I dwell on the appearance, in the mid-19th century, of a formula describing random sampling without replacement under incomplete knowledge and discuss the statistical aspect of this formula and its applications. I also provide illustrations showing that the fairness of sampling without replacement was being questioned.