LAUSR

Pseudasphondylia species (Diptera: Cecidomyiidae) are known to induce fruit galls on Actinidia rufa (Siebold & Zucc.) Planch. ex Miq. and finger-like leaf galls on Elaeocarpus sylvestris (Lour.) Poir. in Taiwan, but their taxonomic positions remain undetermined. Based on gall morphology and host plants, they were supposed to be the same or allied species of known Japanese congeners, i.e., P. matatabi Yuasa & Kumazawa inducing flower-bud galls on Actinidia polygama (Sieb. et Zucc.) Maxim and P. elaeocarpi Tokuda & Yukawa inducing finger-like leaf galls on E. sylvestris. Species identifications of these Taiwanese species provide us an opportunity to study biogeographical aspects and transition of ecological features in these Pseudasphondylia species distributed in East Asian Arc. Morphological comparisons and species delimitation by molecular analysis indicated that the cecidomyiid on the fruit of A. rufa is distinct from P. matatabi and thus it is described as a species new to science, P. kiwiphila sp. nov. Lin, Tokuda, & Yang. The leaf galler on E. sylvestris was identical to P. elaeocarpi, whose southernmost distribution range extended to Taiwan, a new record of its distribution. COI-based phylogenetic tree (Bayesian inference and IQ tree) of Pseudasphondylia suggested that leaf galling habitat and univoltine life history are ancestral, whereas fruit or flower-bud galling and multivoltine life history are derived. In addition, the monophyletic Actinidia-associated species lineage is sistered to the clade including the remaining Japanese fruit and flower-bud gallers, suggesting that Pseudasphondylia has colonized on the host genus Actinidia once and later speciated on different plant species of the host genus. As a biogeographical aspect of P. elaeocarpi, 2.7% of the COI distance between Japanese and Taiwanese individuals indicates that they have diverged around 1.2 mya, which corresponds to the last but second separation of Taiwan and Japan in the Pleistocene. As for Actinidia-associated Pseudasphondylia species, the two valid species are allopatric and have distinct areas of origin, suggesting they may have speciated allopatrically. Nevertheless, there is still the possibility of ecological speciation due to the following reasons: (1) Host species (and varieties) and unidentified congener of Actinidia-associated Pseudasphondylia are occurring China, revealing potential occurrence of these gall midges. (2) The divergence time (2.2-2.9 mya) of the two known species corresponds to the late Pliocene to Pleistocene, when China, Taiwan, and Japan were part of the East Asian continent. During this period, their host species were sympatric in southeast China. (3) The host of two named Actinidia-associated Pseudasphondylia species each belong to different plant groups with distinct fruit features. These presume that the speciation might have been caused via sympatric host shift.