On page 307 of Vaught's paper "Denumerable models of complete theories", theorem 2.1.2 states that there is a denumerable model $\frak{A}$ of $T$ such that if, for each $j\in \omega, \:P_j$ is a non-principal prime ideal (now called ultrafilter) of $F_{p_j+1}(T)$, then there is a denumerable model $\frak{A}$ of $T$ such that $P_0(\mathfrak{A}), P_1(\mathfrak{A}), \cdots$ are all empty. Then Vaught said that 2.1.2 was proved by Ehrenfeucht.
I wonder which of Ehrenfeucht's papers contains theorem 2.1.2.