The interactions between the plant community and their floral visitors might vary through space, causing changes in the structure and composition of the communities of floral visitors. Although numerous studies have addressed the variation in the communities of floral visitors through space, none of them have analyzed how the structure of the community of floral visitors for all the co-flowering annual plant species changes between sites. We describe how the community of floral visitors to annual plants varies through space in a xeric community in central Mexico. We collected all the insects visiting the flowers of the annual plants growing in 2 sites with contrasting plant density (low vs. high). We determined species richness, abundance, and diversity for the insect communities in each site. We established the similarity in the composition of the insect communities and the importance of environmental variables on the dynamics of the communities of floral visitors. Abundance and diversity of floral visitors were significantly higher in the low plant density site. The composition of the insect communities differed between sites (similarity = 32%). Changes in abundance of a few groups of insects seem to be related to environmental factors such as disturbance (Myrmecocystus mexicanus Wesmael [Hymenoptera: Formicidae] and Tiphiidae sp. 1 [Hymenoptera]), humidity (Curculionidae sp. 1 [Coleoptera]), and temperature (Dorymyrmex grandulus (Forel) [Hymenoptera: Formicidae]). Our results indicate that vegetation density has a strong effect on the community of floral visitors of annual plants, which in turn, might affect their reproduction and their interactions with other organisms within the ecosystem.
Numerous studies have shown spatial variation between interacting species of plants and their floral visitors (Horvitz & Schemske 1990; Gilbert et al. 1996; Fenster & Dudash 2001). These studies have documented the variation in composition (Grombone-Guaratini et al. 2004; Moeller 2005; Gibson et al. 2006), abundance (Traveset & Sáez 1997; Steffan-Dewenter et al. 2001; Moeller 2005), species richness (Gibson et al. 2006), and diversity (Herrera 1988) of floral visitors across space. For example, it has been shown that the abundance of floral visitors increases with the closeness to a source of water or to a semi-natural habitat, with the presence of a facilitator species, with higher vegetation coverage, and with lower altitude (Traveset & Sáez 1997; Steffan-Dewenter et al. 2001; Moeller 2005). Similarly, Gibson et al. (2006) and Herrera (1988) have documented higher species richness and diversity at sites with agricultural management, and close to a source of water, respectively.
Most studies addressing the spatial variation in the structure of the community of floral visitors are focused on a subset of species within the community. Thus, most studies are focused on the community of floral visitors or pollinators to 1 or 2 species of plants (Moeller 2005; Gibson et al. 2006; Campbell & Husband 2007). In addition, some studies are focused on specific groups of visitors (Dupont & Skov 2004; Moeller 2005; Rader et al. 2012). Sampling a sub-set of species within a particular community might limit the scope of the results obtained. Thus, generalizations about the environmental factors driving the distribution and responses of floral visitors should be made with caution.
Although numerous studies have documented the existence of spatial variation in different aspects of the communities of floral visitors (Gilbert et al. 1996; Fenster & Dudash 2001; Gibson et al. 2006), none of them have conducted a thorough analysis of the spatial variation in the structure of the community of floral visitors for all the coexisting plant species with the same growth form. In this study, we analyze the variation in the structure of the community of floral visitors of annual plants in a xeric environment in Central Mexico.
Material and Methods
This study was conducted within the Tehuacán-Cuicatlán Reserve, in the locality of Zapotitlán Salinas, Puebla, Mexico (18.3333°N, 97.4666°W), at an altitude of 1460–2600 masl (Zavala-Hurtado 1982). Mean annual temperature is 18–22 °C. Precipitation is scarce throughout the year, with mean annual precipitation of 400–450 mm (Zavala-Hurtado 1982). Climate in the locality is arid to semiarid, with a summer rainy season (Arias-Toledo et al. 2000). The predominant vegetation at the study site is xeric scrubland (Rzedowski 1978), with columnar cacti as the most important physiognomic element (Valiente-Banuet et al. 1997).
At this locality, 2 sites with contrasting vegetation density were selected to conduct the study. Plant density was estimated by counting the number of plants within an area of 50 m2 at each site. Because of the differences in plant density, atmospheric temperature and humidity also are different between sites (Table 1). Temperature and humidity were measured with a digital thermo-hygrometer every h during the d when insects were collected (see the Floral Visitors section below). Maximum temperature was higher at the low than at the high plant density site. Also, on average, temperature at the low plant density site was 2.4 °C higher than at the high plant density site (Table 1). In contrast, minimum RH was higher in the high plant density site than in the low plant density site, and average RH was 9.6% higher in the high than in the low plant density site (Table 1). Moreover, these sites differ in level of disturbance (Table 1). Local people use the low plant density site as horse racing track and it is also a common trail for livestock. Thus, soil compaction and disturbance were higher at the low plant density site. We acknowledge that conducting the study at only 2 sites with no replication might limit our conclusions; however, replications were logistically impossible. Many other studies analyzing spatial variation on different aspects of floral biology (including floral visitors) have been conducted at only 2 sites (e.g., Steffan-Dewenter et al. 2001; Grombone-Guaratini et al. 2004; Campbell & Husband 2007; Stucky et al. 2012; Czarnecka & Denisow 2014).
Environmental characteristics of the 2 sites in which collections of floral visitors were conducted.
THE COMMUNITY OF ANNUAL PLANTS
The study was conducted using the complete community of annual plants growing on the study sites, which totaled 11 species. Eight of them were located at both the high and low plant density sites: Allionia incarnata L. (Nyctaginaceae), Bouteloua barbata Lag. (Poaceae), Eragrostis mexicana (Hornem) (Poaceae), Flaveria ramosissima Klatt. (Asteraceae), Florestina pedata (Cav.) (Asteraceae), Parthenium bipinnatifidum (Ortega) (Asteraceae), Zinnia peruviana L. (Asteraceae), and Gomphrena decumbens Jacq. (Amaranthaceae). In addition, 3 other plant species were found in only 1 of the sites: Dalea humilis G. Don. (Fabaceae) and Portulaca pilosa L. (Portulacaceae) were present only at the high plant density site, whereas Kallstroemia rosei Rydb. (Zygophyllaceae) was present only at the low plant density site. Therefore, the community of annual plants at the high and low plant density sites was composed of 9 and 10 species, respectively. Morphological and ecological traits of the annual species are described in Table 2.
We collected all the insects visiting the flowers of the annual plants growing in 2 sites with contrasting plant density. Insects were collected on 20 and 21 Sep 2011 from 7 to 19 h, and on 22 and 23 Sep 2011, from 20 to 23 h. Collections of floral visitors were made on 2 floral patches of 1 m2 for each plant species at each site (Table 3). Sampling areas were chosen based on the abundance of flowers of each plant species. Collections were conducted simultaneously at both sites for 15 min each h, except in A. incarnata, K. rosei and P. pilosa, which have short periods of anthesis. Therefore, the collection of floral visitors of A. incarnata was conducted from 8 to 15 h, whereas in K. rosei and P. pilosa it was from 10 to 17 h. Thus, total time of insect collection varied among species from 210 min (A. incarnata), and 240 min (K. rosei and P. pilosa), to 510 min (all other plant species) at each site. All flying and crawling floral visitors were collected with insect nets and ethyl acetate-killing chambers, respectively. Nocturnal collections were made using hand lamps covered with red translucent paper in order to avoid collecting insects attracted by white lights instead of only those that were attracted by the inflorescences of the studied species. Weather conditions on the sampling dates were mostly sunny (mean temperature: 24.4 °C, mean atmospheric humidity: 50.9%). All insects collected were sacrificed and stored in glassine paper bags or 70% alcohol. Collected insects were identified to the lowest possible taxonomic level.
Biological traits of the herbaceous plant species that constitute the community of annual plants at the study site.
Floral density (number of flowers or inflorescences/m2) of each annual plant species at 2 sites with contrasting plant density. ND= species not distributed at this site.
Total abundance of each insect species, abundance per insect group, total abundance of floral visitors, total species richness, per group-species richness and the Shannon-Wiener diversity index were determined for the communities of annual plants at each site (high versus low plant density). To establish if species richness, per groupspecies richness, per group-abundance, and total abundance differed between sites, c2 tests were applied (Zar 1999). Abundance tests at the insect species level were conducted only on those cases in which the total number of individuals collected (pooling data from both sites) was equal or higher than 5. Thus, we conducted abundance tests for: Curculionidae sp. 1 (Coleoptera); Dasytinae sp. 1 (Coleoptera: Melyridae); Exoprosopa sp. 1, Lepidophora sp. (both Diptera); Nysius sp. 1, Lopidea sp. 1, Miridae sp. 1, Harmostes sp. 1 (all Hemiptera); Apis mellifera L., Exomalopsis pueblana Timberlake, Dorymyrmex grandulus (Forel), D. insanus (Buckley), Myrmecocystus mexicanus Wesmael, Tiphiidae sp. 1 (all Hymenoptera); Ascia monuste (L.), and Nathalis iole iole Boisduval (both Lepidoptera). Therefore, a total of 28 χ2 tests were conducted to compare insect abundance (total, per-group , and perspecies ) and richness (total and per-group ) between sites. In order to prevent the increase of the likelihood of committing a type I error, a sequential Bonferroni procedure was conducted on this set of 28 tests of abundance and species richness (Rice 1989).
Relative abundance curves (James & Rathbun 1981) were used to compare the patterns of species abundance between sites with contrasting plant density. Relative abundance was estimated as ni/N, where ni = number of individuals of the ith insect species, and N = total number of individuals from all species including both high and low plant density sites. Then, relative abundance was plotted on a log 10 scale against the rank from the most to the least common species.
Differences in the Shannon-Wiener diversity indices of floral visitors to annual plants in each site were explored by applying Hutcheson t tests (Hutcheson 1970). Diversity and t tests analyses were carried out in PAST v. 2.17 (Hammer et al. 2001).
The composition of floral visitors was compared between the communities at each site with the Sørensen similarity index. In addition, a cluster analysis with abundance data was performed using the unweighted pair-group average as grouping method, and the r- Pearson as distance. The analysis was performed with the software Statistica ver. 7.1 (StatSoft 2006. STATISTICA. Data analysis software system and computer program manual. Version 7.1. StatSoft Inc., Tulsa, Oklahoma).
Finally, we analyzed the dynamics of the communities of floral visitors in relation to environmental variables. We conducted 2 canonical correspondence analyses, 1 at the order level and the other at the specific level for those species whose abundance was significantly different between sites. Temperature, humidity, and disturbance were the environmental variables considered within each canonical correspondence analysis. Measurements of both temperature and humidity were taken each h during insect collections, then mean values were estimated and used in the analyses. These analyses were conducted in the program PAST v. 2.17 (Hammer et al. 2001).
A total of 113 species belonging to the orders Coleoptera, Diptera, Hemiptera, Hymenoptera, and Lepidoptera were collected visiting the flowers of the annual plants at both sites (Table 4). The communities of floral visitors of annual plants at the high and low plant density sites were composed of 55 and 81 insect species, respectively. Species richness of floral visitors did not differ significantly after the Bonferroni procedure (χ2 = 4.97; df = 1; P > 0.05).
The insect groups with the highest species richness were Hemiptera and Diptera at the high and low plant density sites, respectively (Tables 4, 5). In contrast, Hymenoptera and Coleoptera were the groups with the lowest species richness at the high and low plant density sites, respectively (Tables 4, 5). However, there were no significant differences in species richness for any of the insect orders (3.52 ≤ χ2 ≥ 0.13; df = 1; P > 0.05).
A total of 571 individual insects were collected visiting the flowers of the community of annual plants across the 2 study sites. The number of individual insects visiting the flowers of annual plants at the low plant density site (339) was significantly higher than the number of floral visitors at the high plant density site (232; χ2 = 20.05; df = 1; P < 0.001; Tables 4, 5).
Coleoptera and Hymenoptera were the most abundant groups of insects visiting the flowers of annual plants at the high and low plant density sites, respectively (Tables 4, 5). In contrast, Lepidoptera was the least abundant group of floral visitors at both sites (Tables 4, 5). Although the total abundance of floral visitors from each insect group was different between sites, this difference was only significant for Hymenoptera (χ2 = 79.53; df = 1; P < 0.001). There were significantly fewer hymenopterans at the high plant density site (16) than at the low plant density site (120)
Abundance of insects collected visiting the flowers of annual species in 2 sites with contrasting plant density in Zapotitlán Salinas, Puebla, Mexico.
Species richness (abundance) per group of insects collected visiting the flowers of the communities of annual plants in 2 sites with contrasting plant density in Central Mexico.
At the specific level, Curculionidae sp. 1 (Coleoptera), and Dorymyrmex grandulus, Myrmecocistus mexicanus, and Tiphiidae sp. 1 (all Hymenoptera) were significantly more abundant at the low than at the high plant density site (32.0 ≥ χ2 ≥ 18.18; df = 1; P < 0.001; Table 4). All other insect species with abundance equal or higher than 5 across both sites did not differ significantly between sites (6.48 3 χ2 3 0.2; df = 1; P > 0.05).
Relative abundance curves showed that Dasytinae sp. 1 (Coleoptera) was the dominant species at both sites (11–12%; Fig. 1). Miridae sp. 1 (Hemiptera) was the second most dominant species at the high plant density site (6%), and it was also a relatively dominant species at the low plant density site (6%; Fig. 1).
DIVERSITY AND SIMILARITY
The Shannon diversity index at the low plant density site (3.26) was significantly higher than that obtained for the high plant density site (2.75; t = 3.705; df = 475; P < 0.001). The composition of the community of floral visitors was different between sites (Sorënsen similarity index = 32.35%). Only 23 insect species (4 Coleoptera, 3 Diptera, 9 Hemiptera, 4 Hymenoptera, and 3 Lepidoptera) were shared between sites (Table 4).
The cluster analysis indicated 2 groups of annual plants (Fig. 2). The first group comprised the communities of floral visitors of P. bipinnatifidum, F. ramosissima, and F. pedata from the low plant density site (Fig. 2). A second group consisted of the communities of floral visitors of all other annual species. Within this group, F. ramosissima, P. bippinatifidum (from the high plant density site) and B. barbata (from the low plant density site) clustered together. Likewise, D. humilis from the high plant density site and Z. peruviana from both sites constitute a second subgroup (Fig. 2). The communities of floral visitors of P. pilosa from the high plant density site and K. rosei and A. incarnata from the low plant density site show the lowest values of linkage distance (Fig. 2).
ENVIRONMENTAL FACTORS AND THE COMMUNITY OF FLORAL VISITORS
At the order level, the canonical correspondence analysis showed that ordination axes 1 and 2 explain 84.3% and 15.7%, respectively, of the total variation on the abundance of floral visitors (Fig. 3a). Disturbance and temperature were associated with axis 1 (analysis scores: 0.988 and 0.984, respectively). Among the insect groups, Hymenoptera was positively associated with axis 1 (analysis score: 1.641), whereas Diptera was negatively associated with that same axis (analysis score: -0.808; Fig. 3a).
At the specific level, ordination axes 1 and 2 explained 53.7% and 46.3% of the variation in the abundance of floral visitors, respectively (Fig. 3b). Disturbance was negatively associated with axis 1 (analysis score: -0.742), whereas temperature and humidity were associated with axis 2 (analysis scores: -0.798 and 0.709, respectively). Myrmecocystus mexicanus (Hymenoptera: Formicidae) and Tiphiidae sp. 1 were associated with axis 1; however, whereas M. mexicanus appears to be associated with disturbance (analysis score: -1.017), Tiphiidae sp. 1 seems to avoid it (analysis score: 1.384; Fig. 3b). Curculionidae sp. 1 and D. grandulus were associated with axis 2, but to different environmental variables. Curculionidae sp. 1 was associated with humidity (analysis score: 1.784), whereas D. grandulus with temperature (analysis score: -1.231; Fig. 3b).
Coleoptera, Diptera, Hemiptera, Hymenoptera, and Lepidoptera are the groups of insects that compose the insect community visiting the flowers of annual plants at the sites studied. In general, those same groups of floral visitors have been recorded previously for the annual species studied or other species belonging to the same plant families (Zimmerman 1977; Adams et al. 1981; Townsend 1993; Osorio-Beristain et al. 1997; Nores et al. 2013; Bizecki 2014; Moore 2014). In addition, we found that abundance, diversity, and composition of the community of floral visitors of annual plants vary through space. Likewise, other studies have recorded spatial variation in species richness, abundance, diversity, and composition of the community of floral visitors of different plant species (Herrera 1988; Steffan-Dewenter et al. 2001; Moeller 2005; Gibson et al. 2006; Czarnecka & Denisow 2014). Because of these differences, the similarity of the communities was relatively low, as was shown by both Sørensen similarity indexes and cluster analysis. The communities of floral visitors to the annual plants do not seem to be grouped by site, floral color, or plant family. According to the cluster analysis, the most similar communities of floral visitors were those of P. pilosa (high plant density site), K. rosei, and A. incarnata (both low plant density site). These species have similar growth form (i.e., creeping plants), and short periods of flower anthesis. In addition, P. pilosa and A. incarnata flowers are similar in color (Turner 1994; Spellenberg 2001; Rzedowski & Rzedowski 2005).
The differences observed in the structure of the community of floral visitors between sites were caused by significant changes in the abundance of 4 insect species: Curculionidae sp. 1 (Coleoptera), Dorymyrmex grandulus, Myrmecocystus mexicanus (both Hymenoptera: Formicidae), and Tiphiidae sp. 1 (Hymenoptera). All of these species were significantly more abundant at the low plant density site than at the high plant density site. However, it seems that their distribution is being determined by different environmental traits. Our results showed that the ants D. grandulus and M. mexicanus seem to be associated with high temperature and disturbance, respectively (Fig. 3). Both of these species have been recorded at sites with low plant cover (Cole 1966; Chew 1987, 1995; Graham et al. 2008). Moreover, colonies of M. mexicanus usually are located at unshaded sites (Cole 1966; Chew 1987, 1995), because low plant cover favors the decrease in temperature needed for these ants to forage at dusk (Chew 1987, 1995). In addition, the canonical correspondence analyses indicate that Curculionidae sp. 1 prefers sites with high humidity (Fig. 3b). Other studies have recorded longer survival, higher reproductive success (eggs laid and eggs hatched) as well as lower mortality at medium to high levels of humidity (Weissling & Giblin-Davis 1993; Al-Ayedh & Rasool 2010). Finally, our results showed that Tiphiidae sp. 1 was negatively associated with disturbance (Fig. 3b). Likewise, it has been shown that this group of wasps is more abundant at undisturbed habitats (Quinn 2004). Thus, it seems that the higher temperature (average temperature: 27.3 °C; maximum temperature: 37 °C) and level of disturbance found at the low plant density site favor the occurrence of D. grandulus and M. mexicanus, but not that of Curculionidae sp. 1 and Tiphiidae sp. 1. Instead, the higher abundance of these 2 insect species at the low plant density site must be determined by other factors that were not evaluated in this study.
Dasytinae sp. 1 (Coleoptera: Melyridae) and Miridae sp. 1 (Hemiptera) were the 2 dominant species of floral visitors to the community of annual plants, and their abundance did not differ between sites. Dasytinae is a group of beetles commonly associated with flowers, where they feed on nectar and pollen (Mawdsley 2003; Hoebeke & Wheeler 2013). Miridae is a group of bugs that lives on plants, feeding on their foliage, seeds, and pollen, and laying their eggs within the stems and petioles of their hosts (Eubanks et al. 2003; Cassis & Schuh 2012). It seems that these 2 insect species have a wide niche, because they were able to survive and be relatively abundant under a range of conditions. Moreover, it seems that although annual plants are available for a short period of time, they are an important resource for these 2 insect species. This might be particularly true at disturbed sites where annual plants might be very abundant (Whittaker 1975; Bazzaz & Morse 1991), providing sufficient nectar, pollen, foliage, and seeds to insects that search for these resources, and favoring their population growth and dominance within the communities.
One might expect that the dominant insect species associated with the flowers are important pollinators for the annual plants. Both Miridae and Melyridae have been recorded as common floral visitors of diverse plant species (Mawdsley 2003; Ishida et al. 2009; Fiala et al. 2011; Pendleton et al. 2011). Actually, Melyridae is considered an important group of pollinators of flowering plants (Mawdsley 2003). However, although it has been proved that mirids can transfer pollen successfully (Ishida et al. 2009; Fiala et al. 2011), some studies have shown that these bugs do not always carry pollen on their bodies (Nagamitsu & Inoue 1997); thus, their importance as pollinators is ambiguous.
Similarly, one might wonder if changes in the abundance of Curculionidae sp. 1 (Coleoptera), D. grandulus, M. mexicanus (Hymenoptera: Formicidae), and Tiphiidae sp. 1 (Hymenoptera) might affect the reproductive success of the annual plants studied. Curculionids, ants, and tiphiids have been recorded as floral visitors and important pollinators of diverse plant species (Brown 1998; Ratnayake et al. 2007; Jones et al. 2010; Ibarra-Isassi & Sendoya 2016). Therefore, they might be important pollinators for the plant species studied, especially for the self-incompatible F. ramosissima (Asteraceae), E. mexicana, and B. barbata (Poaceae) (Powell 1978; de Nettancourt 2001; McKown et al. 2005). In addition, these insect species might be important pollinators for all the other annual species studied, and changes in their abundance through space might have important effects on the reproductive success of annual plants. Thus, studies focused on how changes in the structure of the community of floral visitors across space affect the reproductive success of annual plants are necessary.
We thank E. Bello, D. Rodríguez, D. Ahuatzin, L. Noyola, S. Flores, P. Hernández, S. Tamayo, E. Pineda, M. Rodríguez, and S. Flores for field assistance. Insect identification was conducted by C. Vergara, C. Sandoval, G. Castaño, E. Arriaga, E. Barrera, F. Escobar, J. Sánchez, and A. Ibarra. This work was supported by Vicerrectoría de Investigación y Estudios de Posgrado of the Benemérita Universidad Autónoma de Puebla (grant number: FICD-NAT14-I and FICD-NAT12-I).