林药复合种植是一种缓解中药资源减少及提高土地利用率的方法。 采用傅里叶变换红外光谱法对70份不同种植模式的滇龙胆进行研究, 通过Omnic8.0软件对原始光谱进行基线校正、 归一化、 二阶导数光谱预处理; 对样品与龙胆苦苷标准品的二阶导数光谱进行分析; 利用各类样品的平均光谱建立两个光谱数据库Lib1和Lib2, 其中Lib1在1 800～600 cm-1光谱范围内进行专家检索, Lib2在全谱范围内进行相关性和平方微分差检索。 结果表明, 样品的红外光谱较为相似, 难以直观分析鉴别; 样品与龙胆苦苷标准品具有多个共有峰, 其中1 611和1 076 cm-1为标准品的特征峰, 各样品在1 076 cm-1处峰形差异较小, 通过对1 611 cm-1处峰面积的比较, 种植于核桃林下的样品龙胆苦苷含量最高, 荒坡种植及与木瓜复合种植的样品有效成分含量最低; 不同种植模式样品中有效成分的不同表现为光谱匹配值之间的差异, 与核桃树、 旱冬瓜、 桉树、 茶树、 杉树复合种植的样品间匹配值的差异小于与木瓜复合种植及荒坡种植的样品, 专家检索法对样品的误判数为11, 相关性检索和平方微分差检索法对样品的误判数分别为4和9, 即在全谱范围内相关性检索法对样品鉴别效果最佳, 正确识别率为94.29%。 红外光谱法结合二阶导数光谱、 光谱检索对不同复合种植模式滇龙胆的鉴别效果较好, 为不同复合种植中药药源鉴别提供理论依据。

Multiple cropping systems are the way to cope with the shortage of traditional Chinese medicine (TCM) resources and enhance the utilization ratio of land. The aim of this study was to investigate seventy Gentiana rigescens Franch samples which were planted in slope and in multiple cropping system mixed with Juglans regia, Alnus nepalensis, Eucalyptus robusta, Alnus nepalensis, Camellia sinensis, Chaenomeles sinensis using Fourier transform infrared (FTIR) spectroscopy. The pretreatments such as baseline correction, normalization and second derivative of all the original spectra were performed with Omnic 8.0 software. Moreover, second derivative spectra of samples and gentiopicroside standard were analyzed. In addition, two spectral databases were set up based on the average spectra of every variety with Omnic 8.0 software. The first library (Lib1) which included the spectra region from 1 800 to 600 cm-1 had the algorithm of expert retrieve while the correlation retrieve and square differential retrieve were dealt with the second library (Lib2) in the full spectral range. The results indicated that it is difficult to identify the infrared spectra of samples directly because they had high similarities. There were multiple common peaks between spectra of samples and gentiopicroside standards. The absorption peaks of 1 611 and 1 076 cm-1 of gentiopicroside were the two main characteristic peaks in the second derivative spectra. The difference of peak shape of every sample in the wave number of 1 076 cm-1 was small. By comparing peak area at 1 611 cm-1, the samples planted with J. regia contained highest content of gentiopicroside while the content of this effective constituent was the lowest in the samples which were planted in the slope and in the multiple cropping systems mixed with C sinensis. The differences of effective components in the samples with different planting patterns were represented by the variances of spectral matching values. The difference of matching values of samples planted with J. regia, A. nepalensis, E. Robusta and A. nepalensis were lower than that with slope and C. Sinensis. The number of misjudgment of expert retrieval, correlation retrieve and square differential retrieve were 11, 4 and 9, respectively. The discrimination result of correlation retrieve was the best in the full spectral range and the correct recognition rate was 94.29%. Fourier transforms infrared (FTIR) spectroscopy combined with second derivative and retrieval had good ability to discriminate G. rigescens with different cropping systems. It could provide a basis for the discrimination of different planting patterns of TCM