Necessary permits to conduct sampling of leaf galls were obtained from the Government of Kerala, India.

This study was carried out in two sites in Malappuram district, Kerala, south India; viz. Kottakkal (10°99’N, 76°00’E) and Vettichira (10°93’N, 76°02’E). The sites have a tropical climate and during the study period the average annual temperature was 26°C and annual rainfall was 2842 mm [23].

Weekly field surveys were carried out from August 2014 to January 2015. Leaf galls induced by the psyllid, Phacopteron lentiginosum Buckton (Phacopteronidae), on shrubs of Garuga pinnata Roxb. were collected at different developmental stages and dissected under a stereomicroscope (Olympus).The number of gall-inducing nymphs and the presence or absence of braconid eggs, and early and mature larvae were recorded. Digital photographs of dissected galls and braconid wasp developmental stages were taken in situ with a Canon IXUS 255 HS digital camera.

Alcohol-preserved specimens were processed with hexamethyldisilazane and later card-mounted. Images of the holotype and the egg, larva and pupa of the braconid wasp were taken with a Leica DFC 295 camera attached to a Leica S8 APO Stereozoomtrinocular microscope (Leica, Heerburg, Switzerland). Image stacks were combined into a single image and measurements of the holotype were done using Leica Application Suite V4.2. Images were edited using Photoshop CS8 (Version 6.1) (Adobe Inc.).

Morphological terminology employed in the description follows van Achterberg [24, 25] except wing venation nomenclature which follows Quicke [7]. Terms for sculpturing follow Eady [26] and Harris [27].

Larvae of different stages were dried with hexamethyldisilazane, mounted on entomological minuten pins and glued with epoxy resin on to standard electron microscopy stubs, sputter-coated with gold and studied under a JEOL JSM-5410LVmicroscope.

Larval head capsules were prepared from the gold-coated specimens as well as from uncoated material by macerating in 0.2M aqueous KOH to remove soft tissues, washing in dilute acetic acid, followed by dehydration through to xylene and mounting in Permount®.

A specimen was sequenced for the barcoding 5’ fragment of the mitochondrial cytochrome oxidase gene and the nuclear 28S rDNA D2–D3 region following the methods of Zaldivar-Riverón et al. [28, 29]. Sequences were deposited in GenBank with accession numbers CNIN2004 KT343804 and CNIN2004 KT343805.

All statistical analyses were carried out using the software package R [30].

The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature (ICZN), and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank Life Science Identifiers (LSIDs) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is: urn:lsid:zoobank.org:pub:FF0DA8B1-FB69-4203-8D37-0A6BDA5A2B64. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS

Adult females of B. garugaphagae oviposit in moderately well-developed leaf galls (8–12 mm in diameter) on G. pinnata (Fig 1A). Eggs are 0.8–0.9 mm in length (n = 9), pale yellow green, spindle-shaped, having a short tail end, and are deposited on the inner wall of the floor of the gall chamber (Fig 1B) which is often associated with a white waxy substance. Larvae of B. garugaphagae were observed in galls ranging from 8–22 mm in diameter (n = 779). We did not manage to determine the precise number of instars. Maximum larval length was 5.3 mm (Fig 1C) (n = 779), at which stage larvae started to spin a cocoon (Fig 1D). The spindle-shaped, pale brown cocoons (3.5–3.7 mm long; n = 13) are constructed at the base of the gall chamber with the head end facing towards the gall floor. Adult wasps emerge from the gall by chewing an exit hole with their mandibles.

The incidence of B. garugaphagae sp. nov. larvae in galls was monitored from August 2014 to January 2015. The larvae were first observed in galls which were 1–2 weeks old. The percentage incidence of braconid larvae in immature galls of size 8–12 mm in diameter was 6.5% (n = 779) and in mature galls of size 15–22 mm the incidence was 22%. Some of the galls (n = 53) were observed without any gall-inducing nymphs, but with the presence of B. garugaphagae larva.

Fig 5 shows the relationship between the Bracon stages found in sampled galls and the number of psyllids present, both nymphs and adults. Two regression analyses were carried out, one with all galls included, even those that had no Bracon individuals present, and one with only galls containing a Bracon individual. Both relationships were highly significantly negative (all data: GLM with Poisson errors and log link-total data: null deviance = 674.62 on 778 degrees of freedom, residual deviance = 558.65 on 777 degrees of freedom, p<0.0001; only when Bracon present: null deviance = 30.195 on 248 degrees of freedom, residual deviance: 22.693 on 247 degrees of freedom, p<0.0001). Logistic regression of presence or absence of Bracon in galls (n = 779) versus number of psyllids present, showing highly significant negative relationship (GLM with binomial errors: null deviance = 976.24 on 778 degrees of freedom, residual deviance = 920.01 on 777 degrees of freedom, p<5e-12). Gall size and the presence or absence of Bracon were not correlated (GLM with binomial errors, p = 0.699), but there was a significant positive correlation between gall size and the number of psyllids (GLM with Poisson errors and log link: null deviance = 190.78 on 778 degrees of freedom, residual deviance = 188.82 on 777 degrees of freedom, p = 0.0046).

Bracon garugaphagae sp. nov. Ranjith & Quicke, 2015 (Fig 6)

Known only from Kerala, south India.

Description: Female

Length of body 3.6 mm (3.6–5.8 mm in paratypes), of fore wing 2.8 mm (2.8–4 mm in paratypes), and of antenna 2.8 mm (2.8–3.8 mm in paratypes).

Head. Antenna with 24 flagellomeres (24–28 in paratypes). Terminal flagellomere strongly acute. Median flagellomeres normal in dorsal view. First flagellomere 1.2 times length of second and third flagellomeres respectively, first flagellomere 2.3 times as long as wide. Mandible twisted, only a single tooth visible in anterior view (Fig 6B). Inter-tentorial distance: tentorio-ocular distance = 1.5:1.0 (1.31–1.7: 0.82–1.07 in paratypes). Inter-tentorial distance: height of clypeus = 3.4: 1.0 (2.62–3.54: 0.81–1.04 in paratypes). Face slightly rugose in anterior half with smooth posterior half, sparsely setose laterally, smooth area laterally bordered by indistinct longitudinal groove (Fig 6B). Height of eye: shortest distance between eyes: width of head = 1.0: 1.1: 2.2(1.0–1.68: 1.1–1.83: 2.2–3.67 in paratypes). Oculo-antennal groove well-developed. Frons shiny. Stemmaticum triangular forming equilateral triangle. Shortest distance between posterior ocelli: transverse diameter of posterior ocellus: shortest distance between posterior ocellus and eye = 1.45: 1.0: 3.45 (1.45–1.81: 1–2.53: 3.45–7.06 in paratypes).

Mesosoma 1.6 (1.3–1.6 in paratypes) times longer than maximum height, largely smooth, shiny. Mesoscutum sparsely setose laterally (Fig 6C). Pronotum smooth. Notauli only indicated anteriorly (Figs 6C and 7C). Scutellar sulcus narrow, divided by eight carinae. Scutellum smooth (Figs 6C and 7C). Median area of metanotum large, smooth, slightly bulged in lateral view, without carina anteriorly. Propodeum smooth with a strong medial longitudinal carina (propodeal carina weak in paratypes), sparsely setose laterally (Fig 6D).

This new species can be distinguished from the two other known phytophagous Bracon species, both of which are Neotropical, in having 24 flagellomeres (58 in B. phytophagous and 49 in B. zuleideae), median flagellomeres normal in dorsal view (diamond-shaped in B. phytophagus), face slightly rugose anteriorly and smooth posteriorly (smooth in B. phytophagous and B. zuleideae), scutellar sulcus moderately wide (narrow in B. phytophagus), propodeum with a strong medial longitudinal carina (propodeum smooth without carina in B. phytophagous and B. zuleideae), fore wing vein 1RS forming an angle of 70° with vein C+SC+R (80° in B. phytophagous and 100° B. zuleideae). Further, the Neotropical species are generally larger and have the ovipositor exceedingly thin and largely unsclerotised except for the apical part which is nearly black and presumably very hard.

Bracon garugaphagae sp. nov. can be distinguished from B. psyllivorus Achterberg (reared from psyllid-induced leaf galls) in having the frons with a distinct medial longitudinal suture, impressed notauli, scutellum without antero-medial puncture, scutellar sulcus divided by eight carinae (five in B. psyllivorus), metanotum not tuberculate in lateral view, metasomal tergites 3–5 with distinct longitudinal strip and apex of ovipositor with distinct dorsal nodus and ventral serrations.

Considering the size of the subfamily, very few braconines have sequence data available in GenBank. The closest BLAST search matches (conducted on 25/07/2015) for both sequenced gene fragments were members of the Braconinae but only 93% and 98% similar for CO1 and 28S respectively. Bracon phytophagus was not sequenced for CO1 but its 28S sequence is only 94% similar (differing in 29 positions) showing, as expected based on morphology, that the seed-predating New World species are only distantly related to the new Indian species. No DNA data are available for B. psyllivorus for comparison.

The new species belongs to the large spectrum of Bracon species of size between 3–6 mm and differs markedly from the two other known phytophagous Bracon species both of which are Neotropical [14, 15] notably in having the strongly sculptured metasomal tergites. DNA sequence data further indicates that they are only distantly related. Instead, the new species appears to be most closely related to B. psyllivorus, which also attacks gall-forming Psylloidea viz: B. psyllivorus reared from the leaf galls induced by Pauropsylla gibberulosa Li and P. braconae Li (Hemiptera: Triozidae) [18]. Apart from the host record, no further biological details are known for B. psyllivorus [18], and the possibility that it is also entomophytophagous cannot be excluded.

Braconine wasps are almost entirely idiobiont ectoparasitoids of various concealed Coleoptera, Diptera, Hymenoptera and Lepidoptera with one small group of genera, the Aspidobraconina, being idiobiont endoparasitoids on exposed butterfly pupae [7, 31]. Psyllids have only been recorded as hosts of two other species of braconid wasp. Chadwick and Nikitin [32] recorded an unidentified Bracon sp. from a psyllid host in Australia, but no further details were provided. Recently, Li et al. [18] described a Chinese species, B. psyllivorus, as a parasitoid/predator of the psyllids P. gibberulosa Li and P. braconae Li (Hemiptera: Triozidae) that produce galls on the fig tree (Ficus hainanensis Merr. & Shun.; Moraceae).

The negative relationship between the number of psyllid individuals in galls and the developmental stage of B. garugaphagae indicates that the latter sequentially predate the psyllid nymphs before turning to phytophagy. The dietary shift seems likely to be because the psyllids do not provide B. garugaphagae with sufficient food to complete development. We do not know whether B. garugaphagae is obligately or only facultatively phytophagous during its late larval stages. In contrast to the previous explanation, the absence of B. garugaphagae larvae from galls with higher numbers (7–8 nymphs) of psyllids, may simply be that because the parasitoid larvae have to compete with psyllid nymphs for plant tissue the adult female parasitoids selectively avoid such galls.

Predatory behaviour is known in several parasitoid wasp taxa including cryptine ichneumonoids in spider egg sacs [7]. However, entomophytophagy whilst well-known in several eurytomid chalcidoids, which develop initially as parasitoids and complete their life cycle as phytophagous insects [33], was previously unknown in the Braconidae. Within the sister family Ichneumonidae there is evidence that at least one pimpline ichneumonid, a Calliephialtes species, does the same [Kenji Nishida, cited in 34]. Members of the labenine ichneumonid tribe Groteini consume the pollen ‘cake’ of their bee hosts after consuming the bee larva [35] and at least some Poecilocryptus species in the labenine tribe Poecilocryptini chew host gall tissue to reach other gall cells [36], but it is not confirmed whether any or all gain nutritive value from gall tissue though it seems likely. Within the Braconinae, there are a few other examples of predatory behaviour. One Neotropical species of Compsobraconoides is predatory on Azteca ants and their brood [37], an Afrotropical Trigastrotheca species also consumes ant broods in ant plant domatia [38]; and an Australian Bracon species consumes broods of gall-forming fergussoninid flies [32]. All these braconid species that prey on broods are contained within a single swollen plant structure, either an ant domatium or an insect-induced gall. It is also likely that a similar biology prevails in another braconine, Ficobracon brusi Achterberg & Weiblen, which has been reared from figs of Ficus wassa Roxb., but it is not clear whether it is a parasitoid/predator of the pollinating agaonid wasps (Chalcidoidea) or other fig inquilines or even whether it might be at least partly phytophagous [39]. Several other members of Braconinae, Doryctinae and Mesostoinae have also been associated with plant galls as inquilines or parasitoids on gall inducers [11, 39, 40] and the possibility that some of these are also partly phytophagous cannot be excluded.

Final instar larval cephalic structures, notably the long and relatively heavily sclerotised mandibles with two/three robust accessory teeth are very different from those of purely ectoparasitoid braconines. The latter have a very robust, almost square mandibular base and a short blade furnished with a series of small, closely spaced (comb-like) teeth [41]. The final instar larval mandibles of the entirely phytophagous Bracon phylacteophagus are even more heavily sclerotised, have even more closely-spaced robust ancillary teeth and the hypostomal spur is very well developed. In contrast, the mandibles of the gall-forming Mesostoinae are short-bladed, robust and lack ancillary teeth [42]. Clearly there are multiple possible types of larval head capsule adaptations for feeding on plant gall tissue.

All three known phytophagous Bracon species, viz., B. phytophagus, B. zuleideae and B. garugaphagae are associated with members of the plant family Burseraceae [14, 15]. The Neotropical species do not seem to induce gall tissue in the seeds they feed upon and their larval mandibles are far more highly modified and hardened than in the Indian species described here. The Neotropical species are not reliant on another insect having previously damaged the seed, and instead the apex of their ovipositors are extremely heavily sclerotised almost certainly as an adaptation to penetrating a much harder seed-coat. No such ovipositor modification is displayed by B. garugaphagae which only has to penetrate relatively softer gall tissue to reach its host.