Transmission rates of extra chromosomes found in the full set of trisomics of pearl millet (Pennisetum americanum) (2n = 14) were estimated by examining the progeny of selfed trisomics and the progeny of trisomics crossed to disomics. When the trisomics were selfed, 'dark green' and 'tiny' had the highest transmission rate (23.8% and 23.3%, respectively) and 'pseudonormal' the lowest (13.8%). Other trisomics had an intermediate rate of transmission. When the trisomics were used as females in crosses with disomics, both 'dark green' and 'tiny' again had the highest transmission rate and 'pseudonormal' the lowest. When the trisomics were used as males in crosses to disomics, no trisomic was transmitted to the progeny except for 'spindle', and this occurred with a very low frequency (2.0%). A variation in transmission rate was observed from plant to plant and season to season for the same trisomic type. A study of the transmission rate of the extra chromosomes indicated that the following factors were probably contributing to the lower rate of transmission: small- or light-weight seeds tended to have a higher proportion of trisomics than heavier seeds; lighter seeds had a lower percentage germination; a positive and significant correlation was noticed between trivalent frequency and transmission rate. Plants with reduced vigour produced a higher frequency of trisomics. Though trisomics involving longer extra chromosomes showed a high degree of pollen and ovule sterility, they were highly transmissible. This has resulted in a close relationship between gametic sterility and transmission rate of extra chromosome.