1: \begin{abstract}

2: We consider

3:  properties of determinants of some random symmetric

4: matrices issued from multivariate statistics: Wishart/Laguerre

5: ensemble (sample covariance matrices), Uniform Gram ensemble (sample

6: correlation matrices) and Jacobi ensemble (MANOVA). If $n$ is the

7: size of the sample, $r\leq n$ the number of variates and $X_{n,r}$

8: such a matrix, a generalization of the Bartlett-type theorems gives

9: a decomposition of $\det X_{n,r}$ into a product of $r$ independent

10: gamma or beta random variables. For $n$ fixed, we study the

11: evolution as $r$ grows, and then take the limit of large $r$ and $n$

12: with $r/n = t \leq 1$. We derive limit theorems for the sequence of

13: {\it processes with independent increments} $\{n^{-1} \log \det X_{n

14: , \lfloor nt\rfloor}, t \in [0, T]\}_n$ for

15:  $T \leq 1$ : convergence in probability, invariance

16: principle, large deviations. Since the logarithm of the determinant

17: is a linear statistic of the empirical spectral distribution, we

18: connect the results for marginals (fixed $t$) with those obtained by

19: the spectral method. Actually, all the results hold true for log

20: gases or $\beta$ models, if we define the determinant as the product

21: of charges. The classical matrix models (real, complex, and

22: quaternionic) correspond to the particular values $\beta = 1,2,4$ of

23: the Dyson parameter.  

24: \end{abstract}

25: