1: \begin{abstract}

2:   We prove a sequence of limiting results about weakly dependent stationary and regularly varying

3:   stochastic processes in discrete time. After deducing the limiting distribution for individual

4:   clusters of extremes, we present a new type of point process convergence theorem.  It is designed

5:   to preserve the entire information about the temporal ordering of observations which is typically

6:   lost in the limit after time scaling.  By going beyond the existing asymptotic theory, we are able

7:   to prove a new functional limit theorem. Its assumptions are satisfied by a wide class of applied

8:   time series models, for which standard limiting theory in the space $D$ of \cadlag\ functions does

9:   not apply.

10:   

11:   To describe the limit of partial sums in this more general setting, we use the space~$E$ of

12:   so--called decorated \cadlag\ functions. We also study the running maximum of partial sums for

13:   which a corresponding functional theorem can be still expressed in the familiar setting of space

14:   $D$.

15:   

16:   We further apply our method to analyze record times in a sequence of dependent stationary

17:   observations, even when their marginal distribution is not necessarily regularly varying.  Under

18:   certain restrictions on dependence among the observations, we show that the record times after

19:   scaling converge to a relatively simple compound scale invariant Poisson process.

20: 

21:  {Keywords: point process  \and regular variation \and

22:   	invariance principle \and functional limit theorem \and record times}

23: \end{abstract}

24: