Holotype ♂: COSTA RICA: Cartago: San Ramón de Tres Ríos, 1600 m, (09°56'18"N, 083°58'38"W), ex. Bocconia frutescens, larva exited 5–9.vii.2010; adult emerged 21–26.vii.2010, Kenji Nishida (LEM).

Paratype: same data as holotype (9 ♂; 8 ♀: LEM); same except larva exited: 25–28.vi.2010, adult emerged: 20–22.vii.2010 (14 ♂; 15 ♀: INBio); same except larva exited 29.vi–4.vii.2010, adult emerged 15–24.vii.2010 (8 ♂; 7 ♀: NMNH); same except adult emerged 21–25.vii.2010 (4 ♂; 4 ♀: CNC); same except along main road, (09°56'20"N, 083°58'55"W), 1500 m, collected 19.xii.2008, emerged 16–20.i.2009, K. Nishida & T. Johnson (1 ♂; 5 ♀: MZUCR). San José: San Isidro de Coronado, Centro. 1420 m, (09°53'18"N, 084°00'22"W), ex. Bocconia frutescens, adult emerged 17.vi.2010, Kenji Nishida (1 ♀: LEM). Puntarenas: Monteverde. Estación Biológica Monteverde. 1538 m, (10°19'09"N, 084°48'32"W), mating on Bocconia leaf. 18.vii.2010, Kenji Nishida (1 ♂; 1 ♀: MZUCR).

This species can be distinguished from other Neotropical species of Liriomyza by its completely yellow head and anepisternum, mesonotum almost completely brown to margin of scutellum, usually 2 + 1 dc, legs completely yellow, calypter brown on apical half with margin and fringe brown, and by the shape of the male genitalia and the shape of the anterior and posterior larval spiracles.

Frons width 0.25 mm; ratio of frons width to eye width 2.3; orbit 0.23 times width of frons at midpoint; frons slightly projecting above or in front of eye in profile (Fig. 8), forming a distinct ring (cheek) below eye; 2 reclinate ors and 2 inclinate ori (Fig. 9) (lower ori sometimes reduced or missing on one side); orbital setulae reclinate, varying in number from about 4–8; first flagellomere rounded, not enlarged in males, with slight apical pubescence; arista 0.30–0.40 mm, with short but dense pubescence; gena deep, slightly extended at rear (Fig. 8); gena height at midpoint: 0.44 times maximum eye height. Eye oblique, bare. Normally 2+1 dc (except 4 specimens with 2+0 dc and 3 specimens with 3+1 dc); acrostichals in about 4 irregular rows; prescutellar acrostichal bristles absent; 2 notopleural bristles; 1 strong postpronotal bristle with 1 or 2 small setulae; anepisternum with 1 strong bristle on posterior margin at midpoint, sometimes with a few extra setulae; katepisternum with one strong bristle on posterodorsal corner, on yellow ground. Fore and mid-tibia without lateral bristle. Wing length 1.50–1.95 mm in male and 1.85–2.20 mm in females; M₁₊₂ ending at wing tip; costa extending to M₁₊₂; last section of CuA₁: 1.5–1.9 times length of penultimate. Cross-vein r-m located at midpoint of cell dm. Stridulatory mechanism apparently absent.

Colour. Head (including frons, orbit, face, antenna, palp) entirely bright yellow. Hind margin of eye black for a small section beyond vte; both vt on yellow ground. Occiput black. Eye sometimes with a slight bluish or greenish reflection (not as pronounced as in Liriomyza prompta, Fig. 29); mesonotum almost completely dark brown except for narrow yellow margin posteriorly (Fig. 10), prescutellar area and intra-alar area sometimes slightly paler brown resulting in a weakly defined banded pattern on thorax, most visible in teneral specimen (Fig. 11); scutellum completely yellow with small brown patches laterally. Basal scutellar bristles on brown ground (but at the limit of yellow). Postpronotum, notopleuron and anepisternum completely yellow (at most with a very small pale brown patch on one or two of the sclerites). Katepisternum mostly brown except for upper margin yellow. Calypter brown on apical half, margin and fringe also brown; halter completely white. Legs completely yellow. Abdominal tergites pale brown.

Male genitalia. Distiphallus in the form of two narrow tubules, slightly diverging apically in ventral view (Fig. 13). Mesophallus widest apical section (Fig. 13a), about 1.5–2 times larger than basal narrower tubular section (Fig. 13b). Mesophallus in lateral view with small indent at midpoint (Fig. 12). Surstylus absent. Epandrium without chitinized margin and without spines. Ejaculatory apodeme (Figs 14, 15) weakly sclerotized, symmetrical or sometimes asymmetrical with blade more expanded on one side.

Early stages. Larval length (at maturity): 3.2–4.0 mm, slightly larger than Liriomyza prompta larva. White to creamy white with an internal orange spot at head (live specimens, Figs 24, 25, 27). Anterior spiracles about 0.13–0.23 mm distance from each other; fan-shaped and each with 5 small openings in a single row (Fig. 18). Posterior spiracles divided into 3 subequal projecting bulbs (Figs 16, 17). Cephalopharyngeal skeleton with wide arms (Fig. 19), Each mandible with 2 large teeth. Puparium pale brown to transparent (Fig. 48).

Bocconia frutescens L. (Papaveraceae).

The larvae feed on spongy parenchyma and other tissues of primary veins and petioles of large, relatively old, mature leaves. Most of the larvae were in leaves of >30 cm long, with >10 mm petiole width and >10 mm thickness (n=120). The larvae were more frequently found mining in the thicker part of the primary vein including the petiole (i.e. less frequently near the leaf apex). One larva was found mining inside of a 2.2 mm width primary vein near the leaf apex. A few larvae were mining thick secondary veins. The mining appears to occur longitudinally, mostly near the upper leaf surface; the larvae left some brown to reddish brown scars along the leaf blade where the vein and blade join (Figs 20, 21, 23). These scars were more easily seen with a strong transmitted light (Fig. 21). The mines (internal tunnels) can also be distinguished by narrow pale lines (Fig. 22). When infested veins were longitudinally dissected, usually one to a few white Liriomyza prompta larvae were observed mining singly and scattered (n=12 leaves) (Figs 23, 24). Some solitary parasitoid wasp pupae were also found among the spongy parenchyma (Fig. 50). The mature fly larvae exited from either the upper side or underside of the veins (n=12 holes) (Fig. 25), each larva making a small oval-shaped hole of 1.1–1.3 mm wide (n=12) (Fig. 26). The tissue around the old exit holes was brown to reddish brown (n=5). The newly emerged larvae wiggled around in rearing plastic bags/cases for a couple of hours to a few hours before settling (Fig. 27) and starting to form a puparium. The larvae readily pupated on the plastic surfaces. In general, the puparia (Fig. 48) were more translucent (translucent pale brown) than those of Liriomyza prompta (translucent brown to dark brown) and the pupa inside was visible. Duration of the larval stage was not recorded. The larvae that exited from veins pupated between 26 to 29.vi.2010 and the adults emerged between 20 to 22.vii.2010, i.e. the pupal stage lasted approximately 25 days (n=29). A mating pair was observed on the underside of a leaf around 7:00 am (site 13). No oviposition behavior was observed for this species.

Two species of Pteromalidae: Pteromalinae: sp. 01 from sites 5, 10, 15, parasitizing late instar larva, pupating inside the leaf vein (Fig. 50); Pteromalinae sp. 02 from sites 5, 10, 13, parasitizing larva and pupating inside the host puparium; one species of Braconidae: Opiinae: Opius sp. from site 10, parasitizing larva and pupating inside the host puparium.

Liriomyza mystica is most similar to Liriomyza prompta described below and to the Neotropical species Liriomyza commelinae (Frost) and Liriomyza robustae Spencer, especially in the form of the phallus with paired tubules. But these two latter species differ from Liriomyza mystica in a number of characters, including their host plants (both known from plants in the family Commelinaceae); mesonotum with a distinctive black and yellow pattern; surstylus with a distinct spine; third antennal segment enlarged in males; shape of both anterior and posterior spiracles and pupation occurring inside the mine (Silva and Oliveira 1952, Spencer 1984, Valladares 1984). The anterior and posterior spiracles of Liriomyza mystica are most similar to those of Liriomyza caesalpiniae Valladares reared from a Caesalpiniaceae, Caesalpinia gilliesii Benth. (Valladares 1984: figs 14, 15).

Most adult specimens of Liriomyza mystica were reared from site 5, but it was also found at other sites, up to an elevation of 2765 m (Table 1). Considering that Liriomyza mystica larvae feed inside primary veins and petiole of large, mature leaves, it made it difficult to establish Bocconia arborea as possible host due to the problems in studying large trees with large leaves. In sapling trees of ca. 1 m tall (n=2) at Santo Domingo de Heredia (site 9), no larvae or evidence of feeding was observed.

The species name is derived from the Latin mysticus (secret, mystic), referring to the hidden and inconspicuous leaf mines in primary vein and petiole.

Holotype ♂: COSTA RICA: Cartago: San Ramón de Tres Ríos, 1600 m (09°56'18"N, 083°58'38"W), ex. leaf mine Bocconia frutescens, larva exited 5–9.vii.2010, adult emerged 21–26.vii.2010, Kenji Nishida (LEM).

Paratype. same data as holotype (2 ♀: LEM); same except: ex. leaf mine Bocconia frutescens, emerged 1.vi.2010, Kenji Nishida (1 ♀: LEM); same except larva exited 29.vi.2010, adult emerged 20.vii.2010 (1 ♀: LEM); same except larva exited 29.vi–4.vii.2010, adult emerged 15–24.vii.2010 (7 ♂; 1♀: LEM); same except larva exited 25–28.vi.2010, adult emerged 20–22.vii.2010 (1 ♀: NMNH); same except larva exited 29.vi–4.vii.2010, adult emerged 21–25.vii.2010 (5 ♂; 1 ♀: NMNH); Cartago: Cervantes, 1500 m (09°52'46"N, 083°49'07"W), ex. Bocconia frutescens, emerged 11.vi.2010, Kenji Nishida (1 ♂: CNC); San José Province: San Isidro de Coronado, Centro 1420 m (09°58'18"N, 084°00'22"W), ex. Bocconia frutescens, adult emerged 11.vi.2010, Kenji Nishida (1 ♀: MZUCR); same except puparia formed 10–12.v.2010, adult emerged 12.vi.2010, Kenji Nishida (1 ♀: MZUCR); same except emerged 10.vi.2010 (1 ♂: INBio); same except pupation 6.vi.2010, emerged 26.vi.2010, Kenji Nishida (1 ♀: INBio); same except collected 15.v.2009, emerged 8.vi.2009, K. Nishida (4 ♂; 6 ♀: MZUCR), same except (1 ♂; 3 ♀: CNC); Puntarenas Prov., Monteverde, Estación Biológica, 1538 m (10°19'09"N, 084°48'32"W), mating on Bocconia frutescens leaf, 30.v.2009, Kenji Nishida (1♂; 1♀: INBio).

This species can be distinguished from other Neotropical species of Liriomyza by its completely yellow head and anepisternum, mesonotum almost completely brown to margin of scutellum, usually 3 + 1 dc, legs completely yellow, calypter brown on apical half with margin and fringe brown, and by the shape of the male genitalia and the shape of the anterior and posterior larval spiracles.

As in Liriomyza mystica Boucher and Nishida (described above) except as follows: arista shorter (Figs 28, 29), 0.23–0.3 mm; normally 3+1 dc (except one male specimen with 2+1 dc); generally smaller: wing length 1.4–1.7 mm in males and 1.6–2.2 mm in females; eye often with a more extensive bluish reflection (Fig. 29).

Male genitalia. Similar to Liriomyza mystica, but tubules of distiphallus more sclerotized, slightly wider, parallel sided (not diverging in ventral view). Mesophallus widest apical section (Fig. 33a), more than twice as large as narrower basal tubular section (Fig. 33b). Mesophallus in lateral view with a prominent curve near midpoint. Surstylus absent. Ejaculatory apodeme (Figs 34, 35) slightly more sclerotized than in Liriomyza mystica, and with blade slightly less expanded.

Early stages. Larval length: 1.95–2.70 mm, slightly smaller than Liriomyza mystica larva. White to creamy white with an internal orange spot at head (Fig. 47). Anterior spiracles about 0.1 mm distance from each other. Similar to Liriomyza mystica: fan-shaped with 5 small openings. Posterior spiracles with apparently 3 bulbs, but two very small and 1 much longer, curving toward anal segment (Figs 36, 37). Cephalopharyngeal skeleton (Fig. 39) more elongated with side arms narrower than in Liriomyza mystica. Puparium translucent brown to dark brown (Fig. 49).

Bocconia frutescens L. and Bocconia arborea S. Watson (Papaveraceae)

At site 15 in late October 2012 a few males were perched on small young leaves at the apical shoot of a 2 m host tree in the morning between 6:00 and 7:00 am. The males were either sitting or flying and perching on apical young leaves. A mating couple was observed on the underside of host leaf at 7:00 am (site 13, Fig. 6). Oviposition was observed a few times by two females at site 2. At 3:00 pm (17.vi.2011), overcast with slight drizzle, the two females were walking on the upper side of leaves of a Bocconia frutescens sapling, ca. 50 cm tall. The females either oviposited in the leaf tissue at the edge of the leaf margin (n=2) or along narrow leaf veins on the upper side of the leaf blade (n=3) (Fig. 7). It was difficult to locate the eggs because of their small size and translucent colour (n=1). After oviposition, the leaf was collected and a small larva was found on 28.vi.2011. Also at site 5, three leaves were randomly collected on 24.vi.2010, and newly started mines were observed on 2–3.vii.2010 (n=6 mines). These observations suggest that the duration of the egg stage is approximately 10 days. The larvae mine mesophyll of the leaf blade either singly or gregariously (up to 7 larvae) (n=70 mines) (Figs 40–45). Mines were found on very small mature leaves of ca. 1.5 cm (n=5, at site 13) to 90 cm long (n=10, at site 10). The mature mines with a single larva were narrow and more or less linear (Figs 20, 40, 41), and mines with multiple larvae were blotch-shaped (Figs 40, 42). These mines were conspicuous on the upper side of leaves. With regard to location of mines on leaves, no patterns were noticed; mines appear to occur on any part of the leaf blade — some mines were found near the primary vein, leaf apex, or anywhere in between (uncounted). Presence of the larva in the mine can be recognized by the orange spot of the larval anterior end (Figs 43–45, 47). The early part of the mine becomes brown and the frass is seen in dark green to black linear dots (Figs 42–44). The mature larvae each made an elliptical exit hole 1.1–1.3 mm wide on the upper side of the leaves, (n=7 holes) (Fig. 46). Under rearing conditions, from the time of recognizing very early mines, the larvae completed mining within 3 to 4 days, and on the 5th day the puparia were usually formed on the plastic surface of rearing bags or containers. Young pupae were observed within a day or two after forming of the puparium (n=16). The puparia (Fig. 48) were less translucent than those of Liriomyza mystica. The pupa becomes dark 2–3 days prior to adult emergence. A cohort of larvae from site 5 which pupated on 20–23.xii.2008 emerged as adults on 16–20.i.2009 (n=6, 1 ♂, 5 ♀), other data as follows: pupation 28.vi.2011, adult emergence 24.vii.2011 (a cohort, n=4, 2 ♂, 2 ♀, site 2); pupation 12–13.ii.2012, emergence 8.ii.2012 (n=2, site 12); pupation 20.iv.2012, emergence 14.v.2012 (n=2, site 15); i.e. the pupal stage lasted nearly a month under the rearing conditions. A small population of Bocconia arborea was found at site 9; however, it was only possible to access two saplings, which were about 1 meter tall and infested by Liriomyza prompta larvae (mostly old mines). Also at sites 7, 9 and 15, populations of weedy Papaveraceae, Argemone mexicana L. were found in close proximity to Bocconia frutescens and Bocconia arborea; however, no leaf mines of Liriomyza prompta were observed.

Three species of Eulophidae: Entedoninae: sp. 01 and sp. 02 from sites 5, 6, 13, and 15, Entedoninae sp. 03 from site 15, all these were parasitizing larva and pupating inside the host puparium; Eulophinae: sp. 01 from sites 13 and 15, parasitizing late instar larva, pupating inside the host leaf mine; and two species of Pteromalinae as mentioned in Liriomyza mystica, sp. 01 from site 5 and 15, and sp. 02 from site 13 and 15. A species of Braconidae, probably an Opiinae: Opius sp. was observed parasitizing a mature Liriomyza prompta larva at site 13 (Fig. 51).

An important character differentiating larvae of this species from Liriomyza mystica described above, is that the posterior spiracles each have an elongated, somewhat hook-like bulb. These posterior spiracles are similar to those found in the Neotropical species Liriomyza commelinae and Liriomyza robustae, but the uniformly coloured mesothorax, absence of surstyli, and unforked anterior spiracles, differentiate this new species from these other Neotropical species. This species appears to be more common than Liriomyza mystica with a wider elevation range (Table 1).

The species name is derived from the Latin promptus (visible, apparent), referring to their conspicuous and common leaf mines.