Bamboo species have the long-term clonal growth period until flowering to die (extended semelparity). The period of vegetative growth (flowering interval) has been reported to show a clear geographic cline from tropical (short interval) to temperate (long interval) region (Campbell 1985). This trend is correlated with the difference in rhizome system that also shows a clear geographic trend from tropical (short rhizome: called pachymorph) to temperate region (long-rhizome: called leptomorph) (Makita 1998).
We hypothesize that the difference in flowering interval is determined as a consequence of evolutionary game and depends on the rhizome system. By using a spatially explicit mathematical model, we explored the evolutionarily stable flowering interval and tested the effect of the rhizome system as well as seed and pollen dispersal ranges.
The analysis of the model showed that the flowering interval became longer as the rhizome length increased. The evolution of flowering interval also depended on the seed dispersal range. As the seed dispersal range increased the flowering interval decreased. However, if these ranges were extremely large, the evolutionary consequences of flowering interval did not change with the parameters. These results might be caused by sib-competition due to spatial structure of genets. If rhizome is short, each genet forms a clumping spatial structure by vegetative growth and suffering from severe sib-competition. Hence early flowering might contribute to reduce the competition. When the seed dispersal range is shorter, kin-competitions among seeds were severer, and hence the delayed flowering might be better strategy. When these parameters are extremely large, the effect of sib-competition is sufficiently reduced and we did not see any dependence on the parameters. We also discuss the effect of inbreeding depression on the evolution of flowering interval.