Food, Agriculture & Environment; Vol.1(1), January 2003 52 Introduction The Prunoideae, a subfamily of Rosaceae, includes several species producing edible drupes with economic importance. In 2001, worldwide annual production of Prunoideae exceeded 28.3 million metric tons, including almost 13.5 million tons of nectarines and peaches (Prunus persica (L.) Batsch), 9 million tons of plums (Prunus domestica L.), 2.7 million tons of apricots (Prunus armeniaca L.), 1.8 million tons of sour and sweet cherries (Prunus cerasus L. and Prunus avium L. respectively) and 1.3 million tons of almonds (Prunus amygdalus Batsch = Prunus dulcis (Miller) D.A. Webb) (Table 1) 1 . Prunus species are characterized by developing only one ovary in which two ovules typically form; one of them degenerates soon after anthesis. The fruit is a drupe where the mature stony endocarp together with the seed forms a propagation unit comparable to a botanical seed surrounded by its protective testa.Breeding practices in Prunus must address the challenges resulting from the narrow genetic background of commercial cultivars 2 , a long juvenile period, and the differences in trait expression between juvenile and mature trees 3 . In the last decade, many techniques such as in vitro culture, and the use of molecular markers have become available for Prunus crop improvement. These methods are now moving from laboratory evaluation to field application. Simultaneously, different aspects of the utilization of DNA markers, including marker-assisted selection and genome mapping, as well as the impact of altered gene expression on Prunus spp., have now been extensively reviewed 3,6 . This article offers an overview of the current strategies being developed to improve the traits and germplasm in Prunus and to optimise breeding efficiency. These strategies include germplasm improvement, propagation techniques and use of molecular markers. New approaches to Prunus tree crop breeding Pedro Martínez-Gómez 1 , Gabriel O. Sozzi 2* , Raquel Sánchez-Pérez 1 , Manuel Rubio 1 and Thomas M.Gradziel 3 1 Departamento de Mejora y Patología Vegetal, CEBAS-CSIC, Murcia, Spain. 2 Cátedra de Fruticultura, Facultad de Agronomía, Universidad de Buenos Aires, Avda. San Martín 4453, C1417 DSE – Buenos Aires, Argentina. 3 Department of Pomology., University of California – Davis, One Shields Avenue, Davis, CA 95616, USA. *email: gsozzi@agro.uba.ar Received 2 September 2002, accepted 20 December 2002. Abstract New strategies for Prunus improvement, including germplasm and molecular marker development and improved propagation techniques, are described. In germplasm improvement, the introduction of genes from related Prunus species conferring several traits including self-compatibility, growth habit, drought resistance, and kernel quality are being pursued. Twin seeds (two embryos within the same seedcoat) have produced seedlings useful for genetic studies. Promising propagation methods include in-vitro techniques for the evaluation of plant material and in-vivo micrograft techniques that allow the early propagation of high-risk genotypes. In addition, the growth of seedlings in controlled environments, including the induction of an artificial rest period in cold chambers, provides a useful strategy for obtaining vigorously growing plants year round. Molecular markers have also become an essential tool in Prunus breeding studies. Different types of molecular markers, including isoenzymes, RFLPs, RAPDs, AFLPs and SSRs, have been employed for the genetic characterization of germplasm, the establishment of genetic relationships between cultivars and species, and the construction of genetic maps. Methodologies for the analysis of marker-assisted selection include the use of mapping populations segregating for desired characters and bulk segregant analysis. Key words: Fruits, germplasm, propagation techniques, molecular markers. Germplasm Improvement Related Prunus species: The available germplasm in Prunus is diverse and the origin and dissemination for several species have been extensively reviewed 7,11 . It is still possible to find considerable genetic variation for these species mainly in the mountainous areas of Central Asia from the Tian Shan region in China to Kurdistan, including Turkestan, Afghanistan, Iran and Iraq (Fig. 1). However, in cultivated germplasm, a limited gene pool restricts production to specific areas and conditions 2 . The introduction of genes from related species through interspecific hybridisation has been used in several breeding programs throughout the world to develop better-adapted cultivars and rootstocks. Rootstock breeding programs using interspecific hybridisation have introduced useful traits including size control, adaptation to the new environments and pest resistance. Interspecific crosses between Prunus species (primarily peach × almond, but also P. webbii × peach, and others) have been widely utilized in almond rootstock breeding in France 12 , USA 13 , Spain 14 , and Yugoslavia 15 . The introgression of almond germplasm from related species, including P. webbii (Spach) Vieh., P. argentea Lam, P. persica, P. bucharica Korshinsky, P. mira Koehne and P. scoparia Batal. (Fig. 1) has allowed transfer of several useful traits including self- compatibility, fungal and pest resistance, and frost and drought tolerance 7,16,18 . P. davidiana (Carr.) Frans. has recently been reported to be a source of plum pox virus (PPV) resistance for peach 19 , while the introgression of Prunus mandshurica (Maxim.) Koehne genes to apricot have improved frost resistance in Eastern and Central European programs 20 . In almond, the absence of extensive crossing barriers in either the initial hybridisation or subsequent backcrosses demonstrates a direct accessibility of this rich germplasm to breeding 18,21 . Potential barriers to successful interspecific gene introgression include male sterility, poor germplasm WFL Publisher Science and Technology www.world-food.net                              Food, Agriculture & Environment Vol.1(1): 52-63. 2003