The five epiphytic species we found belong to the Bromeliaceae family, the second with most epiphytic members (1770 spp, Zotz 2016), and to the genus Tillandsia, the richest in epiphytic species (635 epiphytic species) among the Bromeliaceae (Zotz 2013). This over representation of bromeliad species, especially Tillandsia species in urban ecosystems has been already reported for America (Furtado and Neto 2015; Santana et al. 2017; Alvim et al. 2021). It is probably owed to drought adaptations in Tillandsia species, together with the tolerance that some species of this genus have for air pollution (Benzing 2000; Graciano et al. 2003; Miranda et al. 2016; Piazzetta et al. 2019).
Furthermore, there is indication that cities and other human-altered ecosystems, have an over representation of xerophytic species that are tolerant to conditions related to human activity which cause microclimatic changes such as higher solar radiation, temperature and wind velocity, as well as lower humidity compared to undisturbed environments (Padmawathe et al. 2004; Wolf 2005; Cascante-Marín et al. 2006; Werner and Gradstein 2008).
The fact that only T. recurvata indicated viable populations structures could be linked to this species having one of the broadest distribution ranges of all the bromeliads, ranging from Florida to Argentina, with presence in almost every ecosystem, even in semi-desert environments (Smith and Downs 1977). Different adaptations allow this species to thrive in the driest environments (e.g CAM photosynthetic metabolism; presence of trichomes on both sides of the leaves that protect stomas preventing water loss (Loeschen et al. 1993; Piazzetta et al. 2019)) explaining why the species is considered the most xerophytic species among all epiphytes (McWilliams 1992; Chaves and Rossatto 2021). For instance, Pérez-Noyola et al. (2020) reported 100% vivipary in T. recurvata seeds evaluated in the Chihuahuan desert. Vivipary is germination inside the fruits and has been associated with thermoregulation, parental care, conspecific nursing and rapid seedling establishment, constituting an adaptation to harsh environments (Cota-Sánchez and Abreu 2007; Pérez-Noyola et al. 2020). So, the presence of this trait could improve the establishment of T. recurvata in the harsh environment represented by Oaxaca city. However, to confirm this idea it is necessary to test the presence of vivipary in Oaxaca city, because Pérez –Noyola et al. (2020) found that the trait is not widespread in all T. recurvata populations. Another adaptation that could explain the dynamic population structure of this species in Oaxacan parks, is its ability to accumulate and filter air pollutants in cities without any decrement in its development (Graciano et al. 2003; Miranda et al. 2016; Piazzetta 2019).
Additionally, T. recurvata is one of the few vascular epiphytes classified as a weed (Claver et al. 1983), partly due to its huge reproductive capacity, both sexually (self-compatible, large number of fruits and seeds, relatively short periods for fruit ripening and size to first reproductive event) and asexually via vegetative propagation (Fernández et al. 1989; Orozco-Ibarrola et al. 2015; Chaves et al. 2021). This was reflected in the population structures recorded in Oaxaca city, with large proportions of seedlings and infants, evidencing good recruitment of new individuals. Such reproductive capacity together with the high abundance of T. recurvata already present in Oaxaca city, act synergistically to increase its abundance, producing heavy seed rain. According to different authors, epiphytic communities are reflections of the abundance and composition of propagule rain (Yeaton and Gladstone 1982; Cascante-Marín et al. 2009; Zotz 2016).
In the case of T. schiedeana, despite an abundance of 95 individuals, the observed structures indicate a senescent or declining populations, with little evidence of recruitment. Even when the species has adaptations to deal with dry environments (e. g. CAM metabolism, peltate trichomes, CO2 recycling (Loechen et al. 1993)), T. schiedeana has a smaller distribution range than T. recurvata and has lower abundances when growing simpatrically (Orozco-Ibarrola et al. 2015). This could result from the synergetic effect of: a) T. recurvata inhibiting the germination of other Tillandsia species, including T. schiedeana (Claver et al. 1983; Valencia-Díaz et al. 2012) and b) schiedeana producing less fruit and with longer ripening periods than T. recurvata. In our study site, T. recurvata could be limiting the germination of T. schiedeana, since almost all trees had individuals of T. recurvata, and we observed that T. schiedeana required a year from fruit production to seed dispersal, while T. recurvata only required six months (per. Obs.). But more studies are necessary to understand why the populations of T. schiedeana shows a type III structure.
The presence of only one individual of T. ionantha, T. makoyana and T. sp. could be the result of fortuitous long distance dispersal events, that could result in the establishment of new populations of these species. As many other epiphytic species T. ionantha and T. makoyana have mixed breeding systems that allow a single individual to produce viable seeds which can develop into adult individuals (Mondragon et al. 2015). However, these isolated dispersal events could result in populations which do not prosper, as in the case of T. schiedeana.
The lack of differences found in the structures of populations growing on native vs. exotic trees for both Tillandsia species, may be related to the fact that, within each host category, there was a mix of tree species with different morphological, physiological and chemical characteristics, which can impact the demographic behavior of epiphytic individuals (Einzmann et al. 2015; Martins et al. 2020; González and Ceballos 2021; Ramirez et al. in press). This agree with results by Martins et al. (2020) showing a lack of preference between exotic and native hosts by vascular epiphytes in a Brazilian urban area. This lack of an effect of tree origin on associated species agrees with results by Berthon et al. (2021) who examined the relationship between native plants and animal biodiversity in urban areas. The authors found that the resources provided by the plants were more important predictors than their origin, but when in doubt, nativeness was a good surrogate of whether a plant would provide food for local animals. In our case, even when our result suggests that there are no differences in population structures related to tree origin, we recommend taking those results in consideration and developing an experiment to assess the effect of the origin of phorophytes over the demography of vascular epiphytes.