IV. The tribe Triticeae in Japan; studies after1930 concerning species and scope of their distribution

1. Classification
Table 1 shows a list of those species described in reports,which treated mainly Japanese floras. Regardless the change of the scientific name used by the authors, the same kind of plant is arranged in the same line and given Japanese vernacular names, which is rather consistently indicating the same kind of plant.

Apparently after an awareness of the distribution of Japanese endemic and indigenous Triticeae species, our own nomenclature with new scientific names was started by contemporary Japanese, young taxonomist. Among them Honda (1930) published his study for grasses that Ohwi (1941) praised as Honda's great work (voluminous work), that is Monographia Poacearum Japonicarum (1930), with good index and precise and detailed synonyms. During the coming ten years, studies had proceeded beyond Honda. Ohwi reported Gramina Japonica I~IV in Acta Phytotaxonomica et Geobotanica from 1941 to 1942. He amended, supplemented Honda's identification and scientific names, and Ohwi himself also amended and changed the name of Agropyron species in Gramina Japonica I, in the same series of papers. The name of Agropyron tsukushiense was changed; the species name A. Mayebaranum Honda in 1941 was changed to A. Kamoji Ohwi in 1942. Further changes were made and finally the scientific name A. tsukushiense Ohwi var. transiens Ohwi was given to a plant "Kamoji-gusa". Similar changes of a given species name had made by renaming the present-day A. humidorum (Japanese name Mizuta-kamoji). At first the species was named A. Mayebaranum (Honda 1930), and then Ohwi and Sakamoto (1964) named it A. humidum. Later Ohwi (1965) corrected to A. humidorum in his revised edition of Flora of Japan. The confusion could have been resulted from the relatively frequent hybridization between A. tsukushiense var. transiens and A. humidorum. The present author has observed frequently that the hybrid plants now called "Tariho-no-ohtachikamoji", A. x Mayebaranum, were more vigorously and conspicuously growing than the parents. And often A. humidorum plants were completely absent from nearby areas. Furthermore, since A. humidorum shows highly plastic nature that under unfavorable condition on the poor soil such as on the ridge of paddy field, plants become as small as only 15 cm with relatively short spikes or so. For the characteristic ha propagation aspect of A. humidorum in comparison with A tsukushiense see Kimata and Sakamoto (1982). These were probably the reasons that Matsumura S (1941) described a line (2n = 6x = 42,A. humidorum at present) as a dwarf species from Kyushu, without having proper identification a the time. Similarly, for A. ciliare the scientific name Agropyron racemiferum Koiz. is adopted in "New Flora of Japan" of Ohwi, revised by Kitagawa (1983), while A. ciliare is used in the English edition published in 1984. In Fig. 4, A. ciliare var. minus is shown and in Fig. 5, A. humidorum and A. gmelinii var. tenuisetum. A. humidorum is often found in humid places and near by water. Fig.6 is a photograph of A. humidorum growing at the water's edge.

From field observations widely made for species in Japan, the tendency of variations is considered to be genetic within species and that variation is always very littlein the Elymus group. It is so high in A. tsukushiense that in many cases the plants are different within an area of some square meters; frequently neighboring plants show different characters. This tendency is also observed in A. ciliare, though there are fewer variables than in A. tsukushiense.

This fact may indicate that the Japanese islands, or at least the Far Eastern area, are a center of variation for both A. tsukushiense and A. ciliare. This would be a reason why taxonomy had been faced a difficulty to choose a typical plant among these two species, and reflected in an erroneous adoption of species names throughout a period of this 60 years. Therefore, one of the problem in classification has been natural variation of morphology and - accordingly- hasubdivision of the species; one of this was A. racemifer (A. ciliare) that was divided into several varieties and formas (Kitagawa 1967).

Since those indigenous species are very variable at least in morphological characters it is evident that we are facing diversification of a species. Consequently,the present author proposes those groups to make into one large group as indicated in Table 1. In the Turesson's paper, a sense of genetic evolution is involved and coenospecies is a group of paralogous, sympatric species, which differentiated by evolution (Turesson 1922). A. ciliare and A. tsukushiense are thought to be differentiated at Japan, at least in the Far East. Subdivided species and varieties are undoubtedly closely related within a large species. The two species. A. tsukushiense and A. ciliare are each making the group under a sense of "coenospecies". Also I propose that A. yezoense and A. gmelinii are possible member of a coenospecies, since they hybridize and the F₁ showed normal fertility (Muramatsu 1993).

Sakamoto (1964) tried to cIassify Agropyron species from the basis of cytogenetics, and then wrote a comprehensive paper on patterns of phylogenetic differentiation of the tribe Triticeae (Sakamoto (1973). Although in the report wide genera on the worldwide scale were included, the ecological difference among the domostic Japanese species was omitted when he divided the distribution of the Triticeae genus into Mediterranean-Central Asiatic regions and Arctic-temperate group. Agropyron was included in the latter one. In five genera of the group in Agropyron and Elymus high level of ploidy, from 4x to 12x, is known and natural intergeneric and interspecific hybrid formation is extensive. He further subdivided the group into (1) Agropyron- Elymus- Sitanion- complex (2) Asperella (Hystrix) and (3) Hordeum. The Agropyron-Elymus-Sitanion complex is divided into five groups. The second group of the complex, which comprised of three genera Agropyron, Elymus and Sitanion, is represented by the European A. caninum P. Beauv. Most species of section Roegneria of the genus Agropyron are included here. The common genome of the group is that of A. spicatum Scribn. et Sm. or A. libanoticum Hack. Their genome symbol is S by Dewey (1972).