We focus on the task of multi-hop reading comprehension where a system is required to reason over a chain of multiple facts, distributed across multiple passages, to answer a question. Inspired by graph-based reasoning, we present a path-based reasoning approach for textual reading comprehension. It operates by generating potential paths across multiple passages, extracting implicit relations along this path, and composing them to encode each path. The proposed model achieves a 2.3% gain on the WikiHop Dev set as compared to previous state-of-the-art and, as a side-effect, is also able to explain its reasoning through explicit paths of sentences.