Study On Mathematic Classification Of Morphologic Character About 48 Species In Rubus

In this study, The mathematic classification of morphologic character was studied using Rubus materials from Sichuan, in order to make a classification in Rubus with the new technology. Thus it would be helpful for better understanding classification of Rubus, and it also would be helpful for further developing and utilizing resources of Rubus. the major methods and results are as follows.(1) 36 morphologic characters including 13 dualistic characters,16 multi-characters and 7 quantitative characters, were selected and analyzed using SPSS 17.0. Squared Euclidean distance coefficient was used in case clustering and Pearson correlation was used in variable clustering by within-groups linkage.(2) Results obtained through Q cluster analysis showed that the 48 species of Rubus were divided into four caboodles, which were Sect. Idaeobatus, Sect. Malachobatus, Sect. Dalibardastrum and Sect. Cyclactis by case clustering. The different groups were formed under four caboodles and the groups were similar to the traditional classification. Squared Euclidean distance coefficient among the different species of Rubus show that Sect. Idaeobatus had the closest genetic relationship with Sect. Malachobatus, then Sect. Dalibardastrum and Sect. Cyclactis, while Sect. Idaeobatus had the farthest relationship with Sect. Cyclactis. The conclusion was identical with the morphology classification, However, the results showed R.pinfaensi was ascribed to Subsect. Stimulantes were supported by their similarity in the study. In addition, there were dispute between R. ellipticus and R. ellipticus var. obcordatus, it was also between R. inopertus and R. inopertus var. echinocalyx.(3)Results obtained through R cluster analysis showed that the characters were decided in different groups. Size of stipules, accrete or sejunct of stipules, hollow or solid of fruits, leaf types, inflorescence types and inflorescence types had the bigger correlation coefficient among the 36 characters.(4)In principal component analyses (PCA), the accumulative contribution of the fifth principal component just up to 60.948%, which showed that some important characters can be used in classification of Rubus, such as Size of stipules, accrete or sejunct of stipules, hollow or solid of fruits, leaf types, inflorescence types and inflorescence types. However, foliage length/width, petiole length, thorn on petiole, colour of crown and Style of aggregate fruits etc have a small contribution on classification. The results of PCA were in according with that in case clustering, which reflected that we should pay more attention to choose character in classifying.