We are interested in the mechanisms that generate the mature cerebral cortex. We used bromodeoxyuridine (BrdU) to label cortical cells as they were being born. We followed the fates of specific sets of cortical precursors in normal mice and in mice in which other groups of cortical progenitors had been destroyed with the antimitotic agent methylazoxymethanol acetate (MAM Ac). In normal mice, most cells destined for the cerebral cortex were produced from embryonic day 12 (E12) to E16 in the expected inside-to-outside sequence (deep layers first, superficial layers last). Injection of MAM Ac at E13 killed cells that would normally have contributed to the deep cortical layers. As a consequence, the cortex was thinned by approximately 25% at postnatal day 21 (P21). However, all laminae were present and had normal connections with subcortical structures, although all were proportionately thinner. BrdU injected on E16 labelled a normally sized complement of cells that spanned a larger proportion of the depth of the thinned cortex. Thus, the deep cortical layers comprised many cells that were born several days later than normal. At embryonic ages prior to E12, a transient set of cells is produced in the early telencephalon. After injection with MAM Ac at E10, the cortex appeared histologically and histochemically normal at P21. However, many cells that would normally have contributed to superficial cortex (born on E15) were significantly deeper than normal. These results suggest that, during the early stages of cortical development, the nervous system is sufficiently plastic to compensate to some extent for the destruction of specific precursor cells by altering the fates of neurons born later. They indicate that the embryonic date on which a cortical cell is born does not necessarily determine its eventual phenotype.