The necessary proscription against direct investigation of human tumorigenesis limits the types of studies that can be done. Consequently most investigations of human tumor development necessarily bear only indirectly on causative factors. One such indirect approach involves determining the number of cells from which tumors arise, thereby providing important clues to their mode of origin. For example, a neoplasm that results from a rare, more or less random, event like" spontaneous" somatic mutation would be expected to arise in a single cell. On the other hand, multicellular origin might be seen for a tumor caused by cell-to-cell spread of a virus with transformation of infected cells. The number of cells from which tumors arise and several related questions can be investigated in subjects with at least two genetically distinct types of cells, ie, in subjects with cellular mosaicism. To illustrate this point, assume that a person has two types of cells, A and B. If the event which initiates a tumor occurs in only one cell (eg, in an A cell), all the neoplastic cells will be of one type (A), since they are all descended from the one progenitor cell. On the other hand, if the oncogenic event occurs in many cells, it is likely that both A and B types will be affected and the tumor could then contain descendants of both cell types. Several mosaic systems have been employed in studies of human tumorigenesis. For example, chromosomal mosaicism, a condition in which some cells contain a readily identifiable chromosomal abnormality while other cells of the same tissue are normal, has been utilized. However, this system has limited applicability. A second type of marker system is dependent upon immunoglobulin synthesis and therefore is restricted to cells of the immune system. The most generally applicable cell marker system is dependent upon X-chromosome inactivation mosaicism. In this communication studies done using chromosomal and immunoglobulin cell markers are