Japanese knotweed is an aggressive alien species in Europe, North America, and Australia, causing a range of environmental problems. Eradication of Japanese knotweed is proving to be a difficult task, because the plant is able to propagate generatively by intra- and interspecific hybridization, and vegetatively from shoot and tiny rhizome pieces. Despite the economic consequences of Japanese knotweed on natural and built environments, its physiology is not yet fully understood; especially important are sink-source relations between old and young parts of the rhizome and growth of lateral and latent rhizome buds. Current methods of chemical control include three types of phloem-mobile herbicides, such as glyphosate, imazapyr, and synthetic auxins. These herbicides have limitations on their use, and all fail to eradicate the plant completely, for the reasons discussed in this review. Our aim is to suggest prospective approaches to enable chemical eradication: use of signals to induce controlled growth and development of quiescent rhizome buds; use of phytohormones, sugars, and light to increase allocation of phloem-mobile herbicides to the rhizome; use of xylem-mobile herbicides to exterminate the old rhizome parts; and use of different phloem-mobile herbicides at different growth stages.

The Ostrinia latipennis group contains two species, O. latipennis (Warren) and O. ovalipennis Ohno. These two species commonly utilize perennial knotweeds (Fallopia spp.) as their host plants, which are serious invasive weeds in Europe and North America. Ostrinia latipennis is widely distributed across north-east Asia including Japan whereas O. ovalipennis is restricted to north Japan (Hokkaido Is.) and highland areas of central Japan (Nagano Prefecture in Honshu Is.). To estimate the phylogenetic relatedness and geographical differentiation of the two species, mitochondrial COII gene sequences were determined for specimens covering their distribution ranges in Japan. The uncorrected sequence divergence between O. latipennis and O. ovalipennis was 0.6–0.7%, supporting a close relationship. According to the standard molecular clock proposed for arthropod mtDNA, the two species are speculated to have diverged about 0.3 Myr ago. A single COII gene haplotype was found in O. latipennis irrespective of collection locality. In contrast, two haplotypes were found in O. ovalipennis, and their frequencies were significantly different between the Hokkaido and Honshu populations. The patterns of geographical variation in the COII gene within the two species were in agreement with previously reported patterns of geographical differentiation in morphology of the two species in Japan. The present results support the hypothesis that gene flow among local populations of O. ovalipennis has been limited by geographical isolation.