Morphological data and locality information were taken from the personal collections of EJT, GRC, and AEG, the ecological studies mentioned above, and from specimens sent to EJT for identification. Although the new species occurs in Ecuador and Peru, the distribution model was restricted to Ecuador since the majority of collection data is from the northeast and southeast Andean slopes of Ecuador. Collection density was very high in Napo Province, Ecuador (where much of the project activity was focused) relative to other parts of the species range, and to avoid biased results due to this imbalance, locations were filtered so that the model was based on an even sample from across the species’ range. The distribution model was calculated using the maximum entropy method (Maxent software package; Phillips et al. 2006) from 26 presence locations marked with GPS and 19 climatic parameters at the 25km² scale, which were downloaded from WorldClim (www.worldclim.org; Hijmans et al. 2005). The model was calibrated using elevations at which the species is modeled to be present, against elevations at which the species was observed in the field. The methods are described fully in Rodríguez-Castañeda et al. (2012).

Study sites were located in the northeastern Andes of Ecuador in the provinces of Napo and Sucumbíos along an elevational transect from Amazonian lowland rain forest (200 m) to Andean páramo (3200 m). Mean temperature along this gradient drops linearly with elevation at a rate of approximately -0.5°C per 200 m elevation gain, and precipitation ranges from 3900–4500 mm yr^-1 (Rodríguez-Castañeda et al. 2010). This area includes three mountain ranges: the Sumaco, Huacamayos and Reventador, as well as continuous lowland to montane rain forests (sensu Holdridge 1967) that are part of the Cayambe–Coca, Antisana, and Gran Sumaco nature reserves.

To document trophic interactions associated with the new Piper species, standard 10 m diameter plots were established inside the forest at the sampling sites (Dyer et al. 2010a). A mean of 9 ± 0.81 plots were sampled at 200 m intervals along an elevational range from 280–3200 m; the new species was encountered from 1200–2300 m along this transect. Inside each plot, the leaves and stems of all individuals of Piper present were harvested (leaving the roots intact) kept in a separate bag identified by plot number and plant number, and transported to laboratories at Jatun Sacha (400 m), Yanayacu Biological Station (2200 m), or Oyacachi (3200 m) for further study. At the laboratory, all leaves were carefully surveyed for caterpillars or other insects. When the plants had ants or other arthropods inside the petioles, these were collected and placed in 75% ethanol in labeled vials. Each caterpillar found chewing on Piper leaves was assigned a unique number, identified to morphospecies, and reared to an adult moth for identification, or in the case of parasitized caterpillars, to obtain parasitism rates and identify the parasitoids.

Ecological niche model performance was high (avg. train AUC: 0.99; avg. test AUC: 0.99; max. probability range: 0–0.86). Results predict that suitable habitat for Piper kelleyi is more extensive than the distribution reported here, which is based on available collections and observations (Fig. 3B). The species favors montane elevations (ca. 1000–2500 m) and according to the model, important determinants of its habitat were intolerance to freezing temperatures and affinity for temperate climate (coldest month temperatures of ~12.5°C). Further habitat suitability increases with high levels of precipitation during the coolest months (>1000 mm). The species seems to favor river or stream banks and sandy soils (pers. observation by GRC). The model predicts suitable habitat on both sides of the Andes, but Piper kelleyi has never been recorded from localities west of the continental divide despite intensive Piper collection by EJT and surveys of Piper for caterpillars by AEG at several localities on the western slopes of the Andes. These collections included seven species of Piper sect. Macrostachys, but did not include Piper kelleyi. It is possible that additional collecting will reveal the presence of the new species in these areas, but based on current collections it appears that Piper kelleyi is restricted to limited localities on the eastern slopes of the Andes.

Piper kelleyi is a focal component of a study aimed at understanding the influence of plant secondary chemistry on herbivores and their associated parasitoids (Wilson et al. 2012). The genus Piper produces a high diversity of multiple classes of secondary compounds, including biologically active amides and imides (Dyer et al. 2004), which are known to negatively affect generalist herbivores (Dyer et al. 2003). Preliminary evidence suggests that Piper kelleyi is not palatable to generalist herbivores based on assays conducted by GRC and AEG during studies spanning 2005–2008 and 2011–2013 respectively (unpubl. data).

Plot data resulted in herbivory and parasitism rates, and measures of herbivore and parasitoid richness. Mean herbivory rates for the plant are close to 20%, and the only herbivores that are regularly found on Piper kelleyi are species of the specialist herbivore Eois. Even in the absence of caterpillars, the presence of Eois can be identified by the characteristic feeding markings that they leave behind (see the herbivory key in Dyer et al. 2010b). These caterpillars scrape small portions of the under sides of leaves, leaving unpigmented windows (Fig. 4A) that are eventually lost, resulting in characteristic holes in the leaves. Apart from Eois, other lepidopteran herbivores reared from this host plant include members of t the families Erebidae, Hesperiidae, Noctuidae, Pyralidae, and Tortricidae. Over 3, 000 caterpillars have been reared from Piper kelleyi, yet only single individuals have been recorded from these other lepidopteran families (Dyer et al. 2013) suggesting that, unlike Eois, they are not likely to be Piper specialists.

Figure 4.

A Leaf of Piper kelleyi with characteristic herbivory marks or “windows” made by the specialist herbivore Eois species of Eois that specialize on Piper kelleyi include B Eois viridiflava (Dognin) C Eois aff. viridiflava (Dognin) D Eois ignefumata (Dognin) E Eois planetaria (Dognin) F Eois aff. pallidicosta (Warren) G Eois encina (Dognin) H Eois aff. encina (Dognin).

Project collaborators have reared nearly 30, 000 caterpillars and parasitoids from over 100 Piper species, and Piper kelleyi is the host species with highest density and species richness of Eois found thus far. Since many of the Eois species and nearly all of the parasitoids reared from Piper were previously undescribed, the taxonomy of these insect groups currently lags behind that of their host plants. Nevertheless, Piper kelleyi hosts at least 11 morphospecies of the specialist herbivore Eois and these include Eois cancellata (Warren), Eois encina (Dognin), Eois aff. encina (Dognin), Eois goodmani (Schaus), Eois ignefumata (Dognin), Eois aff. necula (Druce), Eois olivacea (Felder & Rogenhofer), Eois aff. pallidicosta (Warren), Eois planetaria (Dognin), Eois viridiflava (Dognin), and Eois aff. viridiflava (Dognin) (Fig. 4B–H; Rodríguez-Castañeda 2009, Dyer et al. 2013: www.caterpillars.org ). Seven of these species were recognizable from adult characters as distinct species, whereas additional cryptic species were recognized from a combination of molecular data and larval characteristics (Wilson et al. 2012). These results are not surprising since molecular identification has increased Geometrid estimates of diversity by 50% in the SE Andes of Ecuador (Strutzenberger et al. 2010).

The parasitism rate for all of the Eois collected from Piper kelleyi was 8.63%. Parasitic wasps were the most frequent and included Eulophidae (Hymenoptera: Chalcidoidea) at 3.5%, followed by Braconidae (Microgastrine) at 2.7%, and Ichneumonidae at 1.5%. Tachinid flies (Diptera: Tachinidae) had the lowest parasitism rate at 0.9%. Within the subfamily Microgastrine (Braconidae), species of the genera Cotesia (Cameron), Glyptapanteles (Ashmead), Mesochorus (Gravenhorst), Parapanteles (Ashmead), and Protopanteles (Ashmead) have been reared out of the Eois species feeding on Piper kelleyi. The determined Tachinid species include Erythromelana abdominalis (Townsend), Eois cryptica (Inclán & Stireman), and Eois jaena (Townsend) (Inclán-Luna 2010, Dyer et al. 2013, Inclán-Luna and Stireman 2013).

Thus, to date, Piper kelleyi is host to at least 17 species of specialist and generalist herbivores, and nine described and an estimated 20–30 undescribed species of parasitoids (unpubl. data). Moreover, Piper kelleyi acts as a microcosm for other insect species. The sheathing leaf petioles occasionally host nests of Pheidole (Westwood) (Hymenoptera: Formicidae; Fig. 5), other ants, or small families of earwigs. Several species of the Macrostachys clade of Piper have specialized associations with ants that nest primarily within the petioles (Risch et al. 1977, Letourneau 1998, Tepe et al. 2004); however, when the large sheathing petioles of other species form a sufficiently closed chamber, a diversity of opportunistic, arboricolous ants take advantage of these chambers for nesting sites (Tepe et al. 2007). Also, predatory coccinellid beetle larvae and salticid spiders are frequently found on the leaves of Piper kelleyi (GRC, personal observation).

Figure 5.

A colony of Pheidole ants nesting inside a petiole of Piper kelleyi.

Empirical evidence has been accumulating that supports the hypothesis that plant biochemistry helps drive the evolution of biodiversity, not only among herbivores that feed directly on plants, but among higher trophic levels as well (Ehrlich and Raven 1964, Forister and Feldman 2011, Wilson et al. 2012). Piper kelleyi represents a hotspot of these interactions. In fact, our group has surveyed lepidoptera feeding on Piper species for over 20 years, and Piper kelleyi supports the highest diversity of Eois recorded to date. We have documented a large number of parasitoids as well, but, because parasitism rates are low and discoveries are still occurring, we expect the number of parasites to be similarly high.

Our description of Piper kelleyi is part of a larger effort to combine traditional taxonomy and natural history with newer integrative approaches. The purpose of this project is to successfully contribute to taxonomy and systematics while characterizing evolutionary patterns via molecular phylogenetics, studying community ecology, and increasing our understanding of functional diversity by quantifying specialized consumer-resource relationships and interaction diversity. The importance of traditional taxonomy and natural history data for these studies is essential (as argued in Dyer et al. 2010a) and descriptions of new species, such as Piper kelleyi, are a necessary component of thorough research programs in ecology and evolutionary biology.