At surgery, the gland appeared indurated, without evidence of abscess. Cultures were obtained. Aerobic, anaerobic, acid-fast bacterial, and fungal cultures were negative. Interestingly, the conventional tube culture, derived from grinding the tissue obtained by biopsy and culturing on cell lines, was positive for adenovirus but negative for cytomegalovirus, HSV, varicella-zoster virus, and enterovirus. Histologic sections of the lacrimal gland biopsy specimen documented the presence of a necrotizing dacryoadenitis, with a dense infiltrate composed of neutrophils; histiocytes; small, cytologically bland lymphocytes;and lesser numbers of plasma cells and eosinophils. This infiltrate was associated with extensive glandular destruction (Figure 3). The viral cytopathic change typical of adenovirus infection—amphophilic inclusions with peripheral clearing or "smudge cells"—was not identified. Immunoperoxidase studies performed on paraffin-embedded material with standard methods and commercially available antibodies (HSV-1 and -2, DAKO Corp, Carpinteria, Calif, polyclonal, product id B0114, 1:250; cytomegalovirus, DAKO Corp, clone DDG9 and CCH2 cocktail, 1:25; adenovirus, Research Diagnostics Inc, Flanders, NJ, clone M58 and M73 cocktail, 1:50) were positive for HSV (Figure 4) and negative for cytomegalovirus and adenovirus. Flow cytometric analysis of the lymphoid component identified a dominant population of reactive CD3+ T cells, with negligible numbers of reactive CD19+ B cells. When taken together, the morphologic finding of inclusions, the positive immunostain results, and the negative lymphoma workup (flow cytometery) all supported the diagnosis of herpetic dacryoadenitis.