The genus was described by Meyrick in 1897 on the basis of a female, C. thysanota from Australia. It was compared with Stathmopoda Herrich-Schäffer, 1853, then included in the family Elachistidae, now included in the family Stathmopodidae (De Prins and De Prins 2011–2025). Meyrick probably confused Stathmopoda for the long and bristly scales on the mid and hindlegs and a similar peculiar posture of the adult. Subsequently Vári (1961) redescribed the genus, arguing that it should be included in the family Gracillariidae, near Cyphosticha Meyrick, 1907. For the description of the male characters, he used a species from India, Phrixosceles trochosticha Meyrick, 1908, the type genus of Phrixosceles Meyrick, 1908 which he considered as a synonym of Cuphodes (Fletcher 1929). For females he used both C. thysanota and P. trochosticha. This synonymy was proposed by Vári (1961) based on the similarity of the wing venation and the bristly hairs on the legs. De Prins et al. (2025), studying the Australian fauna, introduced new characters and considered Cyphosticha as a synonym of Cuphodes. They also moved all Australian species of Phrixosceles to the genus Cuphodes, but left the former genus valid with two species, P. trochosticha Meyrick, 1908 (type genus, India) and P. fibulatrix Meyrick, 1922 (Fiji) (De Prins et al. 2025).

The genus Cuphodes currently comprises 28 species distributed across Australia, Africa and southeastern Asia (De Prins et al. 2025). The nine species with known biology are leaf miners associated with Ebenaceae, Fabaceae, Putranjivaceae and Rhamnaceae. This total excludes Cuphodes paragrapta (Meyrick, 1915) from South America, which upon examination, does not appear to belong to Cuphodes but to an unknown genus. In addition to the described species, six putative species were reported by Kawakita et al. (2010), comprising leaf miners on Ebenaceae and Rhamnaceae from East Asia (Table 1), whose morphology has not yet been described.

Cuphodes has frequently been included in molecular phylogenetic studies. Kawakita (2010) identified it as a putative sister taxon to Epicephala, concluding that it is probably one of the closest relatives of that genus. More recently, De Prins et al. (2025) found that mitochondrial data consistently support the monophyly of the Australian Cuphodes species, recovered as sister to Toowoomba De Prins, Sruoga & Zwick, 2025 + (Parectopa Clemens, 1860+Polysoma Vári, 1961), albeit with low to moderate statistical support. In contrast, Cuphodes diospyrosella (Issiki, 1957) and C. wisteriella Kuroko, 1982, included in the analyses of Kawahara et al. (2016) and Li et al. (2022), were recovered as sister to all other Ornixolinae. In those studies, Parectopa was placed either very distantly (Kawahara et al. 2016) or was recovered as strongly polyphyletic (Li et al. 2022).

The characters reported by De Prins et al. (2025) for Australian Cuphodes are a shining-white backround with a subtle yellow ochreous pattern on the forewing and exceptionally pilose hindlegs, equipped with erected sharply pointed spiniform scales. In male genitalia subscaphium+gnathos strongly developed; costa of valva very strongly sclerotised; subapical part of valva with a hanging semi-round flap or other sclerotisations of different sizes and shapes; transtilla complete but in case of absence the function of transtilla is taken by very long basal apodemes of valvae or sclerotised basal bow of vinculum. In female genitalia apophyses anteriores either absent or rather short; papillae anales strongly fused and deeply nested in the cuticle of segment VIII; antrum/colliculum sclerotised; signa small; bulla seminalis usually situated at anterior margin of segment VII or in segment VII, small with convoluted ductus seminalis.

Examination of African species treated by Vári (1961), specifically those with known males (C. melanostola (Meyrick, 1918), C. diospyri Vari, 1961, and C. dolichocera Vari, 1961), reveals important differences in the male genitalia compared with the characters reported by De Prins et al. (2025), particularly in the shape of the valva, which is simply rounded apically and lacks notable sclerotization. In the newly described C. spermotrophus sp. nov., two bristles are present at the apex of tegumen (Figs 1, 2). This feature is characteristic of several genera in other subfamilies of Gracillariidae, such as Spulerina (Acrocercopinae) and Spanioptila (Ornixolinae), and is particularly significant within Lithocolletinae. Among Australian Cuphodes, apical tegumen bristles are prominent in the pyrochroma clade, which includes C. pyrochroma (Turner, 1894), C. pandoxa (Turner, 1913), C. callimacha (Meyrick, 1920), but are apparently absent in the other groups. Notably, such bristles are not visible in the South African Cuphodes illustrated by Vári (1961) and are not mentioned in his descriptions. The wing venation does not differ from that of the type species C. thysanota Meyrick, 1897, in particular due to the absence of R1 and R2 in the forewing (Figs 4, 5).

Adults of the new South African species exhibit a typical resting posture for Cuphodes: the head is held downward, the abdomen raised and supported by the hindlegs, and the fore- and midlegs extended forward and held close together (Fig. 14) (Kawahara et al. 2016).

Based on these morphological and behavioural traits, we provisionally place spermotrophus sp. nov. in Cuphodes, following Vári’s interpretation, until further molecular or morphological data allow a more definitive placement.

• ♂, South Africa, Western Cape, Woodville, Duif street, e.l. 27/ix/2023, CLV49591, GRAEL7889-25, R Swart leg. slide TRB4550♂. Coll. TRB.

1♂, 7♀♀ (coll. TRB); • 1♂, South Africa, Western Cape, Woodville, Duif street, e.l. 27/ix/2023, R Swart leg., slide TRB4546♂; • 1♀, same locality, f.c. 18/iv/2024, e.l. 11/v/2024, RP306, R Swart leg., slide TRB4545♀, TRB4549♀ [wings]; • 1♀, same locality, f.c. 24/x/2023, e.l. 30/ix/2023, RP180, R Swart leg., slides TRB4547♀, TRB4543♀ [wings]; • 1♀, same locality, f.c. 21/viii/2023, e.l. 18/ix/2023, R Swart leg., slides TRB4553♀; • 1♀, same locality, f.c. 18–27/ix/2023, CLV49590, GRAEL7888-25, RP122, R Swart leg., slide TRB4544♀; • 1♀, South Africa, Western Cape, Garden Route National Park, -33.969966, 23.538344, 202 m, f.c. 22 April 2026, e.l. 23 April 2026, RC Swart leg., Afromontane Forest edge, Pupa on leaf of Rhamnus prinoides, bottom of leaf along midrib, RHAM26-H230426, SAM-LEP-A053741 (SAMC); • 1♀, ditto, except for: e.l. 27 April 2026, RHAM26-H270426, SAM-LEP-A053742 (SAMC); • 1♀, ditto, except for: e.l. 28 April 2026, RHAM26-H280426, SAM-LEP-A053743 (SAMC).

Derived from Greek sperma- (seed) and -tropho (to feed), referring to the seed-feeding habit of the species, a masculine noun in apposition.

Ground colour of forewing pattern dark brown with dorsum marked by a thin white line with an irregular internal margin (7A-C); apex with a small ochre-white area with a central dark brown dot (Fig. 7D). This pattern makes this species easily identifiable from other African Cuphodes but it is very similar to the Australian Diphtheroptila glochidiella De Prins, Sruoga & Zwick, 2025, a problematic species that lives on Glochidion (Phyllanthaceae) (De Prins et al. 2025). However, the latter differs in the different apical pattern of forewing and in the female genitalia while male genitalia and wing venation are unknown. Male and female genitalia of the new species show some affinity with C. diospyri from which it differs for the much shorter valva, costo-apical corner protruded and in the female ductus bursae with convolutions. Adults have strongly setose mid and hindlegs (Fig. 6).

Forewing length 4,5 mm (Fig. 7C).

Head (Fig. 7E). Vertex and occiput white, with slightly raised scales, frons of the same colour; labial palpus white, curved upwards, almost as long as fore tibiae, first and second segment slightly spotted brown above; maxillary palpus smooth, 4 segmented, about a third as long as the median segment of labial; antenna slightly longer than forewing, with scape and pedicel brown above and white below, pecten present with about ten thin bristles, light ochre filament with barely evident ringing.

Thorax. White, tegula dark brown. Foreleg femur, tibia and tarsomeres (Fig. 6) whitish yellow, brownish above; midleg white, femur and tibia brownish above, tibia with a white area subapically, covered with thin white, pointed, spine-like scales continuing onto the first two tarsomeres, first stained dark brown subbasally, last two stained dark brown; hindleg white, tibia stained dark brown apically, median and apical spurs striped with brown, less densely covered with thin white and pointed scales which continue on first two tarsomeres, terminal tarsomere with fuscous apex.

Wings. Forewing elongated (Fig. 7A-C), dark brown with a sparse sprinkling of white scales: a white dorsal band irregularly margined in dark ending at the apical 2/3 of wing, followed by small, irregular light spots characterising the apical area; apex with a small ochre-white area with a central dark brown dot and a double fringe line, the innermost one with strong blue-green highlights. Cilia brownish along tornus and then lighter. Hindwings uniform greyish brown with light ochre-grey cilia.

Abdomen (Fig. 1). Male abdomen brownish above, brownish ochre on sides with five short oblique dark brown streaks, white below. Segment VI slightly shorter than segment V, normally scaled, anterior sternal margin with a subtriangular projection; segments VII and VIII weakly membranous, bare, each bearing a pair of coremata consisting of long, piliform scales, posterior slightly longer than anterior and about the length of segment VI; segment VII reduced into a thin sclerite, with a pair of sublateral oval androconial plates consisting of thin specialized scales; posterior margin of tergum VIII with a subtriangular projection covered with piliform scales about 2/3 the length of the anterior corema. In females, the pattern of abdomen is lighter.

Male genitalia (Figs 2, 3). Tegumen broad, conical, with rounded apex on which two setae are present about 1/5 as long as posterior coremata, basal joint broad and sclerotised; tuba analis apically microsetose, just longer than tegumen and with a wide subscaphium covered with minute setae, more ventrally the diaphragm presents a densely microsetose protrusion; valva simple, with parallel edges then tapering towards the apex, without appendages, barely visible the sacculus and with a strongly sclerotised costal margin with exception of cucullus, setae of varying length along the entire ventro-apical margin and, to a lesser extent, along the dorsal, transtilla absent; vinculum wide, with large lateral lobes, sutures very evident, with a long saccus about 0.6 length of valva; phallus straight, tubular, about 4× length of valva, without cornuti, phallobase about 2/3 long, inception of bulbus ejaculatorius at 0.2 length of phallus, aedeagus with crest-shaped apex.

Female genitalia (Fig. 8). Papillae anales fused dorsally, apophyses posteriores longer than apophyses anteriores; segment VIII very reduced, apophyses anteriores present; segment VII with a deep concavity along the ventro-apical margin which also involves the intersegmental membrane; sterigma membranous, with a weakly sclerotised rounded area into which a small ostium bursae opens; along the sub-basal margin two thin protrusions covered by long and thin scales; antrum weakly sclerotised, ductus bursae membranous with two convolutions, slightly thickened in its distal third; corpus bursae elongated, sub-cylindrical, without signa; bulla seminalis at sub-posterior part of segment VII, smaller than the corpus bursae, ductus seminalis not perceptible; ductus spermathecae spiraling into 4/5 convolutions before entering the vesica.

This species represents the first record of a seed-feeding Gracillariidae from Africa. It attacks the berries of Rhamnus prinoides, a shrub or small evergreen trees (Fig. 9) in the Rhamnaceae family that can reach 6–8 m in height. It is found primarily in East Africa, from Ethiopia to South Africa, but also in the central-western regions, including Cameroon, Angola, and the Democratic Republic of Congo. It grows preferably in regions with an altitudinal range of 1,400–3,200 m, on wooded mountain and hill slopes, at the edges of forests, and on the banks of streams, often among rocks. Its flowers develop into fruits/red berries, which, as they mature, produce ripe, dark-purple, and softer berries, which are ready for harvesting and seed extraction (Negasse 2021).

When seeds of R. prinoides were examined, their mean length was 4.76 mm and their mean width was 3.02 mm. In fruits from which C. spermotrophus emerged, a distinct exit hole is visible on the fruit surface, close to the calyx (Fig. 10). Beneath this opening, an empty seed can be observed (Fig. 11). Only a single exit hole was ever observed per fruit. Cocoons were recorded in January, February, March, May and July while adult emergence occurred in September, October, February, April, May and June. Attacked fruits were found on shrubs growing along forest edges, where the host plant grows naturally. The presence of the species can also be confirmed by examining the underside of the leaf midrib, where cocoons are formed after larval emergence from the seed (Fig. 12).

For reasons that remain unclear, the larva enclosed within its cocoon covers it with frothy “bubbles” produced from the anus and expelled through small slits in the cocoon wall, where they remain attached to the external surface (Fig. 13). This unusual behaviour has also been reported in other Cuphodes species (De Prins et al. 2025) as well as in Chileoptilia, Marmara and related genera (Wagner et al. 2000; Vargas and Landry 2005).

Besides C. spermotrophus, the only frugivorous insect reared from R. prinoides was a single adult belonging to the Drosophila suzukii Matsumura, 1931, a recent invader in South Africa (Steyn et al. 2025).

In total, only one eulophid (Tetrastichinae) emerged from hundreds of berries of R. prinoides. The occurrence of the eulophid is noteworthy, as nothing is currently known about the parasitoids of Cuphodes. Among genera with feeding habits similar to spermotrophus, only a few species of Conopomorpha are known to be parasitized by Braconidae, Ichneumonidae, Encyrtidae, and Trichogrammatidae (De Prins and De Prins 2011–2025).

The micromoth so far has only been found between George (33°57.337'S, 22°31.765'E) and Nature's Valley (33°58.200'S, 23°32.301'E) in the Western Cape province. The distribution of the hostplant in Africa suggests that this gracillariid might also have a wide distribution but it has just been overlooked due to its cryptic life cycle.

The two barcodes (GRAEL7888-25; GRAEL7889-25) form a unique BIN BOLD:AHB8508 (Suppl. material 1, Fig. 15) at 9.66% from the nearest Gracillariidae barcode in BOLD. Our phylogenetic analysis shows that Cuphodes spermotrophus sp. nov. occupies an isolated branch. Six Rhamnaceae feeding specimens belonging to three Cuphodes species form a single clade suggesting host use conservatism. Cuphodes spermotrophus does not come out in this clade but internal nodes are poorly supported and it is possible that with more sequence data and taxa the Rhamnaceae feeding Cuphodes will be monophyletic (Fig. 15). Another possible scenario is that the new species evolved independently from Ebenaceae feeders. At the current level of resolution, the tree does not allow us to discriminate between these two hypotheses. The still unknown biology of some African Cuphodes may perhaps provide further information.