enzymes and are not likely to be destroyed during preparation of the specimens. We cannot explain the absence of organisms in gastric specimens stained with GMS and PMB, except that the TBO slides contained less extraneous background material. Although the median age of the pneumonitis group was lower than the control group, we do not consider the difference significant for comparison purposes, since the incidence of P. carinii in immunosuppressed patients is not age dependent. The absence of P. carinii organisms from the gastric contents of cancer patients in the control group indicates that this organism is not encountered as a part of the gastric flora in patients without PCP. Therefore, if P. carinff organisms are found in the gastric contents of an immunosuppressed patient with diffuse alveolar disease, the diagnosis of PCP is justified. On the other hand, the absence ofP. carinii organisms in such cases by no means excludes this as the causative agent. Since gastric lavage is a simple method for diagnosis of PCP, it may be useful in cases where lung aspiration or biopsy are not available or the patient is too ill to undergo such procedures.