1: \begin{abstract}

2: We analyse a trimmed stochastic process of the form 

3:  ${}^{(r)}X_t= X_t - \sum_{i=1}^r \Delta_t^{(i)}$,

4:  where $(X_t)_{t \geq 0}$ is a driftless subordinator on $\R$ 

5: with its jumps on $[0,t]$ ordered as

6: $ \Delta_t^{(1)}\ge  \Delta_t^{(2)} \cdots$.

7: When  $r\to\infty$, both $\trimX_t\dto 0$ and $\Y_t\dto 0$ a.s. for each $t>0$,  and it is interesting to study the  weak limiting behaviour of $\bigl(\trimX_t, \Y_t\bigr)$ in this case.

8: We term this ``large-trimming" behaviour.

9: Concentrating on the case $t=1$, we study joint convergence of

10: $\bigl(\trimX_1, \Y_1\bigr)$  under linear normalization,

11: assuming extreme value-related conditions on the L\'evy measure of $X$

12: which  guarantee that $\Y$ has a limit distribution with linear

13: normalization.  

14: Allowing $\trimX$ to have random centering and scaling in  a natural way, we show that $\bigl(\trimX_1, \Y_1\bigr)$  has a bivariate normal

15: limiting distribution, as $r\to\infty$; but 

16: replacing the random normalizations with natural deterministic ones produces non-normal limits which we can specify.

17: \end{abstract}

18: