Distinguishing peritonsillar abscess from cellulitis is an important clinical problem, particularly in children, who may require a general anesthetic for drainage of these abscesses. In order to identify those clinical factors most significant for peritonsillar abscess, we did a prospective study of 21 patients who presented with sore throat, fever, trismus, and tonsillar bulge; all symptoms that are consistent with the diagnosis of peritonsillar abscess. On admission, the following parameters were recorded: patient age, duration of sore throat, fever, white blood cell count, drooling, the degree of trismus (measured exactly as incisor-incisor distance), the degree of pharyngotonsillar bulge, and change in voice. After 24 to 48 hours of parenteral antibiotics, 12 patients (57%) had improved sufficiently and were continued on antibiotics until resolution (cellulitis group). Nine patients (43%) had no improvement and underwent surgery for drainage of the peritonsillar abscess (abscess group). At the end of the 18-month study period, the cellulitis and abscess groups were compared. On admission, no significant difference was found in age, duration of sore throat, fever, or white blood cell count. The pharyngotonsillar bulge was mild in 58% and moderate in 42% of the cellulitis group, while in the abscess group, the pharyngotonsillar bulge was mild in only 33% and moderate in 67%. After 24 to 48 hours of parenteral antibiotics, all patients in the cellulitis group had improvement of at least one symptom; whereas, all patients in the abscess group had no change or worsening of at least one symptom, including trismus, dysphagia, voice change, drooling, or pharyngotonsillar bulge. On admission, the precise measurement of trismus was not significantly different in the two groups (24.7 mm in cellulitis group vs. 22.5 mm in abscess group). However, after 24 hours of antibiotics, trismus averaged 7 mm more in the abscess group versus the cellulitis group (p less than 0.05).