https://doi.org/10.5061/dryad.v15dv424c

Summary of findings and dataset contents:

The files contained herein include fruit biomass data, camera trap data, and spatial and habitat data obtained from display courts of white-bearded manakins (Manacus manacus) in northwestern Ecuador. The goal of the project was to investigate the role of fruit resource availability in shaping sexual selection processes (e.g., male display rate, female visitation) in a population of manakins. We found that leks and courts varied significantly in ripe fruit biomass, and the mean biomass of fruit surrounding courts predicted male display rates at the court during the study period. In turn, male display rate was the strongest predictor of female visitation. Other spatial and habitat variables (nearest neighbor distance, centrality, canopy cover) were not predictive of sexual selection indices. Causal modeling supported the hypothesis that local fruit indirectly affects female visitation via direct effects on male display rate.

Description of the data and file structure

Behavioral data: Video data was collected using motion-activated camera traps that were placed 1-2 meters from male display courts. Videos were then transferred from SD cards to hard drives and subsequently annotated for behaviors of interest by trained observers. Videos collected at courts that were monitored for < 3 weeks and/or using a different brand of camera were excluded from analyses. Displays by juvenile males were also excluded from analyses.

• Located in the following files:
  • Lek-6_Video-Review_Dec21-Jan22_11.07.23.csv
  • Lek-8_Video-Review_Dec21-Jan22_11.07.23.csv
  • Lek-14_Video-Review_Dec21-Jan22_11.07.23.csv
  • Lek-M_Video-Review_Dec21-Jan22_11.07.23.csv
  • Lek-C_Video-Review_Dec21-Jan22_11.07.23.csv
• NB: There are many blank cells throughout this file, and all indicate 0 counts. The raw file is cleaned up in R (code attached to this submission).
• Definitions of columns in video review datasheets
  • Folder -- the name of the broadest hard drive folder in which videos were contained (can ignore)
  • Lek -- lek ID (changed for clarity in manuscript; in the manuscript, Lek 6 == Lek 1, Lek 14 == Lek 2, Lek C == Lek 3, Lek M == 4, and Lek 8 == Lek 5)
  • Pista -- the display court name ('pista' is Spanish for display court)
  • DateRange_Folder -- the date range during which the SD card was present at the display court; the folder on the hard drive from which this datasheet was generated was named with this date range
  • FileName -- name of the video file corresponding to the row of data
  • Male -- male ID, typically a three-letter string of letters representing his colored leg-band combination
  • PresentYN -- whether a manakin was present in the video (yes/no)
  • DisplayYN -- whether a manakin was displaying in the video (yes/no), defined as producing wing sonations of some type
  • Snaps -- number of "jump-snap" displays produced in the video
  • quietSnaps -- number of "jump-snip" displays produced in the video
  • Grunts -- number of "grunt-jump" displays produced in the video
  • Rolls -- number of "roll-snap" displays produced in the video
  • Calls -- number of vocalizations/calls produced in the video
  • FemVisitation -- number of suspected females present in the video; not as definitive as "courtshipSuccess", defined below; not used in analysis
  • courtshipSuccess -- number of females that joined the male in the courtship dance during the video; this column was used to denote a definitive female visitation in analyses
  • Copulation -- number of copulations observed in the video
  • JuvPresent -- number of juvenile males in the video; excluded from analyses of male display rate on courts
  • Gardening -- number of distinct attempts a male made (successfully or unsuccessfully) to move debris off of his display court
  • vidLength -- length of the video; cameras were initially set to record 60s videos, but were switched to 30s after initial monitoring revealed that the vast majority of activity happened in the first 10-20 seconds of the video; some camera brands (e.g., Reconyx) recorded shorter videos (3-9 seconds), and thus videos recorded by those cameras were excluded from all analyses
  • Observer -- three-letter initials of observer recording the behavioral data

Fruit survey data: Circular, 10-m radius fruit survey plots were centered around male white-bearded manakin display courts. Both ripe and unripe fruits were counted during surveys, although only ripe fruits were used in analyses. Fruits were identified to morphospecies, and representative samples of fruits were obtained and weighed to obtain an average wet weight per berry for each morphospecies. Courts were surveyed twice per month, on a fortnightly basis, after their establishment or discovery.

• Located in the following files: Pistas-Frutas_06.14.22_cleaned.csv
• NB: There are a small number of rows for which fruit weights were not known. In all of these cases, these ultimately did not affect analyses due to one or more of the following reasons: (1) the species of fruit was likely too large for manakins to eat; (2) the court at which the unknown fruit was observed was excluded from analysis for other reasons; (3) the unknown fruit was unripe, so was not included in any analyses.
• Definitions of columns in fruit survey datasheet
  • Observadores -- initials of the observer(s)
  • Fecha -- the date of the fruit survey (mm/dd/yy)
  • SurveyCode -- we aimed to conduct 2 visual fruit surveys at focal courts each month; these codes correspond to surveys #1 and 2 during each month (some were monitored for more or less time depending on when monitoring of the court began)
  • Lek -- lek ID
  • Pista -- display court ID
  • Familia -- plant family
  • Morphospecies -- morphospecies of the fruit plant, as written in the field
  • MorphospeciesClean -- morphospecies names were sometimes written inconsistently in the original datasheet (although the morphospecies ID was always clear); this column was created to standardize each morphospecies name for analysis and organizational purposes
  • numMaduros -- number of ripe fruits (Spanish: "maduros") on a given plant
  • numPintones  -- number of unripe fruits (Spanish: "pintones") on a given plant
  • PesoFruta -- the average weight of the fruit, used to calculate the weight per plant (and, eventually, the total biomass of fruit in each plot during each survey)
  • TotalPeso -- total biomass of ripe fruit on a given plant (i.e., numMaduros x PesoFruta)
  • Notas -- any notes about the observation

Spatial and habitat data for display courts: UTM coordinates for each display court were gathered using a handheld GPS, and points were then uploaded into QGIS for calculation of centrality and nearest neighbor distances. Canopy cover at display courts was estimated using Canopeo, a mobile application that rapidly assesses the proportion of green cover from photographs.

• Located in the following files: masterDisplayFruitCentralityNNCC.csv (note: NN == nearest neighbor distance; CC == canopy cover)
• Definitions of columns in display court datasheet
  • Pista_Code -- the unique ID for each display court; composed of the lek ID and display court ID, separated by a period
  • meanFruit -- average fruit biomass (g) per survey at a given court
  • totalFruit -- total fruit biomass (g) across all surveys of a given court
  • Lek -- lek name
  • UTM1 -- UTM x coordinate; obtained with handheld GPS in the field
  • UTM2 -- UTM y coordinate; obtained with handheld GPS in the field
  • Elevation -- elevation; obtained with handheld GPS in the field
  • totalVisits -- total number of visits to a given court during the study period; based on courtshipSuccess column in behavioral datasheet
  • totalDisplays -- total number of display components (i.e., snaps, grunts, quietSnaps, and rolls) produced at a given court during the study period
  • daysMonitored -- the total number of days that a given court was monitored using a Browning Recon Force Elite HP4 camera during the study period
  • displayRate -- the number of displays produced per day on a given court (i.e., totalDisplays/daysMonitored)
  • visitRate -- the number of female visitations per day at a given court (i.e., totalVisits/daysMonitored)
  • hubDist_mcpCentroid -- the distance of the court, in meters, from the centroid of the lek MCP; calculated in QGIS
  • Centrality -- the value in hubDist_mcpCentroid multiplied by -1; done so that smaller distances would correspond to greater centrality and vice-versa
  • nn_distance -- the distance, in meters, to the next closest male display court; calculated in QGIS
  • CanopyClosedness6in -- fractional green canopy cover, obtained from mobile app Canopeo; standardized photos taken on an upward-facing at the center of cleared court areas, 6 inches (15 cm) above the ground to approximate the height at which displays are occurring (used in analyses of canopy cover)
  • CanopyClosednessBH -- fractional green canopy cover, obtained from mobile app Canopeo; standardized photos taken on an upward-facing at the center of cleared court areas, at breast height; strongly correlated with CanopyClosedness6in (not used in analyses of canopy cover)

Code/Software

To see R code used to generate results from the above data files, please see the included R markdown file titled: "Biology_Letters_Code_09.03.24.Rmd." Annotations therein describe about how data was cleaned, processed, and analyzed for the paper.