﻿An updated list of butterflies (Lepidoptera, Papilionoidea) of two Guatemalan seasonally dry forests

﻿Abstract Guatemala has a great diversity of butterflies, although there have been few intensive surveys on Lepidoptera in the country so far. We present an updated list of 218 species in 149 genera, 19 subfamilies, and six families of butterflies sampled at two seasonally dry forests in the Salamá and Motagua valleys in central and eastern Guatemala, by integrating new data from field surveys conducted in 2014–2021 into our previously published data (Yoshimoto et al. 2018, 2019), with Amblyscirteselissaelissa Godman, 1900, Repensflorus (Godman, 1900), and Niconiadesnikko Hayward, 1948 (Hesperiidae: Hesperiinae) as new country records. We collected a hairstreak species, Chalybshassan (Stoll, 1790) (Lycaenidae: Theclinae), at the Motagua Valley site, representing the second record for Guatemala since the early 20th century, after we rediscovered it at the Salamá Valley site in 2011 and 2012 (Yoshimoto and Salinas-Gutiérrez 2015). Nymphalidae and Hesperiidae had larger numbers of species than the other four families at both sites. In Pieridae and Nymphalidae, species composition was similar between the sites, whereas in Lycaenidae, Riodinidae, and Papilionidae it differed more greatly between the sites. These results confirm the relatively high lepidopteran diversity of Guatemalan dry forests, noteworthy for the small areas that comprise the study sites, and represent marked similarities and differences in butterfly fauna and phenology within these forests.


Introduction
Neotropical seasonally dry forests are rich in flora and fauna (Pennington et al. 2006;Dirzo et al. 2011), although their ecosystems have been deteriorating because of various anthropogenic disturbances, such as deforestation due to agricultural expansion (e.g., Chazdon et al. 2011). Dry forests in Guatemala also harbor high lepidopteran diversity as well; our previous studies documented more than 150 and 100 butterfly species at the two small forest reserves in central and eastern Guatemala, respectively (Yoshimoto et al. 2018(Yoshimoto et al. , 2019. These species lists, however, are still incomplete, and obviously, more species remain to be sampled at these sites. Moreover, we detected marked seasonal patterns in butterfly species richness and several conspicuous differences in the lepidopteran fauna between the two sites (Yoshimoto et al. 2019). Thus, it was apparent that additional field surveys were needed to make quantitative between-site comparisons of species composition, in order to enhance our understanding of butterfly fauna and phenology of these forests.
In Guatemala, approximately 400 species of Hesperiidae and nearly 700 species of the remaining families of Papilionoidea have been reported (Austin et al. 1998; Barrios et al. 2006;Salinas-Gutiérrez et al. 2009, 2012Salinas-Gutiérrez 2013). Despite such high lepidopteran diversity, Guatemala's butterfly fauna has been studied less intensively compared to neighboring countries; for example, in Mexico, exhaustive species lists for the whole country and for several states have been published (de la Maza et al. 1989(de la Maza et al. , 1991Luis-Martínez et al. 2011, 2016Llorente-Bousquets et al. 2014), whereas Guatemala has had little research on Lepidoptera and few published inventories since the 20 th century (but see Austin et al. 1996 andYoshimoto et al. 2021). Continued field surveys in various parts of Guatemala are thus important to fill a gap in our knowledge of the Neotropical butterfly fauna, which will in turn contribute to biodiversity conservation in the country.
Here, we present an updated and integrated list of papilionoid species (including Hesperiidae; van Nieukerken et al. 2011) for the same dry forest sites where we conducted our previous studies (Yoshimoto et al. 2018(Yoshimoto et al. , 2019, by adding the new data from subsequent field surveys performed in 2014-2021, correcting identification errors, and modifying some of the species names based on taxonomic changes. Additionally, we examine between-site differences in butterfly fauna by comparing species composition at the family level, and identify seasonal patterns at the species level.

Materials and methods
This study was conducted at the Los Cerritos Municipal Park (hereafter, Los Cerritos; Fig. 1a) in the Salamá Valley (a subwatershed of the Chixoy region) of Baja Verapaz Department in central Guatemala (15°05'N, 90°18'W, 960-1160 m a.s.l., 69 ha), and at the Heloderma Natural Reserve (hereafter, Heloderma Reserve; Fig. 1b) in the Motagua Valley of Zacapa Department in eastern Guatemala (14°53'N, 89°47'W, 510-790 m a.s.l., 58 ha). The rainfall patterns are similar between the two areas, in which the rainy season usually begins in late May and ends in October; these six months were accordingly defined as the rainy season and the remaining months (November-April) as the dry season. This climatic trait fits the definition of seasonally dry tropical forests (4-6 months with rainfall being < 100 mm; Dirzo et al. 2011); see fig. 1 in Yoshimoto et al. (2018) Fig. 1b). By contrast, none of these species have been reported from Los Cerritos (M. R. Álvarez, pers. comm.), where there are fewer high arboreal species and abundant shrubs and herbaceous plants (thus commonly called a spiny bush or scrub), thereby the columnar cactus being prominent in its forest landscape (Fig. 1a).
Field surveys were conducted on 21 days from July 2014 to August 2021 at Los Cerritos, and on 19 days from October 2017 to November 2021 at Heloderma Reserve. We collected adult butterflies with an insect net or photographed them in the daytime (09:00-17:00) at each site and in neighboring areas (a small garden at the foot of Los Cerritos and on a farm road adjacent to Heloderma Reserve). The individuals collected were mounted as voucher specimens and were deposited at the Colección de Artrópodos, Laboratorio de Entomología Sistemática, Universidad del Valle de Guatemala. All the individuals collected or photographed were identified to species or subspecies according to Warren et al. (2017). We did not include data for specimens that were not identified to species, except for Calephelis spp. (Riodinidae) and Bolla sp. (Hesperiidae: Pyrginae); see the footnotes of the Appendix 1 for the rationales for the inclusion of these data. We added all these data to our previous data (Yoshimoto et al. 2018(Yoshimoto et al. , 2019, corrected identification errors, and modified scientific names of some of the species based on taxonomic changes, in order to compile an updated and integrated species list of the two sites. Note that the sampling methods were partially different in the previous surveys; only netting was done at Los Cerritos from January 2011 to November 2012, whereas at Heloderma Reserve between February 2016 and March 2017, data were obtained through netting, photographing, and observation; see Yoshimoto et al. (2018Yoshimoto et al. ( , 2019 for detailed information on the sampling methods for each site. The site-level estimated species richness was calculated by using the Chao II index (Chao et al. 2005;Gotelli and Colwell 2011), after pooling the data across observation dates for each month for each site. The total estimated species richness was similarly obtained after pooling these data across both sites. The Jaccard dissimilarity index was used to quantify the between-site similarity in species composition; this index was calculated for all data and for each of the six families (Papilionidae, Pieridae, Lycaenidae, Riodinidae, Nymphalidae, and Hesperiidae). All the analyses were performed using R 4.1.2. (R Development Core Team 2021) with the package Vegan (Oksanen et al. 2020).

Results
By integrating our previous data (Yoshimoto et al. 2018(Yoshimoto et al. , 2019, a total of 218 species (including one unidentified taxon and 107 subspecies) in 149 genera from 19 subfamilies of six families were recorded at the two sites (Appendix 1). Hesperiidae was the richest family (71 species), followed by Nymphalidae, Lycaenidae, Pieridae, Riodinidae, and Papilionidae (66, 36, 20, 16, and 9 species, respectively). Los Cerritos had 166 species in 117 genera, and Heloderma Reserve had 139 species in 107 genera (Appendix 1), 16 and 41 species of which had been newly recorded in the subsequent surveys, respectively (Fig. 2). The estimated species richness (mean±SE) of each site based on the Chao II index is 216.27±16.35 and 187.35±17.74, respectively, indicating that approximately 76.8% and 74.2% of the species inhabiting each site were sampled. The total estimated species richness for both sites is 272.80±17.85 (79.9%).
We detected identification errors for 20 individuals (identified as 11 species in our previous studies) and determined them to represent 13 species in this study; ten individuals from Los Cerritos and ten from Heloderma Reserve have been determined to number eight and six different species, respectively, with one species, Cissia themis, shared between sites (Table 1). Additionally, Yoshimoto et al. (2019) incorrectly listed Piruna (Hesperiidae: Heteropterinae) in the subfamily Hesperiinae.
The following three skipper species (Hesperiidae: Hesperiinae) were recorded for the first time in Guatemala:  Table 1. The specimens were deposited in the Colección de Artrópodos, Laboratorio de Entomología Sistemática, Universidad del Valle de Guatemala, and are being cataloged (Fig. 3c). Distribution: Southwestern Mexico (Warren et al. 2017).
Eighty-six species were shared between Los Cerritos and Heloderma Reserve (Table 2), which amounts to 51.8% and 61.9% of the species sampled at each site (the Jaccard dissimilarity index is 0.606). At both sites, species richness of Nymphalidae and Hesperiidae was greater than that of the other four families, although family-level species richness differed greatly between the sites (Table 2; Fig. 4). In particular, the proportion of Lycaenidae was much higher at Los Cerritos (18.7%) than at Heloderma Reserve (8.6%), which was mainly due to differences in the subfamily Theclinae (26 and 7 species, respectively: Appendix 1).
Family-level species composition also differed between the sites, and the magnitude of this difference varied among the six families ( Table 2). The dissimilarity indices for Riodinidae, Lycaenidae, and Papilionidae were considerably larger, indicating that species composition differed more greatly between the sites in these families. Pieridae and Nymphalidae, by contrast, had smaller indices with many shared species, demonstrating that their species composition was relatively similar between the sites.
Ninety-three species (42.7%) occurred in both dry and rainy seasons, whereas 103 (47.2%) appeared only in the rainy season and 22 species (10.1%) only in the Table 1. Butterfly species that were sampled at Los Cerritos and Heloderma Reserve (abbreviated as LC and HR, respectively) and were misidentified in Yoshimoto et al. (2018Yoshimoto et al. ( , 2019   were included in the species count but excluded from the Jaccard index analysis (see the footnote 7 of the Appendix 1 for its rationale).

Discussion
A total of 218 species were recorded at the two dry forest sites during our 10-year field surveys, which confirms the relatively high lepidopteran diversity of Guatemalan seasonally dry forests for the small areas that comprise the study sites (<70 ha each). The estimated species richness suggests that nearly a quarter of the species inhabiting each site have yet to been recorded. The number of the additional species yielded in the subsequent surveys was more than twice greater at Heloderma Reserve than at Los Cerritos. The proportion of newly recorded species was much higher in Lycaenidae and Riodinidae; seven lycaenid species were added to the list for Los Cerritos, and seven lycaenid and four riodinid species were added to that of Heloderma Reserve, which nearly doubled the species richness of each family at this site (six lycaenid and five riodinid species in Yoshimoto et al. 2019). Among these species, the record of Chalybs hassan (Stoll, 1790) at Heloderma Reserve is highly important (Fig. 3a), as this species had not been reported for more than 100 years in Guatemala before we collected four individuals at Los Cerritos in 2011 and 2012 (Yoshimoto and Salinas-Gutiérrez 2015). These results highlight the importance of continuing butterfly surveys at both sites to create more exhaustive inventories, especially on small and taxonomically difficult taxa such as Lycaenidae and Riodinidae. Moreover, it is important to conduct research in other dry regions (e.g., the Nentón Valley in northwestern Guatemala) and to make quantitative among-site comparisons of species richness and composition as well. All these studies will contribute to a comprehensive understanding of Neotropical butterfly fauna and distribution, and would serve as a scientific baseline for biodiversity conservation in Guatemalan dry regions.
More than half of the species sampled at each site were shared between the sites, suggesting that species composition is partially and moderately similar between Los Cerritos and Heloderma Reserve. Importantly, between-site similarity greatly differed among the six families. Higher similarity in Pieridae (especially in Coliadinae) would likely be associated with the distribution and abundance of their host plants, considering that coliadine larvae mostly feed on fabaceous plants such as Senna (e.g., DeVries 1987) and that these plants appear to be abundant at both sites.
In Lycaenidae and Riodinidae, species composition largely differed between the sites; in particular, Theclinae had considerable differences in species richness and composition (Appendix 1). In addition, most of these thecline species tended to be highly seasonal, as 25 out of 30 species were sampled only in the rainy season. In contrast to their marked seasonal pattern, Strymon megarus (Godart, [1824]) and S. rufofusca (Hewitson, 1877) occurred frequently also in the dry period at Heloderma Reserve; three and five individuals of each species were collected in both December and January at this site (Appendix 1). It should also be mentioned that Hechtia guatemalensis Mez (Bromeliaceae), one of the dominant bromeliad species at Heloderma Reserve (Fig. 1b), may be a possible foodplant for S. megarus at this site, as the larvae of this species are known to feed on bromeliads (Robbins 2010). Examination of abundance and distribution of host-and nectar-plants, as well as of larval and adult feeding behavior in relation to their phenology, would be an initial step to elucidate the bionomics of these species. Such surveys may also identify factors underlying the regional similarity and dissimilarity in the butterfly fauna.
We recorded Amblyscirtes elissa elissa Godman, 1900, Repens florus (Godman, 1900, and Niconiades nikko Hayward, 1948 (Hesperiidae: Hesperiinae) for the first time in Guatemala (Fig. 3c, d, e). Austin et al. (1998) listed A. e. elissa and N. nikko as species with a potential distribution in Guatemala. Repens florus could have been included in this category as well, as it is known to be distributed in the adjacent countries (Mexico, Belize, and Nicaragua; Warren et al. 2017). These results indicate that there still exists a gap in our knowledge of geographic distribution of Neotropical skipper species, again emphasizing the importance of more intensive research in Guatemala to bridge this gap.
Four individuals of Piruna aea (Dyar, 1912) (two in the previous survey and two in the subsequent one: Figs 2g, 3b) were collected at Heloderma Reserve. This is an interesting result, since most species in this genus are distributed in humid areas at higher elevation (1000-2700 m; Warren and González-Cota 1998). As Yoshimoto et al. (2019) pointed out, the wing pattern of these individuals is somewhat different from Mexican P. a. aea (Dyar, 1912), implying that Piruna cingo sombra Evans, 1955, described from Guatemala and currently considered a synonym of P. a. aea, may be a valid subspecies-level taxon. At present, this is difficult to determine, as very few specimens of this species have been sampled in Guatemala (Barrios et al. 2006