Worldwide, rice blast, caused by Magnaporthe grisea (Hebert) Barr. (anamorph, Pyricularia grisea Sacc.), is one of the most economically devastating crop diseases. Management of rice blast through the breeding of blast-resistant varieties has had only limited xuccess due to the frequent breakdown of resistance under field conditions (Bonman etal., 1992; Correa-Victoria and Zeigler, 1991; Kiyosawa, 1982). The frequent variation of race in pathogen populations has been proposed as the principal mechanism involved in the loss of resistance (Ou, 1980). Although it is generally accepted that race change in M. grisea occurs in nature, the degree of its variability has been a controversial subject. A number of studies have reported the appearance of new races at extremely high rates (Giatgong and Frederiksen, 1968; Ou and Ayad, 1968; Ou et al., 1970; Ou et al., 1971). Various potential mechanisms, including heterokaryosis (Suzuki, 1965), parasexual recombination (Genovesi and Magill, 1976), and aneuploidy (Kameswar Row et al., 1985; Ou, 1980), have been proposed to explain frequent race changes. In contrast, other studies have shown that although race change could occur, its frequency was much lower than that predicted by earlier studies (Bonman et al., 1987; Latterell and Rossi, 1986; Marchetti et al., 1976). Although questions about the frequency of race changes in M. grisea remain unanswered, the application of molecular genetic tools to study the fungus, ranging from its genes controlling host specificity to its population sturctures and dynamics, have begun to provide new insights into the potential mechanisms underlying race variation. In this review we aim to provide an overview on (a) the molecular basis of host specificity of M. grisea, (b) the population structure and dynamics of rice pathogens, and (c) the nature and mechanisms of genetic changes underpinning virulence variation in M. grisea.