Let $X_1, X_2$ be two independent variables with means $\theta_1, \theta_2$. Then two examples are given (one binomial, one normal) in which an estimator, depending only on $X_2$ is admissible for estimating $\theta_1$.
"A Paradox in Admissibility." Ann. Statist. 5 (3) 544 - 546, May, 1977. https://doi.org/10.1214/aos/1176343853