Let {X(t),a ≦ t ≦ b}
and {W(t),0 ≦ t < ∞} be a Gaussian process and the standard Wiener
process, respectively. Investigating covariance structure of X(t), the paper gives
various representations of X(t) in terms of W(t), including stochastic integral
representations. Some of these representations are useful in finding hitherto unknown
barrier-crossing probabilities of X(t).