Mesothelin is a cell surface antigen of unknown function that is strongly expressed in mesothelial cells. Although it was reported in 1992 that immunostaining with the K1 anti-mesothelin antibody could be very useful in distinguishing between epithelioid mesotheliomas and pulmonary adenocarcinomas, no further studies have been published on the value of this marker in the diagnosis of mesotheliomas. To determine whether mesothelin can assist in discriminating epithelioid mesotheliomas from lung adenocarcinomas or from other carcinomas metastatic to the serosal membranes, 55 mesotheliomas (44 epithelioid, 3 biphasic, and 8 sarcomatoid), 48 carcinomas of the lung (31 adenocarcinomas, 17 squamous carcinomas), and 86 nonpulmonary adenocarcinomas (14 ovary, 5 peritoneum, 9 endometrium, 11 pancreas, 4 stomach, 16 colon, 12 breast, 9 kidney, 4 thyroid, and 2 prostate) were investigated for mesothelin expression using the recently available 5B2 anti-mesothelin monoclonal antibody. Reactivity was obtained in all 44 (100%) of the epithelioid mesotheliomas, 12 (39%) of the lung adenocarcinomas, and 42 (49%) of the nonpulmonary adenocarcinomas (14 [100%] ovary; 5 [100%] peritoneum; 6 [67%] endometrium; 10 [91%] pancreas; 2 [50%] stomach; 5 [31%] colon; and in none  of the breast, kidney, thyroid, or prostate). Three (18%) of the squamous carcinomas of the lung, but none of the sarcomatoid mesotheliomas, exhibited positivity for this marker, nor was any reactivity seen in the spindle cell component of the biphasic mesotheliomas. It is concluded that despite the low specificity of mesothelin for discriminating between epithelioid mesotheliomas and adenocarcinomas, immunostaining for this marker may have some utility in those instances in which the results obtained with the standard panel of immunohistochemical markers used for the diagnosis of mesotheliomas are equivocal. Because mesothelin is a highly sensitive positive marker for epithelioid mesotheliomas, a negative staining for this marker is an indication against such a diagnosis; however, because of its limited utility, it is not recommended for inclusion in the standard panel of immunohistochemical markers used in the distinction between mesotheliomas and adenocarcinomas.