Pollination syndromes poorly predict main flower visitors in a disturbed Atlantic forest fragment

Abstract Pollination syndromes are often used to guide pollinator conservation and ecosystem restoration, but their effectiveness in predicting main pollinators is contested. This study characterized floral traits and pollination syndromes of 69 plant species in a Mixed Ombrophilous Forest fragment and assessed whether observed floral visitors matched the expected syndromes. Floral attributes (anthesis, color, odor, size, shape, symmetry, orientation, stamen position, nectar guides, resources) and plant characteristic (habit, sexual system, plant status, and occurrence) were recorded. Syndromes were inferred using literature, and insect visitors were collected and identified. The main visitor group (bee, beetle, butterfly, moth, fly, or wasp) was determined by highest visitation frequency. Logistic regression, subsampling, and sensitivity analyses were used to test the effects of monitoring time, the stability of main visitor groups, and alternative treatments of ties. Only 39.4% of species showed correspondence between observed main visitors and predicted syndromes, with melittophily achieving the highest match (55.6%). Sphecophily and phalaenophily could not be properly evaluated. Monitoring time was not significantly associated with correspondence, and the main visitor group was stable for most species, while alternative treatments of ties yielded similarly low correspondence. The predominance of bees and anthropogenic vegetation changes likely contributed to the limited correspondence. Such factors are common in conservation and restoration contexts. Pollination syndrome correspondence is a poor predictor of main floral visitors in disturbed Mixed Ombrophilous Forest fragments. Therefore, restoration programs in the Atlantic Forest should treat syndromes as preliminary hypotheses to be validated by empirical field surveys of local pollinator assemblages.