We comment on a curious relation between Siegel's model of random latt
ice strings and type IIB matrix model. The comparison of the two theor
ies suggests that there may exist extra terms in the latter which are
overlooked in the weak string coupling limit.