1: \begin{abstract}

2: Regular variation of a \sid{multivariate} measure \sid{with a Lebesgue density }

3: implies the regular variation of its density \sid{provided} the density satisfies

4: some regularity conditions. 

5: \sid{Unlike the univariate case,}

6: the converse \sid{also requires} regularity conditions. 

7: We extend these arguments to discrete mass functions and their

8: associated measures \sid{using the concept} that the 

9: the mass function can be embedded

10: in a continuous density function. We  give two different

11: conditions, monotonicity and \wtd{convergence on the unit sphere}, both of which can make the

12: discrete function embeddable. \sid{Our results are} then applied to

13: the preferential attachment network model, and we conclude that the

14: joint mass function of in- and out-degree is embeddable and thus

15: regularly varying. 

16: \end{abstract}

17: