In clinical laboratories of Beijing, China, no known entero-pathogen but almost pure E. coli strains could be isolated from more than 60% fecal samples of diarrheal patients, which had been recognized as normal flora E. coli and dismissed. We suspected that some of the so-called normal flora E. coli strains might be virulent. To prove our idea. We collected 172 strains of E. coli isolated from diarrheal patients from whom no other enteric bacterial pathogens had been identified, including EPEC, EIEC, ETEC. With plasmid DNA analysis, Hep-2 cell adherence assay and 10 DNA probe hybridization, we found that the so-called normal flora E. coli was abnormal, 44% of them were virulent, of which 16 (9.3) were EHEC, 8 EPEC (4.7%), 11 EAggEC (6.8%). Fifty-four of 172 strains were hybridized with INV and SLT1 or SLT2 probes, which had never been reported. These strains could invade Hep-2 cells, but were lack of ipaB gene, a key gene of invasiveness gene cluster of Shigella species and EIEC. The aggregative adherence to Hep-2 cells was observed, but the strains were not hybridized with EAggEC specific DNA probe. The purified toxin protein and cell filtrate were toxic to vero cells. Based on the data obtained, we believed that this is a new category of diarrhea-genic E. coli, named as entero-SLTs-producing and invasive E.coli (ESIEC). ESIEC occupied 31.4% of the strains tested, the isolation of it was probably higher than those of ETEC, EPEC in P R China.