Collection and isolation of blast isolates. In total, 1,200 blast isolates (P. oryzae Cavara) were collected from the infected leaves and panicles of 37 rice (O. sativa L.) cultivars, including landraces, high-yielding inbred lines, and hybrid rice. Sampling covered all of the administrative divisions of Bangladesh (Rangpur, Rajshahi, Sylhet, Dhaka, Khulna, Barisal, and Chittagong) where rice is cultivated and two major rice-growing ecosystems for rice cultivation (rainfed lowland and irrigated lowland) from 2009 to 2012. Single spores were isolated from infected leaves or panicles incubated on moist filter paper in a petri dish at room temperature for 24 h in accordance with the protocols of Hayashi et al. (2009). Colonies from single conidia were grown on water agar for 5 to 7 days; two or three cut pieces of single colony were then transferred to sterile filter paper placed on rice straw agar medium. Finally, to enable repeated access to the original isolates, the fungi were grown on filter paper and stored aseptically in filter paper at−20°C after the necessary drying. Of the isolates collected, 331 were arbitrarily selected as representatives of single fields and used in pathogenicity test. In the 3 years in which this study was done, blast infection was observed in the rainfed lowlands during the Aman season from July to December, and the irrigated lowlands during the Boro season from November to May. It was not present in the Aus season, which occurs from April to July.DV.An international set of DV comprising 23 monogenic lines (Kobayashi et al. 2007; Tsunematsu et al. 2000) carrying 21 resistance genes, two NIL with the LTH genetic background (Telebanco-Yanoriaet al. 2010), and LTH as susceptible check was used. The virulence of the isolates was determined based on the reaction of the 25 DV and LTH. Inoculation and scoring of infection. To produce inocula, stock isolates (on paper discs) were cultured on oatmeal agar medium. The inoculated plates were incubated at 25 to 28°C for 12 to 14 days. The culture was scraped with a sterilized toothbrush, and the plates were then exposed to continuous light for 4 or 5 days to induce heavy sporulation. Conidia were dislodged by gentle rubbing with a paintbrush from incubated plates into sterilized distilled water with 0.01%Tween 20. Spore suspensions were filtered through four layers of gauze mesh; the concentration was adjusted to 10 × 104 conidia/mlby using a hemocytometer (Hayashi et al. 1998; Tuite 1969). Inoculation of the 25 DV and LTH (susceptible check) with the blast isolates was performed under greenhouse conditions following the methods of Bonman et al. (1986) and Hayashi et al. (2009). Seed of these varieties were treated using 0.5% Benlate T20 and then soaked for 1 to 2 days in water. Imbibed seed were sown in a plastic cell tray at 3 seeds/cell (w16 by 25 mm, 5 by 7 cells) and were maintained in the greenhouse. Inoculation was done by placing a seedling cell tray containing 20- to 21-day-old seedlings (at about the four- or five-leaf stage) on a rotating plate, spraying the seedlings uniformly with 10 ml of spore suspension (1 × 105spores/ml) per tray using an electric fine atomizer (nozzle size 0.3 mm; Airtex) at 0.1 MPa. In-oculated seedlings were incubated in a dew chamber at 25°C for 20 hand then transferred to a greenhouse maintained at 25 ± 1°C and 70 to80% relative humidity. The experiment was done twice. The reaction of each DV and LTH with each blast isolate was evaluated 7 days after inoculation using six scores from 0 to 5, where 0 = no evidence of infection; 1 = brown specks smaller than 0.5 mm in diameter, no sporulation; 2 = brown specks about 0.5 to 1 mm in diameter, no sporulation; 3 = roundish to elliptical lesions about 1 to 3 mm in diameter, with gray centers surrounded by brown margins, lesionscapable of sporulation; 4 = typical spindle-shaped blast lesions capable of sporulation, 3 mm or longer, with necrotic gray centers and water-soaked or reddish brown margins, little or no coalescence of lesions; and 5 = lesions as in 4 but about half of one or two leaf blades killed by coalescing lesions. Disease reactions that scored0 to 2 were classified as resistant (R) while those scored 3 to 5 were classified as susceptible (S), with the exception of IRBLta2-Pi forPita-2(resistant = 0 to 3) and IRBL5-M forPi5(t) (resistant = 0 to1), in accordance with Hayashi and Fukuta (2009). Race designation.Blast isolates based on the reaction patterns of25 DV and LTH were designated by following the designation system proposed by Hayashi and Fukuta (2009). The DV and LTH were categorized into five DV groups (“U”,“i”,“k”,“z”, and“ta”), where group U included LTH, IRBLa-A, IRBLsh-B, IRBLb-B, IRBLt-K59, and LTH; group i included three lines with the Piilocus onchromosome 9; group k included seven lines with the Pikregionon chromosome 11; group z included four lines with the Pizregionon chromosome 6; and group ta included seven lines with the Pitaregion on chromosome 12. The resistance genes in four of the DV groups (i, k, z, and ta) were multiple alleles or were located in the same chromosome regions. The remaining four genes—Piaon chromosome 11, Pishand Piton chromosome 1, and Pibon chromosome2—were located in different chromosome regions independently. In this study, we replaced the two monogenic lines used by Hayashi and Fukuta (2009) (namely, IRBLkh-K3 and IRBLk-Ka) with two LTHNIL (IRBLkh-K3[LT] and IRBLk-K[LT], respectively). In this way,each reaction type consisted of one to three units, to each of which we allocated three DV or three blast resistance genes. To designate the reaction type, we applied designation codes 1, 2, and 4 for susceptible reactions of each of the three respective DV to the blast isolates. The sum of these codes of each unit in each DV group designated a reaction type, representing the reactions of the DV in each group. Blast races were designated by the combined sum of the codes in each DV group, in accordance with Gilmour’s method (Gilmour 1973).Isolates classified this way were designated by reaction types within each DV group such as“U63”and“U73”for DV group U or“k100”and“k177”for DV group k, and as races by using the set of reaction types of all DV groups (U63-i0-k100-z00-ta403).Classification.Ward’s hierarchical method (Ward 1963) and JMP (JMP 11.2 for Windows, 2014; SAS Institute, Inc.) were used to per-form a cluster analysis on the basis of the data on the blast isolate infection scores of the 25 DV and LTH. The relationships between the pathogenicities of the cluster groups and the ecosystems and geo-graphical distributions were evaluated. Pathogen diversity was calculated by using Simpson’s diversity index (Simpson 1949). The index value varied from 0 to 1, where 0 represents no diversity and 1 is maximum diversity.