Nine populations were derived in the selection phase of experiments designed to evaluate a recurrent selection scheme in hurley tobacco (Nicotiana tabacum L.). We produced an unselected population of heterozygotes (Co), three populations by recurrent selection for increased weight of green leaves, and five populations by recurrent selection for decreased percent total alkaloids (% TA) of green leaves. Evaluations of the experimental material were made at three locations. The objectives wre to determine the effects of phenotypic recurrent selection on green tobacco and to ascertain the correlated effects of such selection on the yield and alkaloid content of cured tobacco.In response to selection, green leaf weight increased somewhat consistently from C0 to C3 and % TA declined linearly from Co, to C3 The correlated response of cured leaf weight to selection on a green weight basis was a linear increase, averaging 8.1 g per plant per cycle. Percent total alkaloids of cured leaf also declined linearly with successive cycles of selection on a green leaf basis. Five cycles of selection resulted in nearly 23% reduction in percent TA.The genetic correlation coefficient between green leaf weight and percent TA was −0.50 and −0.62 at two locations. This negative relationship may be a limiting factor in practical usage of phenotypic recurrent selection to increase yields without the aid of a selection index.