In this paper, 21 species representing 13 genera were studied by means of humerical taxonomic methodes. One geographical and 52 morphological characters were used. The correlation coefficients were computed by standardized data, and the various clustering methods were performed on the correlation matr x. The UPGMA clustering method was selected as the optimal one and its results were shown in the form of dendrograms. We present a simple method to construct the joint and broken lines by which the boundary of the genera, subtribes and tribes in the dendrogram is determined. By means of numerical taxonomic methods, we can easily work out a systematic dendrogram and the following taxonomic treatments are easily proposed: (1) Sasamorpha sinica (Keng) Koidz. should be reffered to the genus Sasa Makino & Shibata. (2) Pseudosasa amabilis (McClure) Keng f. should belong to the genus Pseudosasa Makino and should not be referred to the genus Arundinaria Michaux. (3) The genus Brachystachyum Keng should be considered as a separate one. (4) The genus Pleioblastus Nakai should not be combined with the genus Arun-dinaria Michaux, but kept as an independent one.

The genus Burmannaia is one of the largest genera in the Burmanniaceae, of which 12 species have been recorded in China. It is mainly a tropical genus. The species in China are all confined to the region south of Yangtze River. They are distributed chiefly in the provinces Guangdong (9 species) and Yunnan (6 species). After having studied the areas of all the species in China, we are able to classify them into following 4 area-types: 1. Area-type of Tropical Asia to Tropical Australia. The two non-saprophytic species (Burmannia disticha, B. caelestis) and one saprophytic (B. championii) belong to this area-type. It is an ancient type. The plants of this type mostly have a wide ecological amplitude, for example, B. disticha may be found in tropical and subtropical regions. The plants occur not only in evergreen forests, in bushs, but also in rather arid herbosa and on the side of streams (Fig. 2). 2. Area-type of Tropical SE Asia. In the type are 3 saprophytic species i.e. B. oblonga, B. wallichii and B. nepalensis. 3. Area-type of E. Asia. (Fig. 3) Burmannia in China with E. Asian distribution is poor in species. There are only 2 saprophytic species. B. cryptopetala is distri-buted in Haina (China), Kyushu and Honshu (Japan); B. itoana occurs in Taiwan (China), Riukiu and Kyushu (Japan). They are known only on the islands of E. Asia. Such a pattern of distribution may suggest connection of these islands once in the prehistoric time in spite of their present isolation. 4. Endemic area-type. (Fig. 4). Here are 3 saprophytic species and one variety with green leaves. B. nana occurs only in E. Taiwan. One of the two new species described by present author in this paper, B. fadouensis, is known from Xichou Xiao, S. E. Yunnan, to Longzhou Xian of the province Gaunxi; the other one, B. pingbienensis occurs only in Pinbien Xian of S. E. Yunnan. The last species is endemic to China. B. pusilla var. hongkongensis is non-saprophytic and known from the province Guangdong and its bordering islands. Both B. fadouensis and B. pingbienensis are characterized by the axillary bulbils, which enable them to adapt to rather arid and cold condi-tions in northern part of the tropical region.

The pollen grains of 33 species representing 11 genera of the family Berberidaceae, mostly from China, were examined with the light microscope and scanning electron microscope. Their characteritic details can be used for generic diagnosis. A pollen key to the genera based on these observations is presented. Based on the morphology, the pollen grains can be grouped into the following three types: 1. The tetrad pollen type found only in the genus Sinopodophyllum. 2. The spiraperturate pollen type found in the genera Berberis and Mahonia. 3. The tricolpate pollen type found in the genera Diphylleia, Jeffersonia, Nandina, Dysosma, Caulophyllum, Leontice and Epimedium. A diagnostic key to the pollen grains of genera in the Berberidaceae. 1. Pollen grains single 2. Pollen grains spiraperturate .................... Berberis L., Mahonia L. 2. Pollen grains tricolpate 3. Exine with spinose sculpture ........................ Diphylleia Michx. 3. Exine with non-spinose sculpture 4. Exine with striate or striate-reticulate sculpture ...... Jeffersonia Barton 4. Exine with reticulate sculpture 5. Exine around colpus with distinct thickening ........ Nandina Thunb. 5. Exine around colpus without thickening 6. Pollen grains larger (45—50)×(32.5—37.5)μ 7. Colpus with membrane ................ Dysosma R. E. Woodson 7. Colpus without membrane ................ Caulophyllum Michx. 6. Pollen grains smaller (25—550)×(20—527.5) μ 7. Pollen grains prolate-perprolate .................... Leontice L. 7. Pollen grains spheroidal-prolate...Epimedium L., Podophyllum L. 1. Pollen grains tetrad ............................ Sinopodophyllum Ying

The pollen grains of Cochlidiosperma (Veronica) hederifolia and C. (Veronica) cymbalaria were examined under SEM and TEM. They differ vastly from those of all the others in Veronica (sensu Elenevskij, 1977, 1978). The taxonomic relationship between the group and the other Veronica spp. is discussed with respect to both gross morphology and pollen characters. Justification for the restoration of the genus as a valid taxon is argued and, finally, a number of new combinations are proposed. The group under consideration was sometimes treated as a separate genus but more often given different ranks in the genus Veronica, which has been a controversial group as to its concept and subdivision. No detail discussion has been made before on the relationship between the group and Veronica. Main purpose of the present mork is to examine pollen morphology and gross morphology of the group and to discussits relationship with Veronica.

Diplopanax is a genus described by H. Handel-Mazzetti in 1933 according to a flowered specimen. It was originally considered as belonging to the family Araliaceae. But fruit characters of this genus such as very large in size, oblong-ovoid in shape, and containing only one seed in a locule which has a curved embryo and very hard when dry etc., differ greatly from all other araliaceous plants, and evidently resemble those of the cornaceous genus Mastixia Bl. In addition, the fruit is very similar to the cornaceous fossil genus Tectocarya F. Kirchh., from which it may be distinguished by the absence of 2 germinate pores in the endocarp. Consequently, the present author tends to consider that the genus Diplopanax Hand.-Mazz. is best placed in Cornaceae, not in Araliaceae.

The genus Palaeosmunda was established by R. E. Gould in 1970 based upon some Late Permian Osmundaceous trunks with well-developed leaf gaps and rhomboidal sclerotic ring within petiolar base seen in cross section. As he thinks that the latter character is more important than the former, this genus could not be assigned to any subfamily of Osmundaceae. However, the leaf gap is one of the most important characters in the structure of the fern stem, so the author suggests that this genus should be assigned to subfamily Osmundoideae and its diagnosis must be emended as follows: The genus Palaeosmunda is represented by some rhizomes (or trunks), roots and leaf bases of ferns which structurally are preserved, resembling Osmundacaulis but which can’t be assigned to any group of this genus. Stem containing an ectophloic dictyoxylic siphonostele; if tracheids present in the pith, they being multiseriate scalariform pitted; pith or cortex sometimes contain ing groups of secretory cells or sclerenchyma; number of leaf traces seen in a tran sverse section of cortex more than 30; leaf traces adaxially curvature, rarely oblong shaped; petiolar bases with or without stipular expansion, containing a C-shaped vascular strand; root diarch. Type species——Palaeosmunda williamsii. According to this diagnosis some primitive osmundaceous species with the leaf gaps, which have already found in Upper Permian and Lower Triassic, could be assigned to this genus. Two of them are P. williamsii Gould and P. playfordii Gould, and Osmundacaulis beardmorensis, which was from Lower Triassic of Antarctica in 1978, should be assigned to the genus Palaeosmunda. In this paper two osmundaceous new species: P. primitiva and P. plenasioides were found in the coal balls of Upper Permian age from Wangjiazhai of Shuicheng of Guizhou Province, China. P. primitiva is represented by two trunks; stem about 4 cm in diameter; stele actophloic dictyoxylic siphonostele; pith cavity about 3—4 mm in diameter, contianing parenchyma and tracheids; xylem cylinder thin, less than 10 tracheids in radial thickness, dissected by leaf gaps. Inner cortex about 1.5 cm thick, mainly parenchymatous, but sometimes containing a few sclerenchymatous; number of leaf traces seen in a transverse section about 50—60; leaf traces departing at 35—45º,open C-shaped at point of departure, gradually becoming shallow C-shaped or V-shaped in different parts; protoxylem in base of leaf traces single, endarch; when leaf traces pass through inner cortex, protoxylem biturcating. Petiole bases without stipular expansion, probablyloosely embracing the stem; xylem strand of potiole trace shallow C-shaped, surrounded by selerenchyma; sclerotic ring round, connected with single sclerenchyma mass in the concavity of the petiole trace. Root arising singly from leaf trace, diarch, with inner and outer cortex. P. plenasioides is represented by a rhizome; stem more than 4 cm in diameter; stele actophloic dictyoxylic siphonostele; xylem cylider with about 20 tracheids in radial thickness, dissected by leaf gaps; xylem bundle U-, O-, or crosier- (i.e. query-) shaped; pith and inner cortex parenchymatous, with many groups of secretory cells; leaf trace C-shaped, its base containing two endarch protoxylem groups; root diareh,with inner and outer cortex, arising singly from leaf trace or its base.

In the present paper, the somatic chromosome numbers of 2 varieties of Actinidia chinensis Planch. are counted as follows: A. chinensis var. chinensis 2n=58(2x), A. chinensis var. hispida 2n=ca. 174(6x). Both numbers are reported for the first timefor the species.

Three new combinations, one new synonym, 3 unperfectly known species and 8 distribution maps of 11 species of Chinese and Indo-Chinese Archidendron are presented in this article, as a supplement of the subject “Notes on the genera Archidendron F. V. Mueller and Pithecellobium Martius in Mainland S. E. Asia” published in Adansonia, ser. 2, 19(1): 3—37. 1979. I am indebted to prof. Wu, head of Taxonomy laboratory of South China Institute of Botany for translating the article into Chinese and adding some distributivepoints of Chinese species of this genus on the maps.

The tuberous root of Hemsleya, namely “jin-gui-lian”, is commonly used as a traditional Chinese medicine. In recent years, it is chiefly used for curing flatulence,diarrhea, dysentery, cardialgia, bronchitis, uteritis with a satisfactory effect. In order to make sure of its medical effect and explore its abundant resources, we have made a survey of the genus Hemsleya in Sichuan Province. The genus is mainly distributed in the southern and southwestern parts of China, with a concentration in Sichuan Province. According to our preliminary survey, there are 9 species and 6 varieties (including 5 new species and 5 new varieties) in Sichuan Province. A key to species and the descriptions for the new taxa are provided in the presentpaper.

The present paper aims at validating a number of new taxa of the fern families Aspidiaceae, Lomariopsidaceae and Bolbitidaceae from tropical regions of China as part of Flora Reip. Pop. Sin. Vol. 5. The types of all taxa described here are preserved in the Herbarium,Institute of Botany, Academia Sinica, Beijing (PE).