The diagnosis of feline infectious peritonitis (FIP) is hampered by the characteristic etiopathogenesis of the disease. 9 The coronavirus responsible for the disease (FIPV) originated by a mutation in the widespread feline enteric coronavirus (FECV). 9,14 These coronaviruses are both phenotypically and genotypically identical, as demonstrated by polymerase chain reaction (PCR). 14 Both feline coronaviruses (FCoVs) are able to trigger the production of antibodies. High antiFCoV titers are also detectable in healthy cats in FECVendemic catteries, 9,10 and the formation of immune complexes could results in negative serology in symptomatic cats. 8 Furthermore, false-positive results occur from cross-reactivity with coronaviruses of other species, such as canine coronavirus (CCV) or transimissible gastroenteritis virus (TGEV), or with extraneous noncoronavirus proteins that are copurified with the FCoV and included in commercially available diagnostic enzyme-linked immunosorbent assay kits. 1 Hematology, serum chemistry, and serum protein electrophoresis give only a strongly suggestive but nonpathognomonic pattern of FIP. 9,12,13 In the effusive forms, diagnosis could be improved by cytology 3 and protein analysis of the fluid. 9,11-13 Although FCoV RNA has been demonstrated in the plasma of healthy cats in FECV-endemic catteries, 6 FCoVs detected from areas other than the intestinal lumen should be interpreted as FIPV, particularly if the viruses are detected in lesions 15 or effusions. 2 In this study, the results obtained using 3 different diagnostic methods (anti-FCoV direct immunofluorescence test, cytology, and protein analysis) for 110 cats with effusions over a 5-year period were compared to determine the utility of these methods for diagnosing FIP from analysis of the effusion alone. Cats with clinically detectable effusion were examined (Table 1), although in some cases clinical and laboratory data (serum chemistry, ultrasonography, macroscopic appearance of the effusion) suggested diseases other than FIP.Approximately 2 ml of effusion was withdrawn from each cat before death (72 cats) or within 2 hours after death (38 cats). The sampled fluid was put in a tube containing ethylenediamine tetraacetic acid. The protein content of the effusions of 51 cats was evaluated in a discrete autoanalyzer a by the biuret method. b Fifty to 100 l of fluid was cytocentrifuged c within 15 hours at 130 ϫ g for 10 minutes. Two Received for publication August 25, 1997.slides were obtained from each fluid sample; 1 slide was routinely stained with May Grünwald-Giemsa and the other was used immediately or after storage at Ϫ20 C for no more than 1 week for direct immunofluorescence (DIF) tests as previously described. 2 The slides were fixed and dehydrated in acetone : methanol (3:1) for 20 minutes and incubated for 30 minutes at 37 C in a moist chamber with 100 l of a feline polyclonal fluorescein-conjugated antiserum d detecting FCoV biotypes I and II and cross-reacting with TGEV and CCV. After washing 4 times for 10 minutes e...