1: \begin{abstract}

2: Multivariate distributions are explored using the joint distributions

3: of marginal sample quantiles.

4: Limit theory for the

5: mean of a function of order statistics is presented. The results

6: include a multivariate central limit theorem and a strong law of large

7: numbers. A result similar to Bahadur's representation of quantiles

8: is established for the mean of a function of the marginal quantiles.

9: In particular, it is shown that

10: %

11: \[

12: \sqrt n \Biggl( \frac{1}{n} \sum_{i=1}^n \phi\bigl(X_{n\dvtx  i }^{(1)},

13: \ldots,

14: X_{n\dvtx  i }^{(d)}\bigr) - \bar{\gamma} \Biggr) = \frac{1}{\sqrt{n}} \sum

15: _{i=1}^nZ_{n, i} +

16: \mathrm{o}_{P}(1)

17: \]

18: %

19: as $n \rightarrow\infty$, where $\bar{\gamma}$ is a constant

20: and $Z_{n, i}$ are i.i.d.~random variables for each $n$.

21: This leads to the central limit theorem.

22: Weak convergence to a Gaussian process using equicontinuity of

23: functions is indicated.

24: The results are established under very general conditions.

25: These conditions are shown to be satisfied in many commonly occurring

26: situations.

27: \end{abstract}