Impacts of moderateness and cobalancedness in seed mass and coat thickness on seed dispersal effectiveness of Quercus variabilis by rodents

By measuring seed mass (representing attraction trait to rodents) and coat thickness (representing defense trait to rodents), we examined the impacts of moderateness (medium defensive and attractive trait values) and cobalancedness (positive association between defensive and attractive trait values) in seed mass and coat thickness on seed fates of Quercus variabilis seeds hoarded mainly by rodents.Three secondary forest patches (codes A, D, and H) in the study area were used to releas and track tagged seeds of Q. variabilis from October to December 2019 and 2021. We randomly weighted 1200 intact seeds (seed size, as measured by fresh seed mass) and released them into three patches (400 seeds per patch). 10 seed stations were spaced 20 m apart in each patch. At each station, 40 individually tagged seeds were placed on the surface of the ground. The tagged seeds were tied with small, light red plastic tags (3.6 × 2.5 cm, < 0.1 g) through a 0.6 mm diameter hole using a 10 cm long thin steel wire; each tag was coded with a serial number using a marker pen. After seed release, the fate of each target seed was surveyed at 1, 3, 5, and 7 days for the first week, and then checked every week until almost all seeds had been harvested (eaten or removed) by animals. Once the seeds were eaten in situ by animals, we searched for seed coats on the remaining seed tags and measured their coat thickness with a Vernier caliper. Once the seeds were removed by animals, we searched a 25 m radius around each station with equal effort (2­3 h by two people for each patch) and recorded the fate of the tagged seeds. If the seed was found with only the tags, we searched for a seed coat nearby and measured the side of the seed coat thickness. If the seeds were scatter-hoarded, we then took them back to the lab to separate the coat and measured the side of coat thickness also using a Vernier caliper. Each coat was measured at three different locations and the average value was used as seed coat thickness for statistical analysis.Seed fates at each station were categorized as intact in situ (IIS), eaten in situ (EIS), or removed; and those removed from seed stations were categorized as scatter-hoarded (i.e., seeds buried in the surface soil or beneath the leaf litter, SH), eaten after removal (i.e., leaving only tags and fragments, EAR), missing (i.e., seeds were not found within our search area or more likely being larder-hoarded in burrows or tree cavities, M). Those data (178 seeds in 2019 and 569 seeds in 2021) were excluded from the analysis because there was no data on seed coat thickness for those seeds that were missing or eaten after dispersal by rodents or birds. We used a proxy representing the seed dispersal effectiveness (SDEp), which was defined as the probability of final seed survival after seeds were scatter-hoarded during the study period: SDEp = (the number of seeds scatter-hoarded by rodents)/(the number of seeds released at the seed stations) × 100%.