multihop communication with retransmissions is order optimal for a
wireless network, and the exact preconstants can be computed.