Two sets of Egyptian cotton varieties were evaluated under early and late plantings. The first set included 16 cultivated and obsolete varieties and were evaluated for two seasons at Assiut Univ. Exper. Farm. The second set was selected from the first one; four varieties tolerant to late planting and four susceptible. The second set of varieties was evaluated in the third season at Shandaweel Res. Stn. Cott. Res. Ins. ARC. The combined analysis of variance of each set indicated insignificant interaction of years × dates, years × varieties and years × varieties × dates. However, the main effects of dates and varieties were significant (p ≤ 0.01). The genotypic and phenotypic correlations were estimated among seed cotton yield / plant, lint yield / plant, boll weight, number of bolls / plants and seed index. Phenotypic and genotypic correlations coefficients were very close and sometimes similar. Under early planting of the first set of varieties; lint yield / plant showed the highest genotypic correlation (0.989) with seed cotton yield / plant followed by number of bolls / plant (0.856) and boll weight (0.296). However, the genotypic correlation of seed index with seed cotton yield / plant was very small (0.042). The results indicated negative correlation between number of bolls and boll weight. The phenotypic and genotypic correlations under late planting and for the second set of varieties showed nearly the same trend. The study of path analysis in the two sets of varieties under early and late plantings indicated that lint yield / plant mostly showed the large direct effect on seed cotton yield, and when its direct effect was negative, it affected seed cotton yield via number of bolls and boll weight.  The direct effect of number of bolls / plant on seed cotton yield / plant was high, and followed by boll weight. The results revealed that simultaneous selection based on lint yield / plant, number of bolls / plant and boll weight could be promising to breakthrough in seed cotton yield / plant.