Field sampling.
To test for differences in soil properties, plant attributes, and theR. nudiflora -AMF interaction (i.e. colonization rate) between urban and rural environments (questions 1 and 2), and to explore the relationship between plant and soil attributes on colonization rates (question 3), we collected R. nudiflora individuals, including their roots and their rhizospheric soil in urban and rural environments (Fig. 1, Fig. S1). Specifically, for the urban environment, we sampled in the oldest parts of Mérida city, including sampling sites at least 3 km inside the ring highway (Fig. S1). We reduced the high environmental heterogeneity of the city creating two extreme categories where R. nudiflora can be found: deep urban sites (hereafter DUS), and open urban sites (hereafter OUS). DUS are sites with a high degree of impervious surface and low plant density in which R. nudifloracan be found growing in cracks of sidewalks and road asphalt (Rauppet al. , 2010);meanwhile OUS include sampling sites in parks and road medians strips where R. nudiflora can also be found (Fig. 1). Rural sites (hereafter, RS; Fig. 1) were sampled in rural areas as far away from human settlements as possible, and sampling sites were situated at least 3 km from the city ring highway (Anillo Periférico; see Fig S1). In urban (DUS and OUS) and rural environments (RS) the spatial distribution of the sampling sites were designed to be equally distributed across each cardinal orientation; however due to lack of plants in certain areas or difficult access (e.g. lack of roads in the rural environment) the sampling design was not fully balanced. We ensure that the distance between the nearest sites was at least 300 meters (Fig. S1). For each environment (DUS, OUS and RS) five plants were sampled in 20 sites (3 environments/20 sites/5 plants) for a total of 300 sampled plants.