We describe a new pest of guava (Psidium guajava L.), Hesperolyra guajavifoliae Stonis & Vargas, sp. nov., that was recently discovered in western Colombia. Hesperolyra van Nieukerken is a small, Neotropical genus of pygmy moths (Nepticulidae). We re-examine and document the complex morphology of the male genitalia of the generic type species, H. diskusi (Puplesis & Robinson). We discuss the diagnostics and composition of the genus and provide a simple pictorial differentiation scheme for all currently known representatives of the genus. The new species is illustrated with photographs of the adults, some of the immature stages, male and female genitalia, and leaf mines. A link to the COI barcodes of H. guajavifoliae sp. nov. is provided and the relationship of Hesperolyra to other genera is discussed.

New species, pygmy moths, guayaba, Psidium guajava, leaf mines, taxonomy, Colombia

Guava or guayaba (Psidium guajava L.) is an important shrub or small tree cultivated for its fruit in many tropical countries in Asia, Africa, South America and the Caribbean. The fruit can be eaten raw or cooked, but, it is mostly known for its processed fruit products, and can be an integral part of local and international cuisine. The leaves and fruits are also fed to livestock (Heuzé et al. 2017). In some countries, this plant is also used in local traditional medicine (ethnopharmacology) to deal with numerous medical issues such as inflammation, diabetes, hypertension, tooth decay, wounds, ulcers, fever, diarrhea, lung ailments, rheumatism, and as a pain-relief remedy (Gutiérrez et al. 2008). Guava usually grows in areas below 1000 m, annual mean temperatures ranging from 23 to 28 °C, and 1500–2000 mm rainfall (Heuzé et al. 2017). In some tropical areas the plant can occur at altitudes up to 2000 m, in temperatures of 15–45 °C, and quiescent trees can even survive light frosts (Heuzé et al. 2017). Although the origin of this widespread species is not clear, it is believed to be native to countries in tropical America (Germplasm Resources Information Network 2019).

In late January to early March 2019, we conducted fieldwork in the Departamento de Valle del Cauca, northwest of Dagua in southwestern Colombia (Figs 2–6), where Psidium guajava is a common plant cultivated in orchards and widespread in the wild in various anthropogenic or natural habitats. We expected to find Ozadelpha guajavae (Puplesis & Diškus) (Lepidoptera, Nepticulidae), a guava-feeding nepticulid species described a few degrees south from the same western tropical slopes of the Andes in equatorial Ecuador (Puplesis et al. 2002a). It was later recorded in large numbers in the Andes of southern Ecuador near the Peruvian border (Remeikis et al. 2014). However, during our fieldwork in western Colombia, we discovered another species producing leaf mines in mass quantities on P. guajava. It appeared to be a new and distinctive species belonging to the recently erected, Neotropical genus Hesperolyra van Nieukerken. Including the newly discovered species described below, Hesperolyra now comprises six species that occur from Central America to the Atlantic coast of Brazil. Prior to our study, H. molybditis (Zeller, 1877), of which the host plant is unknown, was known to occur in central Colombia (Fig. 1). Previously, only one species of Hesperolyra had host-plant family data; H. saopaulensis van Nieukerken 2016 was reared from an unidentified Myrtaceae plant (van Nieukerken et al. 2016b).

Figures 1–6. 

Distribution map of currently known Hesperolyra species and habitat of H. guajavifoliae Stonis & Vargas sp. nov. 1 Distribution map (the map base, courtesy of Tom Patterson, USA) 2, 5, 6 El Naranjo, 3°46'46"N, 76°43'63"W, 550 m 3, 4 Cisneros, 3°46'27"N, 76°44'40"W, 450 m.

Nepticulidae , or pygmy moths, are miners (occasionally gall inducers, e.g., van Nieukerken et al. 2016b) of assimilative tissues of plants. Some species have been included on lists of cultivated plant pests (e.g., Kuznetzov and Puplesis 1994). A general characterization of this family was provided by several authors, notably Scoble (1983), van Nieukerken (1986), Johansson et al. (1990), Puplesis (1994), Puplesis and Robinson (2000), Puplesis and Diškus (2003), Diškus and Stonis (2012), and recently van Nieukerken et al. (2016b). Nepticulidae are distributed worldwide and occur in almost all terrestrial habitats. Because of species endemism and a high degree of stenophagy, these tiny lepidopterans may serve as tools for monitoring the biodiversity richness of habitats and provide data on the evolutionary processes (Diškus and Stonis 2012; Remeikis 2017). Being some of the smallest moths, adults remain rare in many scientific collections, especially in the tropical countries of Latin America. However, in the field, leaf mines produced by nepticulid larvae are usually distinctive and easy to spot. Sometimes these leaf mines may appear in very large numbers, dramatically reducing the assimilative tissue of mined host plants. Sometimes hundreds or thousands of larvae may feed upon a single tree at the same time, as in the case of the Holarctic Ectoedemia occultella (Linnaeus) and East-Asian E. picturata Puplesis (Puplesis 1994).

Below, we provide a description of the new species, documentation of leaf mines, adults, and their male and female genitalia. We also provide comments on the diagnostics of Hesperolyra, with a simple pictorial tool for identification of the currently known Hesperolyra species.

In the first review of the Neotropical Nepticulidae (Puplesis and Robinson 2000) we noticed that some nepticulid species from Belize and Colombia looked different, mostly because of the long transtilla and horn-like processes in the male genitalia. We hypothesized that these species could belong to Fomoria Beirne, but the observed difference in the forewing venation in comparison to Fomoria (Puplesis and Robinson 2000: fig. 64) was incorrectly hypothesized to be a case of partial venational reduction. Later, a formal species group within Fomoria was erected for species with these male genitalic and forewing venational characters (Puplesis et al. 2002b). In the most recent review and global catalogue (van Nieukerken et al. 2016a, 2016b), the molybditis group was excluded from Fomoria and erected as a separate genus.

For this study, we re-examined the type species of Hesperolyra, H. diskusi (Puplesis & Robinson, 2000), that is characterized by a complex morphology of the male genitalia (Figs 65–86). The horn-like processes are the most remarkable genitalic features of this species. Characterization of H. diskusi was provided in the original description (Puplesis and Robinson 2000) and later in the redescription by van Nieukerken et al. (2016b). In the current study, we found two, not three, large processes (Figs 83–86), and observed that they are attached to the valva (Figs 78–83), not the anellus as was supposed earlier (van Nieukerken et al. 2016b); we did not observe the presence of an anellus. We also found that the transtilla in H. diskusi is more flexible and movable in comparison to most Nepticulidae that possess a transtilla; it can be lifted slightly dorsally at an angle even if the valvae are fixed (Figs 65, 77), and the connections of the transtilla to the valva are unusually weak, easy to break since they are supported by slender, little chitinized arms (Fig. 67).

We provide photographic documentation of the genital structures at different angles (Figs 65–86); they, including the spine-like thickening on the ventral lobe of the valva (Fig. 84) or the ventrally-bent spines of the phallus (Fig. 74), are usually not available for observation or appear different in permanent mounts. After our examination, we became convinced that none of the large, horn-like processes are carinal processes of the phallus.

Upon comparison of the male genitalia, we found that Hesperolyra guajavifoliae sp. nov. fundamentally differs in morphology from H. diskusi. We discovered that in H. guajavifoliae the horn-like processes are connected not with the valva itself, but are fused with the transtilla (Figs 50, 55), and, in contrast to H. diskusi, the anellus is present in the male genitalia. We found that the anellus is comprised of a weakly chitinized dorsal membrane (Figs 49, 55), strongly thickened lateral arms (Figs 49, 51, 55), and a ventral membrane with an elaborate thickening (Figs 55–57) surrounding the phallus from all sides (Fig. 50, best seen in Figs 38, 43).

Currently, there are six species of Hesperolyra distributed from Central America (Belize) to the Atlantic coast of Brazil (Fig. 1); the species from Brazil was described from a female only (van Nieukerken et al. 2016b). All species are distinctive; therefore, diagnostics of Hesperolyra species, including the Brazilian H. saopaulensis, is clear (see Fig. 87). However, it also raises some questions: do all the species really belong to the same genus, and what is their relationship to other genera?

The wing venation of Hesperolyra guajavifoliae sp. nov. (Figs 28–33) is almost identical to H. saopaulensis (see van Nieukerken et al. 2016b: fig. 116) and similar to H. diskusi (see Puplesis and Robinson 2000: fig. 64); the venation of remaining species is unstudied. It is important to note that unique, special scales hidden under the forewing fold are characteristic of both H. diskusi and H. guajavifoliae sp. nov. We have concluded that Hesperolyra most likely represents a separate, monophyletic taxon, characterized by a wide forewing, with more or less uniform, but unique, simplified venation, extended, lyre-shaped transtilla, elaborate valva, and the presence of horn-like processes in the male genitalia, and possibly the feeding on Myrtaceae (host plant known for only two species). Hesperolyra was also supported by a multi-gene molecular analysis by Doorenweerd et al. (2016), that grouped it with Neotrifurcula van Nieukerken and Bohemannia Stainton. Neotrifurcula was subsequently synonymized with Glaucolepis Braun (Stonis et al. 2017).

During our study, ten sequences of 657 bp and 1 sequence of 609 bp of the mtDNA COI gene belonging to three Nepticulidae and one Opostegidae species were successfully obtained (Table 1). These data were supplemented by the sequences of other species downloaded from the BOLD website (Table 2). In these sequences, 184 parsimony-informative sites were detected. The overall mean distance between analysed species estimated using the same mtDNA sequence was 14.8 ± 1.0. The interspecific pairwise distances between the pairs of the studied species varied from 5.6 ± 1.0% (between Etainia albibimaculella (Larsen) and E. capesella (Puplesis)), and 22.1 ± 2.5% (between H. guajavifoliae sp. nov. and Pseudopostega sp.). The smallest interspecific distance from H. guajavifoliae sp. nov. to any other species was 15.3 ± 1.9% (i.e., between H. guajavifoliae sp. nov. and Fomoria eriki) (Table 2). This indicates that COI can be used as a useful diagnostic tool for the identification of this new species. On the other hand, intraspecific divergence in H. guajavifoliae sp. nov. has not been observed yet; however, all studied specimens were from the same locality, and additional specimens from different localities would certainly enrich our knowledge about divergence within the species.

Depending on the combination of species set, several versions of the Neighbour-Joining tree with different topology were obtained; some of them are presented in Figs 88–90. In our preliminary analysis using only the COI barcode fragment, Hesperolyra always appeared as a separate clade. Hesperolyra guajavifoliae sp. nov. always clustered at a distance from H. diskusi + H. saopaulensis (Fig. 88). In most of our numerous, different attempts, the Hesperolyra clade consistently grouped either with Fomoria + Ectoedemia (Figs 89), or Fomoria + [Acalyptris + Etainia] (Fig. 88), or [Fomoria + Etainia] + [Acalyptris + Ectoedemia] (Fig. 90), or even only with Fomoria, but never with Glaucolepis or Bohemannia, or other genera, as presented in Doorenweerd et al. (2016). We found that the relationships between approximately half of the clades remain unsupported according to bootstrapping results; however, according to Nieukerken et al. (2012), bootstrap support values for the Neighbour-Joining similarity tree are not necessary. NJ trees are never robust due to the nature of the method; therefore, adding bootstrap supports to indicate the robustness does not add much. NJ trees are useful for indicating pairwise differences between clusters (of species) and estimating whether COI can be used as a diagnostic marker, what has been shown in the case of H. guajavifoliae sp. nov. Although our data are far from complete, the tendency of Hesperolyra to group with Fomoria or other genera causes us to re-evaluate their relationships; i.e., it may be possible that Hesperolyra is related to these taxa. It is interesting to note that, unexpectedly, Acalyptris most often clustered with Etainia (e.g., Figs 88, 90); this was also found by other workers (Doorenweerd et al. 2016).

Our molecular analysis did not show a close relationship between the guava-feeding Hesperolyra guajavifoliae sp. nov. and other Myrtaceae-feeding Nepticulidae, including the South American guava-feeding nepticulid species, Ozadelpha guajavae, which was recently barcoded by us; the sequence is available in the BOLD database: ADH4024.

So far, Hesperolyra guajavifoliae sp. nov. is the only Nepticulidae pest discovered in western Colombia. However, during our fieldwork we observed a couple of old leaf mines on guava with a wider gallery that did not extend into an obvious blotch at the final stage of development. Although these differently looking leaf mines may belong to Ozadelpha guajavae, there is no confirmed evidence that both species H. guajavifoliae sp. nov. and O. guajavae occur together in western Colombia.

Figure 87. 

A pictorial tool for quick diagnostics of the currently known Hesperolyra species. Note: the morphological structures are drawn at different scales.

Figures 88–90. 

Fragments of different versions of Neighbour-Joining tree of Hesperolyra and other Nepticulidae genera (the full molecular phylogeny will be published elsewhere; a general phylogeny of Nepticulidae is not presented or discussed here). The divergence was calculated using the Kimura 2-parameter model based on 657 bp mtDNA COI sequences. Bootstrap values below 50 are not shown. Pseudopostega sp. (Opostegidae) was used as outgroup.

Contributions to this research are as follows: JRS: concept and design of the research and fieldwork, discovery and rearing of the adults from the mining larvae, photographic documentation of leaf mines and habitats; writing the manuscript and technical preparation of all plates of illustrations, and discussion on results of the molecular research and diagnostics of Hesperolyra; AR: preparation of the material collected in Colombia in 2019, dissection and photographic documentation of H. guajavifoliae sp. nov., molecular research, discussion on diagnostics of Hesperolyra and general Nepticulidae phylogeny, and writing of comments on the results of molecular research; AD: dissection and photographic documentation of type species (H. diskusi), discussion on morphology and diagnostics of Hesperolyra, compiling of list of cited references; SO: molecular research, discussion on molecular Nepticulidae phylogeny, writing of comments on the results of DNA research, and deposition of sequences in GenBank; SA: fieldwork in Colombia, assistance in rearing some H. guajavifoliae sp. nov., obtaining of research permits (together with Igor Dimitri Forero Fuentes, see Acknowledgements), various contributions to the new species description, specimen deposition at MPUJ; MAS: manuscript writing, scientific expertise of the data, elaboration of the concept, and discussion on the results.

We are very grateful to Professor Dr. Igor Dimitri Forero Fuentes, coordinator of the Biological Collections (Departamento de Biología, Pontificia Universidad Javeriana, Bogotá, Colombia), for facilitating a permit to collect material in Colombia and for his previous scientific collaboration. For the collecting permit we also thank the Autoridad Nacional de Licencias Ambientales, Bogotá, Colombia. We are also indebted to our field assistant, Franklin Jesus Galindo Meza (Cali), for his generous help during our fieldwork in Valle del Cauca.

This research was partially funded by a grant (S-MIP-19-30, “DiagnoStics”) from the Research Council of Lithuania. The study was carried out using the research infrastructure of the Nature Research Centre under the Lithuanian open access network initiative.

Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by USDA. USDA is an equal opportunity provider and employer.