1: \begin{abstract}

2:     Let $(\veps_{t})_{t>0}$ be a sequence of independent   real random

3:     vectors of $p$-dimension   and let $X_T=

4:     \sum_{t=s+1}^{s+T}\veps_t\veps^T_{t-s}/T$ be the lag-$s$ ($s$ is a

5:     fixed positive integer) auto-covariance matrix of $\veps_t$.

6:      This paper   investigates the limiting behavior of

7:     the singular values  of $X_T$ under the so-called {\em ultra-dimensional regime} where

8:     $p\to\infty$ and  $T\to\infty$ in a related way such that

9:     $p/T\to 0$. First, we show that the singular value  distribution

10:     of  $X_T$ after a suitable normalization  converges to a nonrandom

11:     limit $G$ (quarter law) under the forth-moment condition. Second, we

12:     establish the convergence of its largest singular value to the right edge

13:     of $G$. Both  results  are derived using the moment method.

14:   \end{abstract}

15: