中国梅花草属植物的叶表皮特征及其系统学意义

利用光学显微镜和扫描电镜对梅花草属Parnassia 30种植物的叶表皮进行了观察。结果表明:气孔器普遍存在于叶的下表皮,少数种的上表皮也有分布,均为无规则型。叶表皮细胞形状为多边形或不规则形;垂周壁式样可区分为近平直、浅波状和波状。在扫描电镜下,叶表皮气孔器外拱盖内缘为近平滑、浅波状或波状;一些种的保卫细胞两端有加厚;角质膜条纹状,有的条纹隆起,有的条纹上附有颗粒或小孔穴。气孔器类型及下表皮细胞形状的一致性表明梅花草属是一个自然分类群;sect. Saxifragastrum叶表皮特征具有多样性显示该组可能是一个复合群;突隔梅花草P. delavayi属于subsect. Xiphosandra,其气孔下陷,与其细胞学特征相似,支持独立为一组;此外,气孔器的分布、保卫细胞两端加厚、气孔器外拱盖内缘形态以及角质膜等特征对该属部分种的区分有一定的参考价值。

Comparative morphology of leaf epidermis in Parnassia (Parnassiaceae) from China

Epidermal characters of mature leaves in 30 species representing all the nine sections of Parnassia (Parnassiaceae) were investigated under both light microscope (LM) and scanning electron microscope (SEM). The stomata were anomocytic and existed on abaxial epidermis in all the species examined, and on the adaxial epidermis in some species. The leaf epidermal cells were usually irregular or polygonal in shape. The patterns of anticlinal walls were slightly straight, repand or sinuate. Under SEM, the inner margin of the outer stomatal rim was nearly smooth, sinuolate or sinuous, and the cuticular membrane of the leaf epidermis was striate, sometimes striate to wrinkled, occasionally granular or foveolate. Stomatal and other epidermal features in Parnassia appear to be constant within species, and thus can be used for distinguishing some species. Leaf epidermal features show that Parnassia is a quite natural genus. The previous reports that the stomata are anomocytic and occur only abaxially in Parnassia, yet occur both adaxially and abaxially in Lepuropetalon are not confirmed by this study, which, based on more extensive study, has shown that some species of Parnassia also exhibited stomata on both adaxial and abaxial sides. Evidence from leaf epidermis, together with that from floral anatomy and pollen morphology as well as biogeography, suggests that section Saxifragastrum may be heterogeneous. Parnassia delavayi should be treated as a section of its own, i.e. section Xiphosandra based on morphological, cytological and leaf epidermal data.