In near-isogenic lines of winter wheat (Triticum aestivum L. cv. Maris Huntsman) grown at 20° C under long days the reduced-height genes, Rht1 (semi-dwarf) and Rht3 (dwarf) reduced the rate of extension of leaf 2 by 12% and 52%, respectively, compared with corresponding rht (tall) lines. Lowering the growing temperature from 20° to 10° C reduced the rate of linear extension of leaf 2 by 2.5-fold (60% reduction) in the rht3 line but by only 1.6-fold (36% reduction) in the Rht3 line. For both genotypes, the duration of leaf expansion was greater at the lower temperature so that final leaf length was reduced by only 35% in the rht3 line and was similar in the Rht3 line at both temperatures. Seedlings of the rht3 (tall) line growing at 20° C responded positively to root-applied gibberellin A1 (GA1) in the range 1–10 μM GA1; there was a linear increase in sheath length of leaf 1 whereas the Rht3 (dwarf) line remained unresponsive. Gibberellins A1, 3, 4, 8, 19, 20, 29, 34, 44 and 53 were identified by full-scan gas chromatography-mass spectrometry in aseptically grown 4-d-old shoots of the Rht3 line. In 12-d-old seedlings grown at 20° C, there were fourfold and 24-fold increases in the concentration of GA1 in the leaf expansion zone of Rht1 and Rht3 lines, respectively, compared with corresponding rht lines. Although GA3 was present at a similar level to GA1 in the rht3 (tall) line it accumulated only fivefold in the Rht3 (dwarf) line. The steady-state pool sizes of endogenous GAs were GA19 ≫ GA20 = GA1 in the GA-responsive rht3 line whereas in the GA non-responsive Rht3 line the content of GA19≈ GA20 ⋘ GA1. It is proposed that one of the consequences of GA1 action is suppression of GA19-oxidase activity such that the conversion of GA19 to GA20 becomes a rate-limiting step on the pathway to GA1 in GA-responsive lines. In the GA-non-responsive Rht lines it is suggested that GA19 oxidase is not downregulated to the same extent and GA1 accumulates before the next rate-limiting step on the pathway, its 2β-hydroxylation to GA8. The steady-state pool sizes of GA19, 20, 1, 3 and 8 were similar in developmentally equivalent tissues of the rht3 (tall) line growing at 10° C and 20° C despite a 2.5-fold difference in the rate of leaf expansion. In contrast, in the Rht3 (dwarf) line, the extent of accumulation of GA1 reflected the severity of the phenotype at the two temperatures with slower growing tissues accumulating less, not more, GA1. These results are interpreted as supporting the proposed model of regulation of the GA-biosynthetic pathway rather than previous suggestions that GA1 accumulates in GA-insensitive dwarfs as a consequence of reduced growth rates.