利用主成份分析法探討水稻穗重與其構成因素間的關係

Abstract

利用32個供試稉稻品種(系)以主成份分析法探討穗重與其六個構成因素間的關係，依據萃取之前四個主成份的分析結果，並由第一主成份及第二主成份座標平面上品種材料描點之分佈圖，依各種穗重高低程度及其在圖上距離遠近的分佈，可將參試品種區分成六個品種群，第I群至第VI群，群間差異明顯，且一次枝梗平均粒數，二次枝梗數、二次枝梗總粒數、每穗粒數及千粒重愈高的品種，穗重有愈重的趨勢。另相關分析中則發現，除了二次枝梗數外，穗重與其餘5個構成因素間皆呈現正相關關係，此一結果與主成份分析法探討所得部份不同，而以後者較能探究及掌握穗重與其構成因素間的訊息。 由於穗重與單株量間有正相關關係，卻與穗數間為負相關關係，因此若選育高穗重品種作為提升水稻產量之育種目標時，仍須維持一定的單位面積穗數，方能達到目的。 Thirty two rice varieties (lines) were used as materials in this study, and the principal component analysis was performed to elucidate the relationships between the panicle weight and its six components. According to the results of analysis, the first four components were extracted for study. From the scatter diagram of 32 varieties (lines) on plane defined by the first and second principal components, classification of rice varieties (lines) was made based on the value of panicle weight and the distance of varieties (lines) on the diagram. All varieties (lines) were classified into six groups: I-VI. The distinction among groups was very clear. The higher average number of spikelets per primary branch, number of secondary branches, total number spikelets of secondary branches, number of spikelets per panicle and weight of thousand grains of varieties (lines) were, the higher their panicle weight were. Positive correlations between panicle and its components were found, except number of secondary branches. The conclusion from correlation analysis was a little different from that of the principal component analysis's. The latter was superior for the study of the relationships between panicle weight and its components. The panicle weight was positively correlated to grain yield per hill, but it was negatively correlated to panicle number. Therefore, if we want to increase rice yield by breeding varieties with high panicle weight, the panicle number per unit area has to be kept not too low.