Four experiments are reported which studied the effects of date of sowing on the development of winter and spring barley. The first three experiments (1979–80, 1980–1, 1982–3) were carried out at Trefloyne, Tenby and used the varieties Sonja (winter) and Jupiter (spring). The final experiment (1983–4) was carried out in Cambridge using the varieties Igri (winter) and Triumph (spring). A wide range of dates of sowing (August-June) was used and there were considerable differences in temperature between seasons.

In both Sonja and Igri, delaying sowing from August to mid-February reduced the number of mainstem leaves but further delay in sowing increased the number of leaves. In Jupiter, delaying sowing until late October had little effect but further delay consistently reduced the number of leaves. In Triumph, delaying sowing reduced final number of leaves over the whole range (October-May).

In all experiments rate of leaf emergence was characterized by long periods when leaves emerged linearly with time, despite considerable variation in temperature. The only consistent changes in rate of leaf emergence occurred shortly after each equinox, rates slowing in the autumn and increasing in the spring. Rates of leaf emergence were similar in different seasons and sites for similar periods despite differences in temperature.

Initiation of spikolet primordia proceeded linearly with time for most dates of sowing. For winter varieties rates of spikelet initiation increased with delay in sowing provided emergence occurred before the spring equinox. Further delay in sowing decreased the rate of spikelet initiation. In contrast in Jupiter and Triumph the rate increased with each delay in sowing. During periods of low temperature, rate of spikelet initiation slowed but frequently increased to faster rates in succeeding increasing temperatures. Such increasing temperatures had no effect on rate of initiation of later-sown treatments which had not experienced the low temperatures. Effects on spikelet initiation were found when no effects on leaf emergence could be detected.

Initiation of spikelet primordia in all varieties appeared to begin when a fixed number of leaves (different in some varieties) had emerged and end when a fixed number of leaves were left to emerge. Delaying sowing in winter varieties until February and in Jupiter over the whole range reduced the number of leaves emerging during spikelet initiation and was the developmental mechanism through which large differences in date of sowing became small differences in date of anthesis (and grain maturity).

Attempts to relate development to temperature were unsuccessful and the major influence on development appeared to be daylength. Number of emerged mainstem leaves was the most consistent measure of development over sites, seasons and varieties and for winter varieties was influenced by the direction of change of daylength at emergence. Two positive linear relationships between number of leaves and daylength at emergence were found for the periods from summer to winter solstice and from spring equinox to summer solstice. The slope of the latter relationship was greater than the former. For the period from winter solstice to spring equinox a negative linear relationship between number of leaves and daylength at emergence was found. In Triumph number of leaves decreased with decreasing daylength in the autumn and decreased further with increasing daylength. The effects in Jupiter were less clear but there was evidence of the effect of both daylength at emergence and direction of change on number of leaves.

In view of the control of number of leaves exercised by daylength, close negative linear relationships between time from sowing to anthesis and date of sowing were found. The significance of the results for relating agronomic practice to development is discussed.