1: \begin{abstract}

2: We consider Ewens random permutations of length $n$

3: conditioned to have  no cycle longer than $n^\beta$ with 

4: $0<\beta<1$ and to study the asymptotic behaviour as $n\to\infty$. 

5: We obtain very precise information on the joint distribution of 

6: the lengths of the longest cycles; in particular we prove a 

7: functional limit theorem where the 

8: cumulative number of long cycles converges to a Poisson 

9: process in the suitable scaling. Furthermore, we prove convergence 

10: of the total variation distance between joint cycle counts 

11: and suitable independent Poisson random variables up to a 

12: significantly larger maximal cycle length than previously known. 

13: Finally, we remove a superfluous assumption from a central limit 

14: theorem for the total number of cycles proved in an earlier paper.

15: 

16: 

17: 

18: % We consider uniform and Ewens random permutations of length $n$

19: % conditioned to have 

20: % %\begin{color}{red}

21: %  no cycle longer than $n^\beta$ with $0<\beta<1$.

22: % %\end{color}

23: % %\begin{color}{blue}

24: % %no cycle above a certain sub-macroscopic length,

25: % %\end{color}

26: % We study the asymptotic behaviour as $n\to\infty$ and extend the results obtained in the paper 'Random permutations without macroscopic cycles'.

27: % 

28: % 

29: % Extend to the Ewens measure.

30: % 

31: % We study here the behaviour of the long cycles.

32: % 

33: % Remove some technical assumptions.

34: % 

35: % Improve the total variation distance.

36: \end{abstract}

37: