【正文】 . 1984. The CIMMYT maize advanced unit. in . UniversitiesCIMMYT Maize Conference on Collaboration toward Nutual LDC Production Objectives. Mexico, August 914, 1984. . 15) Peng Zebin and Chen Zehui, 1993. Current status of maize hybrid breeding and its strategies in China. Proceedings of the Fifth Asian Regional Maize Workshop. Hanoi, Vietnam. November 1520, 1993. .16) Pollak L. 1997. United states LAMP final report, in Final Report LAMP, .17) Rood . and . Major, 1980. Responses of early corn inbreds to photoperiod. Crop Sci. 20:679682. 18) Russell, . and C. W. Stuber, 1983. Inheritance of photosensitivity in maize. Crop Sci. 23:935939.19) Russell, . and C. W. Stuber, 1985. Genotypephotoperiod and genotypetemperature interactions for maturity in maize. Crop sci. 25:152158.20) Salhuana W., et al. 1997. Final Report LAMP. .21) Struik P. C. et al. 1986. Environmental effects on flowering characteristics and kernel set of maize (Zea mays L.). Netherlands J. Agric. Sci. 34:469484.22) Taba, S. (ed.). 1995. Maize Genetic Resources. Maize Program Special Report. Mexico, D. F., CIMMYT.23) Taba, S. 2002. Maize Genetic Resources: Conservation and Utilization. 23) in: External Review of the CIMMYT Maize Program. 2002, 9. Mexico, D. F., CIMMYT.24) Troyer, . and . Brown, 1972. Selection for early flowering in corn. Crop Sci., 12:301304.25) Troyer, . and . Brown, 1976. Selection for early flowering in corn: Seven late synthetics. Crop Sci., 16:767772.26) Troyer, . 1978. Corn yield as influenced by flowering date and season. Proc. Annu. Corn Sorghum Res. Conf., 33:115.27) Troyer, . and . Larkins, 1985. Selection for early flowering in corn: 10 late synthetics. Crop Sci., 25: 695697.28) Troyer, . 1986. Selection for early flowering in corn: 18 adapted F2 populations. Crop Sci., 26:283285.Utilization of tropical and subtropical germplasm in hybrid maize breeding programsS. H. Zhang, D. Q. Shi and L. BaiInstitute of Crop Breeding and Cultivation, Chinese Academy of Agricultural Sciences, Beijing, 100081, P. R. China Germplasm is critical for continued maize breeding activities in the ing century. Lacking of germplasm resources is the ‘bottle neck’ of hybrid maize breeding programs in China. Five predominant hybrids covered 53% of the total maize acreage, and more than 60% of the maize area relied heavily on five inbred lines in 1995. Tropical and subtropical germplasm should be used to broaden the genetic background of hybrid maize breeding programs. Selection for improvement of adaptation should be done to adapt tropical and subtropical germplasm under longday conditions in temperate. Mass selection is usually effective for this purpose. The adaptation of tropical and subtropical maize in temperate is always related with early flowering, and the adaptability of maize genotypes is associated with the photoperiod sensitivity under longdays in the temperate. The CAAS introduced 2 subtropical pools from CIMMYT, and improved them for early silking using biparental mass selection method for 4 cycles under longday condition. The photoperiod sensitivity was calculated using the formula: RD(%)=[(LS)/S]100,where L is the leaf number under longdays, S is under shortdays. RD was reduced from and to 23 and in the improved populations respectively. Four cycles of biparental mass selection for early silking increased grain yields significantly in both populations, with average gains of 627 and 625 kg/ha/cycle respectively when evaluated under longdays. As a sharp contrast, the yield responses to selection were minus when evaluated under shortdays. The minus responses under shortday conditions indicate that the yield gains of improved exotics under longdays should be due to the effects of desensitizing to photoperiod. The ears per plant, grain weight per ear, and harvest index increased as 4 cycles of mass selection. Biparental mass selection is an effective method to improve the adaptability of tropical and subtropical maize populations to longday conditions in temperate. 玉米的輪回選擇與群體改良張 世 煌（中國(guó)農(nóng)科院作物育種栽培研究所 農(nóng)業(yè)部作物遺傳育種重點(diǎn)實(shí)驗(yàn)室 AMBIOENT中國(guó)實(shí)驗(yàn)室 北京 100081）玉米的群體改良是通過(guò)對(duì)基礎(chǔ)群體進(jìn)行輪回選擇(Recurrent selection)，逐漸提高群體中有利基因的頻率，以改進(jìn)群體的表現(xiàn)，從而創(chuàng)造和改良玉米育種的基礎(chǔ)材料。被改良的對(duì)象可以是開放授粉品種、雜交種、綜合種、群體、復(fù)合品種或基因庫(kù)。改良的產(chǎn)物可以是選育自交系的基礎(chǔ)材料，也可以直接用于邊遠(yuǎn)地區(qū)的農(nóng)業(yè)生產(chǎn)。歐美工業(yè)化國(guó)家和玉米生產(chǎn)商業(yè)化程度較高的發(fā)展中國(guó)家通過(guò)輪回選擇方法創(chuàng)造育種群體等基礎(chǔ)材料，而許多發(fā)展中國(guó)家直接把改良群體用于農(nóng)業(yè)生產(chǎn)。可見(jiàn)，無(wú)論是在玉米生產(chǎn)商業(yè)化程度較高的國(guó)家還是在邊遠(yuǎn)不發(fā)達(dá)地區(qū)，用輪回選擇方法改良群體都可以發(fā)揮重要作用。在對(duì)玉米不同性狀的改良中，各種輪回選擇方法都有一些成功的例子。群體內(nèi)輪回選擇方法通常對(duì)改良加性基因作用的性狀和一般配合力有較好的效果，但隨著雜種優(yōu)勢(shì)的普遍利用，既能開發(fā)一般配合力又能開發(fā)特殊配合力的相互輪回選擇變得越來(lái)越重要，已發(fā)展成為當(dāng)今玉米群體改良的主流技術(shù)。第一節(jié) 數(shù)量遺傳學(xué)基本概念一、數(shù)量性狀（quantitative characters）。在群體中個(gè)體間呈連續(xù)變異并且能夠度量的性狀稱為數(shù)量性狀。栽培植物的許多重要經(jīng)濟(jì)性狀，例如產(chǎn)量、配合力、蛋白質(zhì)含量、含油量、植株高度和抗逆性等都是數(shù)量性狀。數(shù)量性狀通常由微效多基因控制，容易受環(huán)境影響。 對(duì)孟德爾式遺傳的質(zhì)量性狀來(lái)說(shuō)，表現(xiàn)型與基因型一致，因而被稱為簡(jiǎn)單遺傳性狀。但對(duì)于數(shù)量性狀來(lái)說(shuō)，群體內(nèi)個(gè)體間的差異一般呈連續(xù)分布，難以在個(gè)體間明確地劃分，所以很難根據(jù)個(gè)體的表型值判斷基因型。通常把群體的表型值描述為遺傳效應(yīng)與非遺傳效應(yīng)兩個(gè)成份之和，即： Pj=Gj+Ej通常我們只能度量個(gè)體或群體的表型值（Pj），而無(wú)法直接度量基因型值(Gj)和非遺傳效應(yīng)，即環(huán)境效應(yīng)(Ej)。數(shù)量性狀的表型值通常符合正態(tài)分布（normal distribution）。一般假定非遺傳效應(yīng)（Ej）呈現(xiàn)正態(tài)分布，表型值Pj也呈正態(tài)分布。在選擇理論中，通常認(rèn)為基因型值Gj也呈現(xiàn)正態(tài)分布。正態(tài)分布有兩個(gè)重要的特征參數(shù)，即群體平均數(shù)和方差。在數(shù)量遺傳學(xué)中經(jīng)常涉及這兩個(gè)參數(shù)，而且用于選擇理論。假設(shè)一個(gè)位點(diǎn)上有2個(gè)等位基因：A和a。雜交后代分離出3種基因型： Aa224。AA，Aa，aa。根據(jù)哈代溫勃格平衡定律（HardyWeinberg Law），在一個(gè)群體內(nèi)假設(shè)兩個(gè)等位基因的頻率是p和q ，則：p+q=1。在HardyWeinberg平衡群體中，一個(gè)位點(diǎn)、兩個(gè)等位基因的遺傳模型、基因型頻率和基因型值表述如下：基因型基因型頻率模型基因型值A(chǔ)Ap2m+2aaAa2pqm+a+bdaaq2ma群體均數(shù) X=(pq)a+2pqd . 加性方差 sA2=2pq[a+(qp)d]2. 顯性方差 sD2=4p2q2d2. 從上述三個(gè)公式看出，群體的平均值X、加性方差sA2和顯性方差sD2都受基因頻率的影響。而且遺傳方差中的加性方差具有線性特征，而顯性方差則表現(xiàn)為非線性方差。 基因型 aa Aa AA 頻率 q2 2pq p2圖1 基因頻率與基因型值的關(guān)系回歸線代表加性遺傳方差的平方和（即育種值），三種基因型值的點(diǎn)沒(méi)有落在回歸線上。顯性方差就是這些基因型值的點(diǎn)與育種值間的離差平方和。在全部遺傳方差中包括sA2和sD2兩部分。在一個(gè)基因位點(diǎn)上，既有sA2又有sD2，則遺傳方差 sG2=sA2+sD2。但實(shí)際上還常常包括上位效應(yīng)，即sI2。在群體的遺傳方差分量中，sA2是關(guān)系到選擇作用是否有效的最重要的遺傳方差成份，也是育種家希望得到的方差。在植物性狀遺傳中，加性方差s