A review of the New World species of the parasitoid wasp Iconella (Hymenoptera, Braconidae, Microgastrinae)

Abstract The New World species of Iconella (Hymenoptera: Braconidae, Microgastrinae) are revised. Iconella andydeansi Fernández-Triana, sp. n., Iconella canadensis Fernández-Triana, sp. n., and Iconella jayjayrodriguezae Fernández-Triana, sp. n., are described as new. Iconella isolata (Muesebeck, 1955), stat. r., previously considered as a subspecies of Iconella etiellae (Viereck, 1911), is here elevated to species rank. All species have different, well defined geographic distributions and hosts. Taxonomic keys are presented in two formats: traditional dichotomous hardcopy versions and links to electronic interactive versions (software Lucid 3.5). Numerous illustrations, computer-generated descriptions, distributional information, host records (mostly Lepidoptera: Crambidae and Pyralidae), and DNA barcodes (where available) are presented for every species. Phylogenetic analyses of the barcoding region of COI indicate the possibility that Iconella is not monophyletic and that the New World species may not form a monophyletic group; more data is needed to resolve this issue.

and Pyralidae), and DNA barcodes (where available) are presented for every species. Phylogenetic analyses of the barcoding region of COI indicate the possibility that Iconella is not monophyletic and that the New World species may not form a monophyletic group; more data is needed to resolve this issue.

Keywords
Iconella, Microgastrinae, New World, taxonomic review, host caterpillars, DNA barcoding, Area de Conservacion Guanacaste, parasitoid wasps introduction The genus Iconella was erected by Mason (1981) to accommodate a group of Apanteles species with a sinuated vein cu-a in the hind wing, a character he interpreted as plesiomorphic among Microgastrinae. Iconella includes all species of Nixon's sundanus-group and part of the merula-group (Nixon 1965, Whitfield 1997. It is cosmopolitan, with 33 described species, most of them found in the Palearctic (Yu et al. 2012). All species with known hosts are solitary parasitoids of microlepidoptera, especially concealed feeders (Whitfield 1997); one species, Iconella isolata (Muesebeck, 1955) has been extensively studied in the tropics for biological control of a pest caterpillar (e.g. Bennett 1960, Bartlett et al. 1978).
Until now, only one species was known from the New World, Iconella etiella (Viereck, 1911). However, ongoing research on the Microgastrinae fauna of Area de Conservacion de Guanacaste (ACG) in northwestern Costa Rica (Smith et al. 2008, Janzen et al. 2009) and study of specimens housed in the Canadian National Collection of Insects (CNC) in Ottawa, Canada, have both revealed several new species for the Western Hemisphere. They are described here, and a key to all species in the New World is provided.
It should be noted that Achterberg (2002) did not consider Iconella as a valid genus and transferred its species (as well as those of other genera) back to Apanteles. However, that decision has not been universally accepted (e.g. Fernández-Triana 2010, Broad et al. 2012) and is still a subject of debate. In this paper we treat Iconella as a valid genus on its own.

Methods
Iconella is rare in collections, and even when present, specimens tend to be categorized as "unidentified Microgastrinae". In the New World, with the exception of one species, the genus seems to be rare in nature as well. Even long-term and comprehensive rearing programs -such as those in ACG, Costa Rica-have recovered relatively few specimens. This study is mostly based on the examination of unidentified Iconella specimens housed in the CNC, representing close to 40 specimens; 11 specimens reared by the ACG inventory -and housed in the CNC and the Illinois Natural History Survey, Champaign, Illinois, United States (INHS); one specimen from the Laurentian Forestry Center, Ste.-Foy, Quebec, Canada (LFS); and the holotype of I. etiellae from the Smithsonian Institution, Washington DC, United States (NMNH).
Morphological terms and measurements of structures are mostly as used by Mason (1981), Huber and Sharkey (1993) and Whitfield (1997). Non-morphological characters are also provided in the key whenever available (e.g., host species, geographical distribution). Those features are included in brackets at the end of the corresponding couplet and are intended as supplementary information that can help with identification.
The species descriptions are based mostly on the holotype female, with other specimens studied (when available) for intraspecific variation. When the holotype could not be examined (I. isolata) or it was a male (I. etiellae), female specimens were used for these redescriptions.
Lucid 3.5.4 (http://www.lucidcentral.com/) software was used to generate automatic descriptions of the species and to prepare Lucid identification keys. A dataset of 20 characters and 126 character-states was used to provide uniform descriptions for all species treated. Description format includes one sentence per character, with the character (in italics) mentioned first and the character-state following after a colon, e.g., "Pterostigma color: mostly brown, with yellowish-white spot at anterior 0.2 ×".
A map with the distribution of all New World species of Iconella was generated using SimpleMappr (Shorthouse 2010).
In the taxonomic treatment of species, "Specimens Examined" presents the specimen's information in the following format: "Number of females/males, acronym of the storing institution between parentheses, COUNTRY: State/Province, city, other locality details, coordinates (in Decimal Degrees, abbreviated as "Lat:" and "Long:"), date, collector name, biological information on host (starting with "ex"), ACG database codes (in the format "yy-SRNP-xxxxxx" or "DHJPARxxxxxxx", with the former being the voucher code of the host and the latter being the voucher code of the parasite). For states of the United States and for Canadian provinces/territories, acronyms consisting of two capital letters are used, following Canada Post (http://www.canadapost.ca/tools/pg/manual/PGaddress-e.asp).
DNA barcodes for all specimens that were barcoded were obtained using DNA extracts prepared from single legs using a glass fibre protocol (Ivanova et al. 2006). Extracts were re-suspended in 40 μl of dH2O, and a 658-bp region near the 5' terminus of the COI gene was amplified using standard primers (LepF1-LepR1) following established protocols (Smith et al. , 2007(Smith et al. , 2008. If the initial 658 bp amplification was not successful, composite sequences were generated using internal primers (primers are as detailed in Smith et al. (2008)). All available DNA barcodes for Iconella specimens from the New World and the Palearctic are available on the Barcode of Life Data System (BOLD, www.boldsystems.org/), along with sequences from Apanteles and Dolichogenidea, which were used as outgroups in the phylogenetic analyses (dx.doi.org/10.5883/ DS-ASICON1). Collection information and accessions (BOLD and GenBank) for all sequences were already published in a previous paper (Smith et al. 2013).
DNA barcode sequences were aligned in Geneious Pro 6.0.5 (Drummond et al. 2011) using the default settings for a MUSCLE alignment. Because of the high AT content characteristic of Insect and Hymenoptera mitochondrial DNA, sequence divergences were calculated using the TN93model (Tamura and Nei 1993) and a neighbor-joining (NJ, Saitou and Nei 1987) tree of distances was generated using Geneious Pro 6.0.5 (Drummond et al. 2011) to provide a graphical representation of the species divergences.
Many of the Iconella specimens in BOLD were collected before 1979. Characteristic of such moderately-aged specimens, the COI fragments generated were less (~160bp) than the full length characteristic of the standard DNA barcode (>600bp) (Fernández-Triana et al. 2011). Therefore, more than half of the sequences used in the analyses described below were less than 300bp, and 44% were less than 200bp. Phylogenetic comparisons of fragments this small, and involving cases where there is even lower overlap will be compromised. However, small COI fragments (<200bp) are not devoid of phylogenetic signal and have been used to successfully identify species (Fernández-Triana 2010, Fernández-Triana et al. 2011, Hajibabaei et al. 2006. To investigate the phylogenetic relationships of the species, the aligned dataset was analyzed in MrBayes v. 3.2.1 (Ronquist and Huelsenbeck 2003). Model selection was based on the Akaike Information Criterion as implemented in JModelTest v.2.1.1 (Darriba et al. 2012). Two independent analyses with 4 chains each were run in parallel for 10 million generations under a GTR+I+G model. The parameter trace files of each run were observed in Tracer v.1.5 (Rambaut and Drummond 2009) to verify that the runs had converged on the same stationary distribution, and to decide on the appropriate number of generations to discard as burn-in. A 50% majority rule consensus tree was constructed from the 18 million post-burn-in generations in Geneious Pro 6.0.5. The above protocol was followed for additional analyses in which 1) all 3 rd codon positions were removed from the dataset, 2) characters with more than 2% missing data were removed, and 3) sequences that were less than 547 base pairs long were eliminated. A Maximum Likelihood (ML) analysis was also run in RAXML v. 7.3.4 (Stamatakis 2006) under a GTR+I+G model. 1000 bootstrap (BS) replicates were run and the BS support values were then drawn onto the best-scoring ML tree. A parsimony analysis was conducted in TNT v.1.1 (Goloboff et al. 2008) using the "aquickie.run" script provided with the program.

Results
As with many other taxa, the generic status of Iconella will only be solved when a comprehensive phylogeny of Microgastrinae is carried out. In the meantime we think is best to keep it as a valid genus, based on the available morphological evidence. Mason (1981) already mentioned the sinuated vein cu-a in the hind wing as a plesiomorphic character that suggests the unique status of Iconella among similar genera. Besides that, we also consider the presence of a median longitudinal carina on the propodeum (or the secondary loss of that carina, which occurs in some Paleartic species but not in the New World species) as a strong support for the generic status of Iconella -in contrast with Apanteles, which (sensu Mason 1981) almost always has carinae defining a complete or partial areola on the propodeum, but never has a median longitudinal carina on the propodeum.
Phylogenetic analyses of the COI DNA, however, are inconclusive as to whether the genus Iconella is monophyletic (Figs 1 and 2). All Bayesian (only the tree resulting from analysis of the full dataset is shown in Fig. 1), ML (tree not shown), and parsimony (tree not shown) analyses failed to recover monophyly of the genus. However, the model-based methods do not offer strong support against monophyly either. The highest posterior probability (PP) supporting non-monophyly was 0.55 when 3 rd codon positions were removed from the Bayesian analysis, and BS support against monophyly in the ML analysis was only 0.21.
We recognize five species of Iconella in the New World, including the three new species described in the present paper, and one that was previously considered as a subspecies (Muesebeck 1955) but is here elevated to species. We are also aware of an additional sixth species, represented by two specimens in poor condition from British Columbia (Western Canada), deposited in the CNC. They were mentioned in Fernández-Triana (2010) as "Iconella sp. 2" and are different from all other New World species of the genus. However, until more material is found, it is impossible to describe that species.
All of the described species have a different, well delimited geographic range (Fig. 3) and also differ in their known hosts. Most of the host species in the Western Hemisphere belong to the Lepidoptera families Pyralidae and Crambidae (with appropriate caution that what was called Pyralidae in the past is in part called Crambidae at present). The NJ tree (Fig. 2) as well as Bayesian (Fig. 1), ML, and parsimony based phylogenetic analyses of the barcodes support the validity of all New World species recognized herein (there is no molecular data available for I. isolata). However, the barcoding dataset did not contain a strong enough phylogenetic signal to resolve the phylogenetic relationships among Iconella species.
While these DNA barcodes and short COI fragments (mini-barcodes (Hajibabaei et al. 2006, Meusnier et al. 2008) were sufficient to unambiguously separate the species of Iconella, more collections and longer DNA sequences will be needed to test the monophyly (or lack thereof) of Iconella.

1
Propodeum mostly smooth and polished (as in Fig. 21), except for sparse punctures on the anterior 0.2 × of propodeum, and rather small carinae ra- Pterostigma mostly transparent or whitish, with only thin brown margins (Fig. 24); humeral complex yellow to white; profemur mostly yellow, dark brown area limited to anterior 0.2 or less; interocellar distance 2.1 × or less posterior ocellus diameter; mediotergite 2 width at posterior margin 5.0 × or more its length (Fig. 26); smaller size, body length (head to apex of metasoma) 3.0 mm or less, and fore wing length 3.
Distribution. Only known from the holotype locality, in Sector Rincon Rain Forest of ACG at 410 m.
Comments. The species has been reared only in one place from three caterpillars collected at the same time on the same species of food plant, out of 12,000+ rearings of ACG Pyralidae of more than 200 species. A single additional specimen, identified by DNA barcoding, has been reared in 2012 from the same species of host caterpillar in the same place and on the same food plant, but was not available for study. Iconella andydeansi is sympatric with I. jayjayrodriguezae in ACG, the latter being a species with a slightly larger ACG rain forest distribution but equally narrow food plant record, and also known from Chiapas, Mexico. Those are the only two species of New World Iconella that are known to be sympatric, and that was revealed through extensive collecting in ACG and the use of DNA barcodes. It is likely that further collecting in other areas of the Neotropics, as well as the barcoding of more fresh specimens, will reveal additional species.
Male. As female. Molecular data. We analyzed three short 164 bp COI sequences from the DNA barcode region.
Distribution. Eastern Canada: NB, ON, QC. Comments. The holotype has a label stating that it emerged from either Acrobasis betulella (Pyralidae) or Rheumaptera hasta (Geometridae). Based on the known biology of the genus Iconella in the world, the second alternative is unlikely, and thus we consider A. betulella as the potential host in that case. However, the pyralid host cannot be taken as definitive until more reared specimens confirm the decision.
Male. As female. Molecular data. No DNA barcode sequence was available for this species, the only one among the New World species without molecular data.
Comments. Until now, I. isolata had been considered to be a subspecies of I. etiellae. After study of numerous specimens, we have found consistent and significant differences in morphology, hosts and geographical distribution and thus consider I. isolata as a different species on its own. Cayman Islands and Saint Kitts & Nevis are new locality records, based on CNC specimens. The specimen from Cayman Islands represents the westernmost locality, and it expands considerably the previously known distribution of the species in the Caribbean islands.