甘藷是世界重要經濟作物之一。本試驗藉由甘藷‘台農27號’(TN27)×'Nancyhall' (NH)及‘台農57號’(TN57)ב清水紫心’(CP)兩組正、反交，獲得產量與品質之母系效應資訊。兩組正、反交於同一地點、分兩年完成，調查TN27×NH119株、NH×TN27112株、TN57×CP200株以及CP×TN57200株之經濟性狀之遺傳性、及性狀間之相關性。結果顯示上述4個F1分離族群各性狀之平均值及變異係數均有相當差異。地上部重及塊根重具有母系效應，而地上部重、塊根重及粗蛋白則具有雜種優勢。地上部重，塊根重及塊根數之分布圖則出現偏向低產量特性之歪峰分布。地上部重、塊根重及塊根數之間、塊根形狀與地上部重及塊根重之間、皮色及肉色之間、以及澱粉含量及amylose含量之間均呈現顯著正相關。澱粉含量、amylose含量與糖含量及粗蛋白含量之間有顯著負相關，可藉由選拔低澱粉及amylose含量改良糖含量及粗蛋白含量。選用母系親本可以有效改良甘藷品質及產量相關性狀。
Sweet potato is one of the most economically important crops in the world. This study obtains information on maternal effect on yield and quality traits of sweet potato under the following designed reciprocal crosses: 'Nancy Hall' (NH) × 'Tainung 27' (TN27), and 'Chingshey purple' (CP) × 'Tainung 57' (TN57). The reciprocal crosses were conducted at the same location in different years. Four populations of 119, 112, 200 and 200 plants form NH×TN27, TN27×NH, CP×TN57, and TN57×CP crosses, respectively, were used to investigate the inheritance of commercial traits. The correlations between traits were also analyzed. The results indicate that substantial differences of mean and coefficient of variation in the measured traits existed in the F1 populations. Maternal effects on top weight and root weight between progeny groups have been observed using NH and TN57 as the seed parents. Top weight, storage root weight, and crude protein content of the hybrids were heterotic. The distributions of top weight, weight and number of storage roots were skewed toward low yield characteristics. Significant positive correlations were detected among top weight, weight, and storage root number, between storage root shape and top weight and storage root weight, between skin color and flesh color of storage root, and between starch content and amylase content. Selecting lower starch or amylose content may be useful to improving sugar or crude protein content in view of the significant negative correlation between these traits. Choosing a maternal parent of the breeding hybrid for the yield and quality should effectively improve the selection scheme.
