Eosinophilic pleural effusion accounts for 5 to 10% of all clear liquid pleurisies. Its pathogenic significance is unclear and its relation to the causative diagnosis of pleural eosinophilia is diversely evaluated. In this study, 86 cases of eosinophilic pleural effusion observed at the Ariana Pneumophthisiology Hospital over a 5-year period are reviewed. At first aspiration, the proportion of pleural fluid eosinophils varied from 12% to 85% (54% on average). Blood eosinophilia was present in 60% of the patients. The main cause of effusion was tuberculosis (39.6%, but many other diseases were observed; despite numerous investigations, no cause could be found in 23.2% of the cases. The discovery of a pleural eosinophilia has an impact on the probable cause of the pleurisy, and this can be determined by applying Bayes' rule. The probability of tuberculous being the cause falls from 70.3% for clear liquid pleurisies generally to 44.6% for eosinophilic pleurisy. The probability of cancer as a cause falls from 4.5% to 0.66%; whereas that of "idiopathic" pleurisy rises from 13.5% to 59.56%. These findings are concordant with Adelman's conclusions. In other words, the finding of a pleural eosinophilia decreases the probability of tuberculous or malignant pleural effusion and increases the probability of benign or "idiopathic" effusion. Clinicians confronted with an eosinophilic pleural effusion should be particularly careful and accurate since this diagnosis may spare the patient an unnecessary exploratory thoracotomy and an unwarranted antituberculous treatment.