We discuss higher order terms in the effective actions of 10-dimension
al SO(32) heterotic and type I superstrings from the point of view of
duality between the two theories. Some of these terms do not receive h
igher-loop corrections being related by supersymmetry to 'anomaly-canc
elling' terms, while others (like R(4) tree-level super-invariant) sho
uld be multiplied by a non-trivial function of the dilaton in order fo
r the duality conjecture to be consistent.