# Who first observed that Conjugate Gradient for Symmetric Positive Definite linear systems is a Krylov method?

Conjugate gradient was originally presented in the 50's before the modern understanding of Krylov subspaces (and the resulting iterative methods) was fully realized. As such, the method was derived using different tools and language. My question is, who was the first person to observe that Conjugate Gradient is a Krylov subspace method? I have heard Youssef Saad given this credit, but when I search for references to back up the assertion, I cannot find relevant references. I figure someone must know or that perhaps the understanding came gradually and cannot be attributed to any one author (or even a small group).

-