In this paper we build sequence data types that can efficiently support two important operations for par allel programming: partition and concatenation. We present a common chunk-based (instead of element based) interface that enables good performance in practice, and consider two alternative implementa tions: ropes [2] and skip list arrays (based on skip lists [21]). We parallelize a run-length encoding algo rithm as a motivating example, and show that, com pared to dynamic arrays, the proposed data struc tures significantly increase the scalability of parallel algorithms that include both partitioning and con catenation operations.