RESEARCH ARTICLE
New species of fossil butterfly (Nymphalidae: Limenitidinae) from the Upper Pliocene to Lower Pleistocene Teragi Group, Hyogo Prefecture, Japan
2025 年 29 巻 p. 76-86
詳細
2025 年 29 巻 p. 76-86
This study describes a new fossil butterfly species, Tacola kamitanii sp. nov., from the Upper Pliocene to the Lower Pleistocene Teragi Group in Hyogo Prefecture, Japan. The new fossil is characterized by remarkably large wings, with an estimated forewing length of 48 mm and a wingspan of 84 mm. This new fossil species belongs to the genus Tacola based on the following characteristics: both discal cells open, smoothly curved humeral veins, and a thick thorax and abdomen. However, it does not identify with any modern relatives of Tacola with a small discal cell, straight 1A+2A anal vein of the forewing, or long hindwing median vein. Therefore, the fossil was identified as a new species of the genus Tacola and one of the largest species in the subfamily Limenitidinae. The modern relatives of Tacola are distributed in the subtropics and tropics, while this fossil species may have survived in the temperate zone. This is the first named Limenitidini fossil and the youngest example of an extinct butterfly.
ZooBank registration: urn:lsid:zoobank.org:pub:FCF699C6-B28F-4916-91BE-B9C93A0BB0DC
Butterfly fossils are extremely rare (e.g. Sohn et al., 2015; Aiba et al., 2023). Butterfly bodies and wings are fragile and buoyant, making them less likely to be preserved as fossils than other insect groups (Duncan, 1997). The butterflies of the tribe Limenitidini (Limenitidinae, Nymphalidae) comprise seven genera with over 170 species (Wahlberg et al., 2009; Dhungel and Wahlberg, 2018; Wu et al., 2018; Wahlberg, 2023), and have a disjunct distribution, mainly in America and East to Southeast Asia. However, the only known fossil of the tribe Limenitidini is Limenitis sp., from deposits in Willershausen, the Late Pliocene of Germany (Branscheid, 1977). This specimen has only forewings and part of the thorax and has not been identified to species.
Because of the limited distribution of Konservat-Lagerstätten, few Late Pliocene to Early Pleistocene butterfly fossils are known. The only named butterfly fossil is Neptis kabutoiwaensis Aiba et al., 2023 from the Upper Pliocene “Kabutoiwa Formation” (= part of the Maisawa Formation; ca. 3.5 Ma) in Gunma Prefecture, Japan (Aiba et al., 2023).
Until now, no butterfly fossil has been found from the Early Pleistocene. Subsequent fossil records correlate with extant species or their close relatives. For instance, Hestina japonica (Felder and Felder, 1862) and Papilio cf. maackii Ménétriès, 1858 were described from the Shiobara Group (Middle Pleistocene, ca. 0.3 Ma) of Tochigi Prefecture (Fujiyama, 1968, 1983).
Here, we describe a new fossil species of the genus Tacola (Limenitidini, Limenitidinae, Nymphalidae) from the Upper Pliocene to Lower Pleistocene Teragi Group in Hyogo Prefecture, Japan. This is the first named fossil species in Limenitidinae and, meanwhile, the youngest example of known extinct butterflies to date.
The Teragi Group (Wadatsumi and Matsumoto, 1958), distributed around the border between Hyogo and Tottori prefectures (Figure 1), was formed by caldera (called Teragi Cauldron)-fill deposits. The group is composed of volcanic, volcaniclastic, and sedimentary rocks (Haji et al., 2023). Furuyama (1989) divided it into five units as the Basal Conglomerate, the Utaosa Rhyolite, the Yudani Conglomerate, the Haruki Mud, and the Terada Volcanics.
The Haruki Mudstone (= Haruki Mud of Furuyama, 1989) consists of thin alternating layers of sandstone and siltstone with numerous inserted layers of rhyolitic and pumice tuff, yielding many plant and insect fossils (Kinugasa et al., 1968; Furuyama, 1989). However, the study of insect fossils is insufficient. Kinugasa et al. (1968) identified five species and briefly described the following two: Asilidae gen. et sp. indet. and Camponotus sp. Arita and Yamana (1970) described a fossil as Philodromidae gen. et sp. indet. Kinugasa (1981) reported eight species and five genera belonging to three ephemeropteran families. Fujiyama (1982) described two species of Cicadidae, Graptopsaltria sp. and Meimuna sp. Kinugasa (1983) figured Camponotus sp., Coreidae gen. et sp. indet., and Asilinae gen. et sp. indet. Inoue (1986) listed 75 species belonging to 11 orders with photographs, however, no descriptions were provided.
The depositional age of the Haruki Mudstone is ca. 2.8–2.2 Ma (the Late Pliocene to Early Pleistocene) based on K-Ar dating (Furuyama et al., 1998; Furuyama and Nagao, 2004) and by U-Pb dating (Haji et al., 2023) of the different layers of the Terada Volcanics, which are contemporaneous heterotopic facies of the Haruki Mudstone.
This specimen was discovered by Kiyoshi Kamitani on July 31, 1988, near the outcrop where insect fossils occurred abundantly (Kamitani, 1990). The rock slab of the specimen consists of tuffaceous siltstone, which is consistent with the lithofacies of the Haruki Mudstone in the area. Kamitani (1990) provided photographs and a simple sketch. He only mentioned that the fossil was a close relative of Limenitidinae without a detailed description. The fossil was housed at the Museum of Unique Insect Fossils (Omoshiro Konchu Kasekikan) in Shin’onsen Town, Hyogo Prefecture, Japan, with repository number SOU-001.
The fossils were observed under a Leica M205 C microscope (Leica Corporation, Wetzlar, Germany). Photographs were captured using a Leica MC170HD microscope camera (Leica Application Suite Version 4.1.3). The images were sharpened, and their contrast and tonality were adjusted using Adobe Photoshop TM Version CS6 (Adobe Systems Incorporated, San Jose, CA, USA). The nomenclature of the veins follows that of Nielsen and Common (1991), and the terminology for morphological characteristics follows that of Leite et al. (2013), with the following abbreviations: A = anal vein, Cu = cubital vein, CuA = cubitus posterior, dcm = median discocellular, dcs = upper discocellular, M = median vein, R = radial vein, Rs = radial sector, Sc = subcostal, t 1–t 8 = tergum I–VIII, Sc II = mesoscutellum, St III = metascutum, Sc III = metascutellum.
Order Lepidoptera Linnaeus, 1758
Superfamily Papilionoidea Latreille, 1802
Family Nymphalidae Rafinesque, 1815
Subfamily Limenitidinae Behr, 1864
Tribe Limenitidini Behr, 1864
Genus Tacola Moore, 1898
Tacola kamitanii sp. nov.
[New Japanese name: Kamitani-onimisuji]
ZooBank lsid: urn:lsid:zoobank.org:act:E3957321-2AEA-4BDA-AA20-2E73A74A3C51
Holotype.—SOU-001; impression of incomplete thorax, forewing and hindwing, midleg and hindleg, and abdomen, preserved in grayish, fine-grained siltstone, consisting of two parts: a concave impression showing the ventral side (SOU-001-A: Figures 2A, 3A, D) and a convex counterpart exhibiting the dorsal side (SOU-001-B: Figures 2B, 3B, C, E).
Type locality and horizon.—The Umigami, Shin’onsen Town, Hyogo Prefecture, Japan (Figure 1). However, exact locality remains unknown. Based on the lithofacies of tuffaceous siltstone, the stratigraphic horizon is correlated with the Upper Pliocene to Lower Pleistocene, Haruki Mudstone of the Teragi Group (Kinugasa, 1981; Kamitani, 1990).
Etymology.—After Kiyoshi Kamitani, a fossil collector.
Diagnosis.—Large nymphalid butterfly (estimated forewing length 48 mm, estimated wingspan 84 mm). Both wings with completely open discal cells. Forewing discal cell relatively small. The forewing has cell streaks and an upper postdiscal band with a straight anal vein 1A+2A. Vein R2 arises from the cell. Hindwing has smoothly curved humeral vein and longer vein M. Thorax and abdomen thick.
Description.—Impression of incomplete body preservation. Head and anterior parts of thorax not preserved. Left forewing and hindwing partially preserved. Scales preserved in several wing parts, but details unknown. Hairs not preserved.
Thorax (Figure 3E): Part of metathorax preserved, 4.85 mm wide, brown in color, slightly sclerotized, scutellum II, scutum III, and scutellum III visible. Scutum III almost completely divided into two lateral regions, but details not preserved.
Forewing (Figures 2A, B, 3A–C, and 4K): Basal, median, and costal margins of left forewing preserved. Apex and outer margins absent. Inner margin invisible. Preserved part of forewing 42.74 mm long. Wing markings indistinct, but cell streaks and upper postdiscal bands faintly visible (Figure 3C). Discal cell relatively small, 15.42 mm in length, 6.23 mm at the widest point, completely open. Vein Sc starts in axillary region and ends at middle of costal margin. Vein R, parallel to vein Sc, forks into veins R1 and Rs. Vein Rs divided into veins R2, R3, R4 and R5. Vein R1 parallel to vein Sc and ends before apex. Vein R2 arises from discal cell, slightly curved, close to common stalk of veins R3–R5, almost same length as vein R1, ending before apex. Short cross vein dcs present. Cross vein dcm convex and roundly curved towards base (Figure 3C). Veins M1 and M2 almost straight and parallel. Vein CuA begins at axillary region. Contact point of vein M3 and vein CuA1 slightly distal to bifurcation of vein M1 and vein M2. Vein M3 curves towards vein M2 in its anterior half. Veins CuA1 and CuA2 straight and parallel. Vein 1A+2A arises from axillary region and straight. Lengths of visible veins: Sc (26.45 mm), R (13.62 mm), R1 (22.45 mm), R2 (22.54 mm), and CuA (9.38 mm).
Hindwing (Figures 2A, B, 3A, B, D and 4L): Left hindwing entirely well preserved, 34.55 mm long but costal margin indistinct and inner margin not preserved. Outer margins partially preserved. Wing markings indistinct. Discal cell completely open. Humeral vein (Figure 3D) smoothly curved toward proximal region of costal margin, arising from separation of veins Sc+R1 and Rs. Vein Sc+R1 relatively long and gently curved. Vein M relatively long, 1.5 times as long as vein Rs+M. Vein Rs longer than vein Sc+R1 and nearly parallel to Sc+R1. Veins M1 and M2 slightly curved in submarginal region. Vein M3 curved towards vein M2 at its anterior half. Veins CuA1 and CuA2 almost straight. Widths of Sc+R1, Rs, and cell M1 almost equal. Cells M2 and CuA1 also of same width but wider than those above. Two anal veins, 1A+2A and 3A preserved. Vein 3A runs along abdomen. Lengths of visible veins: Sc+R1 (26.78 mm), humeral vein (4.12 mm), Sc+R (3.21 mm), Rs+M (1.79 mm), M (2.68 mm), CuA (6.41 mm) and M3 (19.46 mm).
Legs: Mesofemora and metafemora faintly preserved (Figures 2A and 3A), but details not preserved.
Abdomen (Figure 3E): Very thick, brown in color and slightly sclerotized, 12.70 mm long, 5.53 mm maximum width. First to eighth abdominal terga preserved; first abdominal tergum contracted in width. Second tergum wider, almost as long as first; third half-length of second one but widest; gradually narrowing from fourth to eighth; and eighth one incompletely preserved. Abdominal sternum not visible (Figure 2A).
The fossil is included in the subfamily Limenitidinae because the forewing veins are not swollen at the base, and the hindwing humeral vein arises from the separation of veins Sc+R1 and Rs (Inomata, 1990; Gallo and Della Bruna, 2013). In addition, the fossil can be identified as a species belonging to either the tribe Neptini or the tribe Limenitidini based on wing markings and open forewing and hindwing discal cells.
We measured and compared the forewing length and thorax width of several extant species within the tribes Limenitidini and Neptis alwina, one of the largest species in the tribe Neptini (Table 1). Because the fossil is considered female based on its large body size and thoracic thickness, all specimens compared were female.
| species | Number of specimens | Forewing length (L) | Width of thorax (W) | W/L | Locality of specimens examined | ||||
|---|---|---|---|---|---|---|---|---|---|
| range (mm) | mean (mm) | SD | range (mm) | mean (mm) | SD | ||||
| Athyma asura Moore, 1858 | 11 | 32.6–41.6 | 37.5 | 2.96 | 3.2–4.9 | 4.0 | 0.58 | 0.11 | Sarawak, Borneo, E. Malaysia; Langogan, Palawan, Philippines; Bangka, near Sumatra, Indonesia; Bintan, Indonesia; Lam Dong, C. Vietnam; Khanh Hoa, C. Vietnam; Da Nang, C. Vietnam; Lao Cai, N. Vietnam; Ha Giang, N. Vietnam |
| A. jina Moore, 1858 | 5 | 29.0–36.4 | 32.8 | 3.15 | 1.8–3.2 | 2.6 | 0.53 | 0.08 | Jiuloanshan, Guangdong, China; Xam Neua, N. Laos; Pulchoki, Nepal; Ha Giang, N. Vietnam |
| A. perius (Linnaeus, 1758) | 9 | 29.5–35.7 | 32.4 | 2.18 | 2.4–3.6 | 2.9 | 0.45 | 0.09 | Da Nang, Vietnam; Lam Dong, C. Vietnam; Yen Bai, N. Vietnam; Lam Dong, C. Vietnam; Bin Thuan, S. Vietnam; Taiwan; Gopeng, Perak, W. Malaysia; Weh, near Sumatra, Indonesia |
| A. selenophora Kollar, 1844 | 10 | 30.7–36.7 | 34.2 | 1.57 | 2.8–4.8 | 3.8 | 0.63 | 0.11 | Ishigaki, Okinawa, Japan; New Taipei city, Taiwan; Ha Giang, N. Vietnam; Oudomxay, N. W. Laos; Bolikhamsai, C. Laos |
| Auzakia danava (Moore, 1858) | 7 | 37.5–48.5 | 42.4 | 3.87 | 2.8–3.8 | 3.5 | 0.34 | 0.08 | Yen Bai , N. Vietnam; Kon Tum, C. Vietnam; Ha Giang, N. Vietnam; Lao Cai, N. Vietnam; Kathmandu, Nepal; Uttarakhand, N. India |
| Limenitis populi (Linnaeus, 1758) | 5 | 40.9–43.7 | 41.8 | 2.03 | 3.3–4.3 | 3.6 | 0.36 | 0.09 | Nagano, Japan; Hokkaido, Japan |
| Moduza procris (Cramer, 1777) | 7 | 33.3–35.4 | 34.5 | 0.73 | 2.8–3.8 | 3.4 | 0.34 | 0.10 | Da Nang, C. Vietnam; Lam Dong, C. Vietnam; Kbal Spean, Cambodia; Langkawi island, W. Malaysia; Weh Island, near Sumatra, Indonesia; Chiang Mai, N. Thailand |
| Neptis alwina (Bremer et Grey, 1853) | 8 | 36.2–42.0 | 39.1 | 1.87 | 2.5–3.4 | 2.9 | 0.36 | 0.07 | Yamanashi, Japan; Nagano, Japan; Beijing, China; Yunnan, China; Sichuan, China; Gyeonggi, S. Korea |
| Neurosigma siva (Westwood, 1850) | 8 | 38.2–42.6 | 40.4 | 1.90 | 2.5–3.6 | 3.0 | 0.43 | 0.07 | Lam Dong, C. Vietnam; Ha Giang, N. Vietnam; Karen, Myanmar; Chiang Mai, N. Thailand; Chiang Mai, N. Thailand |
| Tacola eulimene (Godart, 1824) | 3 | 47.5–50.8 | 49.6 | 0.50 | 4.7–5.7 | 5.2 | 1.82 | 0.10 | Seram, Moluccas, Indonesia |
| T. larymna (Doubleday, 1848) | 10 | 38.8–44.7 | 42.1 | 0.60 | 3.5–5.4 | 4.3 | 1.69 | 0.10 | Sarawak, Borneo, E. Malaysia; Sabah, Borneo, E. Malaysia; Lam Dong, Vietnam; Dak Nong, C. Vietnam; Bin Thuan, S. Vietnam; Banyak, near Sumatra; Kbal Spean, Cambodia; Palawan, Philippines |
| T. kamitanii | 1 | – | 48* | – | – | 4.9 | – | 0.10 | This study |
The fossils share characteristics of the tribe Neptini, as both wings have completely open discal cells (Bozano, 2008). The tribe Neptini includes five genera: Aldania Moore, 1896; Lasippa Moore, 1898; Neptis Fabricius, 1807; Pantoporia Hübner, 1819; and Phaedyma Felder, 1861 (Eliot, 1969). Recently, the genus Phaedyma was treated as a synonym of Neptis based on a molecular phylogenetic study (Ma et al., 2020).
These genera are different from the fossil species in the following. (1) The hindwing humeral veins of these genera is straight, angular, or branched (Chermock, 1950). For example, Neptis alwina Bremer and Grey, 1853 has a short and straight humeral vein (Figure 4C). In contrast, the humeral vein of the fossil is long and curved toward the proximal region of the costal margin (Figures 3D and 4L). (2) The forewing anal vein 1A+2A of the fossil is noticeably straight (Figures 3A and 4K). In contrast, veins 1A+2A of these genera are curved (Figure 4B). (3) The fossil has a branching point of veins M3 and CuA1 slightly distal to the bifurcation of veins M1 and M3 (Figure 4K). While in those genera the branch is more distal (Figure 4B). (4) The fossil is larger in body size. For example, the forewing length of N. alwina (Figures 4A, 5 and Table 1), one of the largest species in the tribe Neptini, is approximately 40 mm or less (Bozano, 2008; Kudo, 2022). Based on the above, the fossil cannot be identified to the tribe Neptini.
The tribe Limenitidini has the same characteristics as the fossil, as the forewing vein R2 arises from the discal cell and the hindwing humeral vein is smoothly curved (Gallo and Della Bruna, 2013), which includes the following seven genera: Adelpha Hübner, 1819; Athyma Westwood, 1850; Auzakia Moore, 1898; Limenitis Fabricius, 1807; Moduza Moore, 1881; Neurosigma Butler, 1869; and Tacola Moore, 1898 (Wahlberg, 2023). Among these, the discal cells of both wings are open in all members of Tacola and in a few species of Athyma, such as Athyma perius Linnaeus, 1758 and Athyma lyncides Hewitson, 1859.
The genus Athyma includes 43 species (Wahlberg, 2023). Gallo and Della Bruna (2013) mentioned that the forewing discal cell is closed in most Athyma species, except Athyma asura Moore, 1858, Athyma fortuna Leech, 1889, Athyma jina Moore, 1858, and A. perius. However, our observations showed that the discal cells of all four species, except A. perius, are closed. Athyma perius can be distinguished from the new fossil species by the following characters. (1) Forewing discal cell is longer than in the fossil. (2) Forewing anal vein 1A+2A is slightly curved, whereas that of the fossil is straight. (3) Athyma perius is smaller than the fossil: the forewing length of A. perius is less than 40 mm (Figures 4D, 5, and Table 1). Athyma lyncides also has open discal cells (Tsukada, 1991). This species was once included in the genus Lamasia Moore, 1898 (Tsukada, 1991) but was recently transferred to the genus Athyma (Dhungel and Wahlberg, 2018; Wahlberg, 2023), and is like the fossil in that the forewing discal cell is small and the 1A+2A vein is straight. However, A. lyncides is clearly different from the fossil by having an acute dcm cross vein and the absence of a cell streak (Tsukada, 1991). Based on the abovementioned factors, the fossil cannot be identified as a member of the genus Athyma.
The genus Tacola includes three species: Tacola larymna Doubleday, 1848, Tacola magindana Semper, 1878, and Tacola eulimene Godart, 1824. The genus Tacola was once included in the genus Athyma (e.g. Gallo and Della Bruna, 2013) but was recently separated from the latter based on differences in larval and pupal morphology and molecular phylogenetic analysis (Wolfe, 2018; Wu et al., 2018). Tacola is characterized by a thick thorax and long forewing. The fossil is the most like those of Tacola, particularly in agreement with the range of T. eulimene (Figure 5, Table 1).
Tacola larymna is the type species and like the fossil in that both discal cells are completely open, the thorax and abdomen are thick, the humeral vein curves smoothly, and the body is large (Figure 4G–I). However, it is distinguished from the fossil by its distinctly larger forewing discal cells and slightly curved 1A+2A vein (Figure 4H). These morphological differences are also observed in two other species, T. magindana and T. eulimene. The abdominal size could not be compared because abdominal parts of butterflies are soft and easily deformed, making it difficult to measure the exact size of the abdomen of the present new species. A thick thorax and a long forewing are the major morphological features of the genus Tacola, which are shared with the fossil. Fossil characteristics may indicate the possibility of a new genus of Limenitidini. However, because of the incomplete preservation of the fossil, we tentatively include it here in the genus Tacola.
Furthermore, the new species Tacola kamitanii may be the largest example in the subfamily Limenitidinae (excepting some Euthalia spp.) with reconstructed forewing length more than 48 mm. Limenitis populi Linnaeus, 1758, the type species of the genus Limenitis of tribe Limenitidini, is one of the largest congeners of Limenitidinae (excepting some Euthalia spp.) butterflies (Gorbach et al., 2010; Gallo and Della Bruna, 2013), with a forewing length of 34–48 mm, and is distributed in modern Japan (Figure 4, Table 1) (Kudo, 2022). The size of T. kamitanii sp. nov. was comparable to that of the largest specimen (female) of L. populi.
Paleoecological considerationsThe distribution of extant Tacola ranges from subtropical to tropical climate regions of Southeast Asia, including Myanmar, Thailand, Indonesia (Sumatra, Java, Borneo, Celebes, Moluccas), Malaysia (Borneo), and Philippines (Palawan, Sulu Archipelago) (Tsukada, 1991). The occurrence of the present species from Japan adds further to the distribution of Tacola, and suggests that during the Late Pliocene to Early Pleistocene the genus was widely distributed not only in Southeast Asia but also in East Asia.
The early Late Pliocene (ca. 3.3 to 3 Ma) is considered a period when the average global temperature was 2°C to 3°C higher than today (Haywood and Valdes, 2004). Then, during the Late Pliocene to Early Pleistocene, the global average temperature became cooler (e.g. Chandler et al., 1994; Haywood and Valdes, 2004; Robinson et al., 2008). Analysis of ostracod fossils indicates that the Japanese Islands were in transition from a warm to a cold climate from 2.8 to 2.55 Ma (Yamada et al., 2005). The results of plant fossil analysis also indicate that it was temperate at that time (Momohara, 2008). Therefore, this past species of Tacola may have lived in warm-temperate to temperate zones.
We would like to thank Kiyoshi Kamitani for collecting the fossils. We thank Kenta Nakai of the Shin’onsen Town Board of Education and Tadatsugu Uemura of the Museum of Unique Insect Fossils for providing us with the opportunity to study this topic. We thank Kumiko Handa of the Museum of Nature and Human Activities, Hyogo, for providing information regarding fossil localities. We are grateful to the Keio Gijuku Fukuzawa Memorial Fund for the Advancement of Education and Research for financially supporting this study. Finally, we thank Lang Son-Yun of the Chongqing Museum of Natural History, Adam Miles Cotton and one anonymous reviewer and the editor-in-chief for their constructive comments.
H. A. initiated the study and wrote the manuscript and made all figures. Y. T. revised the manuscript and discussed the geology of the Teragi Group. K. S. revised the manuscript and taxonomic aspects. All authors contributed to the writing of the paper.
