The breakdown rates of natural branch wood were studied in 2 streams (lowland and upland), and the current knowledge on breakdown rates of wood in streams was reviewed. The breakdown coefficient of natural alder branches, k = 0.0650/y, in an unpolluted upland stream corresponded well to the few published data on breakdown rates of natural wood substrates in streams. Breakdown rates of Scots pine (Pinus sylvestris, k = 0.0698/y), bark-covered black alder (Alnus glutinosa, k = 0.0859/y), and debarked black alder branches (k = 0.0717/y) in a nutrient-rich sandy lowland stream indicated that breakdown rates differ among tree species. Data from the present and previous studies on the breakdown of natural wood and commercial wood products (wood veneer, tongue depressors, ice-cream sticks) showed that breakdown is slower for natural wood than for commercial wood products, in part because the commercial wood products used in the studies had high surface to volume ratios. A strong relationship exists between surface to volume ratio of the substrate and breakdown rates. The breakdown rates of wood in streams were associated with several abiotic factors, including the physical and chemical properties of the wood substrates used, dissolved nutrients, and water temperature. Overall, the review of published breakdown rates emphasized that the current knowledge on the breakdown rates of wood in streams, especially the temporal patterns of microbial activities, is very limited, and heterogeneous experimental designs in previous studies prevent general conclusions. Nevertheless, measuring wood breakdown rates could be a promising tool for practical application in assessment of the ecological integrity of streams.