This project bred HHB 67-like hybrid combinations of improved versions of the parental lines of popular elite pearl millet [Pennisetum glaucum (L.) R. Br.] hybrid HHB 67 that not only have superior host-plant resistance to pearl millet downy mildew disease (the trait originally targeted in the marker-assisted breeding program implemented in this project) caused by the oomycetic plant pathogen Sclerospora graminicola (Sacc.) J. Schröt., but that also have improved resistance to rust caused by the unrelated basidiomycetic plant pathogen Puccinia substriata Ell. & Barth. var. indica Ramachar & Cumm., and 25-30% higher mean grain yield than the original hybrid. This has been achieved without altering hybrid growth duration by more than a day or so. Thus these new versions of hybrid HHB 67 still well match the demands of farmers' producing a rainy season pearl millet grain and stover crop in dryland double-cropping systems on over 400,000 ha of marginal semi-arid lands in Rajasthan and Haryana. This surprisingly positive achievement is equivalent to the grain yield gains expected from a conventional hybrid breeding program of circa 10 years, and was accomplished despite unanticipated difficulties in generating the planned number of linkage group and segmental substitution lines having genomic material from ICMP 85410 introgressed into the genetic background of elite dwarf seed parent maintainer line 843B, and genomic material from 863B introgressed into the genetic background of elite tall seed parent maintainer line 841B. These difficulties occurred while trying to convert the marker-assisted breeding program in this project from more costly, slow, and tedious, hybridization-based, 32P-labelled restriction fragment length polymorphism (RFLP) markers to less expensive, more rapid, polymerase chain reaction (PCR)-based simple sequence repeat (SSR) and sequence-tagged site (STS) markers that can be visualized by silver staining. Unfortunately, the levels of genome coverage for and marker polymorphism between the donor parents (ICMP 85410 and 863B) and recurrent parents (843B and 841B), was markedly less for the PCR-based markers than for the RFLP markers that we had been using previously. Hence we had to continue making use of the slower, more costly RFLP markers and this reduced the total number of plants that could be genotyped and backcrossing generations that could be completed during the course of this project. Despite these difficulties, we were able to confirm (by marker-assisted backcross transfer followed by disease reaction assessment of the substitution lines produced) the occurrence of previously mapped downy mildew resistance QTLs on pearl millet linkage group 1 (LG1) and LG4 of ICMP 451-P6; LG1, LG4, LG6, and LG7 of ICMP 85410, and LG1 of P7-3. We also identified the presence of a previously unmapped downy mildew resistance QTL on LG5 of 843B. Two hybrids initially tested in this project during the rainy season of 2001 have been selected by our collaborators at CCS Haryana Agricultural University for entry into national trials during the rainy season of 2002. Results from evaluations in these national trials, combined with those from more extensive on-station and on-farm trials in Haryana that will be conducted simultaneously, are expected to be used to support a state-release proposal for one or more of the improved versions of HHB 67 early in 2003. With this substantial quantities of Certified Seed of the new versions of HHB 67 should reach farmers by the rainy season of 2004, in time to replace the original HHB 67 before it succumbs to a downy mildew epidemic. If successful, as it should be, this will be the first time that a farmeraccepted replacement for a widely popular pearl millet hybrid has been adopted before its predecessor succumbed to an epidemic. Direct losses to pearl millet-producing farmers in India due to a major downy mildew epidemic on this popular hybrid, which can be avoided if the improved versions of HHB 67 reach farmers in time, would be in the range of £3,300,0001.