烈香杜鹃为常用藏药, 具有止咳、 祛痰、 平喘、 清热解毒、 健胃消肿之功效, 藏医常用于治疗类风湿性关节炎, 多数为野生药材; 为有效鉴别烈香杜鹃的真伪、 产地和品质差异, 采用红外光谱法对13个不同产地的烈香杜鹃进行了红外光谱图的识别分析。 红外光谱的扫描范围为4 000~400 cm-1, 实验发现其红外光谱相似; 对红外吸收谱带归属进行判别分析, 建立了烈香杜鹃红外指纹图谱, 其特征吸收峰位于3 404, 2 921, 2 852, 1 734, 1 625, 1 449, 1 374, 1 266, 1 060和534 cm-1等处; 在1 517, 1 316, 1 161, 825, 779和594 cm-1附近, 峰数目、 峰位置与峰强度有差异。 计算共有峰率和变异峰率, 建立共有峰率和变异峰率双指标序列, 利用SPSS软件对其红外图谱数据进行聚类分析。 双指标序列分析法与聚类分析法的分析原理与角度不同, 分组结果基本一致, 表明这两种方法可靠, 可用于分析烈香杜鹃的产地与品质差异。 双指标序列分析法比较结果显示, 样品的共有峰率≥68.75, 变异峰率≤27.27, 分组结果表明, 产地接近、 气候条件与生长环境相似的烈香杜鹃之间共有峰率较高; 而产地、 气候条件与生长环境相差较大的烈香杜鹃之间变异峰率较高。 聚类分析结果显示, 当欧氏距离为15时, 可聚为三大类, R2, R3和R4为一类, R7, R8, R10, R11和R12为一类, 剩下的归为一类; 当欧氏距离为20时, 聚为两大类, R2, R3和R4为一类, 剩下的归为一类; 当欧式距离为25时, 13个产地的烈香杜鹃聚为一类。 将聚类分析结果与使用ArcGIS软件所做的烈香杜鹃采样点分布图结合起来分析, 能直观地看出烈香杜鹃品质与其产地的关系。 综合以上分析, 建立的烈香杜鹃红外指纹图谱, 结合双指标序列分析法和聚类分析法, 可以为鉴别烈香杜鹃的真伪、 产地和品质差异提供快速、 有效的新方法。

Rhododendron anthopogonoides Maxim., a traditional Tibetan ethnodrug, has been used for antitussive, expectorant, antiasthmatic, heat-clearing and detoxicating, stomachic and swelling agent for a long time, besides, it is also commonly used for treating rheumatoid arthritis. And most of them are wild. Thus, in this study, R. anthopogonoides from 13 different regions were identified in the range of 4 000～400 cm-1 by adopting infrared fingerprint (IR) in order to identify the adulterants, regions and quality of this herbal medicine effectively. Results showed that the infrared spectrum of the samples are similar. And the main IR absorption peaks of the samples were identified and assigned. Then, the fingerprint of R. anthopogonoides was established and the characteristic peaks were at 3 404, 2 921, 2 852, 1 734, 1 625, 1 449, 1 374, 1 266, 1 060, 534 cm-1. However, there were still some differences in the number, position and intensity of the characteristic peaks at 1 517, 1 316, 1 161, 825, 779, 594 cm-1. Moreover, the common peak ratio and variant peak ratio dual-indexes sequential were also calculated and established, respectively. In addition, the cluster analysis was used to analyze the fingerprint data by using SPSS software. What’s more, the grouping results of sequential analysis of dual-indexes and cluster analysis were nearly the same although the analysis principle and aspect of the two methods were different. And it was shown that the two methods are reliable and can be used to analyze the differences in regions and quality of the samples. Results also showed that the common peak ratios of the samples are ≥68.75, and the variant peak ratios are ≤27.27. The common peak ratios are higher when the samples grow in the closer regions with the similar climatic conditions and growing environments, while variant peak ratios are higher when the samples grow in the farther regions with the different climatic conditions and growing environments. The results of cluster analysis showed that when the Euclidean distance is 15, the samples can be clustered into three categories, where R2, R3 and R4 are one class, R7, R8, R10, R11 and R12 are another class, and the rest are classified into the last class. When the Euclidean distance is 20, the samplesare divided into two categories, where R2, R3 and R4 are one class, and the rest are classified into the other class. When the Euclidean distance is 25, the samples from 13 regions are grouped together. So, the relationships between the quality of R. anthopogonoides and their origins can be summarized intuitively by combining the results of cluster analysis with the figure of sampling plots’ location made with ArcGIS. To sum up, fingerprint combined with sequential analysis of dual-indexes and cluster analysis provides a new method which is effective and rapid for the identification of R. anthopogonoides with the adulterants, regions and quality.