The spleen colony technique. I. Correction for the overlap effect and sources of error in CFU-s determination

A linear model for the errors of the 'spleen colony' assay for haemopoietic stem cells has been derived. The components emerging from the model are interpreted and practical recommendations given for interpreting measurements made with this assay. The model permits correction for the effect of overlapping colonies and gives average errors for single measurements of the number of CFU-s. More reliable and more precise information can be obtained using this model. The spleen colony technique detects a population of immature precursor cells designated as CFU-s (Till & McCulloch, 1961). The relative error of measurement is often large when compared with the changes in the phenomena studied. Consequently a better knowledge of the errors of this technique is highly desirable. This paper should be regarded as an extension of the previous analysis of Till (1972). The theory for the errors of the spleen colony technique was applied to 905 determinations of the CFU-s numbers performed on random-bred mice. Data from random-bred mice rather than those from inbred mice have been used because the error components can be expected to be larger and, consequently, more easily detectable. The model of errors has also been validated using data published by Till (1972) and has subsequently been applied to data from several inbred mice strains (Znojil & Necas, 1988).