the original index on the first level, and the `other` DataFrame
        columns on the second level.

        In the case of missing elements, only complete pairwise observations
        will be used.
        Ú