IN normal mouse hæmatopoietic tissue, there is a class of cells which, on being transplanted into heavily irradiated mice, can proliferate and form macroscopic colonies. In the spleen, the colonies formed in this manner are discrete and easy to count1, 2. Microscopically, each colony appears as a cluster of hæmatopoiotic cells, many of which are dividing; and often, within a given colony, the cells which are observed indicate that differentiation is occurring along three lines, into cells of the erythrocytic, granulocytic and megakaryocytic series, respectively1.

In vitro techniques, which permit investigations of clonal populations, have greatly advanced the knowledge of the genetic and physiological properties of cells3. It was therefore desirable to prove whether or not spleen colonies each develop from single cells and hence are clones; for if they are, it would be possible to study clonal populations of hæmatopoietic cells in vivo. Evidence from previous experiments supporting the view that the colonies are clones is as follows:(1) the curve relating the number of nucleated marrow cells that are transplanted to the number of colonies that develop in the spleen is linear, and shows no initial threshold1, 2;(2) the radiation survival curve of cells that can form colonies closely resembles the survival curves obtained for single cells in cell culture4 or tumour transplants5. These observations provide only indirect evidence that the colonies are clones. It seemed desirable, therefore, to attempt to obtain direct evidence for the unicellular origin of the colonies. At first glance the easiest way to accomplish this aim would be to make use of marrow cells that have a known chromosome marker, such as that found in T6/+ mice6, and to transplant into irradiated mice a mixture of marrow cells with a marker and marrow cells with no marker. If the colonies that developed in the spleen were always composed of cells with markers or cells without markers it would, of course, suggest the single-