The author has mentioned the following eleven diatoms as the remarkable members of the southern, tropical elements of the diatom-flora of Japan: 1. Planktoniella sol 2. Triceratium dubium 3. Podocystis spathulata 4. Surirella reniformis 5. Hydrosilicon mitra 6. Hydrosera triquetra 7. Actinella brasiliensis 8. Eunotia monodon var. tropica 9. Eunotia formica var. sumatrana 10. Achnanthes crenulata 11. Pinnularia acoricola Of these diatoms, the former five (No. 1-5) belong to the marine species and the other six (No. 6-11) to the fresh-water or brackish-water species.

View full abstract

In the genus Phytolacca, the flower of P. dioica seems to have the most primitive structure, though it is specialized in being unisexual. This male flower consists of 5 perianths, 20-30 stamens and 3-5 sterile carpels. The stamens are generally arranged in 10 fascicles of which the five fascicles (dst) situated on the dorsal side of the carpels with 1-2 stamens each and the five fascicles (1st) on the lateral side of the carpels with 3-4 stamens each (Figs. 1 and 3:4). One of the observed flowers has 29 stamens which are grouped into 12 staminal fascicles (Fig. 2:2 and 3:3). They are five fascicles of (dst), five fascicles of (1st) and two additional fascicles of (ast) (Fig. 3:3). Since the female flowers have an 8-10-carpellate ovary, it is possible that the male flowers having an 8-10-carpellate ovary and 16-20 staminal fascicles are assumed to be the basic structure. The normal male flower seems to be derived from the basic structure by reduction of staminal fascicles into 10. The two additional fascicles of (ast) seem to be the remains of the staminal fascicles of the assumed basic male flower. In the other flower observed (Fig. 2:1), it has 23 stamens in which each of two fascicles of (dst) consist of 3 stamens as compared with one or two stamens in the normal flower. These facts indicate that the basic structure of the flower in the genus Phytolacca is assumed to have 5 sepals, 16-20 fascicles with 3-4 stamens each and an 8-10-carpellate ovary as seen in Fig. 3:5. In P. japonica the normal flowers have 5 perianths, 7-9 stamens and a 7-9-carpellate ovary. The stamens are usually situated on alternate with the carpels (Fig 4:1). One of the observed flower has 5 perianths, 11 stamens and an 8-carpellate ovary (Fig. 4:2). Among these 11 stamens, the eight are situated on the usual position of alternate with the carpel, two are derived from the division of each of two normal stamens, the remained one (dst) is opposed to the carpel. The other flower has 6 perianths, 13 stamens and a 7-carpellate ovary (Fig. 5). among these 13 stamens, the seven (1st) are situated on usual position of alternate with the carpel, the five (dst) are opposed to the carpel and the remained one is divided from the basic structure of the genus by reduction of the number of stamens in the staminal fascicles of (1st) and entire disappearance of the staminal fascicles of (dst).

View full abstract

The distribution area of the four-carpellate Impatiens species which are also characterized by the connate wing petals was proved to be confined to Southeast Asiatic region (Fig. 2). Their habitat is mostly in the limestone area. Semeiocardium alliensii was treated as a member of the four-carpellate Impatiens species.

View full abstract

The seedlings of the Berberidaceae were examined with special reference to the cotyledonary morphology. Excepting for Jeffersonia and Caulophyllum, seedlings of this family have epigeal cotyledons. Cotyledons are articulated into three parts, laminar part, petiolate part and basal tubulate part. Petiolate parts and tubulate parts are usually much shorter than laminar parts. In Ranzania, laminar parts and petiolate parts are nearly equal in length, and tubulate parts are much longer than laminar parts in Diphylleia and Podophyllum. In Leontice, Gymnospermium and Bongardia, solid bar parts, which are considered to be homologous to the tubulate parts, are recognized. In case where cotyledons have very long tubulate parts or solid bar parts, plumules are situated deeply under ground. The stele in the primary root and hypocotyle is usually diarch, but rarely triarch in Nandina, Berberis and Mahonia. In Caulophyllum, the stele is constantly tetrarch. Based mainly on the cotyledonary morphology and vasculature in the seedling, it is possible to recognize 7 groups of genera; Nandina, Berberis-Mahonia, Ranzania, Epimedium-Vancouveria-Plagiorhegma-Jeffersonia-Achlys, Caulophyllum, Leontice-Gymnospermium-Bongardia, Diphylleia-Podophyllum. This grouping is in accordance with classification system based on the floral morphology and other characteristics. The only discordance is the relationship of Caulophyllum. this genus is distinctive in the subtribe Leonticinae in the seedling morphology, while it is closely related to Leontice and Gymnospermium in floral anatomy and seed coat anatomy.

View full abstract

Several breeding experiments were conducted to clarify the breeding and pollination systems of four Japanese Trillium (Liliaceae) species, i.e., T. kamtschaticum, T. apetalon, T. tschonoskii and T. smallii, in their natural habitats. Based on bagging experiment prior to flowering, all four species showed self-compatibility and also no significant differences in seed set compared to open pollinated individuals. Although no effective wind pollination occurred in all four species, pollination accounted for 46% of seed set in T. kamtschaticum. Field observations of insect visits also confirmed that no insect pollination was taking place in T. apetalon and T. smallii. Several Coleoptera and Diptera were observed visiting T. kamtschaticum and T. tschonoskii. However, their activities were low and infrequent. Accordingly, it was suggested that insect pollination in T. kamtschaticum was imperfect and the observed fecundity level resulted from the complementary contributions of both outbreeding and inbreeding systems. Furthermore, sufficient cross- and self-pollination by hand did not also show any significant differences in seed set from open-pollinated individuals in all four species. This fact indicates that seed set in these species is determined not by the frequency of pollination, including pollinator visiting, but by the resources of the plant itself. Consequently, although these four species substantially possess the possibility of both outbreeding and inbreeding, inbreeding proved to be the dominant breeding system. However, subsequent resource limitation allows only certain levels of fecundity regardless high fertilization levels, and as a result guarantees continuous yearly flowering and fruiting.

View full abstract

Chimaphila umbellata (L.) W. BARTON is an evergreen half-shrub occurring principally in the coniferous forests of the temperate regions of the Northern Hemisphere. This species has been divided into five geographically delimited infraspecific taxa: (1) subsp. occidentalis (RYDB.) HULT.-western North America, (2) subsp, acuta (RYDB.) HULT-Arizona and New Mexico, (3) subsp. mexicana (DC.) HULT.-Mexico, (4) subsp. cisatlantica (BLAKE) HULT.-eastern North America, (5) subsp. umbellata-Eurasia (Fig. 1). In order to clarify taxonomic features of each taxon and relationships among them, the following characters are examined on over 300 herbarium specimens: number of flowers inflorescence length, stem diameter, blade length, blade length/blade width, sepal length, sepal length/sepal width, anther length, capsule diameter, number of serrations, distinctness of leaf veins, shape of leaf apex, degree of pedicel and ovary vestment (Figs. 2, 4, Table 1). furthermore the morphological similarities among 13 local populations of this species are considered, based on a comparison among polygraphs of eight characters and a principal component analysis using nine characters (Figs. 3, 5, 6, Table 2). (1) Subsp. occidentalis is characterized in having shrubby thick stems (2.5-3.5 mm diam., rarely over 5.0 mm), 4-7 flowered corymobose-racemose inflorescences, oblanceolate acute leaves (blade length/width ratio 2.8-4.2, 4-6 cm long) with 8-13 serrations on one side and obscure veins, and depressed deltoid sepals (1.5-2.2 mm long, usually broader than long) (Fig. 2). Among seven populations within this subspecies, California population is somewhat similar to subsp. mexicana, and Wyoming & Utah population shows some similarities to subsp. acuta and subsp. cisatlantica (Figs. 3, 5, 6). (2) Subsp. acuta is slightly distinguished from subsp. occidentalis by having narrowly oblanceolate leaves (blade length/width ratio 4.2-5.3) with few serrations (5-7), deltoid and long sepals (2.0-2.8 mm long, usually longer than broad), large anthers and capsules (Fig. 2). This subspecies may be deviated from subsp. occidentalis in the shape of leaves and the size of flowers and fruits. (3) Subsp. mexicana is slightly distinguished from subsp. occidentalis by having somewhat large leaves (4.2-6.6 cm long) with many serrations (8+16) and definite veins, deltoid and long sepals (1.9-3.0 mm long, usually longer than broad), large anthers and capsules (Figs. 2, 4). This subspecies is allied to subsp. occidentalis and slightly differentiated from it in size of leaves, flowers and fruits. (4) Subsp. cisatlantica stands intermediate between subsp. occidentalis and subsp. umbellata from the morphological similarities (Figs. 2, 4), though it is more closely allied to the former subspecies based on the principal component analysis (Figs. 5, 6). (5) Subsp. umbellata is distinguished from subsp. occidentalis by having thin stems (1.8-2.7 mm diam.), 3-5 flowered umbellate inflorescences, short oblanceolate obtuse leaves (blade length/width ratio 2.5-3.5, 3-4 cm long) with few serrations (6-9) and distinct veins, small anthers and capsules, and densely puberulent pedicel and ovary (Figs 2, 4). Within this subspecies European population is more similar to subsp. cisatlantica than East Asia population in having more acutish leaves with obscure veins and larger anthers (Figs. 3, 5, 6). (6) Relationships among five subspecies are summarized in Fig. 7 based on the morphological similarities. Neighboring two taxa connected by a line in Fig. 7 are slightly differentiated from each other because the variation ranges of the most characters in one taxon overlap with those of the other taxon. On the other hand the two taxa mediated by the other one or two taxa; e.g. between subsp. umbellata and the three subspp. acuta, mexicana, and occidentalis, are more differentiated from each other than the extent that exists between the two taxa connected by a line. (7) The following gradual changes i

View full abstract

1) An endemic species of Mitella in Taiwan, M. formosana, was examined for clarifying its affinity. For this purpose the outer morphology of flowers and seeds, and the chromosomes of the species were observed in detail, through cultivation of the individuals collected at Mt. Alishan in Taiwan. The results are shown in Figs. 1-3. 2) The chromosome number of this species was counted to be 2n=28. Although the number 2n=14 was reported by HSU (1968) for M. japonica var. formosana, we should varify the material to be M. formosana or not. The karyotype shows that the plants are diploidized tetraploids (allotetraploids) characteristic of Japanese species in sect. Mitellaria, and resemble well those of M. furusei var. furusei, M. furusei var. subramosa, M. stylosa var. stylosa and M. doiana. 3) A cladistic analysis was made in order to presume the phylogenetic position of M. formosana in sect. Mitellaria, on the basis of 7 characters listed in Table 1. The result is shown in Fig. 4. 4) M. furusei var. furusei, M. furusei var. subramosa, M. formosana, M. stylosa var. makinoi, M. stylosa var. stylosa and M. doiana are presumed as the members which constitute a monophyletic group. Among them M. furusei var. furusei is inferred to be most primitive and M. stylosa var. stylosa and M. doiana most specialized. 5) The ancestor of this monophyletic group might have occurred in the central Honshu which lie east of Ibuki-Suzuka mountains, and have differentiated into several taxa in the area westward from this mountain range. M. formosana is considered to be one of these taxa, and to be most strongly related to M. furusei var. subramosa and M. stylosa var. makinoi.

View full abstract

Southeast Asian violet group Subsect, Serpentes has about ten species and well characterized by their acaulis stoloniferous habit. It is possible, however, to subdivide them into two distinct groups by some characteristics particularly by those related to seed dispersal. Viola pilosa and its close relative, V. canescens, have such features as possessing decumbent fruiting peduncles, capsules without mechanisms for explosive seed dispersal, large seeds with large elaiosomes (Fig. 2) and styles slightly thickened and downward curved at the apex. The other species are characterized by such features as having erect fruiting peduncles, capsules with explosive mechanism, smaller seeds with smaller elaiosomes (Fig. 2) and styles laterally thickened at the apex. Among these, characteristics related to seed dispersal of the former group are those of purely myrmecochorous species in Viola, V. odorata group, i.e. Subsect. Uncinatae. The shape of style is also somewhat reminiscent in both groups, although the apex does not curve downward so strongly in V. pilosa group as in Subsect. Uncinatae. In some other features such as general habit, chromosome number, both groups seem to share principally the same characteristics. It is doubtless that V. pilosa group and Subsect. Uncinatae are closely related taxa. BECKER's subdivision of Sect. Nomimium should be revised with due regard to these facts. In recent expeditions, specimens of Subsect. Serpentes in Thailand are accumulated in herbarium of Kyoto University (KYO). Their distribution map is shown in Fig. 1. Among them Viola sp. is a taxon common in mountains of Northern and Western part of Thailand and characterized by following features; being almost glabrous, leaves rather coriaceous, with scattered hairs above, elliptic to lanceolate-ovate, up to twice as long as broad, flowers without beards on lateral petals. This taxa exhibits intermediate form between V. glaucescens and V. sumatrana, and seems to be independent species. Specimens of Subsect. Serpentes in Thailand and Malesia deposited in KYO are listed below.

View full abstract

[in Japanese]

View full abstract

Five ploidy forms, 2n=14, 21, 49, 56 and 84, were observed in Duchesnea in Gifu Prefecture, Japan. Triploid Duchesnea was reported from a locality for the first time. In a similar manner as in Toyama Prefecture, the plants with 2n=14 were D. chrysantha (ZOLL. et MOR.) MIQUEL and those with 2n=84 D. indica (ANDR.) FOCKE and thus, those with 2n=49 (heptaploid) and 2n=56 (octoploid) were considered as natural hybrids between the above two species. Among 303 samples, 176 were diploid (58.1%), 1 triploid (0.3%), 2 heptaploid (0.7%), 21 octoploid (6.9%) and 103 dodecaploid (34.0%). The geographical distribution of five types of ploid is shown in Fig. 3 and Fig. 4. Apparently, diploid and dodecaploid plants do not show any differential geographical distribution, but octoploid plants were found in some area of Mint Province.

View full abstract

The taxonomic status of plants called Poa crassinervis HONDA (Fig. 1) was ambiguous, and it was sometimes suggested that the plants would be conspecific with Poa annua L. In order to clarify this taxonomic ambiguity, morphological features and ecological properties of plants referable to P. crassinervis were compared with those of P. annua on the basis of many new collections and field observations. it was confirmed that P. crassinervis can be recognized as different from P. annua in the architecture of panicles (Figs. 2-3), shape of second glume and lemma, and usually in the hairiness of nerves of lemma (Figs. 4-5). Habitats of these plants were quite similar. In their mixed colonies, plants with the intermediate morphological features were often found. The intermediates were highly sterile and were morphologically separable from both P. crassinervis and P. annua, and delimitations of these species were not obscured by the presence of the intermediates. A conclusion was drawn that "Poa crassinervis" may be a good species. A distribution map of this species was presented (Fig 6).

View full abstract

From the floristic regional point of view, horizontal distribution in Hokkaido belongs to "Sino-Japanese Region". Fagus crenata, characteristic species of Japanese temperate zone, distributes up to the Kuromatsunai Depression in south-western part of Hokkaido and further northern part of Hokkaido is covered by mixed forest. The mixed forests in this region, together with those in southern Sakhalin, the Amur, the Usuri, the North East District of China, and North Korea, are called "The mixed forest zone of the Far East" (TATEWAKI, 1955-57) or "Tatewakia". The mixed forest are categorized into "Hokkaido-N. E. China subregion", as a subregion of "Sino-Japanese Region" (MURATA, 1977). This study has been carried out in the Nopporo natural forest (Fig. 1), which belongs to the mixed forest zone, and has phytogeographically analyzed forest stratification. The method to classify forest stratum is developed by WATANABE (1985), based on the "Synusia" concept by H. GAMS (1918) and K. IMANISHI (1937). 7 Synusias (M1-M5, N,L-Sy) for the arboreal species and one Synusia (H-Sy) for the herbaceous species are discriminated, respectively. The determination of the Synusia for certain arboreal species is made as follows; firstly, calculate the ratio of the maximum tree height of a given species is made as follows; firstly, calculate the ratio of the maximum tree height of a given species within the stand to maximum tree height of the stand, and call it a coefficient of the rank of stand precedence (CRSP); secondly, calculate CRSP for each species, utilizing the data from 581 climax forests in Hokkaido; and finally based on the criteria shown in Table 1, the Synusia for each individual species is determined. From the phytogeographical point of view, the arboreal species in the Nopporo natural forest mostly consist of Sino-Japanese, Japanese, N.E. Asiatic elements, except for one species. N. E. Asiatic element dominates in M1-Sy, and Sino-Japanese and Japanese elements are more emphasized in the Synusias of lower strata. H-Synusia of herbaceous plants is characterized by wide distribution elements such as Circumboreal, Eurasiatic, et al., accounting for 27%. In this way, the difference in the composition of distribution elements between arboreal and herbaceous plants becomes more evident (Fig. 2). This should reflect the basic difference in the structure of the forest. Among the stands including whole species, the difference in the composition is not clear, and each stand shows relatively high ratio of Sino-Japanese element, ranging from 30% to 35% (Fig. 3). However, by analyzing the composition of distribution elements using coverage value for each Synusia, diverse distribution patterns for each stand and the stratification of the forest can be clearly identified (Fig. 4). Furthermore, compared with the forest vegetation, majority of the plants in marsh, swamp and pond consist of wide distribution elements, and they differ in phytogeographical pattern.

View full abstract

Frequent occurrence of intermediate forms and poor knowledge on the variability of characters have caused some difficulties in the taxonomy of Trapa in Japan. Thus I made a preliminary analysis on the variation of nuts collected from 21 populations in the Southwestern Japan. Attention was paid to some morphometrical characters of the nut and development of lower spines or "pseudohorns". Each population usually contained different forms of nuts. Among them, however, several entities could be recognized based on the nut shape as follows. 1) Two-spined form: This included nuts of middle size (width 30-50 mm) and ones of big size (width over 45 mm). In case of former ones (Fig. 6-B, C, D) the nuts with pseudohorns of varying degree of development usually occurred together within one population and even on a single plant. I propose to treat them as one taxon, Trapa japonica sensu OHWI (1965), without inventing varieties. But at the same time it was remarkable that the tendency of development of pseudohorns was apparently different from population to population. The bigger ones included two types, that is, one without pseudohorns (Fig. 6-F) and the other with pseudohorns (Fig. 6-G). The former one may be identified as T. bispinosa ROXB., but the latter one has not been described in literature. 2) Four-spined form: The nuts of small size (width of about 20 mm, Fig. 6-A) were well definable and thought to be T. incisa SIEB. et ZUCC.. The nuts of bigger size showed some variations with respect to their size and/or stoutness of lower spines. The big ones (width over 45 mm, Fig. 6-H) may be treated as one taxon, T. natans or its variety. The nuts of middle size (Fig. 6-E) have been named T. natans var. pumila NAKANO. but so far as present materials were concerned, its entity seemed dubious. They might be immature nuts of bigger ones. The different patterns of variation among populations were ascribed to genetic differentiation. Predominance of self-pollination and isolation of habitats were thought to promote genetic isolation and preservation of genetic variations which occurred in each population. But the possibility of hybridization can not be excluded.

View full abstract

The genus Rhodiola, consisting of 49 species, is one of the representative genera dominating in the high altitudes of the region from Pamir-Alai to SW China (mainly Yunnan and Szechuan) trough Himalaya, Tibet and N Burma. Actually excepting 6, the remaining 43 species of Rhodiola are distributed in that region (Fig. 2-13). I noticed three remarkable distribution patterns common among Rhodiola. These were named as Aa, Bb and c, tentatively. The Aa pattern is characterized by the shape that is along the east or the west margin of the Central Asiatic highland. In the Ice Ages the flora of circum boreal or arctic regions might migrate southward (A) to the Himalayan-Tibetan-SW Chinese high altitudes through these margins and vice versa (a). So that these margins of the Central Asiatic highland are regarded as corridors of both boreal or arctic and Himalayan-Tibetan-SW Chinese floras in the past. I have named the Central Asiatic highland corridor to it (OHBA in press). In the case of Rhodiola, the Aa pattern is found in the species of the subgenera Clementsia, Crassipedes and two sections, Rhodiola and Chamarehodiola. It might be considered that the Aa pattern is given rise by the causes such as continentral drift or climatic changes in global level. The Bb pattern can be defined as the east-west extention limited in the region from Pamir-Alai to SW China through Himalaya, Tibet and N Burma. This pattern is characterized by the migration from west to east (B) and vice versa (b). The c pattern is related to the migration and remnant during the Ice Ages.

View full abstract

1. Cytological and cytogenetical characteristics of the genus Chrysanthemum in Japan were summarized. 2. From the cytological viewpoint, some taxonomical and phytogeographical problems were raised here, that is: (1) In Kyushu diploid species with white-ray flowers which is considered as Ch. makinoi could not be found. At lease at present, Kyushu should be excluded from the geographical distribution of Ch. makinoi. (2) There are two polyploids, tetraploid with 2n=36 and hexaploid with 2n=54, in the Ch. indicum group. The hexaploids have been mainly found in the northern part of Kyushu and the eastern part of Chugoku district. This distributional pattern differs from that of Ch. aphrodite which is known as a hexaploid and similar to Ch. indicum. Interrelationships between these two species and taxonomical revision on this complicated Ch. indicum group need further investigations. (3) Ch. weyrichii was considered as a member of Ch. zawadskii by KITAMURA (1981) and described its chromosome number as 2n=54. However the present authors have counted 2n=72 chromosomes in the materials from Oshoro, Hokkaido (unpublished). So far as those found in Japan is concerned cytologically, Ch. weyrichii differs from Ch. zawadskii. 3. From the viewpoint of their crossabilities, the genus Chrysanthemum in the sense of KITAMURA (1940) is considered to be the most acceptable treatment of the grouping. The subdivided system by KITAMURA (1978, 1981) can also be accepted because hybrid sterilities and differences in karyotypes exist between some species in the genus. 4. some questions in the list of chromosome numbers on Chrysanthemum compiled by FEDOROV (1974) were pointed out as follows: (1) 2n=19 for Ch. rupestre as a report by KITAGWA & NAGAMI (1960) might be 2n=18. (2) 2n=54 for Ch. zawadskii as a report by SHIGENAGA (KITAMURA 1957) should be n=9 for Ch. z. var. latilobum. (3) The species name Ch. sibilicum which was reported as 2n=54 by SHIMIZU (1958a) should be Ch. zawadskii var. latilobum. (4) 2n=18 for Ch. hakusanense reported by ISHIKAWA (1916) should be n=27. (5) 2n=54 for Ch. ornalum as a report by DOWRICK (1953) might be 2n=72.

View full abstract