In a recent paper, Wei and Zakhor  presented a novel multipath selection framework for streaming video over wireless ad-hoc networks. We point out some limitations of their approach and suggest techniques to overcome those limitations. We propose that different metrics for multipath computation should be used when different schemes (e.g., Multiple Description Coding and Layered Coding) are used for video encoding. We adopt a more practical measurement technique for predicting the quality of a link. In addition to analytical reasoning for improved performance of our techniques, we conduct extensive experimentation to evaluate the efficacy of our approach. The results demonstrate that our approach leads to significantly better performance when compared to the approach proposed in .