Fig 1.
Distribution of Cryptotympana facialis in the Ryukyu and Japanese islands and C. takasagona in Taiwan.
These species are lowland species as shown for the Japanese islands. Island populations are shown by color highlights. See Discussion; Green anticlockwise curved track: Super typhoon dispersal from Yaeyama (pale green) to Japan (red). The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green) and where the same cicada were released (red). Recent dispersal by human transport is also shown by the shorter counterclockwise path. The Kuroshio warm current is shown by the thin orange path with the arrow indicating northward to northeastward movement. Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Table 1.
Cicada species collected and analyzed in this paper.
Fig 2.
Island populations are shown by color highlights. See Discussion; Green anticlockwise curved track: Typhoon dispersal at 0.27 Ma from Miyako (green) to Okinawa (blue). The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green) and where the same cicada were released (blue). Inset: Original habitat in the southern Okinawa and new habitats (blue). We clarified that only the Yagachi-jima population (green) has the same COI sequence with the southern Iriomote-jima, Yaeyama, and dispersed from there. Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Fig 3.
Island populations are depicted by color highlights. See Discussion; Green anticlockwise curved tracks: Typhoon dispersal from Toku to Daito islands (SW part of map), and also from Honshu to O-shima (much shorter path on NE part of map). The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (red) and where the same cicada were released (orange). Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Fig 4.
Distribution of Meimuna kuroiwae.
Island populations are shown by colored highlights. See Discussion; Green anticlockwise curved track: Super typhoon dispersal from Okinawa (blue) to Bonin (purple, Meimuna boniensis). The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (blue) and where the same cicada were released (purple). The Kuroshio warm current is shown by the thin orange path, and the southward Ogasawara current is also shown. Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Fig 5.
Distribution of Meimuna opalifera, that is absent in the Ryukyu islands.
See Discussion; Green anticlockwise curved track: Typhoon dispersal from Honshu to the northern Izu islands of Hachijo and Aoga-shima. The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (red) and where the same cicada were released (red). Base map from Vector Map (VMap) Level 0.
Fig 6.
BI tree for Cryptotympana, Euterpnosia, Mogannia, and Meimuna.
Vicariantly speciated island populations are shown by different colored text labels. Calibration points and dates are shown by green stars and corresponding large font green text labels. Black numbers on each node: posterior probability, red numbers: age (in Ma). The green curved arrows depict dispersal after the vicariance.
Fig 7.
Haplotype network of Cryptotympana.
C. facialis was vicariantly speciated due to isolation of island groups of Okinawa, Miyako, and Yaeyama, separated by major straits (red double line) formed since 1.55 Ma, and three island group haplotypes or populations were formed. However, the Japan haplotype was included in the Yaeyama haplotype network, suggesting a long-distance dispersal (green curved arrow) from Yaeyama (southern end of Ryukyu) to Japan at 0.58 Ma (Fig 6), crossing Miyako and Okinawa. Orange heavy line: C. facialis is separated by the other Cryptotympana species including C. takasagona.
Fig 8.
Haplotype network of Mogannia.
On each of Yaeyama islands, Mogannia minuta constitutes the distinct haplotype, reflecting vicariant speciation since 1.55 Ma. However, this species was recently recorded from Yagachi-jima (Fig 2 inset), northern Okinawa, and the haplotype is identical to the haplotype of S (southern) Iriomote-jima, suggesting recent long-distance dispersal. The Miyako and the original southern Okinawa populations (Fig 2 inset) were included in the same Miyako haplotype network, and the southern Okinawa population originated from the Miyako population, dispersed at 0.27 Ma (Fig 6). The network of the recently recorded specimens in southern and central Okinawa such as Tsuken (Fig 2 inset) suggests recent dispersal also from the Miyako islands. Red double line: barrier for vicariance. Orange heavy line: M. hebes is separated by M. minuta, but genetically closely related.
Fig 9.
Haplotype network of Euterpnosia.
The haplotypes were divided into E. chibensis okinawana (Okinawa) and E. chibensis networks, and the latter divided into the Amami and Japan networks, reflecting the 1.55 Ma event of vicariant speciation acted on the Ryukyu islands. However, E. chibensis daitoensis from the island of Daito was included in the Amami network, showing particularly close resemblance to the Toku network, suggesting that the Daito population was dispersed from the Tokuno-shima island at 0.72 Ma (Fig 6).
Fig 10.
The haplotype network of Meimuna.
Meimuna kuroiwae and M. opalifera constitute different networks. The M. kuroiwae network divided into the Okinawa and Amami-Tokara-Osumi-Kyushu networks, and the population of the southern end Kyushu may have been derived by dispersal from the southern islands. Meimuna boniensis may have been dispersed from Okinawa in ancient times (Nagata, 2019).
Fig 11.
Track of super Typhoon 19, 1990 (blue fine curved path), and 1000 km inferred transport of Polygonia c-aureum by typhoon winds (green curved arrow).
The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green) and where the same cicada were released (red). Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Fig 12.
Track of Typhoon 17, 2013 (fine blue sinuous path), and hypothetical 1000 km transport of Anax parthenope over one day by typhoon winds (green curved anticlockwise path).
The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green) and where the same cicada were released (red). Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Fig 13.
Track of track of Typhoon 5, June 2002 (medium blue curved path) and track of Typhoon 5, June 2011 (very dark blue sinuous path), and hypothetical 350~500 km transport Mogannia minuta by the marginal typhoon winds (green curved path).
The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green and dark green) and where the same cicada were released (blue). Base map from Vector Map (VMap) Level 0.
Fig 14.
Track of Typhoon 1, May 2001 (brown curved path), that may have transported and then released Mogannia minuta on 14 May when it was lost energy west of Okinawa.
The track of Typhoon 8, May 1997 (orange curved path), is also shown, because this typhoon also followed a path that may have transported M. minuta from the southern Ryukyu islands to Okinawa. The green path depicts this hypothetical transport. The bold short arrows schematically (not to scale) show where the typhoon winds picked up cicada (green and dark green) and where the same cicada were released (blue). Base map from Vector Map (VMap) Level 0, National Geospatial-Intelligence Agency.
Table 2.
Summary of dates of first records of Cryptotympana facialis and transplantation dates for the Amami islands after references shown on this table.
For land formation (human landscaping) dates, documentation by archival aerial photos offered by Geospatial Information Authority of Japan are referenced.
Fig 15.
Geological map and cross section (below) of Tokuno-shima island.
Three terraces consisting of the 1.5, 0.9 and 0.4 Ma limestone show three episodes of uplift or progressive uplift with erosion of wavecut benches during high sea level stands. Note the lack of coastal plains at 0.9 and 0.4 Ma, with development of 100-m-high sea cliffs instead. Background: Digital topographic maps of 1: 25,000 scale, the Geospatial Information Authority of Japan.
Fig 16.
Geological map and cross section (below) of Kikai-jima island.
Three terraces of the 1.5, 0.9 and 0.2 Ma limestone show three episodes of uplift or progressive uplift with erosion of wavecut benches during high sea level stands. Note the lack of a coastal plain at 0.9 Ma (100 m high sea cliff), but a relatively wide coastal plain at 0.2 Ma above a lower sea cliff. Background: Digital topographic maps of 1: 25,000 scale, the Geospatial Information Authority of Japan.
Fig 17.
Geological map and cross section (below) of Okinoerabu-jima.
A terrace developed on 1.5 Ma limestone. Uplifting was apparently slower compared to the examples of Toku-shima and Kikai-jima and has been progressed since 1.55 Ma. This history has resulted in a relatively wide coastal plain lacking high sea cliffs, except for fault scarp cliff at along the western part of northern margin. Miyako-jima and Ishigaki-jima have also only a single terrace in 1.5 Ma limestone, and northeastern end of Amami Oshima is also associated with 1.5 Ma limestone, whereas Okinawa-jima and Kume-jima have low elevatioon terraces developed with ca. 0.2 ~ 0.4 Ma (not dated) limestone. Background: Digital topographic maps of 1: 25,000 scale, the Geospatial Information Authority of Japan.