1: \begin{abstract}

2: In this paper nonparametric methods to assess the multivariate L\'evy

3: measure are introduced. Starting from high-frequency observations of a

4: L\'evy process $\vect X$, we construct estimators for its tail

5: integrals and the Pareto--L\'evy copula and prove weak convergence of

6: these estimators in certain function spaces. Given $n$ observations of

7: increments over intervals of length $\Delta_n$, the rate of convergence

8: is~$k_n^{-1/2}$ for $k_n = n \Delta_n$ which is natural concerning

9: inference on the L\'evy measure. Besides extensions to nonequidistant

10: sampling schemes analytic properties of the Pareto--L\'evy copula

11: which, to the best of our knowledge, have not been mentioned before in

12: the literature are provided as well. We conclude with a short

13: simulation study on the performance of our estimators and apply them to

14: real data.

15: \end{abstract}