Certain strains of bacteria are >1 million-fold more lethal when simultaneously injected with hemoglobin (or other iron compounds), which may supply iron (the lack of which ordinarily prevents replication of most aerobic bacteria in normal body fluids). In E. coli, iron-dependent lethality (the`hemoglobin-adjuvant effect') is specific to only certain strains and most isolates show no iron-dependent enhancement of virulence. We have compared two adjuvant and two non-adjuvant clinical isolates of E. coli. We find: (i) Approximate LD50 for non-adjuvant strains = 3 × 109 CFU ± Hb whereas for adjuvant strains LD50 = 108 -Hb vs. 102 +Hb (intraperitoneal inoculation in mice). (ii) In vitro, iron acquisition and iron-dependent growth are equivalent. Adjuvant and non-adjuvant strains show slow or absent growth in iron-poor medium (serum) and enhanced growth when supplemented with hemoglobin (nephelometric doubling time, 60 min for non-adjuvant vs. 90-120 min for adjuvant in the presence of 5 mg/ml Hb). (iii) None of the test strains produce commonly found toxins (e.g., Verotoxin, shiga-like toxins 1 and 2) and all are negative for enteroinvasion. (iv) However, after intraperitoneal inoculation into mice (without added iron compounds), adjuvant strains persist much longer (t1/2 ≈8 hr) than the non-adjuvant E. coli(t1/2 ≈30 min). (v) Co-injection of equal mixtures of ampicillin-resistant adjuvant E. coli with ampicillin-sensitive non-adjuvant bacteria (or vice versa) indicates that the presence of adjuvant E. coli does not affect the rapid clearance of the non-adjuvant strains. This suggests that adjuvant organisms do not produce diffusable leukotoxins or leukostatic factors. (vi) When injected with hemoglobin, adjuvant strains show rapid in vivo growth whereas non-adjuvant are still rapidly cleared (t1/2 ≈60 min). (vii) Finally, cytospin preparations of peritoneal lavage demonstrate profoundly impaired phagocytosis of adjvuant strains (both + and - Hb) apparently independent of capsular [K] antigen status. We conclude that E. coli strains exhibiting the hemoglobin-adjuvant effect are uniquely able to evade host phagocytosis/clearance and that this single characteristic - coupled with iron-mediated bacterial replication - may underlie the hemoglobin adjuvant effect.