On the “selectivity” of the Golgi‐Cox method

The randomness of the impregnation of layer IV cortical neurons by the Golgi-Cox method (Van der Loos, '56) has been assessed directly in Barrel C-1 of the mouse SmI. All Golgi-Cox impregnated neurons and unimpregnated neurons which were revealed with a Nissl counterstain were counted and measured in ten cerebral hemispheres cut tangential to the pia overlying the barrel field. The percentage of Golgi stained neurons varied considerably in different preparations from 0.73% to 2.26% with an average of 1.29%. The size distributions of both the Golgi impregnated and Nissl stained cells are similar but the difference of the means is statistically significant. However, if the means are equated there is no statistical difference in the two populations. When the Golgi precipitate is removed and the cells re-measured following Nissl staining there is a systematic reduction of the perikaryal cross-sectional area which is compatible with the differences in the means observed for the two populations as a whole. Finally, the frequency with which Golgi impregnated neurons are found in the barrel sides and hollows parallels the frequency with which Nissl stained neurons are observed in these two locations. We conclude that this variant of the Golgi method impregnates barrel neurons randomly. The value of this information for quantitative studies of cerebral cortex is discussed as is the potential of the system for elucidating some of the mechanisms responsible for Golgi impregnation.