In an epidemiological study of human rotavirus (HRV) infections in metro Jeddah, Saudi Arabia, the relationships among subgroups, serotypes, and RNA electropherotypes of the rotavirus isolates were investigated. Of the 523 rotavirus-positive stool specimens, 245 were examined for subgroup, serotype, and electropherotype. Of these, 84 isolates were analyzed for their subgroup and RNA electropherotype specificites, 12 (14.3%) were of subgroup 1, 69 (82.1%) were of subgroup II, and 3 (3.6%) were a mixture of subgroup I and II. Of the subgroup 1 specimens, 5 (41.7%) showed long electrophoretic migration patterns and 7 (58.3%) showed short patterns. In subgroup II specimens, 66 (95.7%) were of long patterns and 3 (4.3%) of short patterns. The relationship between HRVS serotypes and electropherotypes was also determined for the same 245 rotavirus specimens. Of these, 36 (14.7%) exhibited short RNA patterns and 209 (85.3%) exhibited long patterns. The short pattern specimens consisted of serotype 1 (8.3%), serotype 2 (63.9%), serotype 3 and serotype 4 (2.8%) each. The long pattern specimens consisted of serotype 1 (60.3%), serotype 2 (1.4%), serotype 3 (7.2%) and serotype 4 (17.7%). Among the previous 245 specimens, subgroup specificities were available for 51 specimens. All subgroup I were of serotype 2, and all subgroup II were of serotype 1, 3 or 4. RNAs of either subgroup showed both long and short electropherotypes. No relationship could be established between subgroups or serotypes and a particular electropherotype. It seems unlikely that electropherotyping of human rotavirus (HRV) can be used for identifying the subgroups or serotypes of strains.