Given a bipartite connected finite graph G =(V, E) and a vertex upsilon(0)
epsilon V, we consider a uniform probability measure on the set of graph ho
momorphisms f: V --> Z satisfying f(upsilon(0)) = 0, This measure can be vi
ewed as a G-indexed random walk on Z, generalizing both the usual time-inde
xed random walk and tree-indexed random walk, Several general inequalities
for the G-indexed random walk are derived, including an upper bound on fluc
tuations implying that the distance d(f(u), f(upsilon)) between f(u) and f(
upsilon) is stochastically dominated by the distance to O of a simple rando
m walk on Z having run for d(u,upsilon) steps. Various special cases are st
udied, For instance, when G is an n-level regular tree with all vertices on
the last level wired to an additional single vertex. we show that the expe
cted range of the walk is O(log n). This result can also be rephrased as a
statement about conditional branching random walk. To prove it, a power-typ
e Pascal triangle is introduced and exploited. (C) 2000 Academic Press.