Journal of Arid Environments 186 (2021) 104408 0140-1963/© 2020 Elsevier Ltd. All rights reserved. Plant traits interplay to balance pollen limitation in the Brazilian seasonal dry forest: A meta-analysis Thiago Amorim a , Andr´ e Maurício Melo Santos b , Natan Messias Almeida c , Reinaldo Rodrigo Novo d , Ana Virgínia de Lima Leite d , Cibele Cardoso Castro d, e, * a Programa de P´ os-Graduaç˜ ao em Ecologia e Recursos Naturais, Departamento de Biologia, Universidade Federal do Cear´ a, Fortaleza, Cear´ a, Brazil b Universidade Federal de Pernambuco, Centro Acadˆ emico de Vit´ oria, Vit´ oria de Santo Ant˜ ao, Pernambuco, Brazil c Universidade Estadual de Alagoas, Campus de Palmeira dos ´ Indios, Alagoas, Brazil d Programa de P´ os-Graduaç˜ ao em Botˆ anica, Departamento de Biologia, Universidade Federal Rural de Pernambuco, Recife, Pernambuco, Brazil e Universidade Federal Rural de Pernambuco, Unidade Acadˆ emica de Garanhuns, Garanhuns, Pernambuco, Brazil A R T I C L E INFO Keywords: Caatinga Floral specialisation Meta-analysis Plant habit Pollen limitation ABSTRACT Animal pollination is one of the most important ecosystem services that beneft both natural plant communities and plant crops. A failure in pollination service can result in pollen limitation of fruit and seed set. This scenario can be even worse in endangered ecosystems such as tropical dry forests. Pollen limitation is a way of measuring the reproductive performance of plants by comparing the output of natural pollination with that of supplemental hand cross-pollination. Studies have shown that pollen limitation can be moderated by plant traits, such as plant habit, mating system, pollinator dependence and foral specialisation. Herein, we carried out a meta-analysis to determine the magnitude of pollen limitation in 60 plant species from the Caatinga vegetation of the Brazilian seasonal dry forest. We found an overall mean pollen limitation of 0.41 and a prevalence of 35%. Trees, shrubs, and self-incompatible, pollinator dependent and generalist plants were pollen limited. We found a signifcant interaction between plant habit and foral specialisation. Specialist and generalist trees and generalist shrubs were pollen limited. Trees and shrubs are mostly threatened in the endangered Caatinga. Species that are not pollen limited, such as lianas and herbs, may also be indirectly threatened by deforestation and intensive farming. 1. Introduction Pollination is a pre-requisite for reproduction in most plants, and directly impacts their population dynamics, community structure, and ultimately their ecosystem functioning (Willmer, 2011). Indeed, most fowering-plants, including crops, need pollinating animals to success- fully reproduce (Klein et al., 2007; Ollerton et al., 2011). Animal polli- nation is, thus, an important ecosystem service that is considered in biological conservation efforts worldwide (Mace et al., 2012). However, animal pollination is a multifactorial process, involving pre- and post-pollination events, as well as the environmental stochasticity that interferes in the population dynamics of pollinators. Consequently, failure in pollination service can result in inadequate and/or insuffcient pollen deposition, causing pollen limitation of seed and fruit set (Aizen and Harder, 2007; Knight et al., 2006). On the other hand, the excessive deposition of self- or heterospecifc pollen can also have pre- and/or postzygotic consequences on plant reproductive success (Aizen and Harder, 2007; Alonso et al., 2013). Although the latter authors argue that poor-quality receipt plays a more important role than a quantitative approach, pollen limitation is often assessed by comparing the plant output (fruit and/or seed set) of natural-pollinated controls with that of supplemental hand cross-pollinated treatments (Larson and Barrett, 1999; Chen and Zuo, 2018; Sun et al., 2018). Using this quantitative approach, the magnitude of pollen limitation has been experimentally determined in plant communities around the world. Pollen limitation has been found in 25% of the species from a temperate deciduous forest (Motten, 1986), in 18% of species from a temperate grassland community (Hegland and Totland, 2008), in 38% of species from the coastal Atacama desert (Gonz´ alez and P´ erez, 2010), and in 52% of species from a tropical montane rain forest (Wolowski et al., * Corresponding author. Programa de P´ os-Graduaç˜ ao em Botˆ anica, Departamento de Biologia, Universidade Federal Rural de Pernambuco, Recife, Pernambuco, Brazil. E-mail address: cibele.castro@ufape.edu.br (C.C. Castro). Contents lists available at ScienceDirect Journal of Arid Environments journal homepage: www.elsevier.com/locate/jaridenv https://doi.org/10.1016/j.jaridenv.2020.104408 Received 15 April 2020; Received in revised form 25 November 2020; Accepted 27 November 2020