Mature and immature seed counts for 84 species on Barro Colorado Island, Panama

Description of methods used for collection/generation of data Data was collected from the 50-ha plot on Barro Colorado Island, Panama (BCI; 9°100 N, 79°510 W). BCI is a 15.6km2 island in the Panama Canal, which supports a lowland, moist tropical forest with a 35m tall canopy. The 50-ha plot is a 1 km × 0.5 km section of old growth forest on a plateau in the centre of the island. Since 1987, weekly censuses of seed rain have been conducted within the 50-ha plot. A first set of seed traps (n = 200) was established in 1987 and has been continuously monitored since then. These seed traps are located at 13.5 m intervals on alternating sides of and 4-10 m from 2.7 km of pre-existing trails within the 50-ha plot. Another 50 traps were added in naturally occurring tree fall gaps within the 50-ha plot between 2002 and 2004 and have been monitored since. A final set of 200 traps were established 2 m from each of the 200 original traps in 2011 and were censused for 23 months. Each trap consists of a 0.8 m tall PVC frame and 1 mm mesh covering 0.5 m2 of forest floor. In weekly censuses, all fruits and seeds encountered in each trap are identified to species, counted, and categorised as mature or immature. The distinction between immature and mature fruits was based on the examination of endocarps: If the endocarp of seeds within a fruit is filled with material that is solid (rather than hollow or filled with liquid or soft substances), the fruit was considered mature. Prior to 1990, aborted fruits (which fall soon after flowering, have a swollen ovule, and often have some flower parts attached) were included in immature fruit counts. We use data from 1990 onwards, since these fruits were likely dropped due to pollination failure rather than as a consequence of enemy attack. For some species, the number of “capsules” is also recorded. Capsules are defined as a part of the fruit which is not dispersed and not eaten by animals. Botanically, capsules could refer to a capsule, pedicel or bract and each count of a capsule corresponds to a single fruit count (i.e., if a fruit exocarp has three valves, three valves from that species would be counted as one capsule). From dropped capsules we can estimate the number of mature dispersed seeds. For example, in wind-dispersed species such as Jacaranda copaia, mature fruits dehisce in the canopy, releasing winged seeds, and therefore mature capsules can indicate how many fruits dehisced. For vertebrate-dispersed species such as Quararibea asterolepis, capsules of mature fruits (in this case bracts) are inedible and discarded by consumers. Methods for processing the data For each species, we estimated the start of the phenological year as the calendar day that best aligned fruiting into a single, temporally coherent season. Specifically, we evaluated all possible start days (0–364) by re-indexing fruiting observations using circular modulo arithmetic and calculating the quantity-weighted variance of fruiting times relative to each candidate start day. Fruit quantity was used as a weight to emphasise periods of greater reproductive output. The start day that minimised this variance was chosen as the phenological year boundary, indicating the division between successive annual fruiting cycles that yielded the greatest temporal concentration of fruiting activity. To assess immature seed mortality rates for each species in each trap, we extracted the total sums of immature and mature seeds in each trap for each phenological year. Seed counts equal the sum of individual seeds plus seeds inside fruits collected in the traps; the latter was obtained by multiplying fruit counts by known species-specific seed-to-fruit ratios. Seed-to-fruit ratios were calculated as the mean values from dissections of five mature fruits from five individuals of each species (n = 25 fruits per species) within the 50-ha plot. For species that shed capsules, seed counts were calculated using the greater of two estimates: (i) seeds contained within fruits plus the number of individual seeds or (ii) seeds contained within fruits plus seeds inferred from capsule counts. This approach was taken to avoid double counting, since some of the individual seeds may have originated from the same fruit as the capsule. Species names were matched to a static copy of The World Flora Online (http://www.worldfloraonline.org) v.2024.12 valid at 2025-04-23 using the function "WFO.match" from the R package {WorldFlora}. The function "WFO.one" was then used to find one unique matching name for each submitted name. using the argument 'priority = "Accepted"', it first limits candidates to accepted names, with a possible second step of eliminating accepted names that are synonyms. Guarea grandifolia (GUA1) and Guarea guidonia (GUA2) are sometimes considered synonyms, but they are distinct species on Barro Colorado Island. Metadata sp4 4-letter ID code of plant species. This code does not change and can be used to track earlier botanical names to compare across publications. sp6 6-letter ID code of plant species which corresponds to codes used in the censuses of the 50-ha plot https://doi.org/10.15146/5XCP-0D46 family Family name genus Genus name species Species name year Phenological year. The start of the phenological year was estimated per species as described above. If the date of the weekly seed trap census was after the start of the phenological year then `year` = calendar year. If the date of census was before the start of the phenological year then `year` = calendar year - 1. quadrat Quadrat ID; created by overlaying a 50 ✕ 50 m grid over the 50-ha plot. Labelled from "quadrat_001" to "quadrat_100", n = 100 trap Seed trap ID. Labelled from "trap_001" to "trap_450", n = 450 x The x coordinate within the plot, meters from the west border of the plot, always in [0,1000]. y The y coordinate, meters from the south border, always in [0,500]. abscised_seeds Number of immature seeds, estimated as described above. Rounded to integers. viable_seeds Number of mature seeds, estimated as described above. Rounded to integers. total_seeds abscised_seeds + viable_seeds proportion_abscised abscised_seeds / total_seeds, calculated with non-rounded values.