The present study was carried out during the five years 1995/96, 1996/97, 1997/98,, 1998/99 and 1999/2000 at Sids Research Station, Beni-Suef governorate, Egypt. The study aimed at evaluating the effectiveness of pedigree, bulk (natural selection), single pod descent (SPD) and mass selection breeding method. Two F₂ populations derived from two crosses were used. Eighteen families derived from each of the four breeding methods in each cross were tested for seed yield (t/ha), seed yield per plant (g) and 100-seed weight (g) in F₆. A randomized complete block design with three replications was used. Significant seed yield differences existed within F₆ families of each cross by applying either of the four methods. The widest ranges in seed yield were obtained by the bulk method. The bulk breeding method produced consistently more superior families in either and over the two crosses with 19, 18, 15 and 14 families for bulk, SPD, pedigree and bulk methods, respectively. The genotypic (d²_G) and phenotypic (d²_Ph) variances estimated from bulk method were the highest compared with other breeding methods over the two crosses. Heritability estimates and expected genetic advance indicated that the bulk method recorded higher values compared with the other three methods. Based on the results obtained it seems that the bulk breeding method was the most effective compared to the other three breeding methods. Considering the partial allogamous nature of the crop, it could be concluded that the bulk method was more efficient and could be less costly in breeding for high seed yield.