Development of sorghum populations for resistance to Striga hermonthica in the Nigerian Sudan Savanna

Six elite sorghum varieties, ICSV 111, ICSV 400, KSV 4, Gaya Early, CS 54, and CS 95 were used to improve Striga resistance through pedigree breeding. SRN 39 and IS 9830 varieties were confirmed resistant to Striga hermonthica (Del.) Benth in field trials. The two Striga resistance lines were crossed with the six elite varieties. Three hundred and sixty Striga-free plants from F2s of good agronomic traits were identified, and selfed to produce F3s. The number of selections varied from 12 (ICSV 400 x IS 9830) to 59 (ICSV 400 x SRN 39) per F2 population. About 50% more plants were selected from the crosses involving SRN 39 as donor parent for Striga resistance than the crosses involving IS 9830. In addition, 58 out of the 100 plants selected were from a cross, ICSV 111 and SRN 39. Crosses from which Striga free plants were obtained were CS 54 x SRN 39 (12 plants selected), CS 95 x SRN 39 (9), ICSV 400 x SRN 39 (6), Gaya Early x IS 9830 (6), Gaya early x SRN 39 (5), and KSV 4 x IS 9830 (4). Crop syndrome reaction score was higher in 1995 (2.3 4.0) than in 1996 (1.0 2.7). Our results suggest that SRN 39 is a better donor parent for Striga resistance than IS 9830. ICSV 00090 NG, a cross between ICSV 111 and SRN 39 gave the highest grain yield of 2.02 t/ha in a replicated trial compared to the two parents, ICSV 111 (1.11 t/ha) and SRN 39 (0.86 t/ha). This variety combines potential for high yield and resistance to Striga. Our data indicates that the elite varieties can be improved for Striga resistance using pedigree breeding. A large F2 population (500 to 1000 plants per cross) is recommended for the selection of transgressive segregants. Further efforts are required to back cross-promising segregants with established Striga resistant variety in order to develop durable Striga resistant varieties with acceptable agronomic traits.


INTRODUCTION
Striga hermonthica (Del.)Benth. is the most destructive parasitic weed on cereals in western Africa, (Sauerborn, 1991).Grain losses on a regional scale average 5 -15%, however, Striga can exert a much greater impact locally, sometimes resulting in total crop failure (Doggette, 1988;Riches and Parker, 1995).Up to 5% and 95% yield losses have been recorded for resistant and susceptible sorghum hybrids, respectively (Obilana, 1980).In Africa 21 million hectares of land are estimated to be infested with Striga, resulting in an annual grain yield loss of 4.1 *Corresponding author.E-mail: i.ezeaku@cgiar.org million tons (Sauerborn, 1991) and an estimated overall loss of 7 billion US dollars in revenues to Africa (M'Boob, 1986).Some of the recommended control options for S. hermonthica include: adequate land preparation, hand pulling, hoe weeding, use of trap and catch-crops, seed treatment, application of appropriate rate of nitrogen fertilizer, herbicide spray, use of biological control as well as host-plant resistance (Lagoke et al., 1988).Among these control methods, host-plant resistance is the most economic control measure since it is affordable to farmers and resistant cultivars can be grown without additional inputs (Hess and Ejeta, 1992).Striga has coevolved with sorghum whose origin is Africa.The crop has developed various complex mechanisms of resistance to the parasite as a survival strategy (Riches, 1999).For example, the mechanism of resistance of SRN 39 to Striga species was observed to be a combination of low stimulant production, and restricted parasite development following attachment to host roots.These complex mechanisms have facilitated durable resistance of the cultivar in the field and it is currently being recommended for use in Sahel agro-ecological region of Western and Central Africa (WCA) (Bailey et al., 1995).The Striga resistant varieties such as SRN 39, and IS 9830 have undesirable traits such as low grain yield, short stature, and small grains, that limit their acceptance by farmers.Among the introductions of Striga resistant cultivars to WCA, Framida exhibited stable resistance to S. hermonthica under diverse conditions in addition to other important traits including stability of yield, good seedling establishment and resistance to grain mold and birds.However, due to its high tannin content and the brown seed color, the variety is not acceptable to majority of the farmers in WCA (Ramaiah, 1991).Crosses between resistance sources and adaptable elite varieties could result in varieties with combined characteristics of high productivity, acceptable grain and plant traits and resistant/tolerant to Striga.This paper describes screening to identify resistance sources to S. hermonthica and the transfer of resistant genes from these sources into elite sorghum lines.

MATERIALS AND METHODS
The trials were conducted in Striga sick plot at Bagauda located in the Sudan Savanna agro-ecological zone of Nigeria (11 o 53'N, 8 o 14'E) at a mean altitude of 440 m above sea level.Soil type is a plinthic luvisol with an average depth of 90 cm.Eighteen sorghum entries/introductions were evaluated in a randomized block design with three replications in 1995 and 20 entries in 1996.Gaya Early and Bagauda Farafara were used as local checks in both years.Trials were established in July in 1995 and 1996.In 1995, the gross plot size consisted of 6 rows of 4 m long ridges, and the net plot size was 12.75 m 2 while in 1996 the gross plot size consisted of 2 rows of 5 m long ridges, and the net plot size was 7.125 m 2 .The number of Striga plants was converted from 12.75 m 2 to facilitate comparison (Table 1).
Two Striga resistant lines (SRN 39 and IS 9830), four varieties from Nigeria (KSV 4, ICSV 111, ICSV 400 and Gaya Early), and two selections from Cameroon (CS 54 and CS 95) were sown at Dadin-Kowa, Gombe State in January, 1996 to produce crosses between Striga resistant lines and elite varieties (ICSV 111, ICSV 400, KSV 4, Gaya Early, CS 54, and CS 95) using double plastic bag technique.Eleven F1 crosses along with female parents were sown at Bagauda during the 1996 main season to identify true hybrids.Ten crosses were found to be true hybrids, and were further advanced to F2s by selfing individual plants panicle.Ten F2 populations, each with 600 plants, were sown in Striga sick plot at Bagauda in June 1997 to identify Striga free plants.Striga count was carried out at maturity.Three hundred and sixty Striga free plants with good agronomic traits were identified, and then selfed to produce F3s.The 360 F3 progenies were sown in Striga sick plot at Bagauda in June 1998 to identify resistant progenies and to advance selected 100 plants from 62 progenies to F4s.The list of crosses, total number of F3 progenies, and selected number of F3 Ezeaku and Gupta 325 progenies are presented in Table 2.A preliminary evaluation trial consisting of 30 F6 Striga resistant lines along with their parents were established in Minjibir in July 2000.The trial was established in a randomised block design consisting of three replicates.Striga seeds were inoculated in planting hills at sowing.Each hill of about 3 cm deep and 5 cm wide received one full coca-cola bottle cap of Striga seeds.In all the trials, intra-row spacing was 25 cm while inter-row spacing was 75 cm.First weeding was done with hoe while subsequent weeding was by hand pulling to avoid tampering with Striga plants.Basal application of fertilizer NPK 15:15:15 was done at the rate of 300 kg ha -1 by broadcasting uniformly and thereafter incorporated into the soil.Calcium ammonium nitrate (CAN) was applied three weeks after sowing at the rate of 100 kg ha -1 .In trials conducted from 1995 to 1998 data collected were Striga count at 90 days after sowing (DAS), grain yield and crop damage score.Crop syndrome reaction score was taken on a scale of 1-9 where 1 = healthy plants and 9 = completely dead plants.In 2000, data collected include grain yield, time to 50% bloom, plant height, panicle length, threshing percentage and grain mass (Table 3).All data were analysed using Genstat 5 release 3.2 statistical packages 1989 (Lawes Agricultural Trust, Rothamstead Experimental station) following standard analysis of variance procedures (Gomez and Gomez, 1984).

RESULTS
The data on Striga count, grain yield and crop damage score at Bagauda during 1995 and 1996 main seasons are presented in Table 1.Crop syndrome reaction score was higher in 1995 ranging between 2.3 -4.0, than in 1996 that ranges from 1.0 -2.7.Framida red, KSV 4 and CS 95 gave lower crop reaction score and produced higher grain yield than local checks in both years.Although, KSV 4 supported high Striga plant in both years, it possesses low crop reaction score.The local checks, Bagauda Farafara, and Gaya Early, and an improved variety, KSV 8, were highly susceptible to Striga.Bagauda Farafara supported more number of Striga plants than Gaya Early.Although Bagauda Farafara supported the highest Striga plants, it registered low crop reaction score of 2.0 in 1996.ICSV 400 produced higher grain yields than the locals in 1996, in spite of supporting high Striga emergence.ICSV 111 supported fewer Striga emergence and produced higher grain yield than the checks in 1996.Similarly, ICSV 901 NG supported fewer Striga emergence than Gaya Early and produced higher grain yield than the two local checks in 1995.The two varieties from Cameroon, CS 54 and CS 95, were moderately susceptible to Striga in both years.S 35 produced higher grain yield and supported lower Striga plants in 1996 when compared to local checks.Six entries; KSV 4, Gaya Early, ICSV 111, ICSV 400, CS 54, and CS 95 were consequently identified as a new varieties which exhibits promise for tolerance to Striga infestation.
In both years, ICSV 1007 BF consistently supported fewer Striga emergence and produced similar grain yield as the local checks.SRN 39 and IS 9830 supported fewer Striga emergence and produced higher grain yield in both years than Bagauda Farafara.Two hundred and fifteen Striga free plants were selected from five F 2 populations involving SRN 39 as donor parent, whereas only 145 plants were selected from five F 2 populations involving IS 9830 as donor parent (Table 2).The F 2 populations derived from crosses involving SRN 39 had lower incidence of Striga than the other crosses.
One hundred plants from 62 progenies were selected out of 360 F 3 progenies during 1998 main season (Table 2).The majority of the plants (58 out of 100) selected were from a cross derived by crossing ICSV 111 and SRN 39.The other crosses from which Striga free plants were obtained included CS 54 x SRN 39 (12 plants selected), CS 95 x SRN 39 (9), ICSV 400 x SRN 39 (6), Gaya Early x IS 9830 (6), Gaya early x SRN 39 (5), and KSV 4 x IS 9830 (4).It is interesting to note that the mean Striga count for total F 3 progenies was similar to the selected progenies for most of the crosses.This could be because at the time of selection, both Striga emergence and agronomic traits were considered.Perhaps one backcrossing will improve the chances of selection of Striga resistance combined with high yield.
Results from Table 3, shows that cross involving ICSV 111 as female parent and SRN 39 as male parent gave higher grain yield and good grain mass than other cross combinations.ICSV 00090 NG gave the highest grain yield of 2.02 t/ha compared to the two parents, ICSV 111 and SRN 39 with grain yield each of 1.11 t/ha and 0.86 t/ha respectively.Among all the crosses ICSV 00090 NG also possess the highest threshing percentage of 78 and look promising because of its good grain yield and generally crosses made with ICSV 111 gave progenies that supported fewer Striga plants.Crosses between CS 54 and SRN 39 gave the lowest grain yield and grain mass.ICSV 00097 NG, a cross between Gaya Early and SRN 39, gave the highest grain mass of 29.In general, crosses generated gave higher grain yield than their respective parents, probably because agronomic scores were considered along with Striga free plant during the selection process.

DISCUSSION
Based on low Striga count, SRN 39, IS 9830, ICSV 1007 BF, and Framida were identified as resistant to Striga.These four entries were reported resistant to Striga by Murty et al. (1995) based on three years results at Bagauda.Framida red produced higher grain yield and supported fewer Striga plants than the two checks in both years.Earlier reports by Ramaiah (1991) and later by Lagoke et al. (1999) have confirmed the stability of resistance in Framida red across locations and years.This variety has been used to develop Striga resistant materials including ICSV 1007 BF.It has been shown that single recessive gene controls resistance in SRN 39 (Hess and Ejeta, 1992).The mechanism of resistance to Striga was also attributed to low germination stimulant production (Hess et al., 1992) and post-infection growth inhibition (Bailey et al., 1995).IS 9830 and SRN 39 have also been reported to have adoption problem in various parts of sub-saharan Africa due to small grain size and low grain yield (Ramaiah, 1991;Lagoke et al., 1994).These varieties were, however, adopted by farmers in Sudan (Obilana and Reddy, 1999).These two varieties were selected as source parents for improvement of the elite varieties because of their stable resistance to S. hermonthica.ICSV 111 supported fewer Striga plants and produced higher grain yield than checks.Murty et al. (1995), also found that ICSV 111 was similar to SRN 39, ICSV 1007 BF, IS 9830, and Framida for resistance to Striga.Our observation that local checks of long duration types supported higher Striga plants than test varieties and possesses low crop syndrome score is in line with earlier reports indicating that local sorghum varieties, especially landraces which have a longer growing season usually support higher Striga incidence because of longer period of exposure of the Striga seeds and plants to the land races than improved short season types (Talleyrand et al., 1991;Kureh et al., 1999).Ejeta et al (1991) noted that Striga tolerant genotypes permit and support as many Striga plants as susceptible genotypes but do not show a concomitant reduction in grain production or overall productivity.
The results suggest that SRN 39 is a better donor parent for Striga resistance than IS 9830.Riches (1999) had indicated that the combination of low stimulant production with post-infection cellular resistance observed in SRN 39 is likely to be durable in the field, unlike in IS 9830 that only produces low germination stimulant.The segregants derived from crosses with SRN 39 appear to be morphologically superior to IS 9830.The results also suggest that ICSV 111 supported fewer Striga emergence than ICSV 400.This finding was also supported by the mean Striga emergence in F 3 progenies derived from crosses involving ICSV 111 and ICSV 400 with SRN 39 and ICSV 9830.CS 54 and 95 were moderately susceptible to Striga, and F 3 progenies derived from crosses involving these lines and SRN 39 supported more Striga shoot.ICSV 00090 NG possesses superior agronomic traits than the two parents.It appears that the elite varieties can be improved for Striga resistance using pedigree breeding with one or two back crossing.Obilana (1984) reported that segregation in later generations allowed resistant types to be recovered following pedegree breeding.In future, marker-assisted selection could be used to accelerate the selection process for improvement of Striga resistance in sorghum varieties.

Table 1 .
Performance of sorghum lines in Striga infested plot at Bagauda, Nigeria, 1995 and 1996 main seasons.

Table 2 .
Number of total and selected F3 progenies (derived from crosses between Striga resistant and elite varieties), sown at Bagauda, 1997 main season.
Numbers of selected plants are given in parenthesis.