Multimedia applications in multihop wireless networks have great market potential. Multiple channels and multiple radios are commonly used for exploring multimedia transmissions in multihop wireless networks. Split transmission allows multiple channels attached to different radios simultaneously to be used, and so to achieve a fundamentally improved transmission capacity. The goal of this paper is to present a theoretical background to justify the improved performance of the split transmission. We theoretically study and prove that, by using the split transmission, the worst-case delay is decreased to $\frac{\sigma\rho_{k-1}}{LC_{m-1}C_{k-1}}$ of that without using the split transmission, the average throughput is increased to $\frac{1}{1-\prod_{j=0}^{k-1}\alpha_j}$ of that without using the split transmission, and the average delay jitter is decreased to $\frac{C_{k-1}C_{\rho}}{C_{m-1}[C_{\rho}+L(\rho+C)]}$ of that without using the split transmission. We believe that this is the first attempt to consider split transmission in theory. We also evaluate the performance of the split transmission in ns-2 simulations. The observed results show that the split transmission achieves shorter worst packet delays, higher average throughput, and smaller average delay jitter as compared to the performance achieved without the split transmission.