浙江杜鹃花属植物种子形态及其分类学意义

SEED MORPHOLOGY OF RHODODENDRON l.(ERICACEAE )FROM ZHEJIANG AND ITS TAXONOMIC SIGNIFICANCE

本文对浙江产杜鹃花属１３个种（录属５个亚属７个组）的种子形态进行了研究。比较观察了种子形状、大小、扁化率，翅和表面结构等形态学特征，井对该属种子形态的演化趋势、种于类型的划分及种子形态在分类学上的意义进行了讨论。结果表明种子形态特征与植物习性和形态特征一定程度上相关，反映了周内类群的分化，因而种子形态在杜鹃花属的系统研究中是不可忽视的性状。

This paper studie the seed morphdogy of 13 species of Rhododendron from Zhejiang.Thespecies are in 5 subgenera and 7 sections.Under stercomicrosCope and SEM,seed characters including shape,size,size index(=length×width×thickness),flat rate[=(Width-thickness)÷width],wing and surface morphology areexamined.The seed morphology of these 13 species has been devided into two types and six sub-types:Type 1,winged type,the seeds winged,7 species in all,their flat rates above 60%,thesurface vvith many conspicuous longitudinal grcove.These 7 species have been devided into 3 sub-types according to the morphology of their wings and surface cells,Type 2,unwinged type,theseeds unwinged,6 species in all,their flat rates smaller than 60%,the surface cells are shorter,thesurface with shallow grooves,protruded occationally.These 6 species have been further devided in-to 3 subtypes.The results in this paper show that the evolution trend of seed morphology of Rhododendron ismost probably as follows:1.seed size:large→small;2.Wing:well-developed→not well-devel-oped→unwinged;3.flate rate:high→low;4.surface cell:long→short;otitside wall of surface cell:falling into inside→plain→protruded.Among the 13 species,the seed morphology of R. molle is similar with that of R,fortuneiwhich belongs to the original group while it is quite different from that of species of subgenus Tsut-susi.It is most probably that there is close relationship between R. molle and subgenus Hymenan-thes.The seed of R. huadingense is large, without grcove on the surface,and the surface cell of theseed is of short and protruding somewhat;all these are regarded to be very specific characters andare conspicuously different from the other species of section Brachycalyx;so it seems to be morereasonable that R. huadingense can be regarded as a subsection of section Brachycalyx.