Floral offer and seasonality shape plant-pollinator networks in tropical gardens (2025)

Floral offer and plant-pollinator interactions (insects and birds) were sampled monthly over 12 months (October 1, 2022 - September 12, 2023) in 10 tropical garden sites in Martinique (Lesser Antilles), including five communal gardens (C-sites; C1 to C5) and five neighbourhood clusters of five private gardens (P-sites; P1 to P5). Sampling was carried out by the same observer, using a transect method with one site surveyed per day and the sampling order of sites and transect sections randomly changed each month. Each survey (floral offer and interactions) was conducted along 150 m (total transect length) per site per month. For each private garden site, this spatial coverage was evenly distributed across the five private gardens, resulting in 30 m of coverage per individual garden. Each month, transect sections were distributed across distinct blooming flower patches to capture the floral diversity of the focal site at that time and to avoid spatial clustering, with positions and lengths adjusted from month to month to accommodate natural and/or human-induced changes in floral availability. Plants and pollinators were identified to the lowest taxonomic level possible and verified with expert taxonomists. For each site, monthly data were paired into six two-month periods (bi-periods) for analysis.

Floral offer survey
20 quadrats (1 × 1 m, up to 3.5 m in height) were randomly placed along transect sections (one per 7.5-m section). Within each quadrat, all blooming plant species were identified along with the number of their open floral units. A floral unit was defined as either a single flower or an inflorescence on which an insect of length l ≥ 0.5 cm could move between flowers without needing to fly. When a plant species had highly abundant and evenly distributed floral units within a quadrat, totals were estimated by counting a subset of the quadrat and scaling up.

Plant-pollinator interactions
1) Plant-insect and 2) plant-bird interactions were recorded separately, along the same transect sections used for sampling floral offer. An interaction was defined as floral visitor behaviour indicative of floral nectar or pollen collection. 1) Plant-insect interactions were monitored over 60 min per site (along the 150 m) per month at a constant speed (v = 2.5 m/min). Interactions were recorded within 1 m on either side of the transect line and up to 3.50 m in height. Insects were captured using a butterfly net and euthanised. 2) Plant-bird interactions were recorded from a distance. As an individual bird could be observed multiple times, each plant-bird species pair was counted only once per site per bi-period, resulting in unweighted (binary) interaction data. Each transect section was monitored for 10 min. To minimise disturbance to pollinators, bird-plant interactions were always recorded first, followed by insect-plant interactions, before conducting floral surveys.