1: \begin{abstract}

2:   An important theme in recent work in asymptotic geometric analysis

3:   is that many classical implications between different types of

4:   geometric or functional inequalities can be reversed in the presence

5:   of convexity assumptions.  In this note, we explore the extent to

6:   which different notions of distance between probability measures are

7:   comparable for log-concave distributions. Our results imply that

8:   weak convergence of isotropic log-concave distributions is

9:   equivalent to convergence in total variation, and is further

10:   equivalent to convergence in relative entropy when the limit measure

11:   is Gaussian.

12: \end{abstract}

13: