Background and aims – Understanding the traits that lead to the invasion potential of invasive alien species (IAS) provides insight for their management. The reproductive traits of IAS help us understand the mechanisms that allow for their invasive potential, and colonization into new ranges. Asphodelus fistulosus is a native Mediterranean species commonly found invading Australia, South-East Asia, South Africa, and North America.

Materials and methods – Two populations of A. fistulosus in the Chihuahuan Desert were monitored for reproductive phenology. Floral visitors and their behaviour were described, and we assessed the breeding system through floral morphological characters and the mating system in controlled pollination experiments.

Key results – Reproductive phenology showed continuous reproduction throughout the year. Floral morphology suggested a facultative autogamous breeding system, but the mating system was mixed with autonomous selfing. Flowers lasted one day, with anthesis lasting 11 h. Floral visitors of A. fistulosus consisted of a variety of taxa including species of Coleoptera, Hymenoptera, and Lepidoptera, the exotic Apis mellifera being the most frequent visitor.

Conclusions – The reproductive traits of A. fistulosus in the invaded range provide the biological potential for further invasion. The continuous production of reproductive structures attracts many diverse pollinators, and the autonomous self-pollination implies that a single plant has the potential to develop a new population, which makes the control of this IAS a global challenge.

breeding and mating system, floral visitors, invasive alien species, phenology

Current severe environmental threats are brought about by changes in land use, climate change, and invasive alien species (IAS; Dirzo and Raven 2003), the latter being identified as one of the leading causes of species extinctions worldwide (Pejchar and Mooney 2009). Several hypotheses have been put forward to explain the success of biological invasions (Essl et al. 2015). For invasive plant species, a subgroup of these hypotheses links successful invasion to traits that allow these species to establish new populations in their invaded range. Traits such as the ability of having sexual and asexual reproduction (Moravcová et al. 2015), multiple phenological strategies (Wolkovic and Cleland 2011), high germination rates (Gioria and Pyšek 2017), and a persistent seed bank (Gioria et al. 2021), facilitate the ability to become invasive and provide insight into the biology of IAS. These traits, many of which were partially described by Baker (1965) in his treatment of the perfect weed, usually confer the advantage of higher population growth than native species (Doody et al. 2009) and provide a working hypothesis expected to be found in IAS. For many plants, the ecological characteristics that favour establishment success are poorly understood (Gallien and Carboni 2017). However, the identification of traits that confer invasive potential (Van Kleunen et al. 2010) are also an important component of risk assessment protocols (Pheloung et al. 1999) and a means of characterizing weedy species (Baker 1965; Sutherland 2004). Understanding the drivers of establishment and spread of IAS beyond their native range is key to predicting new invasions (Pyšek et al. 2020), although results do not always support the predictions.

Variation in mating and breeding systems of invasive plant species is broad (Barrett 2002). Reproductive systems can change after colonization (Ferrero et al. 2020) and Baker’s Law emphasizes a preference of self-compatible over self-incompatible species (Baker 1965; Williamson and Fitter 1996; Barrett 2011). However, successful invasive plant species have a wide variety of mating systems, from self-compatibility (Jacquemart et al. 2015; Redmond and Stout 2018), self-incompatibility (Sutherland 2004; Friedman and Barrett 2008), and even mixed mating systems (Souza et al. 2016). There is also the argument that IAS benefit from non-specialized pollination systems under the assumption that new habitats pose pollinator limitations that can be avoided (Stout et al. 2002) with generalized pollinators found in the invaded range (D’Antonio et al. 2000; Stout and Tiedeken 2017) or in the absence of pollinators, favour species with autonomous self-pollination (Pannell et al. 2015). Even though most plants, including IAS, are pollinated by animals (Ollerton et al. 2011), self-compatibility and autonomous pollination reduce the dependence on biotic interactions (Van Kleunen and Johnson 2005; Van Kleunen et al. 2008) favouring invasion success. The loose interactions with native pollinators can even decrease fitness (Knight et al. 2005; Burns et al. 2013) or may generate new interactions with native pollinators causing a shift in preference and a competition for floral visitors that negatively impact native plant species (Chittka and Schürkens 2001; Mitchell et al. 2009).

Seed output in IAS is thought to be high (Díaz-Segura et al. 2020), and even though IAS invasions benefit from asexual reproduction in clonal spread (Guerra-García et al. 2015), there are successful IAS that rely entirely on sexual reproduction (Forman and Kesseli 2003). The overall behaviour of reproductive traits in terms of seasonality is correlated with overall fitness linked to environmental signals. For example, phenology is a key component of plant reproduction (Cleland et al. 2007), and it provides relevant information to understand the impacts of IAS in invaded ecosystems (Vilà et al. 2011). Hypotheses that help explain their success suggest that IAS shift flowering with climate change compared to native species (Wolkovich et al. 2013), invade habitats by having phenological differences with the native community (Wolkovich and Cleland 2014), and can germinate earlier, faster and respond to broader germination cues in their non-native than in their native range (Gioria and Pyšek 2017). This variation in phenological phases and phenotypic plasticity can increase or decrease niche overlap with native plants species or increase fitness when cued with the prevailing environmental conditions at the site of introduction (Alexander and Levine 2019).

Asphodelus fistulosus L. (Xanthorrhoeaceae) is a native herb from the south of Europe, found in the Mediterranean basin and Macaronesian region (Díaz-Lifante and Valdés 1996), a region that includes a wide variety of climates, mainly classified as temperate with no dry season and a mild/hot summer as well as temperate with dry and warm summers. The average annual precipitation is 1512 mm at the westernmost portion and decreases to 749 mm towards the east (Cropper 2013). It is a nitrophilous species, locally frequent in grasslands, steppes, and coastal sandbanks, preferably on alkaline and occasionally siliceous soils, in secondary habitats linked to human activity (Díaz-Lifante and Valdés 1996). Asphodelus fistulosus is considered invasive in Southeastern USA, Australia, India, South Africa, and New Zealand (Patterson 1996; Boatwright 2012), its introductions have been intentional for ornamental purposes (Bailey and Bailey 1976; Russel 2008) but the species can easily escape from cultivation (USDA 2016) and establish populations in disturbed, over-grazed habitats (Victoria State Department 2016). It impacts vegetation cover (Martínez-Cruz and Téllez-Valdés 2004) forming dense populations in arid and semi-arid environments and disturbed areas (Parsons and Cuthbertson 2001; Cullen 2012). The first reports of A. fistulosus in Mexico were from the northern arid environments (Conzatti 1946a, 1946b; Villaseñor and Espinosa-Garcia 2004), and since then, the species has considerably increased its geographical distribution, and is now found in 15 out of 32 states (Guerrero-Eloisa 2017).

The purpose of this study was threefold: (1) describe the flowering phenology of A. fistulosus at two invaded sites in the Chihuahuan Desert, (2) identify floral visitors and how these change over floral anthesis, and (3) describe the breeding system using morphological floral characters and define the mating system through field-controlled pollination experiments to assess how these traits can favour the invasion potential of A. fistulosus in Mexico.

Reproductive traits in IAS are considered important components in their invasion potential (Baker 1974). Phenological events can be advantageous for invasion in several ways providing information on the success of IAS (Baker 1974): the potential to change phenophases in response to different habitats, anticipated or delayed flowering, and extended or continuous phenophases (Wolkovich and Cleland 2011, 2014; Wolkovich et al. 2013). IAS have been shown to extend flowering periods that confer advantages over native plant species (Pyšek et al. 2003; Pyšek and Richardson 2008), with examples across the taxonomic spectrum such as in Bidens frondosa L. (Yan et al. 2016), Leonotis nepetifolia (L.) R.Br. (Díaz-Segura et al. 2020), and Coreopsis lanceolata L. (Zeng et al. 2021). Using A. fistulosus citizen science data from 170 observations for both sites (QRO = 42 observations, SLP = 128 observations), reproductive phenophases could be seen year-round with a peak in mid-winter for QRO (buds, flowers, and fruits). In SLP, two of the phenophases (flowers and fruits) did not show a clear peak, they were distributed uniformly year-round, but buds did show a marginally significant peak during winter. Comparing the study and citizen science data there is no clear relationship for either site. These differences could be a response to local conditions, frequency of observations, and geographic coverage of observations, which highlights the importance of citizen science data for large-scale phenological data and the relevance of small-scale phenological data that shows the effects of local conditions on phenological stages. The continuous reproductive pattern suggests a wide niche breadth (Wolkovich and Cleland 2011) that enables reproductive events year-round, similar to other IAS found in the Brazilian savannas (Xavier et al. 2019), a trait that enhances the potential for invasion.

A second component in the success of IAS that is reflected in phenological events is phenotypic plasticity. At the study sites in the Chihuahuan Desert, A. fistulosus reproduction peaked during two different seasons (summer in QRO and spring in SLP), while data from other countries suggest flowering peaks in summer in the USA (DiTomaso et al. 2013) and during spring in South Africa (August–October; Boatwright 2012). In the native Mediterranean range of A. fistulosus, flowering occurs between December and June (Boatwright 2012), while congeners are known to flower during March–May for A. albus Mill. (Obeso 1992) and June to September for A. aestivus Brot. in Spain and May to April in Portugal (Díaz-Lifante 1996). There is therefore evidence that phenologically, A. fistulosus and quite likely its congeners follow two strategies: extended flowering periods as well as the potential to easily change phenophases depending on local conditions.

As a possible consequence of extended flowering periods, there is also an enhanced attraction towards floral visitors (Ojija et al. 2019). Extended flowering periods of A. fistulosus and the interaction with climatic variables suggest a response to different habitats that provides an advantage over native species (Pyšek et al. 2003; this study). The floral resources generated by a single individual of A. fistulosus can be significant, producing 30–60 flowers per plant (Oscar Sandino Guerrero-Eloisa pers. obs.) and this abundance of floral resources can potentially divert native pollinator species from visiting native flora (Powell et al. 2011; Yan et al. 2016) and even increase visitor frequency in congeneric sympatric species (Zeng et al. 2021). Generalist pollination systems are thought to favour the invasive potential of IAS (Baker 1974), such that attraction is not confined to a specific group of visitors (Stout et al. 2006). Within the genus Asphodelus, a diverse assemblage of species visits the flowers in its native range, A. mellifera being common but also including Xylocopa Latreille, 1802, Bombus Latreille, 1802, Anthidium Fabricius, 1805, Chelostoma Latreille, 1809, and Megachile Latreille, 1802 (Obeso 1992; Díaz-Lifante 1996). Lara (2009) describes 11 species (ten bees and one beetle) as flower visitors and potential pollinators of A. fistulosus in its native range, including five Bombus species, four Xylocopa species, Apis mellifera, and Agapanthia asphodeli Latreille, 1804. Lara (2015) describes two species as confirmed pollinators of A. fistulosus (Xylocopa cantabrita Lepeletier, 1841 and Apis mellifera), our study showed that A. mellifera was the most common visitor that would mean a first step in invasive species favouring the success of another IAS (invasion meltdown; Simberloff and Holle 1999) but there were also new associations with three insect orders (Hymenoptera, Lepidoptera, and Coleoptera). The presence of A. mellifera impacts pollination systems in invaded ranges because of its negative effect on native plants and positive effects on invasives (Morales et al. 2017). Secondly, A. mellifera is not usually an efficient pollinator of native plants (Santos et al. 2012) and is resilient to disturbance (Winfree et al. 2009). The presence of A. mellifera in Mexico is widespread with feral, Africanized (Barrios et al. 1990) and managed colonies brought about through apiculture (Labougle and Zozaya 1986). In the native range of A. fistulosus, Agapanthia asphodeli (Coleoptera) acts as a floral visitor, while Trichochrous sp. (Coleoptera) was found in QRO. If we consider the extended flowering period, generalist pollination systems, new associations with native visitors, and positive feedbacks with exotic floral visitors, the potential success of A. fistulosus as an IAS is certainly present. These interactions not only benefit the exotic species, but may also cause competition with native plant species for floral visitors (Stout and Tiedeken 2017). Of the native pollinator species found on A. fistulosus in the present study, Lasioglossum sp. (native to America) and Trichochrous sp. have been reported as frequent visitors of Cactaceae in the same area (Briseño-Sánchez et al. 2020). For the butterflies, even though the main activity on A. fistulosus was the collection of nectar, it is known that they can be an important pollinator for plant species (Zhang et al. 2011; Geerts and Adedoja 2021). Lara (2015) describes 17 butterfly species as potential pollinators (four families) in the native range of A. fistulosus and in this study, nine species from five families were present. In both studies, the Nymphalidae and Pieridae families were found in the native range as well as in the places where A. fistulosus invades.

Not only is the length of the flowering period relevant for IAS success but also floral longevity, because these determine on one hand the seasonality in the reproductive periods and on the other the availability of resources at any given time (Janzen 1971). This can be understood as two opposing strategies, short-lived (< 1 d) floral resources that are spread over a long period of time or long-lived floral resources (> 1 d) over a shorter time period usually associated to generalist pollination syndromes (Yan et al. 2016). There are few studies that have addressed floral longevity in IAS (single flower), with evidence in species with anthesis that can last 4–5 d in Bidens frondosa (Yan et al. 2016), 5–6 d in Coreopsis lanceolata (Zeng et al. 2021), 4 d in Stapelia gigantea N.E.Br., and 6–11 d in Kalanchoe daigremontiana Raym.-Hamet & H.Perrier (Herrera and Nassar 2009), and others that are short-lived usually one day such as Leonotis nepetifolia (Díaz-Segura et al. 2020) or less than 24 h (Alegro et al. 2010; this study). Short-lived floral resources that are spread over a long time period can favour recurrent visitors, which maximize the amount and activity of the floral visitors and promotes outcrossing due to the number of receptive flowers in the short time period (Janzen 1971). Even though outcrossing is unnecessary for A. fistulosus, it generates genetic recombination and can potentially increase seed set in some species (Díaz-Segura et al. 2020).

The role played by mating and breeding systems spurred Baker’s law, according to which selfing species would be better colonizers (Baker 1967). The evidence supporting this hypothesis seems to be quite widespread among IAS (Baker 1967). Furthermore, mixed mating systems guarantee offspring in new habitats whereby autonomous self-pollination generates progeny and dispersal without the need of another individual (Cruden 1977). The reproductive success in A. fistulosus is a consequence of an outcrossing and a selfing system, a trait that leads to higher genetic variability and long-term survival (Cruden 1977). However, extreme cases of selfing may hinder future growth (Van Kleunen and Johnson 2007) through inbreeding (Novak and Mack 2005; Sakai et al. 2001) but floral visitors can favour outcrossing, generating more vigorous recombinant seeds. Evidence of this is contradictory, some IAS species are highly successful through entirely clonal reproduction (Corredor-Prado et al. 2015; Guerra-García et al. 2015), while others have even higher genetic variation than the native populations (Wang et al. 2016; Lucardi et al. 2020; Smith et al. 2020). Furthermore, many IAS possess self-incompatibility (17 species in South Africa; Rambuda and Johnson 2004) among others such as Mikania micrantha Kunth (Hong et al. 2007) and Ambrosia artemisiifolia L. (Friedman and Barrett 2008). Self-compatibility would favour continuous seed production, population maintenance, and some degree of dispersal (Herrera and Nassar 2009; Zhang et al. 2011; Zeng et al. 2021) and the continuous availability of floral buds, flowers, and fruits in A. fistulosus and its mixed system could potentially lead to further invasion in arid environments (Janzen 1971; Díaz-Segura et al. 2020).

In addition to the number of reproductive traits found in A. fistulosus that very likely increase its invasive success, the establishment of A. fistulosus in areas that present a high disturbance regime also contributes to their success (Elton 1958; Hobbs and Huenneke 1992). Asphodelus fistulosus forms large patches of vegetation causing a reduction of native species, impacting the biodiversity of the ecosystem (Elton 1958; Levine and D’Antonio 1999). The ornamental use of A. fistulosus (Jeschke and Strayer 2006) and the lack of natural enemies (Keane and Crawley 2002) also favour invasion (Blumenthal 2006), increasing the availability of resources for pollinators and the possibility of pollination by exotic species (Simberloff and Holle 1999). The set of invasive traits described in A. fistulosus shows the invasive potential, especially given the wide range of attributes considered in the ideal weed that are expressed in the invaded range.

All reproductive phenophases of the species in both populations were found throughout the year, providing continuous availability of resources for floral visitors. However, flower and fruit production peaks differed between populations, suggesting that reproductive phenology responded to local conditions. The large number of flowers favoured the presence of native visitors, which ranged from nectarivorous species of Lepidoptera, species of native bees (e.g., Lasioglossum sp. and Ceratina sp.), to exotic bees (Apis mellifera) that collect pollen and nectar. Our evidence supports Baker’s law that self-pollinated species would be better colonizers. Furthermore, the mixed mating system of A. fistulosus guarantees variable offspring and dispersal to new habitats, and through autonomous pollination it generates progeny without the need for another individual. The invasive potential of onionweed within the Chihuahuan Desert is favoured by its mating system and phenological plasticity facilitating its expansion to other areas, prompting an urgent need to establish plans for its control.

The first author is a student of the PhD Program Doctorado en Ciencias Biológicas y de la Salud - Universidad Autónoma Metropolitana Xochimilco (UAM-X), and the paper is part of his dissertation in partial fulfilment of the requirements for the graduate program. This research was funded by Consejo Nacional de Ciencia y Tecnología (CONACyT) that awarded a scholarship to O.S.G-E. (815924). Financial and logistical support was provided by the Instituto de Ecología, UNAM, SEP-CONACyT 221362 and the GEF 00089333 project “Enhancing National Capacities to manage Invasive Alien Species (IAS) by implementing the National Strategy on IAS” to M.C.M. and J.G. We thank Mariana Rojas-Arechiga for logistical support and comments. Isabel Briseño Sánchez, Diana Cárdenas Ramos, Esteban Munguía Soto, Omar Díaz Segura, José Aranda Pineda, Jessica Reyes Tovar, Gerardo Manzanares Villasana, and Linda Martínez Ramos for field assistance. Two anonymous reviewers and the associate editor provided thoughtful insight and their suggestions helped improve the manuscript.