Hydrusa kefersteinii Butler, 1882: 2, by original designation

Antennae in both sexes piliform, each antennomere with pair of prominent, arched setae and additional pair of shorter, straight ones (Fig. 1). In the male genitalia, uncus basolaterally provided with differently expressed, setaceous protrusion; valva biramous, with costa transformed into variable, claw-like or spatula-like protrusion, and sacculus developed into weakly sclerotised lobe.

The most similar to Tritonaclia are other “black-and-white” Madagascan Syntomini, especially from the genera Melanonaclia (in the tree represented by M. nigra Griveaud, 1964 and M. toulgoeti Griveaud, 1964), Vadonaclia Griveaud, 1964 (absent in the tree),

Juliennaclia moerens (congeneric with J. pauliani (Griveaud, 1964), in the tree), as well as Mortinaclia Wiorek, 2026 (monotypic, in the tree) and Tenuinaclia (in the tree represented by T. cf. andapa Griveaud, 1964) (Fig. 2). The genus Melanonaclia (as reviewed by Wiorek et al. 2026) differs in having only intensely yellow, and never white marks on the head, thorax and abdomen, piliform antennae without additional prominent setae, simple male valva without the costal process, and the presence of a pouch-like appendix or diverticulum in the basal portion of the female corpus bursae. Some Melanonaclia species do have a row of cornuti in the male vesica, although differently expressed than in Tritonaclia. The monotypic genus Vadonaclia (V. marginepuncta) differs in intensely orange to reddish head, thorax and abdomen blotches, distinctly bipectinate antennae, and an additional row of small blotches along the fore wing outer margin. The monotypic genus Mortinaclia (M. perplexa), although provided with a row of cornuti in the male vesica, differs from Tritonaclia in much smaller body size, piliform antennae without distinct setae, a differently expressed abdomen pattern extending on its dorsal side, and the presence of distinct lateral arms on the tegumen. The genus Tenuinaclia differs in much smaller body size, a differently expressed pattern of the abdomen: either completely absent or consisting of lateral and ventral rows of at most three small blotches, and only a single distinct cornutus in the male vesica (Griveaud 1964).

The morphological findings are congruent with the phylogenetic tree (Fig. 2). The genus Tritonaclia, represented in our dataset by four out of six species, for which fresh materials were available for molecular study, forms a separate monophyletic clade. It is not closely related to other groups comprising “black-and-white” species, i.e. Melanonaclia, Tenuinaclia, Juliennaclia Wiorek, 2026 and Mortinaclia. The genus forms three main lineages grouped in two clades, well-supported by both molecular and morphological results. In the tree, the first clade comprises T. kefersteinii and T. quinquepunctata Griveaud, 1967 (represented by the former) and it is sister to all remaining species with a high support. The lineage is characterised by: i) the prominent body coverage with elongate, hairy scales, giving the species “fluffy” appearance; ii) pale yellow to creamy white marks on the abdomen with no blotches on sternites, iii) six or seven pale yellow to creamy white, never semi-transparent forewing wing blotches and one basal yellow blotch at the forewing hind margin (Fig. 3), iv) male genitalia vesica armed with a bunch (not a row) of elongate cornuti (Fig. 4), and v) female genitalia corpus bursae signum developed into a longitudinal plaque (Fig. 5), although the female of T. quinquepunctata remains unknown. The second clade comprises two remaining lineages: T. tollini (sensu Keferstein 1870 nec Griveaud 1964) + T. vonifotsy Wiorek, sp. nov. (i.e. T. tollini, sensu Griveaud 1964), and T. stephania (Oberthür, 1923) + T. ombilahy Wiorek & Przybyłowicz, sp. nov., in the tree represented by all species but the latter, and sister to each other with good support (Fig. 2). They are characterised by: i) white marks on the abdomen, both on tergites and sternites, ii) four or five always semi-transparent pure white forewing wing blotches, with basal blotch always absent, iii) forewing blotch m₅ distinctly smaller than m₆ (Figs 6, 7, 8); iv) male genitalia vesica armed with a differently expressed row of prominent, tooth-like cornuti (Figs 9, 10, 11, 12). Additionally, the lineage T. tollini + T. vonifotsy Wiorek, sp. nov. is characterised by patches of scent scales on the male fore- and hindwing, and spirally coiled ductus bursae and double signum in the female genitalia (Figs 13, 14). The lineage T. stephania + T. ombilahy Wiorek & Przybyłowicz, sp. nov. has no wing scent scales, wide and short ductus bursae, and a single signum (Fig. 15) – but the female genitalia of T. stephania remain undescribed.

The genus Tritonaclia (monotypic at the moment of establishment, with T. kefersteinii) was characterised by Hampson (1898: 125) based on the shape of palpi, setation of antennae, and wings venation scheme. Griveaud (1964: 74) sustained this set of characters, indicating additional ones in the male genitalia. According to Griveaud (1964), the forewing veins M₂ and M₃ should be basally stalked in Tritonaclia and completely separate in Melanonaclia (Griveaud 1964: figs 154, 172), which turned out to be a variable character (see Wiorek et al. 2026: figs 25E–H, 34E, F).

The species Tritonaclia inauramacula Griveaud, 1964, was described only from females and placed in the genus based on the alleged forewing venation scheme mentioned above. However, they appeared to be conspecific with the female of Melanonaclia luctuosa (Oberthür, 1911) collected in copula and described in Griveaud (1969) and therefore synonymised in Wiorek et al. (2026). Thus, in the arrangement proposed here, the genus Tritonaclia consists of six species.

Despite the morphological variation and clear division into three lineages, the genus Tritonaclia as a whole is well-defined by the proposed synapomorphies, especially the antennae setation and the shape of the male valva. Thus, to maintain a possibly stable systematics of the group, we refrain from establishing subgenera, especially within a group of six species.

31 specimens, including holotype (MNHN - 9 ♂♂, 1 ♀; NHMUK - 5 ♂♂, 2 ♀♀; PZBT - 12 ♂♂, 2 exx. sex unknown) (Table S1).

♂, NHMUK, NHMUK010621038, by monotypy (in the original description as ♀, in error, see Przybyłowicz and Wiorek 2023).

The most similar species is T. quinquepunctata, the diagnostic characters are provided in the determination keys. For more details, see Przybyłowicz and Wiorek (2023).

The female of T. kefersteinii was first described, and the male genitalia were redescribed by Przybyłowicz and Wiorek (2023).

A single female specimen (NHMUK015109820) collected by DL in 2018 has five forewing blotches (the second medial absent), and it lacks the hindwing distal blotches (Fig. 3C). The specimen, although relatively fresh and unspread before this study, appears faded, with head, thorax and basoproximal portion of forewing dark brown, almost black, and the remaining body parts paler, ochraceous brown. It was unavailable for molecular study, but we interpret it as a female of T. kefersteinii with a reduced wing pattern, since the general shape and arrangement of the remaining forewing blotches, especially distal ones, are typical for that species (see the determination key), rather than T. quinquepunctata (see Fig. 3A, B vs. Fig. 3D). Comparing to the first description of the female genitalia (Przybyłowicz and Wiorek 2023), the specimen is almost identical and differs only in shorter, almost absent ductus bursae, and shorter and wider dorsal pheromone glands (Fig. 5). Still, the taxonomic status of this morphotype requires further study.

(Fig. 16A) Disjunct, in the northern and central east of Madagascar, the Regions of: SAVA (PN de Marojejy), Analamanga (La Mandraka), Alaotra-Mangoro (area of Lakato), and Matsiatra Ambony (Ankafina-Tsarafidy). At around 600–1350 m elevation (possibly up to around 1450 m, since the exact locality in the area of Ankafina-Tsarafidy is unknown). See Przybyłowicz and Wiorek (2023) for further comments on the species distribution.

Attracted to light (UV) at night, there is no explicit data on diurnal activity. Recorded in January, February, August and September. Associated with humid, evergreen lowland and midland forests.

10 specimens, including holotype (MNHN - 7 ♂♂; NHMUK - 2 ♂♂; PZBT - 1 ♂) (Table S1).

♂, MNHN, EL65138, by original designation.

The most similar species is T. kefersteinii. The diagnostic characters are provided in the determination keys. Tritonaclia quinquepunctata differs in the first distal blotch (m₄) located more apically, parallel to the costal margin, and separated from it by only a very narrow ochraceous brown stripe (Fig. 3D). Female unknown.

(Fig. 16B) Endemic to the Tsaratanana massif in northern Madagascar (DIANA Region). Up until now known only from the type locality “below Andohanisambirano”. Here, the species is recorded from two further localities in the southern part of Tsaratanana: Andohananalila and Andohanambatoafo (see discussion). The information about the occurrence of T. quinquepunctata in the Marojejy massif provided in Przybyłowicz and Wiorek (2023) resulted from a mistake and is corrected here. At 1850–2050 m; the elevation of the type locality, in the original labels and description (Griveaud 1967) given at 1900 m, should be corrected to 2050 m (Viette 1967, 1991). The latter is equivalent to a campsite by a seasonally dry lake (see below) at –14.1524, 48.958, where DL camped.

Attracted to light (UV) at night, there is no explicit data on diurnal activity. Recorded in March and December. Associated mostly with humid, midland and montane evergreen forests. The type series was collected in a “matsabory”, a marsh-like open area, periodically drying out and turning into a pond in the rainy season (Viette 1967; Diakonoff 1970), with bamboo thickets nearby (Viette 1967). Such bamboo-dominated habitats are found in Tsaratanana at 2000–2200 m (Goodman et al. 2018), and the species may potentially be associated with this kind of vegetation. It might also dwell in sclerophyllous forest vegetation, occurring in Madagascar above 1300 m on the slopes open to winds (Wells 2003), and in Tsaratanana reaching to 2500 m (2000 m in the other massifs of Madagascar) (Crowley 2004). The occurrence of T. quinquepunctata in ericoid thickets is less probable, as this type of vegetation, although usually commencing around 1800–2000 m elevation (Du Puy and Moat 1996; Crowley 2004), in Tsaratanana occurs above the known elevational range of the species and is currently highly degraded (Goodman et al. 2018).

24 specimens, including lectotype and two paralectotypes of G. tollini, and lectotype of N. melania (ISEA PAS - 3 ♂♂, 2 ♀♀; MLUH-ZNS - 3 ♂♂; MNHN - 5 ♂♂; NHMUK - 6 ♂♂, 1 ♀; PZBT - 4 ♂♂) (Table S1).

Lectotype (of tollini Keferstein, 1870): 1 ♂; [no label], collection data based on the original description: Madagascar, “Tamatave” [Toamasina]; 02 Aug. 1862; C. Tollin leg.; MLUH-ZNS. — Paralectotypes (of tollini Keferstein, 1870): 2 ♂♂; [no label], collection data based on the original description same as for lectotype; MLUH-ZNS, designated here. —

♂; Madagascar, Karianga [doubtful locality]; “sud de Madagascar, 04.1920, Lamberton” NHMUK, NHMUK010620989, designated here.

Diagnostic characters are provided in the determination keys, and in the diagnosis of T. vonifotsy Wiorek, sp. nov. (see below). Males of T. tollini are the most similar to T. vonifotsy Wiorek, sp. nov., and to Melanonaclia toulgoeti – for the details, see diagnosis of the former (below). The females of T. tollini and T. vonifotsy Wiorek, sp. nov. (Fig. 7F–I) appear indistinguishable morphologically, even in the genitalia (but see taxonomic note for the latter), and can currently be told apart undoubtedly only using molecular methods (Fig. 2). Griveaud (1964) described the female of T. melaniasyn. nov. and indicated the female of “T. tollini” (i.e. T. vonifotsy Wiorek, sp. nov.) to be unknown.

(Fig. 9) Tegumen well sclerotised, laterally very narrow, medial portion subtrapezoidally widened towards uncus, anterior margin rounded, not incised; uncus short, multidimensional, at base laterally on each side rounded protrusion densely covered with erect setae, medial portion laterally on both sides with distinct rounded incision, terminal portion bifid, with two distinct, lobate protrusions flattened laterally, opened at around 120 degrees, resembling whale tail, covered with creases and serrate along outer margins, entire lateral and dorsal surface of uncus densely covered with elongate, soft setae; anal cone well developed, membranous , without defined scaphium; vinculum very narrow, well sclerotized, incompletely fused with tegumen; saccus subtriangular, dully terminated, straight; juxta in form of well sclerotised ring, on dorsal side submedially with indistinct, subtriangular protrusion or without any additional structures, ventrolaterally on each side with differently expressed, short, lobe-like protrusion, well sclerotised to entirely membranous; valva elongate, of complex shape, reaching beyond uncus tip, costal margin in basal portion slightly incised, straight, distinctly folded inwards, terminal portion transformed into elongate and narrow, claw-like curved and pointed protrusion; sacculus moderately sclerotised to almost membranous, well sclerotised, basal ⅓ straight, then arched, terminal portion in form of elongate, roundly terminated lobe, not reaching to costal tip, almost membranous, heavier sclerotised only along margin, valva outer surface densely scobinate and covered with easy removable scales; phallus weakly to moderately sclerotised, tubular, bulbous in basal portion, terminal portion narrowed and flat cut; vesica left uneverted to avoid destruction, membranous, with well visible structures, in basal portion with two irregular, heavily sclerotised plaques, and elongate, sinuous and coiled row of around 40 heavily sclerotised, densely arranged tooth-like to spine-like cornuti, touching each other laterally.

(Fig. 16C) Northeast Madagascar, mostly around Antongil Bay, in the Regions of: SAVA (Binara, and at Ambavony and Antafononana rivers), Analanjirofo (Lake Amparahibe, south to Antanambe and Anove), and Atsianana (locus typicus “Tamatave” = Toamasina). At around 0–850 m elevation. The lectotype of T. melaniasyn. nov., labelled as collected in Karianga (SE Madagascar, Region Atsimo-Atsinanana), originates from Charles Lamberton’s materials (“Sud de Madagascar, Reçu de M. Lamberton en avril 1922”), and as such should be treated with caution, considering the general species’ range and the other known cases of Lamberton’s unreliable specimens from “Southern Madagascar” (see Wiorek et al. 2021, 2026). Nevertheless, this locality requires confirmation in light of a similar, disjunctive distribution of T. vonifotsy Wiorek, sp. nov. (see below).

Active by day (specimens provided with detailed label data were caught between 11:00 AM and 5:00 PM, Table S1), there is no data on nocturnal activity or light attraction. Specimens undoubtedly representing the species (i.e. males and DNA barcoded females) were recorded from November to March, in July and August. For information about the phenology of indistinguishable female specimens of T. tollini / T. vonifotsy Wiorek, sp. nov., unavailable for further identification, see the latter species below. Moths are associated with humid, evergreen lowland forest, including riparian habitats. The single specimen from Binara was collected in a gallery forest, in an area generally covered with deciduous, seasonally dry, western forest.

• Hampson (1898) apparently did not examine the original specimens of Keferstein (1870), and most probably did not know where they were deposited. The description in the Catalogue (Hampson 1898) suggests specimens of the species later described as Melanonaclia toulgoeti to be mistaken there for T. tollini, additionally with Melanonaclia lugens (Oberthür, 1893) treated as its synonym (therein as “Thyrosticta tollini ab. 1 lugens”, see also Wiorek et al. 2026). Griveaud (1964) certainly did not know the locality of T. tollini type specimens, as indicated by a question mark (Griveaud 1964: 76). Despite that, he “utilised” the name tollini for the specimens with yellow hindwing medial blotch, even though the said blotches are directly described as “gray” in the original description (Keferstein 1870) – perhaps because the specimens with gray ones had been known to P. Griveaud as Tritonaclia melaniasyn. nov., described by Oberthür (1923). The Lepidoptera collection of Georg A. Keferstein is currently deposited in the Zentralmagazin Naturwissenschaftlicher Sammlungen at Martin-Luther-Universität Halle-Wittenberg in Halle (Saale), Germany (Schneider and Steinheimer 2018). The original series of T. tollini was rediscovered there by MW in 2022, and here the lectotype and paralectotypes are designated (Fig. 6A–C, F–H). The specimens perfectly match the original description and illustration in Keferstein (1870) (Fig. 6E). They have no labels, but the collection data can be extracted from the first part of Keferstein’s (1870) paper, a logbook of C. Tollin’s expedition to Madagascar. According to that, Tollin arrived at Toamasina [at the time Tamatave] in May 1862, and in the next months conducted excursions in the area. There is no information about him moving to more distant regions of Madagascar, and indeed Viette (1963) indicates that Tollin collected in Foulepointe and Vohidotra villages, both in Toamasina Province. Thus, the specimens were collected there on August 2nd, 1862 [“on August 2nd a new species of the genus Glaucopis”]. All three specimens are males and have well-visible, extending genitalia (Fig. 9F, G), identical to T. melania (Griveaud 1964: figs 174, 175). Thus, we treat Tritonaclia melaniasyn. nov. as a junior subjective synonym of T. tollini. Consequently, the specimens with a yellow hindwing blotch, interpreted by Griveaud (1964) as “T. tollini”, belong to a separate species Tritonaclia ombilahy Wiorek & Przybyłowicz, sp. nov., described below. • Griveaud (1964) indicated the type series of T. melaniasyn. nov. to consist of the male holotype and female allotype. However, the original description by Oberthür (1923) is based on one male and three females, with no direct designation of holotype. The single male specimen in question, housed in NHMUK (NHMUK010620989; Oberthür 1923: pl. DLXVI, fig. 4880), has only a “type” label. Thus, here we designate that specimen as the lectotype of T. melaniasyn. nov.. In the collection of NHMUK, we found only one female (NHMUK013325655), which might have belonged to the original series of one male and three females, and the fate of the remaining two specimens is unknown; they were not found in the NHMUK supplementary collection. Additionally, in the same collection, we found a single, old and worn male specimen (NHMUK016251182), of “Thyrosticta” tollini (i.e. following the generic classification by Hampson 1898), correctly identified as such in Paul Mabille’s handwriting (Horn and Kahle 1935–1937). The specimen is provided with two additional labels indicating it was subsequently obtained by Charles Oberthür in 1923, perhaps after Mabille had passed away in April that year, and eventually deposited in NHMUK in 1927. Therefore, we suppose that Oberthür was unaware of Keferstein’s (1870) description of T. tollini, and he received the specimen after he described T. melaniasyn. nov. • The male specimen DL_2983 from Binara, in the tree forms a branch sister to all other conspecific individuals (Fig. 2). Although that northernmost locality is relatively remote from the species’ main range (Fig. 16C) and differs in the type of vegetation, we did not find any differences in the genitalia of the specimen. The taxonomic identity of the Binara population requires further studies.

Holotype: ♂; Madagascar, Makira Ankirindro; 15.2904 S, 49.5474 E; 636 m a.s.l.; 16 Jan. 2003; D.C. Lees leg.; by cuisine; ISEA PAS, DL_1864, genitalia slide S553, hindwing venation slide S554; BOLD Process ID SYNMD008-26, GenBank COI sequence accession number: PX777880. — Paratypes: • 1 ♂; Madagascar Est, Ambodivoanio; 16.1816 S, 49.6631 E; 50 m a.s.l.; Aug. 1965; A. Peyrieras leg.; “Tritonaclia tollini Keferstein ♂ P. Griveaud det. 1966”; MNHN, MNHN, EL74622, préparation ♂ no. 155; • 1 ♂; [collection data unknown]; “acheté à M. Goudot, 1839”, “Tritonaclia tollini Keferstein ♂ P. Griveaud det.”; MNHN, EL74623, slide P. Griveaud No. 286; • 3 ♂♂; “Madagascar Est, s/préf. Vondrozo, forêt Madiorano, 30 km ouest Vondrozo” [30 km W of Vondrozo, Madiorano Forest]; 22.8065 S, 47.043 E, 750 m a.s.l.; Apr. 1973; A. Peyrieras, A. Rakotoarisolo leg.; MNHN, EL74620, EL74621, EL74624–préparation ♂ no. 156; • 1 ♂; same locality and collector as for holotype; 11 Jan. 2003; NHMUK, NHMUK010918661 / DL1816 / DL-SE11, BMNH(E) 2019-70; BOLD Process ID SYNMD003-26, GenBank COI sequence accession number: MK158587 • 1 ♀; Makira, Ankirindro, ridge, SA6675; 15.2931 S, 49.5472 E, 675 m a.s.l.; 10 Mar. 2003; D.C. Lees leg.; collecting time 2:40 pm; NHMUK, NHMUK010918662 / DL1863, BMNH(E) 2019-70, genitalia slide NHMUK010316589, BOLD Process ID SYNMD005-26, GenBank COI sequence accession number: MK158588. — The type series comprises all known specimens undoubtedly belonging to the species (Table S1).

17 indistinguishable females of T. tollini / T. vonifotsy Wiorek, sp. nov., unavailable for further determination with DNA barcoding (MNHN - 7 ♀♀; NHMUK - 3 ♀♀; PZBT - 7 ♀♀) + 1 ♀ record from iNaturalist (Table S1).

Madagascar, Makira Ankirindro.

The diagnostic characters are provided in the determination keys. The most similar species, Tritonaclia tollini, differs in the colouration of the male hindwing medial blotch (Fig. 17). In males of both species, the outer margin of the forewing is slightly incised, and in spread specimens it tends to crease between the veins CuA₂ and 1A+2A (Figs 6A–J, 7A–E, 18A), deepening the incised appearance, and making it similar to Melanonaclia toulgoeti (Wiorek et al. 2026). However, the latter species lacks the hindwing colour medial blotch and terminal bifurcation of the vein 1A+2A, both characteristic for T. tollini and T. vonifotsy Wiorek, sp. nov. (Fig. 18A, B). Females of T. tollini and T. vonifotsy Wiorek, sp. nov. appear morphologically indistinguishable (Fig. 7F–I), but see the taxonomic note below.

(Fig. 7A–E, 18D–F) Head. Proboscis well developed, ochraceous yellow; head entirely brown to dark brown, almost black; ventroterminal portion of postocular area provided with tuft of protruding scales; labial palpus 3 segmented, porrect, around ½ protruding beyond frons, 1^st palpomere with elongate scales directed downwards, 2^nd palpomere 2× longer than 1^st one, 3^rd palpomere of the length of 1^st one; antennae entirely brown, piliform, ventrally densely covered with short erect setae, additionally each antennomere on both margins with pair of distinct setae, one longer and arched, one shorter and straight; Thorax. Patagia inner side brown to almost black, medially with elongate scales, lateral side with pure with blotch of regular scales; tegulae entirely brown to almost black, terminal portion with tuft of long, piliform scales; thorax entirely brown to almost black; Legs. Entirely brown to almost black, in some specimens with admixture of single, creamy to white scales on frontal surface of hindleg coxa, and sometimes also on midleg coxa; arolium present; Foreleg. Tibia of ⅔ of femur length, epiphysis present, of ⅔ of tibia length, ochraceous yellow; Midleg. Tibia with one pair of brown terminal spurs; Hindleg. One pair of brown terminal spurs, and one pair of additional spurs in ¾ of tibia length, shorter than terminal ones; Abdomen. Entirely brown to almost black, tergites 4^th–6^th lateroterminally with well developed, pure white blotch, in some specimens admixture of white scales also on 7^th tergite; sternites 2^nd–6^th medioterminally with differently expressed, well developed, pure white blotch, sometimes with admixture of single creamy scales, sternite 7^th medially with only small white dot, sternite 8^th without markings, with deep U-shaped incision, reaching to around ½ of 8^th tergite length; lateral portion of abdomen unavailable; valvae distinctly protruding from abdomen, outer surfaces densely covered with scales concolorous with abdomen; Forewing. Subtriangular, hind margin rounded, without well-defined tornus; backround brown to almost black, with 5 or 4 pure white, semi-transparent blotches of different size: one or two in medial portion and three in distal portion; first medial blotch (m₂) subsquare to subtrapezoidal, in medioterminal portion of DC, between from R and DC hind margin (touching both); second medial blotch (m₃) medially between base of CuA₂, and 1A+2A, in form of subtriangular to rounded dot, very small and indistinct, up to completely absent; first distal blotch (m₄) subsquare to suboval; second distal blotch (m₅) relatively small, rounded to irregularly suboval; third distal blotch (m₆) the largest one, always much larger than m₅, elongate suboval to subtrapezoidal; wing underside as upperside, but basal and medial portion covered with pale area of dull-white regular scales, along tornus and hind margin reduced in size; medially above 1A+2A elongate, suboval field of pale yellow-creamy scales, terminated before CuA₂; 1A+2A in terminal portion distinctly bifurcated; Hindwing. Suboval to irregularly subtrapezoidal, reaching to ½ of forewing length, costal margin in medial portion arched and widened, outer margin slightly sinuous, medially with indistinct incision; background along outer and hind margins brown to almost black, medial part covered with dull-white, in DC with additional, elongate, irregular blotch of no well-defined margins, covered with regular, elongate, pale yellow-creamy scales; basal blotch in basodistal portion of wing, along DC hind margin and CuA₂, separated from basal part of terminal margin by narrow stripe of brown scales, heart-shaped, slightly semi-transparent, covered with pure white scales of slightly reduced size; first distal blotch beyond outer margin of DC, most often hardly visible on upperside; second distal blotch suboval, of moderate size, medially between CuA₁ and M₂+M₃, covered with slightly reduced, pure white scales; underside with entirely brown to almost black background, basal blotch as on upperside, covered with mixture of regular and slightly reduced pure white scales, first distal blotch well defined, suboval, covered with regular and reduced pure white scales, beyond DC outer margin and below Sc+Rs+M₁ (touching); second distal blotch as on upperside, with mixture of regular and reduced pure white scales; additionally, in some specimens above terminal portion of M₂+M₃ additional, indistinct dot of a few white scales; M₂ and M₃ fused in around ½ of their length, DC outer margin convex, Sc+Rs+M₁ completely fused, but in some specimens slightly bifurcated in the terminal portion; 3A present, in some specimens in basal portion with vestigial, lateral arm.

(Fig. 10) Tegumen well sclerotised, moderately wide, in medial portion subtriangular projected towards uncus, both margins with narrow, rib-like sclerotisation all along; uncus multidimensional, of variable length, from elongate, with subtriangular basal portion, reaching to valva tip, to short and reaching to ⅔ of valva length, basally always provided with subconical, curved lateral protrusion densely covered with short erect setae, terminal portion Y-shaped bifid, with two distinct, lobate protrusions flattened laterally and opened at around 90 degrees, in terminal portion creased, with serrate outer margin; terminal half of uncus laterally covered with elongate, erect setae; anal cone well developed, subconical, originating at basal portion on uncus, with very indistinct scaphium formed by moderate sclerotisation of dorsal wall, sometimes absent; vinculum narrow, moderately to well sclerotized, incompletely fused with tegumen; saccus subconical, dully terminated, projected slightly anteriorly; juxta well sclerotised, in form of ring, lateroventrally provided with differently expressed lobes, entirely or partially sclerotised, dorsoposteriorly provided with two straight, spine-like pointed protrusions of moderate length, of equal or different lengths, ventral portion of juxta distinctly bent posteriorly and touching basal portion of valvae; valva elongate, of complex shape, reaching at least uncus tip, costal margin straight, distinctly folded inwards, terminal portion transformed into spatula-like, slightly curved and roundly to subtriangularly terminated protrusion; sacculus well sclerotised, basal ⅓ arched, terminated with straight, slightly narrowing and pointed protrusion directed inwards and reaching to around ½ of valva length and width, terminal portion lobate, almost membranous, heavier sclerotised only along margin, roundly terminated and almost reaching costal tip; valva outer surface densely scobinate and covered with easy removable scales; phallus well sclerotised, moderately short and broad, slightly irregularly curved, terminal portion narrowed, cylindrical and flat cut; vesica left uneverted to avoid destruction, but with well visible structures, in basal portion with two irregular, heavily sclerotised lobes, and elongate, sinuous row of around 25–30 heavily sclerotised, densely arranged tooth-like to spine-like cornuti, touching each other laterally;

(Fig. 7F–I) Head. (including eye size), thorax, legs, and abdomen colouration and pattern as in the male, but see the comment on the foreleg coxa in the taxonomic note below. Forewing. Subtriangular, slightly more elongate and with more prominent tornus than in male; background brown to almost black, with five pure white, semi-transparent blotches of different size, arranged as in male, second medial blotch (m₃) differently expressed, round to suboval but always present; third distal (m₆) slightly larger than in male; underside as upperside, with no additional colour blotches; 1A+2A in terminal portion bifurcated as in male; Hindwing. Shape and size similar to male, uniformly brown to almost black, with three or four differently expressed, pure white, semi-transparent blotches, and no additional marks or blotches of different colour; basal blotch as in male; first distal blotch beyond outer margin of DC, variable in shape and size, from small round dot to elongate, S-shaped; second distal blotch in form of small dot between bifurcating M₂+M₃, in some specimens absent; third distal blotch as in male, between CuA₁ and M₂+M₃; underside as upperside, but basal and first distal blotches slightly enlarged; venation as in male, in a single female vein 3A asymmetrically developed, bifurcated in left hindwing.

(Figs 13, 14) Anal papilla subtrapezoidal, densely scobinate and covered with elongate, erect, soft setae of different lengths; pseudopapilla narrow, membranous, reaching to valva terminal margin, densely scobinate; apophysis posterioris slightly longer than anal papilla, well sclerotized, needle-like, pointed; apophysis anterioris slightly shorter, needle-like, originating from subtriangular lobe; dorsal pheromone glands in form of two narrow, membranous, densely entangled and furcate tubes; ventral pheromone glands absent; lamella postvaginalis well defined, subtrapezoidal, terminal margin densely scobinate, medioterminally with shallow, rounded incision, lateroterminally on each side with differently expressed, rounded protrusion covered with small grain-like sculpture and elongate, erect setae, absent in some cases; sternite 7^th medially with deep, U-shaped incision; ostium bursae distinct, wide, rounded to suboval, antrum short, cup shaped; tergite 7^th enlarged, of the length of sternites 7^th+8^th, tergite 8^th reduced to narrow plaque, with heavier sclerotised, rib-like basal margin, laterally fused with lateral arms of lamella postvaginalis and forming subtriangular base of apophysis anterioris; ductus bursae tubular, of uniform width, well sclerotised and distinctly convoluted rightwards into three coils; Cervix bursae of around ½–⅓ of corpus bursae length, of width of ductus bursae, slightly curved, entirely membranous with dense longitudinal plicae; corpus bursae suboval, membranous, with dense minute scobinations and moderately wide longitudinal plicae;. Signum bursae located medially, in form of two horizontal, 8-shaped, well sclerotised plaques arranged longitudinally in ⅓ and ½ of corpus bursae length, covered with short, subconical, spine-like protrusions; ductus seminalis in form of narrow, membranous tube originating from lateral diverticulum located at base of corpus bursae, of shape and plication similar to cervix bursae and parallel to it.

From Malagasy vonifotsy, meaning “pale yellow”, referring to the colouration of the male hindwing medial blotch.

(Fig. 16D) Mostly in the northeast of Madagascar, in the Analanjirofo Region (Ankirindro and Ambodivoanio). Three male specimens (Fig. 7D, E, 10C–E) from Madiorano Forest in the area of Vondrozo, Ihorombe Region, collected by A. Peyrieras and A. Rakotoarisolo in April 1974, although distinctly distant from the remaining localities, leave little doubt. At 50–750 m elevation. All remaining, unidentified females of T. tollini / T. vonifotsy Wiorek, sp. nov., unavailable for further identification, were collected or observed in the northeast part of the island, in the Regions of SAVA (at Antafononana River), Analanjirofo (from Ambodivoangy and Fampanambo, south to Antanambe), and Atsinanana (Forêt d’Analalava). At around 25 to max. 1000 m elevation (the highest possible point of Antampona). Associated with humid, evergreen lowland forest.

Specimens undoubtedly belonging to the species (males and one DNA barcoded female) were collected in January, March, April and August. Additionally, the unidentified, non DNA-barcoded females of T. tollini / T. vonifotsy Wiorek, sp. nov. were recorded in January, February, April, June, July and November.

Tritonaclia vonifotsy Wiorek, sp. nov. exhibits the above-described distinct differences in the male uncus length (Fig. 10D–F). The specimen from the remote, southernmost locality of Madiorano Forest has an uncus of intermediate length (Fig. 10E). This character requires further research, including molecular examination, which was limited in our study. The rounded, setaceous protrusions of the female lamella postvaginalis seem to be present in T. vonifotsy Wiorek, sp. nov. and absent in T. tollini (Figs 13A–C, 14B). However, we had only a single DNA barcode-confirmed female genitalia preparation available for each of the two species, whilst the character in question is potentially prone to destruction during dissection. Further, both species seem to stably differ in the very indistinct admixture of pale creamy to white scales in the proximomedial part of the foreleg coxa, allegedly present in both sexes of T. tollini and absent in T. vonifotsy Wiorek, sp. nov., including the three females DNA barcoded in our study (not shown). Although potentially the only morphological character distinguishing the females of both species, it is a very subtle trait requiring further research on a longer series of DNA barcoded specimens. Therefore, here we refrain from interpreting it as a diagnostic character.

43 specimens, including holotype (ISEA PAS - 1 ♂; MNHN - 21 ♂♂; NHMUK - 1 ♀; PZBT - 19 ♂♂, 1 ♀) (Table S1).

♀, NHMUK, NHMUK010620988, by monotypy (in Griveaud 1964 erroneously as male lectotype).

Superficially, the most similar species is Tritonaclia ombilahy Wiorek & Przybyłowicz, sp. nov., described below. The diagnostic characters are provided in the determination keys. — Male genitalia were redescribed in Wiorek et al. (2021).

(Fig. 16E) Longitudinal range along the central east of Madagascar, in the Regions of Alaotra-Mangoro (Vallée d’Ivelona, Ambatondrazaka, Forêt Ambohiboatavo), Analamanga (RS d’Ambohitantely, La Mandraka), Vakinankaratra (Ampolomita), Matsiatra Ambony (Ankafina-Tsarafidy), Ihorombe (Forêt Vakoany). At around 1000–1600 m elevation.

Attracted to light at night, there is no information about diurnal activity. Collected from September to March, in May and July. Associated with evergreen, humid midland forest.

Griveaud (1964) indicated the type specimen of T. stephania (NHMUK010620988), deposited in NHMUK (at the time BMNH), to be a male lectotype, but in 2021 ŁP identified it to be a female. Further, there is no information about the number or sex of the type series specimens in the original description by Oberthür (1923). For the five other Syntomini species described in Oberthür (1923), numbers of specimens larger than one are specifically indicated, and generally congruent with the materials currently housed in NHMUK (verified by MW, but see T. melaniasyn. nov. above). This, by analogy, implies that the abovementioned specimen should have been interpreted as a holotype by monotypy. The only superficially most similar specimen in the collection (NHMUK013702406) belongs to Tritonaclia ombilahy Wiorek & Przybyłowicz, sp. nov. described below. It can be excluded as the potential second specimen from the original series (alleged “paralectotype”), since in the original description Oberthür (1923) directly indicates the small size of T. stephania, whilst T. ombilahy Wiorek & Przybyłowicz, sp. nov. is distinctly larger. • The female “neallotype” of T. stephania (MNHN, EL65139, genitalia slide by P. Griveaud, No. 291), designated by Griveaud (1964), is distinctly larger than the male (Fig. 8A, G). In 2022, ŁP and MW found in MNHN three male specimens with a body size similar to that female, collected in 1970 and 1971, thus after Griveaud’s (1964) “Amatidae” monograph. These specimens turned out to represent a new species, Tritonaclia ombilahy Wiorek & Przybyłowicz, sp. nov., described below. Based on the distinct and discrete wingspan difference between both species, with almost no intraspecific variation, we assign the female “neallotype” of T. stephania to T. ombilahy Wiorek & Przybyłowicz, sp. nov. • During the visit to PZBT in 2023, ŁP found a single female specimen of T. stephania in poor condition (not shown). The specimen, as well as the female holotype in NHMUK, was unavailable for dissection, and the female genitalia of the species remain unknown.

Holotype: ♂; “Madagascar Est, N.O. de Manantenina, Chaines Anosyennes, massif nord” [NW of Manantenina, Anosyennes Mountains, northern massif]; 24.1134 S, 47.108 E, 1050 m. a.s.l.; 22–29 Nov. 1971; P. Griveaud leg.; “Museum Paris, Madagascar Centre mission C.N.R.S.R.C.P. no. 225”; MNHN, EL74649, préparation ♂ no. 157; Paratypes: • 1 ♂; same collection data as for holotype; MNHN, EL74647; • 1 ♂; Ambohimanakana, Manambato, Anova [=Anove]; 16.6055 S, 49.798 E, 10 m a.s.l.; 26 Oct.–4 Nov. 1970; P. Viette, P. Griveaud leg.; “Museum Paris, Madagascar Centre mission C.N.R.S.R.C.P. no. 225”; MNHN, EL74648; • 1 ♀; Madagascar, Fianarantsoa [doubtful locality]; [collection date and collector unknown]; “ex Lamberton 1922”; originally designated by P. Griveaud (1964) as “neallotype” of Tritonaclia stephania; MNHN, EL65139, slide P. Griveaud No. 291; • 1 ♀; Madagascar, Fianarantsoa [doubtful locality]; [collection date and collector unknown]; “ex Lamberton 1922”; “Ex Oberthür Coll., Brit. Mus. 1927-3”; NHMUK, NHMUK013702406. — The type series comprises all known specimens of the species (Table S1).

NW of Manantenina, Anosyennes Mountains, northern massif.

Superficially, the most similar but distinctly smaller species is Tritonaclia stephania. The diagnostic characters are provided in the determination keys.

(Fig. 8D–G) Head. Entirely pale brown, proboscis well developed, ochraceous yellow; labial palpus 3 segmented, elongate, porrect and slightly curved upwards, with around ½ of the length protruding beyond frons; 1^st palpomere short, ventrally covered with elongate scales, 2^nd palpomere around 2× longer than 3^rd palpomere; antennae piliform, ventrally densely covered with short erect setae, additionally each antennomere on both margins with pair of setae, one elongate and arched, another one much shorter and straight; Thorax. Patagia pale brown, with elongate scales, laterally with small dot of clear white regular scales; tegulae entirely pale brown, with elongate scales, and piliform scales in terminal portion; mesothorax and metathorax entirely pale brown; Legs. Entirely pale brown, tarsal claw with additional tooth; Foreleg. Tibia ⅔ of femur length, epiphysis present, of ⅔ of tibia length, tibia and 1^st tarsomere together slightly longer than femur; behind coxa paintbrush-like tuft of fine, elongate, creamy scales. Midleg. Tibia with one pair of pale brown terminal spurs; Hindleg. Tibia with one pair of pale brown terminal spurs, and one pair of additional spurs in ¾ of tibia length, of around ⅔ of length of terminal ones; Abdomen. Pale brown, tergites 4^th–6^th lateroterminally with differently expressed clear white to pale creamy marks on each side, from fully developed blotch to admixture of a few scales; in some specimens white scales present also on tergite 7^th; sternites – 6^th medioterminally with differently expressed clear white blotch, in terminal segments usually reduced to a few scales; in form of scales admixture; 8^th sternite reduced; abdomen laterally unavailable; valvae protruding and well visible, outer surfaces covered with pale brown scales; Forewing. Subtriangular, slightly elongate; background pale brown, with 5 white blotches, two in medial and three in distal wing portion, of various shape and size, all semi-transparent, covered with smaller and scarce scales; first medial blotch (m₂) small, subtriangular to suboval, in terminal portion of DC, from R (touching), terminated in half of DC width, in some specimens distinctly reduced, almost absent, in form of small, but always semi-transparent dot; second medial blotch (m₃) medially between CuA₂ and 1A+2A, smallest one, always in form of indistinct dot; first distal blotch (m₄) of moderate size, suboval to round; second distal blotch (m₅) always relatively small, rounded to elongate, in some specimens very indistinct; third distal blotch (m₆) always the largest one, rounded to irregularly subtrapezoidal; Underside as upperside, but in medial and hind portions almost entirely covered with small, scarce, grain-like scales, giving entire wing slight semi-transparent appearance; hind margin along basal portion with tuft of piliform scales; retinaculum subcostale present, narrow and elongate; Hindwing. Oval, elongate, reaching to ⅔ of forewing length; background pale brow with 1 or 2 white semi-transparent blotches of different shape and size; basal blotch from wing base, between DC hind margin and basal portion of wing hind margin, irregularly pear-shaped, slightly incised in basal and terminal portion along vein 1A+2A, separated from wing margin by very narrow pale brown stripe; distal blotch between CuA₁ and M₃, in form of relatively small, suboval dot, in some specimens very indistinct and reduced up to just a few scales visible only under microscope; underside as upperside; M₂ and M₃ fused in basal portion but at different length, Sc+Rs+M₁ completely fused;

(Fig. 12) Tegumen moderately sclerotised, arms laterally narrow, submedially provided with irregularly subtrianagular, slightly curved lateral lobe covered with tuft of elongate, erect setae; uncus basally broad, then distinctly narrowed, rod-like, arched downwards, in terminal portion slightly widened laterally and dully terminated, dorsally with indistinct, rounded incision, ventrally with sharp horizontal edge; vinculum narrow, moderately sclerotized, incompletely fused with tegumen; saccus indistinct, in form of subtriangular plaque; juxta particularly well developed and heavily sclerotised, in form of slightly suboval ring, ventrally connected with membrane provided with two moderately sclerotised, shield-like plaques, dorsolaterally armed with prominent, sinuous, bull-horn-shaped protrusions directed posteriorly, asymmetrical, right one slightly longer and reaching to around ½ of uncus length; valva elongate, of complex shape, not reaching beyond uncus, costal margin distinctly folded inwards, in basal portion shallowly incised, terminal portion transformed into well sclerotised, claw-like, pointed protrusion, sacculus well sclerotised, distinctly arched, in terminal portion provided with membranous, lobe-like, densely scobinate protrusion, roundly terminated, not reaching to valva tip, covered with elongate erect setae; phallus short and broad, narrowed in terminal half, coecum penis rounded; vesica left uneverted to avoid destruction, but with well visible structures, basal portion with two heavily sclerotised, subtriangular plaques with elongate and slightly twisted terminal portion, along vesica sinuous row of 15 distinct, heavily sclerotised subconical teeth flattened dorsoventrally and leaning one on another;

Figure 15 were described and illustrated as T. stephania in Griveaud (1964: 80, fig. 184).

From Malagasy ombilahy meaning “bull”, after the shape of the dorsal spines of the male juxta, resembling the horns of zebu cattle commonly bred in Madagascar.

(Fig. 16F) Mountain massifs of southeastern Madagascar: Andringitra, at the border of the Regions Matsiatra Ambony and Ihorombe, and the Anosyenne Mountains, at the border of the Regions Anosy and Atsimo-Atsinana. Fianarantsoa (Region Matsiatra Ambony), as the collecting locality of two specimens from C. Lamberton, is doubtful (see similar cases above, and Wiorek et al. 2026). At 1050–1650 m elevation.

Specimens were collected in October and November. Associated with evergreen, humid midland forest.