?Addendum to a minimalist revision of Costa Rican Braconidae: 28 new species and 23 host records

?Abstract Twenty-nine species are treated, most of which have host caterpillar and food plant records, and all but one are new to science. The first host record for the agathidine genus Amputoearinus is given. Gnathopleurajosequesadai Sharkey, sp. nov. is reported as a hyperparasitoid of fly larvae, the first such record for the genus. The following new species are diagnosed primarily using COI barcode data; Sharkey is the authority for all: Agathidinae: Aerophilusdavidwagneri, Aerophilusfundacionbandorum, Aerophilusnicklaphami, Lytopylusdavidstopaki, Lytopylusdavidschindeli; Alysiinae: Gnathopleurajosequesadai; Braconinae: Braconandreamezae, Braconfranklinpaniaguai, Braconrafagutierrezi, Braconguillermoblancoi, Braconoscarmasisi, Braconpauldimaurai, Braconshebadimaurae, Saciremakarendimaurae; Cheloninae: Chelonusminorzunigai; Homolobinae: Homolobusstevestroudi; Macrocentrinae: Macrocentrusmichaelstroudi; Orgilinae: Stantoniagilbertfuentesi; Rhysipolinae: Rhysipolisstevearonsoni; Rogadinae: Aleiodeskaydodgeae, Aleiodeskerrydresslerae, Aleiodesjosesolanoi, Aleiodesjuniorporrasi, Aleiodesrocioecheverri, Aleiodesronaldzunigai, Choreborogasjesseausubeli, Triraphisdoncombi, and Yeliconesmayrabonillae.


Introduction
of specimens of Alabagrus (Braet 2002, Cauich-Kumul 2012 are any better, despite his being the world authority on identification of Agathidinae. The sole publication in which Sharkey did not play a role in identification was one dealing with the Brazilian fauna (Yamada et al. 2006). In this publication 21 species of Alabagrus were identified. Of these, ten of the holotypes are from Mexico or Central America and one is from the United States. The likelihood of any of these occurring in Brazil is extremely low and yet they probably fit the key in Sharkey (1988). We estimate there to be many more than 1,000 species of Alabagrus in the Americas; the probability of the undescribed species fitting the key is therefore obviously high. A key that deals with only 10% of the fauna is all but useless, and if all of the species were described, the key would be more than 1,000 couplets long; impossible to work with. A key of this length would preclude accurate identifications due to user error or location inaccuracies (e.g., a Brazilian specimen that looks like a Mexican specimen has a high probability of being a different species).
In summary, in the 30-plus years since the publication of the morphology-based revision of Alabagrus, only one person other than Sharkey has used the key to arrive at a determination for Neotropical species, and in that instance most of the identifications are probably incorrect. It took Sharkey more seven years to produce the 1988 revision, and it is worse than useless because it is full of misleading information on species limits and species distributions, owing to misidentifications. Some might argue for an integrative approach, such as the revision of Alabagrus by Sharkey et al. (2018), but what is the point of including morphological descriptions, which tend to be lengthy and time-consuming to produce, when the COI barcode is the only reliable source for identification (barring much more expensive and complex multi-gene information)?
There are many genera of ichneumonoids that contain hundreds or thousands of species in the Neotropics, but few of these have been revised for the entire area. One of these exceptions, besides Sharkey (1988), is the revision of Neotropical Mesochorus (Dasch 1974), a genus of hyperparasitoid Ichneumonidae. Dasch treated what he considered to be 245 Neotropical species, and like the Sharkey revision of Alabagrus, few publications have used his keys or descriptions to identify specimens of Mesochorus; we have located three. Of all of the 245 species of Neotropical Mesochorus that he treated, 30 were recorded from Costa Rica. Based on the 172 BINs of Costa Rican Mesochorus presently on BOLD (March 2, 2021), we estimate that there are approximately 688 species in Costa Rica. These species are almost exclusively from the Area de Conservación Guanacaste, from rearings that have been conducted exclusively in the provinces of Guanacaste and Alajuela . Given this estimate, the odds of a Costa Rican specimen being in Dasch's (1974) key is 4.4%. The fact that these large revisions are not useful is not the result of poor workmanship, nor is it that the readers are poorly trained. In fact, the only users are highly trained taxonomists specializing in Ichneumonoidea. The keys are not used simply because they do not work. Not only are the species concepts poor with many species having similar morphology, but the revisions deal with such a small portion of the total number of species that the odds of a specimen in hand being in the key is remote.

Delimiting species
We received considerable critical feedback after the publication of Sharkey et al. (2021a). One of the most common criticisms directed at our approach is that barcodes in general, and BINs in particular, are not capable of delimiting species. This was dealt with at length in Sharkey et al. (2021a); nonetheless in an effort to avoid further confusion, we describe in detail below the process we go through to arrive at species limits or at the least, species central tendencies. BIN is an abbreviation for Barcode Index Number and an article by Ratnasingham and Hebert (2013) describes how and why the BIN algorithm was developed. They describe the BIN system as a means of forming Operational Taxonomic Units (OTUs) based on divergence in COI sequences. In essence, the BIN is like a unit tray of specimens believed to be monospecific by similarity of contained barcodes rather than appearance. They clearly state that no system like this can be perfect, "Any algorithmic approach based on the analysis of sequence diversity in a single gene region will be an imperfect tool for the discrimination of closely related species as they will be overlooked because of their low sequence divergence." (Ratnasingham and Hebert 2013: 2).
We start a revision by grouping our specimens into unit trays based on their BIN placements. The specimens in each tray are then investigated for general morphological consistency, and inconsistent specimens are flagged. This is followed by an inspection of a NJ tree that we generate on the BOLD website using only those specimens with full or almost full barcodes, i.e., barcodes with 500-658 base pairs. We carefully examine the branching pattern of the specimens in each BIN. If there is any clumping or any outliers in the tree of the specimens within a BIN, we look at the rearing host data and microgeography, if it is available. We also look at the morphology of the specimens, to see if they differ significantly and check for concordance between these three data sources. If these data sources are consistent with the hypothesis that any cluster of branches represents a separate species within the BIN, we consider this possibility based on the degree of difference in morphology, sequence divergence, and host use. We then build a new NJ tree that includes shorter barcodes to place those specimens into species formulated in the previous step and to add new species that may not be represented by specimens with full barcodes. Finally, we look at the morphology and host data of the nearest neighbors for each BIN. If these do not differ morphologically, we might consider this to be a case where a pair of BINs split a species. This, of course, depends on the degree of COI divergence. To date, we have found no such case. Specimens that fail to barcode, are contaminated, or are otherwise not barcodeable, are excluded from consideration, but the specimen and its record are retained. There are times when a reared specimen is obviously conspecific with others reared from the same host species but not currently barcodeable, and therefore, they are only retained for ecological analyses, such as what fraction of caterpillars were killed by that parasitoid.
Co-author Janzen estimates that the BIN algorithm lumps two or more sympatric species of Costa Rican Lepidoptera within a BIN at a rate of ~ 10%. And in those cas-es, almost invariably the multiple species are evident by genitalic differences, caterpillar food plant, microgeography, and/or extremely slight differences in coloration. He also has not come across a case in which a species is split into more than one BIN, although this is certainly possible through within-species barcode polymorphisms. Thus, the BIN algorithm can be described as conservative. The 403 new species in Sharkey et al. (2021a) were grouped into 395 BINs with only three "multi-species" BINs, for a 2% BIN "error" rate. Error is in quotes here because the barcodes do separate the species, but the BINs do not in these few cases.
In the following paragraphs we give an empirical example of how we arrive at species delimitations; we do not say "species limits" because these geobiological limits have not been explored further than ACG, or Costa Rica, or the Neotropics. BIN BOLD:ACK7466, treated in Sharkey et al. (2021a), is a BIN with multiple species, and there are also a handful of examples in the literature (e.g., Hebert et al. 2004;Janzen et al. 2017). In this BIN, BOLD:ACK7466, we have what are probably 11 species, nine of which we have described, and one is in the current publication. Each of these 11 species matches a distinctive set of host caterpillars yet are fully sympatric, just as was the case for 6 of the first 11 sympatric species found to be described as a single species, Astraptes fulgerator (Hesperiidae) (Hebert et al. 2004). To help discover potential species within a BIN, the first NJ tree that we build employs only those sequences with complete or almost complete barcodes, e.g., > 500bp. The portion of the tree that contains the specimens of BIN BOLD:ACK7466 is presented in Figure 1. The reason for using almost complete barcodes at this stage is to base our decisions on the highest quality data. A quick look at the tree ( Fig. 1) shows that there are a number of specimens with identical sequences that cluster together on different branches of the tree. We investigate each cluster individually. For example, the two specimens of M. michaelstroudi sp. nov. (branch A at the top of the tree in Figure 1) are consistent with the hypothesis of being a separate species because: 1. They have the same barcode sequence, which is quite divergent from other members of the clade. 2. Their hosts are the same crambid, Phaedropsis leialisDHJ03, and no other specimens in the BIN attack members of this genus. 3. These two specimens are morphologically different from all others in the BIN, the details of which are in the diagnosis in the species treatment; "In the morphological key for the species in this BIN, Macrocentrus michaelstroudi keys to M. gustavogutierrezi. Macrocentrus michaelstroudi differs in having pale basal flagellomeres, contrasting with the melanic basal flagellomeres of M. gustavogutierrezi (Sharkey et al. 2021a).
Within the cluster of specimens identified as Macrocentrus geoffbarnardi (Fig.1, branch B) we have the same situation as in M. michaelstroudi, so we compare with the specimens on the branch with specimens of M. fredsingeri (Fig. 1, branch C). Here we have two clusters (C1 and C2) that are joined on a relatively long branch. Members of branches C1 and C2 are all parasitoids of Neurophyseta clymenalisDHJ03. (N. cly-menalisDHJ03 is the interim name for a species, probably unnamed, that is similar to N. clymenalis). The specimens on C1 cannot be separated from those on C2 on morphological grounds. However, the entity (C1 + C2) can be separated on morphological grounds from all of the other specimens in BIN BOLD:ACK7466. Finally, no other specimens in the BIN are parasitoids of species of Neurophyseta. Therefore, we considered the entire cluster (C1 + C2) as one species. If further examination or data suggest that it is two, then one more will be also described. Similar arguments were used to delimit the other nine species in the BIN (Fig. 1). The specimens highlighted in blue ( Fig. 1) represent probable new species that have not yet crossed the desk of author Sharkey because they are still in the barcoding pipeline.
The next step in our process is to add specimens to the analysis with less COI data, i.e., shorter COI barcodes. This often produces a "noisier" NJ tree. Here we just show a small segment of the NJ tree to make our point (Fig. 2). The highlighted terminals have shorter barcodes and are new to the tree of Figure 1. Neither falls in the large homogeneous polytomy of M. gustavogutierrezi, and this is not uncommon for specimens with shorter barcodes. Specimens of this sort are looked at more carefully both morphologically and biologically and may or may not be included in the paratype series. In this case the two specimens share the same host, which is unique to the species, and do not differ morphologically in a substantive way.
By this stage we have examined the membership of each BIN to determine whether there is one or more species in the BIN. The final step before species description is to investigate the nearest neighbor of each BIN to ensure that they differ morphologically and/ or biologically. To date, all BINs examined for Braconidae have differed from their neighboring BINs. The nearest neighbor can be found on the BOLD website. For example, to find the nearest neighbor of BIN BOLD:ACK6477, we search for that BIN on the BOLD database and are taken to a page that includes the information in Figure 3. In this case the nearest neighbor to BIN:BOLD:ACK6477 is Macrocentrus iangauldi, which occupies BIN BOLD:ABY7812. Specimens in the two BINs are then compared to ensure that they differ morphologically and biologically, which they do. We stated earlier that the BIN algorithm failed to be mono-specific at a rate of 2% in the treatment by Sharkey et al. (2021a), but it is worth noting that the barcode sequences themselves did not fail. Even when there are multiple species within a BIN, the COI sequences differentiated the included species as seen in Figure 1; these results are corroborated by host data and morphology. In contrast to COI barcode diagnostics, we have found cases in which morphology cannot discriminate species that are clearly diagnosed by COI barcodes. Of the 86 species treated in the revision of Alabagrus by Sharkey et al. (2018) there were three species that could not be separated morphologically but were clearly delimited based on host data and COI sequence data. The final couplet for this group from the key by Sharkey et al. (2018) Figure 4 is a portion of the tree of highest posterior probability (from Sharkey et al. 2018) based on COI data showing the relationships of these three species (indicated with a red dot). The NJ tree produced by BOLD indicates slightly different relationships but, as with the Bayesian tree, the three morphologically indistinguishable species are not sister species nor are they nearest neighbors by any definition of that concept. We may have made different decisions if these lineages shared hosts or formed a monophyletic clade and were represented by very few sequences.

Specimens and generic placements
As with those of Sharkey et al. (2021a), most of the species described here were collected by rearing wild-caught host caterpillars in ACG in northwestern Costa Rica (Janzen and Hallwachs 2016). Holotypes of all newly described species are deposited in the insect collection of the Canadian National Collection of Insects, Ottawa. Paratypes and all other specimens are currently deposited in the Centre for Biodiversity Genomics in the Biodiversity Institute of Ontario at the University of Guelph. Identification of specimens to the subfamily level can be achieved using the key by Sharkey (1997). Keys to the genera of the species treated here are found in Sharkey et al. (2021a) and references therein. However, identification to any level is best acquired by obtaining COI barcodes and submitting them to BOLD. Instructions on how to do this are included below.
Some host species treated here are awaiting full identification and are given interim names. For example, Antaeotricha Janzen233 is identified to the genus Antaeotricha by classical morphology-based criteria and to Janzen233 by barcode and ecological information. However, no formal scientific species name is available until a barcode-match is obtained with an existing holotype or until it is described as new, or interim matched morphologically with a described species by a taxonomic specialist, which may take decades. Equally, Antaeotricha radicalis EPR03 is also an interim name based on what the species looks like, however, it is not a scientific name. It temporarily retains the information that this species is recognized by similarity with its look-alike, A. radicalis, before barcoding and associating it with other ecological data. Finally, a name such as gelJanzen01 Janzen407 signifies a caterpillar in the family Gelechiidae for which even a generic name is not obtainable at present.

DNA extraction and sequencing
Molecular work was carried out at the CBG using their standard protocols. A leg of each frozen-then-oven-dried specimen was destructively sampled for DNA extraction using a glass fiber protocol (Ivanova et al. 2006). Extracted DNA was amplified for a 658-bp region near the 5' terminus of the cytochrome c oxidase subunit I (COI) gene using standard insect primers LepF1 (5'-ATTCAACCAATCATAAAGATATTGG-3') and LepR1 (5'-TAAACTTCTGGATGTCCAAAAAATCA-3') (Ivanova and Grainger 2007). If initial amplification failed, additional amplifications were conducted following the established protocols using internal primer pairs: LepF1-C113R (130 bp) or LepF1-C_ANTMR1D (307 bp) and MLepF1-LepR1 (407 bp) to generate shorter overlapping sequences. Amplified products were sequenced using Sanger technology, though the most recent were sequenced by SEQUEL II. Specimens that "failed" barcoding are not included here unless otherwise indicated. When included, they are usually identified by unambiguous morphological and ecological information equally possessed by others from ACG in that species.
Barcode sequences presented in the species descriptions herein are a consensus of the barcode sequences of all included individuals, meaning base pairs that differ between conspecific specimens are replaced by IUPAC ambiguity codes.

Databases
Voucher codes are presented for all holotype specimens (and all other barcoded individuals) treated herein. All host caterpillars are individually vouchered to their individual records (yy-SRNP-xxxxx). Codes beginning with DHJPARxxxxxxx are for the parasite (or hyperparasite) specimens reared from the caterpillar; therefore, each wasp carries two voucher codes, one for the rearing (host) record and one for the wasp itself. The SRNP voucher codes are from the Janzen and Hallwachs' database (http:// janzen.sas.upenn.edu/caterpillars/database.lasso). Specimen voucher codes beginning with BIOUG are from the BOLD database (http://www.boldsystems.org), and most of the specimens obtained from ACG Malaise traps have this prefix. The DHJPAR and their associated SRNP codes can also be found on the BOLD database. The abundant collateral information obtainable from these two databases complements the species treatments. See Sharkey et al. (2021a) for a brief introduction to what to look for and how the two databases supplement the species treatments herein.
The BOLD database can be used to identify specimens using the following steps: 1. Navigate to the identification tab of the BOLD Systems database (http://www.boldsystems.org/index.php/IDS_OpenIdEngine). 2. Paste the COI sequence of the query organism (in forward orientation) into the query box and search against the appropriate library (e.g., All Barcode Records on BOLD, Species Level Barcode Records, etc.). 3. The search results page shows the top hits based on % similarity starting with the closest matches. This page also provides additional information to help verify the identity of a match, such as links to the BIN where specimen data (including images) can be found, a distribution map, and a tree-based identification tool. 4. Use the Tree-Based Identification button to generate a neighbor-joining tree and find the query taxon (name in red). This allows you to visualize how distant the query sequence is from the closest matches.

Agathidinae
A key to the genera of the New World can be found in Sharkey et al. (2021a). Agathidines are cosmopolitan and exclusively koinobiont endoparasitoids of caterpillars. They emerge from the host after the caterpillar is full-grown and has begun to spin or has already spun a cocoon.
Etymology. Aerophilus davidwagneri is named in honor of David Wagner of the University of Connecticut, Storrs, Connecticut, USA, for his recent work as an environmental activist for a healthier global climate and wild biodiversity.

TATTTTATTATTATTATCTTTACCAGTATTAGCTGGGGCTATTCTATATT-ATTAACTGATCGAAATTTAAATCTAGATTTTTTGATCCTTCTGGAGGAG-GAGATCCAATTTTATATCAACACTTATTT
Morphological data. Aerophilus fundacionbandorum keys to A. macadamiae in Sharkey et al. (2011). Aerophilus fundacionbandorum differs in many ways. One of the most obvious is its wide, sharply angled, median propodeal areola (Fig. 6). In A. macadamiae the areola is gradually narrowed anteriorly. A. fundacionbandorum can be morphologically distinguished from its nearest neighbor, A. calcaratus, by its more heavily sculptured first metasomal tergum. It is mostly striate in A. fundacionbandorum and mostly smooth in A. calcaratus (Sharkey et al. 2016: figs 12, 13).
Paratype. Same host species as that of holotype DHJPAR0054547. Etymology. Aerophilus fundacionbandorum is named in honor of the BAND Foundation of the USA, in recognition of its decades of support for growth and development of Costa Rica's Área de Conservación Guanacaste and most recently for adding 85 more hectares to ACG of original forest, Bosque Transición, lying on the nearly extinct fusion of dry forest with rain forest (http://www.gdfcf.org/content/introducingbosque-transición).
BOLD data. BIN: BOLD:ACT7814; nearest neighbor: Aerophilus colleenhitchcockae BOLD:ACA4890; distance to nearest neighbor is 5.16%. Consensus barcode: Morphological data. This species keys to A. jessicadimauroae in Sharkey et al. (2011), but A. nicklaphami differs in many ways. Two of the most obvious are the wide sharply angled median propodeal areola and the sharp lateral longitudinal carinae on the first metasomal median tergite. In A. jessicadimauroae the areola is gradually narrowed anteriorly and the carinae are not sharp.
Etymology. Aerophilus nicklaphami is named in honor of Nick Lapham of the BAND Foundation of the USA, in recognition of his decades of support for growth and development of Costa Rica's Área de Conservación Guanacaste, Costa Rica, and most recently adding 85 more hectares to ACG of original forest, Bosque Transición, lying on the nearly extinct fusion of dry forest with rain forest (http://www.gdfcf.org/ content/introducing-bosque-transición).
BOLD data. There is no BIN for this specimen because the barcode is too short to merit a BIN code. The short barcode follows: Morphological data. This specimen was identified based on morphological criteria from the key in Lindsay and Sharkey (2006).  Reared specimens host data: Dysodia spissicornis (Thyrididae) a leaf-roller feeding on Heisteria concinna (Olacaceae), caterpillar voucher code: 07-SRNP-22487.
Note. This is the first host record for this wasp genus.
Etymology. Lytopylus davidstopaki is named in honor of David Stopak of the Editorial Office of the Proceedings of the National Academy of Sciences (PNAS), in recognition of his decades of editorial understanding and accepting the strange research emerging from the biodiversity inventory of Costa Rica's Área de Conservación Guanacaste.
BOLD data. BIN: BOLD:ACB1289; nearest neighbor: Lytopylus davidstopaki BOLD:ACJ2185; distance to nearest neighbor is 2.56%. Consensus barcode AATTTTATATTTTATATTTGGAATTTGATCAGGAATTTTAGGATTATCAT- Morphological data. This species keys to L. angelagonzalezae in Kang et al. (2017) and differs in many respects. The most evident is that the propodeum of Lytopylus davidschindeli is almost completely smooth with the central areola barely indicated. This species can be morphologically distinguished from its nearest neighbor, Lytopylus davidstopaki, by having the mesosoma and coxae entirely black or dark brown (Fig. 10) compared to entirely tan (Fig. 9).
Etymology. Lytopylus davidschindeli is named in honor of David Schindel of the greater Washington, D.C. area and formerly the US National Science Foundation for his decades of understanding the unconventional traits of the biodiversity inventory of Costa Rica's Área de Conservación Guanacaste.

Alysiinae
The key by Wharton (1997) is outdated but it is the only available key to Alysiinae genera of the New World. Members of the subfamily are parasitoids of cyclorrhaphous Diptera. Diagnostics. Figure 11. BOLD data. BIN: BOLD:AAE0055; nearest neighbor: Gnathopleura sp. BOLD:AEF6891; distance to nearest neighbor is 7.99%. Consensus barcode GTATTATATTTTATATTTGGTATTTGAGCTGGTATAGTAGGGTTATC- Morphological data. Gnathopleura josequesadai keys to G. cariosa in Wharton (1980) but differs in many ways. For example, the first flagellomere is approximately equal in length to the second flagellomere in G. cariosa but much shorter than the second in G. josequesadai. This species can be morphologically distinguished from its nearest neighbor by the carina separating the propodeum from the metapleuron. This is rough and complete in G.josequesadai (Fig. 11) and smooth and restricted to the posterior 1/3 in the specimen in BIN BOLD:AEF6891.
Holotype host data. Hyperparasitoid of the fly Leschenaultia Wood30DHJ01(Tachinidae) which is a primary parasitoid of Pachyliaficus (Sphingidae) feeding on Maclura tinctoria (Moraceae). Including the holotype, five specimens were reared from the fly puparia parasitizing the caterpillar, voucher code 08-SRNP-13289.The host flies were identified from their surviving sibs, one of which is DHJ-PAR0027924 of BIN BOLD:ACE9310.
Etymology. Gnathopleura josequesadai is named in honor of José Ramón Quesada Mora, the manager of the 2020-21 BioAlfa Malaise traps for the Hacienda Baru Wildlife Refuge, Costa Rica.
Note. This is the first species of Gnathopleura confirmed to be a hyperparasitoid.

Braconinae
Braconines are mostly primarily idiobiont parasitoids of Coleoptera and Lepidoptera, but use many other insect orders as well. A key to the Braconinae genera of the New World is in Sharkey et al. (2021a).

TAATTGATTAATCCCTTTAATATTGGGATCTCCTGATATAGCTTTTCCTC-
Morphological data. This species can be morphologically distinguished from its nearest neighbor by having the metasoma dorsally entirely yellow and the mesepisternum dorsally black (Fig. 12) compared to the metasoma dorsally black and the mesepisternum entirely yellow-orange in B. franklinpaniaguai (Fig. 14).
Paratype. Ten specimens, reared from the same caterpillar as the holotype, were mounted and designated as paratypes (DHJPAR0066400 to DHJPAR0066409), depository CNC.
Etymology. Bracon andreamezae is named in honor of Ministra Andrea Meza Murillo of the Ministerio de Recursos Naturales y Energía de Costa Rica (MINAE) in recognition of her taking on this difficult ministerial task mid-government.
BOLD data. BIN: BOLD:ACK6897; nearest neighbor: Bracon alejandromasisi BOLD:AAA5367; distance to nearest neighbor is 4.49%. Consensus barcode: Figure 13. Communal and jointly constructed cocoon of at least 56 sibling larvae of Bracon andreamezae, one of which is DHJPAR0031182, displaying adult exit holes through the tough silk roof of the same consistency as the floor of the chamber; multiple wasps exited through a single hole. This species of wasp has been reared only twice among 1,391 rearings of solitary Yanguna cosyra caterpillars for more than 34 years.

ATATTATATTTTTTATTTGGAATTTGAGCTGGAATAATTGGTTTAT-
Morphological data. This species can be morphologically distinguished from its nearest neighbor by having the mesepisternum entirely orange-yellow, lateral sides of the head orange-yellow, and yellow fore-and mid-tibiae and femora (Fig. 14) compared to the mesepisternum entirely black, head entirely black, and all tibiae and femora black in B. alejandromasisi (Sharkey et al. 2021a: fig. 31).
Paratype. Eight specimens reared from the same host specimen as the holotype were mounted and designated as paratypes (DHJPAR0066410 to DHJPAR0066417), depository CNC.
Etymology. Bracon franklinpaniaguai is named in honor of Vice-Minister Franklin Paniagua Alfaro of the Ministerio de Recursos Naturales y Energía de Costa Rica (MI-NAE) in recognition of his taking on this difficult task mid-government.

GATCGAAATTTAAATACTTCTTTTTTTGATTTTTCTGGTGGAGGGGATC-CTATTTTATTTCAACATTTATTT
Morphological data. This species can be morphologically distinguished from its nearest neighbor by having all femora dark brown and the metasoma dark brown dorsally starting at the third tergum (Fig. 15) compared to all femora yellow and the metasoma yellow to light brown dorsally.
Holotype host data. Cosmorrhyncha albistrigulana (Tortricidae) feeding on Dialium guianense (Fabaceae). This is one of the only two species of Bracon reared by us that is solitary; the ten species treated by Sharkey et al. (2021a) are all gregarious. It was reared from a very small caterpillar; caterpillar voucher code: 12-SRNP-67398.
Etymology. Bracon rafagutierrezi is named in honor of SINAC Director Rafa Gutiérrez Rojas of the Ministerio de Recursos Naturales y Energía de Costa Rica (MI-NAE) in recognition of his taking on this difficult task mid-government. Morphological data. There is only one low-quality image on BOLD for the nearest neighbor, but the p-distance makes it doubtful that it is conspecific.
Bracon pauldimaurai Sharkey, sp. nov. http://zoobank.org/723581AA-45CD-4F42-9AB1-CED38EBEA383 Diagnostics. Figure 19.  Morphological data. This species can be morphologically distinguished from its nearest neighbor by having large portions of the mesoscutum and mesepimeron brown (Fig. 19)   This species differs from the three described species of Sacirema (Papp 2007) in many ways. The easiest to see is that none of the other three species has a predominantly yellow meso-and metasoma. The generic placement of this species is somewhat uncertain as are those of other specimens in the group of braconine genera with a medial area of the face delimited by longitudinal grooves or ridges.
Morphological data. No images of the unnamed nearest neighbor are available, and when it is described, it should be carefully compared to Sacirema karendimaurae.

Cheloninae
Cheloninae are egg-larval parasitoids of Lepidoptera. A key to the genera of the New World is included in Sharkey et al. (2021a).
BOLD data. BIN: BOLD:AEB3509; nearest neighbor: Chelonus jeffmilleri BOLD:ACF0845; distance to nearest neighbor is 4.81%. Consensus barcode: Morphological data. This species can be morphologically distinguished from its nearest neighbor by having the hind tibia entirely black (Fig. 22) whereas that of C. jeffmilleri has a light brownish yellow patch near the base of the hind tibia, which is otherwise black (Sharkey et al. 2021a: 164, fig. 101

Homolobinae
Members of Homolobinae are endoparasitoids of lepidopteran larvae. A key to the genera of the New World is included in Sharkey et al. (2021a). TATCAATAAGAATTATTATTCGAATAGAATTAAGAATACCAGGTAATT-

Homolobus stevestroudi
Morphological data. The specimen keys to Homolobus infumator in van Achterberg's (1979) key and is very similar morphologically. Subtle differences include the shape of vein R1a of the hind wing, which is longer in H. stevestroudi and the size of the basal tooth of the hind tarsal claws, which are longer in H. stevestroudi. The convincing difference can be found by looking at the NJ tree produced from the BOLD website; H. stevestroudi is found in its own BIN, far removed from any other species of Homolobus and particularly distant from specimens identified as H. infumator from Norway. The type locality of H. infumator is England. All nine specimens in the unnamed nearest neighbor are from Canada. They might represent the same species as the Costa Rican specimens, but more sampling will need to be done between Canada and Costa Rica to confirm or refute. There are no obvious morphological differences based on the BOLD images of the Canadian specimens.
Etymology. Homolobus stevestroudi is named in honor of Steve Stroud as the primary supporter of the BioAlfa Malaise trapping at the Hacienda Barú Wildlife Refuge, Savegre, ACOPAC, Costa Rica, as well as decades of support for the Area de Conservación Guanacaste.

Macrocentrinae
Members of all genera are koinobiont endoparasitoids of caterpillars from a wide range of families. Most are solitary, but several gregarious species are known. A key to the genera of the New World is in Sharkey et al. (2021a). Holotype host data. Phaedropsis leialisDHJ03 (Crambidae) feeding on Gouania lupuloides (Rhamnaceae), caterpillar voucher code: 16-SRNP-30230.
Etymology. Macrocentrus michaelstroudi is named in honor of Michael Stroud Bonilla as the primary supporter of the BioAlfa Malaise trapping at the Hacienda Baru Wildlife Refuge, Savegre, ACOPAC, Costa Rica, as well as decades of support for the Area de Conservación Guanacaste.

Orgilinae
Members of all genera are koinobiont endoparasitoids of caterpillars. A key to the genera of the New World can be found in Sharkey et al. (2021a). Morphological data. This species can be morphologically distinguished from its nearest neighbor by having the first metasomal tergite uniformly pale yellowish-orangeand the mesoscutum uniformly yellowish-orange (Fig. 26), contrasting with the first metasomal tergite darkening apically and melanic patches on each of the three mesoscutal lobes in S. miriamzunzae (Sharkey et al. 2021a: 502, fig. 354).

Stantonia gilbertfuentesi
Holotype ♂:  Etymology. Stantonia gilbertfuentesi is named in honor of Gilbert Fuentes of the Organización de Estudios Tropicales of Costa Rica in recognition of his decades of intensive management of the OET library of tropical publications.

Rhysipolinae
Members of the subfamily are thought to be solitary, koinobiont ectoparasitoids of caterpillars. A diagnosis for the subfamily is included in Sharkey et al. (2021a). Morphological data. This species can be morphologically distinguished from its nearest neighbor by having its mesoscutum entirely black (Fig. 27)  Paratype. BIOUG27682-G07. Etymology. Rhysipolis stevearonsoni is named in honor of Steve Aronson of San Jose, Costa Rica, in recognition of decades of concern and involvement with the betterment of Costa Rica's positive relationship with its wild environment, and specifically with providing broadband internet to Área de Conservación Guanacaste as the first Costa Rican Área de Conservación to be so facilitated.

Rogadinae
Members of all genera are koinobiont endoparasitoids of caterpillars from a wide range of families. A key to the genera of the New World is in Sharkey (2021a). Morphological data. This species can be morphologically distinguished from its nearest neighbor by having the base of the stigma brown (Fig. 28), contrasting with yellow in the nearest neighbor.
BOLD data. BIN: BOLD:AEF3944; nearest neighbor: Aleiodes sp. BOLD:AAG1309; distance to nearest neighbor is 8.29%. Consensus barcode: Morphological data. This species can be morphologically distinguished from its nearest neighbor by its pale stigma (Fig. 29), which is mostly melanic in the nearest neighbor.
Holotype Morphological data. The nearest neighbor is a sole specimen from Canada. There is no image available on BOLD but due to the distribution, conspecificity is doubtful.
Holotype ♀:  Morphological data. This species can be morphologically distinguished from its nearest neighbor by its uniformly colored hind legs (Fig. 31), compared to hind femur darker than remaining leg segments in the nearest neighbor.  Diagnostics. Figure 32.
Other material. BMNHE897799, from Belize deposited in the Natural History Museum (London), based on barcode, not viewed and lacking image on BOLD.
Etymology. Aleiodes rocioecheverri is named in honor of Rocio Echeverri of San Jose and Liberia, Costa Rica, in recognition of her lifetime of concern and involvement with the betterment of Costa Rica's positive relationship with its wild environment.

AGATCAAATTATATTATTAGTTTGATCTGTATTAATCACTGCTTTTTTAT-TATTATTATCATTRCCTGTTTTGGCGGGGGCAATTACTATATTATTATTT-GATCGTAAYATTAATAGAACTTTTTTTGATTTTTCAGGGGGAGGGGATC-CTATTTTATTYCAGCATTTATT
Morphological data. This species can be morphologically distinguished from its nearest neighbor by its swollen hind basitarsus (Fig. 34), which is much narrower in the nearest neighbor. Males lack the swollen hind femora.
Other material: BMNHE897774 from Belize is in the same BIN and likely conspecific. There is no image on BOLD and the specimen was not examined.
Etymology. Choreborogas jesseausubeli is named in honor of Jesse Ausubel of Rockerfeller University, New York, USA, for his very strong support of the germination and early development of DNA barcoding as an identification tool. Morphological data. No image of the nearest neighbor is available on BOLD, but the COI distance and geographic distribution suggest that they are not the same species.
Etymology. Yelicones mayrabonillae is named in honor of Mayra Bonilla as the primary supporter of the BioAlfa Malaise trapping at the Hacienda Baru Wildlife Refuge, Savegre, ACOPAC, Costa Rica, as well as decades of support for the Area de Conservación Guanacaste. Figure 39. Yelicones mayrabonillae, remains of host caterpillar, epipajanzen01 Janzen882 (Pyralidae); note mummified host caterpillar curved into a distinctive "C" shape, characteristic of other species of pyralid caterpillars attacked by species of Yelicones.