Clonal plants can propagate both sexually and vegetatively and the balance between the efforts of sexual and vegetative propagation can be determined by inter- and intra-specific competition. This raises the question of whether sexual propagation becomes more suitable than vegetative propagation or vice versa. Some studies suggest seed production is positively correlated with plant density, and others, negatively. To evaluate the effects of local density of ramets on plant propagation strategy, we developed a lattice model where two strategies compete with each other. In the model, a plant changes its strategy based on local density as defined by the proportion of neighboring lattice cells containing plants. One strategy produces seeds at high local density (SEEDING at HLD) and produces ramets at low local density, and the other strategy produces ramets at high local density (RAMETING at HLD) and seeds at low local density. Seeds can spread all cells while ramets can reach only adjacent cells, and if an individual from a seed competes with an individual from a vegetative propagule, the latter always wins. Each plant has different thresholds to change its strategy from sexual to vegetative, or vice versa, along with local density. In the model, a realized density depends on both reproductive rate and mortality; thus, we evaluate the performance of each strategy under different mortalities. When mortality is high, a realized density becomes low. Under this condition, plants reproduce mainly by seeds (either the SEEDING at HLD plants with lower threshold values, or RAMETING at HLD plants with higher threshold values) and show high performance. High mortality results in many vacant lattice cells that do not have neighboring plants. Since vegetative propagules cannot reach these isolated vacant lattice cells while seeds can, plants that propagate mostly or exclusively by seed become dominant. When mortality is low, the local density becomes high and few cells are available for new individuals. Under this condition, SEEDING at HLD becomes dominant. SEEDING at HLD plants with a high threshold value almost always reproduce ramets, but they also produce seeds when neighboring cells are completely occupied. Although SEEDING at HLD plants with a high threshold value and RAMETING at HLD plants have an equal chance to get established by vegetative propagation, SEEDING at HLD plants also produce seeds when there are no vacant neighboring cells (i.e., high density), and those seeds can, although rarely, establish new individuals. In conclusion, the production of ramets at lower densities and the production of seeds at higher densities seem to be a proper strategy in our model.
