FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048541059> ?p ?o ?g. }

• W3048541059 endingPage "1205" @default.
• W3048541059 startingPage "1201" @default.
• W3048541059 abstract "Macroevolutionary analyses have provided powerful insight into the factors underpinning the morphological and species diversity of flowering plants, and have identified interactions between plants and pollinators as a key evolutionary driver of this diversity (Stebbins, 1970; Dodd et al., 1999; Kay & Sargent, 2009; Van der Niet & Johnson, 2012; Barrett, 2013). Macroevolutionary analyses of pollinator-driven diversification require knowledge of phylogenetic relationships among taxa, but also critically depend on knowledge of their pollination systems. During the history of research on pollinator-driven evolution of plants, these two elements have not developed synchronously. Detailed natural history observations, which are required to characterize the ecology of plant–pollinator interactions, have been made for centuries (e.g. Darwin, 1862; Marloth, 1895; Knuth, 1906–1909). The relatively recent breakthrough in phylogenetic systematics associated with DNA sequencing was expected to lead to powerful new studies of pollinator-driven evolution (Armbruster, 1992; Givnish & Sytsma, 1997), but it is becoming clear that ecological data has become the weakest part of the diptych of evidence required for macroevolutionary analyses. In a recent study, Reich et al. (2020) evaluated an important but rarely tested idea in biology, namely whether floral traits and the processes of their evolution differ between ecological specialists and generalists in the context of plant–pollinator interactions. These vary along a continuum of extreme specialization to extreme generalization in terms of the number of pollinator species with which plants interact (Waser et al., 1996; Johnson & Steiner, 2000; Fenster et al., 2004). Using a sample of the large and florally diverse genus Erica as a model system, Reich et al. implemented geometric morphometric methods to quantify variation in floral shape, categorized each species according to three specialized and one generalized pollination syndrome, and used a phylogenetic framework for macroevolutionary analyses. They specifically focused on modules, which are defined as subsets of traits that vary in a coordinated manner and relatively independently from other such subsets (Klingenberg, 2014). Based on their results, Reich et al. conclude that ‘flower shape modularity […] critically depends on pollinator specialization and syndrome’. The validity of results from macroevolutionary analyses of pollinator-driven evolution is contingent on the data and methods used for phylogenetic analyses and for the assignment of pollination systems. The former has been dealt with in several reviews (e.g. Smith, 2010). In this Letter I discuss approaches to characterize pollination systems for macroevolutionary analyses. Based on this theory, and considering what is known about pollination systems in Erica, I identify several problems with the pollinator data used in Reich et al., and use this example to illustrate the broader problem of a paucity of natural history data for analyses of pollinator-driven evolution. If plant species are visited by multiple animal species, characterization of the pollination system requires identification of the most important pollinator (Stebbins, 1970), taking into account visitation frequency and per-visit effectiveness (Fenster et al., 2004). Spatiotemporal variation in relative visitation frequencies (Mayfield et al., 2001; Gomez et al., 2008) implies that determination of pollination systems should include assessment of pollinator importance from multiple sites and years, and incorporate a measure of variance of importance of different pollinator species (Reynolds & Fenster, 2008). In practice, it is often not feasible to implement the methods outlined above for the large number of species that typically feature in macroevolutionary analyses. Pollinator importance can instead be inferred from visitor observations, although this is problematic if not all visitors are effective pollinators (cf. Fenster et al., 2004) or if the most frequent visitor is not the most important pollinator (e.g. Barrios et al., 2016). Observations can be combined with proxies for effectiveness, such as the presence of pollen on pollinator bodies (e.g. Goldblatt et al., 1995). Indirect approaches for determining the most important pollinator include use of selective exclusion experiments, supplemented with visitor observations (e.g. Cozien et al., 2019). Macroevolutionary analyses typically require pollination systems to be coded as discrete characters (but see Smith et al., 2008). Relatively specialized pollination systems, which feature in the majority of macroevolutionary analyses, can be coded as discrete functional pollinator groups (cf. Fenster et al., 2004), although too broad definitions of groups may underestimate specialization. If representatives from multiple functional groups act as pollinators, an arbitrary cut-off point for monomorphic coding can be applied (e.g. the most important pollinator is at least three times as important as the next most important pollinator (cf. Fenster et al., 2004)), or species can be coded as generalists (e.g. Marten-Rodriguez et al., 2010). Information from pollination networks has been used to differentiate species with relatively generalized pollination systems, by identifying to which pollination network module they belong (Gomez et al., 2014, 2015). These modules can be used as characters in macroevolutionary analyses. For most macroevolutionary studies, data of pollination systems are available for only a subset of species (but see Waterman et al., 2011; Valente et al., 2012). Researchers must therefore choose to base analyses on the subset of species for which pollinator data are available (e.g. Johnson et al., 1998), or to infer pollination systems for species lacking data, to increase power but at potential cost to reliability of character coding. These inferences are often based on pollination syndromes, which are suites of floral traits, including rewards, associated with the attraction and utilization of particular pollinators (Faegri & van der Pijl, 1979; Fenster et al., 2004). Syndromes are thought to indicate the selective agents that have driven convergent evolution of correlated floral traits (Fenster et al., 2004). Tests of syndrome hypotheses by analyzing selection on floral traits by specific pollinator groups through male and female function are, however, rare (Reynolds et al., 2010; Zhou et al., 2020). Quantitative analyses of syndrome traits for species with known pollination systems (‘observed species’) can be used to identify clustering in trait space (Wilson et al., 2004; Marten-Rodriguez et al., 2009; Armbruster et al., 2011), and predict pollination systems of species for which traits could be measured, but for which pollinator data are lacking (‘unobserved species’) (Whittall & Hodges, 2007; Navarro-Perez et al., 2013). Use of syndromes for inferring pollination systems has been strongly criticized (Waser et al., 1996; Ollerton et al., 2009), but their predictive power has been verified with empirical pollination studies (Pauw, 2006; Rosas-Guerrero et al., 2014). However, Johnson & Wester (2017) cautioned that the presence of unrecognized pollination systems in a group can result in false inferences. Additionally, inferences based on syndromes may be sensitive to circular reasoning in macroevolutionary analyses (Armbruster, 1993): if pollination systems are inferred based on floral traits alone, evolutionary associations between floral traits and pollination systems become a self-fulfilling prophecy. Studies of pollination ecology in Erica were historically based on syndrome-based inferences (Vogel, 1954; Rebelo et al., 1985), but also include recent empirical characterization of pollination systems. In the most comprehensive analysis to date, Rebelo et al. (1985) assigned 426 species from the Cape Floristic Region of South Africa to one of four pollination syndromes: wind-, bird-, long-proboscid fly- and insect-pollination. These assignments were not verified by published pollination data and Rebelo et al. (1985) specify that the broadly defined insect pollination syndrome (72% of species) in part reflected a paucity of pollinator data to distinguish putatively bee-, butterfly-, fly-, and moth-pollinated species. Studies of pollination ecology of individual species (available for 1.8% of all 824 Erica spp.) indicate that even based on visitation rate data alone, generalization is rare. Most insect-pollinated Erica species appear relatively specialized, predominantly pollinated by a single functional group (Supporting Information Table S1). These studies also reveal that despite their general validity in the South African flora (Johnson & Wester, 2017), syndrome predictions appear problematic for Erica (misclassification for three out of six cases (albeit not a random selection of species), Table S1). Finally, the syndrome categories proposed by Rebelo et al. (1985) have since been shown to not include all known pollination systems in Erica (Table S1). Based on these arguments, the natural history data upon which Reich et al. are based is inadequate to reach any conclusions about patterns of floral modularity between ecological specialists and generalists and pollinator-driven evolution of floral shape, and alternative explanations for their findings should be considered. In particular, the apparent differences in modularity between flowers with long vs short floral tubes are intriguing. Since specialized pollination systems occur in Erica species with both short and long floral tubes (Table S1), differences in behavior between pollinators of short- (bees, short-tongued moths) vs long-tubed (e.g. birds, long-proboscid flies) Erica flowers may underlie this pattern. Birds perch (Siegfried et al., 1985) and long-proboscid flies hover in front of Erica flowers while feeding (Fig. 43.1 in Barraclough, 2017), whereas short-tongued insects typically crawl over flowers while probing them for nectar and collecting pollen (Van der Niet et al., 2014, 2020). These different foraging behaviors likely result in higher precision of pollen placement in long-tubed vs short-tubed flowers, similar to the pattern found between zygomorphic and actinomorphic flowers (Nikkeshi et al., 2015). The behavioral differences and associated differences in pollen placement precision may be intensified by associations between floral orientation and tube length (typically horizontal for long-tubed flowers, vertical upward or downward facing for short-tubed flowers) (cf. Fenster et al., 2009). Less precise pollen placement in short-tubed flowers may explain the absence of functional modules in comparison to long-tubed flowers. Alternatively, the observed absence of functional modules in short-tubed flowers in Reich et al. may be due to lumping of distinct pollination systems, each with their own unique patterns of modularity, within the ‘generalized’ pollination system of Reich et al. Whether the patterns observed by Reich et al. indeed reflect differences between the level of specialization or alternative hypotheses, requires better information on pollinator identity and behavior. Many contributors have stressed the importance of maintaining and even accelerating natural history research in science (Bartholomew, 1986; Greene & Losos, 1988; Greene, 1994; Futuyma, 1998; Grant, 2000; Schmidly, 2005; Bury, 2006; Tewksbury et al., 2014). Several argue that natural history research is central to our understanding of the world, and is important in the light of the biodiversity crisis (Greene & Losos, 1988; Bury, 2006; Tewksbury et al., 2014). However, the number of contemporary biologists who emphasize the importance of natural history research as the fundamental basis for ecological and evolutionary analyses is still disappointingly small, perhaps because this is assumed to be self-evident, but, more worryingly, this could also reflect a misconception that available natural history information is adequate for our scientific needs. Several relatively recent large-scale studies of pollinator-driven evolution are characterized by limited natural history data on plant–pollinator interactions relative to the sample of species included (Abrahamczyk et al., 2014; Roalson & Roberts, 2016; Hoekstra et al., 2018; Kriebel et al., 2019; Lozada-Gilobard et al., 2019). In some cases this may reflect the pervasive practice of using pollination syndromes as proxies for pollination systems. Assignment of pollination syndromes without quantitative pollination data was a tradition in the latter half of the previous century (Vogel, 1954; Grey-Wilson, 1980) and lists of syndrome assignments can provide a quick way of obtaining data for macroevolutionary analyses, especially if geographical disjunction between plant localities (biodiversity hotspots in developing countries) and research institutes (frequently in the First World) limits opportunities for natural history research. Several approaches can be considered to overcome limitations to macroevolutionary studies due to paucity of natural history data. An obvious starting point is to focus on already well-studied taxa. For example, Landis et al. (2018) built on the seminal work on pollination ecology of Polemoniaceae by Verne and Karen Grant (Grant & Grant, 1965) to test ideas of floral evolution. By contrast, Reich et al. omit species for which pollination systems have been studied in favor of species sourced from a botanical garden collection for which no natural history data were available (Tables S1, S2), possibly reflecting pressure of research institutes to prioritize use of collections over optimal sampling. Three further approaches, focused on potential time limits to collection of natural history data, can be considered. First, citizen science projects can massively increase the number of observation hours, although geographic scope and data quality may limit utility of citizen science data for identification of pollination systems (Mason & Arathi, 2019). Second, motion-activated cameras can be used to record floral visitors, and image recognition software can be implemented to analyze footage (Steen, 2017), allowing detailed review of visitor behavior and quantification of visitation rates (Johnson et al., 2020), though not of pollinator effectiveness. Third, the literature on pollination networks worldwide is growing rapidly and datasets of flower–visitor interactions are often publicly available (e.g. Olito & Fox, 2015). Data derived from network studies is particularly promising if measures of visitation frequency and pollinator effectiveness are incorporated (de Santiago-Hernandez et al., 2019). Although technological advances can be co-opted in sourcing natural history data, field-based studies by naturalists cannot be replaced (e.g. Amorim et al., 2020). Improvement of natural history data relies on a combination of research effort and critical training in organismal identification and ecology for young biologists (Schmidly, 2005). Most importantly, however, especially in the context of the current biodiversity crisis, there is an urgent need for scientists and funding agencies to recognize that the paucity of natural history now impedes our progress in understanding the evolution and functioning of biodiversity (Greene, 2005; Page, 2005). The author would like to thank Adam Shuttleworth, Steve Johnson, Ruth Cozien, Scott Armbruster, Mark Rausher, and an anonymous reviewer for constructive comments on an earlier draft of this manuscript. Thanks to Sally Adam for providing the photograph of Erica gracilis and to Karin Sternberg and Ross Turner for providing information on Erica pollinators. Fig. S1 Floral biology and insect visitation rates of Erica gracilis. Notes S1 Statistical analysis and discussion of floral visitor data for Erica gracilis included in the study of Reich et al. (2020). Table S1 Published pollinator data available for Erica spp. Table S2 Source of pollinator data used in the study of Reich et al. (2020). Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article." @default.
• W3048541059 created "2020-08-18" @default.
• W3048541059 creator A5087173606 @default.
• W3048541059 date "2020-08-12" @default.
• W3048541059 modified "2023-10-16" @default.
• W3048541059 title "Paucity of natural history data impedes phylogenetic analyses of pollinator‐driven evolution" @default.
• W3048541059 cites W1975874549 @default.
• W3048541059 cites W1987114042 @default.
• W3048541059 cites W1993275264 @default.
• W3048541059 cites W1996798188 @default.
• W3048541059 cites W2000584407 @default.
• W3048541059 cites W2000844300 @default.
• W3048541059 cites W2003995245 @default.
• W3048541059 cites W2024145336 @default.
• W3048541059 cites W2026959327 @default.
• W3048541059 cites W2031821271 @default.
• W3048541059 cites W2036858463 @default.
• W3048541059 cites W2038386936 @default.
• W3048541059 cites W2042571642 @default.
• W3048541059 cites W2048669176 @default.
• W3048541059 cites W2055721794 @default.
• W3048541059 cites W2063188949 @default.
• W3048541059 cites W2070868301 @default.
• W3048541059 cites W2093970638 @default.
• W3048541059 cites W2096473110 @default.
• W3048541059 cites W2101724634 @default.
• W3048541059 cites W2102274120 @default.
• W3048541059 cites W2104094242 @default.
• W3048541059 cites W2106325139 @default.
• W3048541059 cites W2112238812 @default.
• W3048541059 cites W2117766354 @default.
• W3048541059 cites W2120142743 @default.
• W3048541059 cites W2120872673 @default.
• W3048541059 cites W2125217979 @default.
• W3048541059 cites W2125527835 @default.
• W3048541059 cites W2125623936 @default.
• W3048541059 cites W2125993511 @default.
• W3048541059 cites W2130023019 @default.
• W3048541059 cites W2132962374 @default.
• W3048541059 cites W2133866716 @default.
• W3048541059 cites W2136976058 @default.
• W3048541059 cites W2148650454 @default.
• W3048541059 cites W2152351051 @default.
• W3048541059 cites W2159085415 @default.
• W3048541059 cites W2160508732 @default.
• W3048541059 cites W2168933851 @default.
• W3048541059 cites W2170146285 @default.
• W3048541059 cites W2181277309 @default.
• W3048541059 cites W2219910692 @default.
• W3048541059 cites W2307686301 @default.
• W3048541059 cites W2318247868 @default.
• W3048541059 cites W2322671961 @default.
• W3048541059 cites W2331288455 @default.
• W3048541059 cites W2342033907 @default.
• W3048541059 cites W2402311143 @default.
• W3048541059 cites W2493686749 @default.
• W3048541059 cites W2510110337 @default.
• W3048541059 cites W2587663446 @default.
• W3048541059 cites W2810591225 @default.
• W3048541059 cites W2886753558 @default.
• W3048541059 cites W2888158582 @default.
• W3048541059 cites W2912472507 @default.
• W3048541059 cites W2933068931 @default.
• W3048541059 cites W2939223928 @default.
• W3048541059 cites W2955569012 @default.
• W3048541059 cites W2991468749 @default.
• W3048541059 cites W2991896110 @default.
• W3048541059 cites W3006161256 @default.
• W3048541059 cites W3012834070 @default.
• W3048541059 cites W3036364018 @default.
• W3048541059 doi "https://doi.org/10.1111/nph.16813" @default.
• W3048541059 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/32786085" @default.
• W3048541059 hasPublicationYear "2020" @default.
• W3048541059 type Work @default.
• W3048541059 sameAs 3048541059 @default.
• W3048541059 citedByCount "18" @default.
• W3048541059 countsByYear W30485410592021 @default.
• W3048541059 countsByYear W30485410592022 @default.
• W3048541059 countsByYear W30485410592023 @default.
• W3048541059 crossrefType "journal-article" @default.
• W3048541059 hasAuthorship W3048541059A5087173606 @default.
• W3048541059 hasBestOaLocation W30485410591 @default.
• W3048541059 hasConcept C102789151 @default.
• W3048541059 hasConcept C104317684 @default.
• W3048541059 hasConcept C151730666 @default.
• W3048541059 hasConcept C163276114 @default.
• W3048541059 hasConcept C168197293 @default.
• W3048541059 hasConcept C18903297 @default.
• W3048541059 hasConcept C193252679 @default.
• W3048541059 hasConcept C2776608160 @default.
• W3048541059 hasConcept C2780618852 @default.
• W3048541059 hasConcept C54355233 @default.
• W3048541059 hasConcept C78458016 @default.
• W3048541059 hasConcept C86803240 @default.
• W3048541059 hasConcept C90132467 @default.
• W3048541059 hasConceptScore W3048541059C102789151 @default.
• W3048541059 hasConceptScore W3048541059C104317684 @default.
• W3048541059 hasConceptScore W3048541059C151730666 @default.

• next