Water stored deep in the soil profile is generally considered valuable to crop yield because it becomes available during grain filling, but the value of subsoil water for grain yield has not been isolated and quantified in the field. We used rainout shelters with irrigation to control the water supply to wheat crops that had different amounts of subsoil water available to isolate and quantify the efficiency with which the subsoil water was converted to grain yield. Under moderate post-anthesis stress, 10.5 mm of additional subsoil water used in the 1.35–1.85 m layer after anthesis increased grain yield by 0.62 t/ha, representing an efficiency of 59 kg/ha.mm. The additional yield resulted from a period of higher assimilation 12–27 days after anthesis and was related to an increase in grain size rather than other yield components. Under more severe stress with earlier onset, extra water use below 1.25 m was accompanied by additional water use in upper soil layers and it was more difficult to isolate and quantify the benefit of deep water to grain yield. The additional water used from all layers from the time the stress was imposed was converted to grain at 30–40 kg/ha.mm, but this increased to 60 kg/ha.mm for water used after anthesis. The high efficiency for subsoil water use is 3 times that typically expected for total seasonal water use, and twice that previously estimated for total post-anthesis water use in a similar environment. The results demonstrate that relatively small amounts of subsoil water can be highly valuable to grain yield.

"Haying-off" was studied by comparing wheat responses to applied nitrogen (N) at 3 sites in southern New South Wales, which differed in the amount and timing of rainfall during crop growth. At a site where the crops encountered little water deficit, dry grain yield increased from 607 g/m2 for a low-N control crop to 798 g/m2 for a high-N crop. At a site with severe terminal drought, dry grain yield decreased 24% from 374 g/m2 for the control, to 284 g/m2 for the highest N crop. At the third site, yields increased with small applications of N, whereas greater applications resulted in a negative yield response. At the 2 latter sites, the crops that showed decreased yield with applied N had clearly hayed-off. At all sites, irrespective of water status, N application resulted in increases in biomass at anthesis, spike density, kernels per spike, and kernel number. Kernel weight decreased in response to additional N at all sites, but most markedly at the haying-off sites, where it decreased by up to 38%. Harvest index increased in response to N at the high-rainfall site, but decreased in crops that hayed-off. Grain protein increased in response to N at all sites, with a range from 9% to 18% at the haying-off sites. The apparent retranslocation of assimilates to grain contributed 37-39% of grain yield (depending on N supply) at the high-rainfall site, compared with 75-100% at the haying-off sites. In contrast, when apparent retranslocation was expressed in relation to biomass at anthesis, it remained relatively constant, amounting to 23-26% at the high-rainfall site and 24-28% when crops hayed-off. By anthesis, high-N crops extracted more soil water than the low-N crops. By maturity the most severely hayed-off crop had extracted 10 mm less soil water than a low-N crop, but at the high rainfall site the high-N crops extracted 20 mm more soil water than the control crops. The weather conditions between anthesis and physiological maturity were relatively mild, with no daily maximum temperatures above 30ºC and no sudden increases in evaporative demand. Thus, there appeared to be 3 processes leading to haying-off. Firstly, the results confirm previous studies showing that haying-off was associated with reduced post-anthesis assimilation in response to a lack of soil water. The water deffcit was due to vigorous vegetative growth stimulated by a high level of soil N and was not associated with heat shocks or sudden increases in evaporation. Secondly, the most severely hayed-off crop failed to extract soil water fully, leading to a further reduction in post-anthesis assimilation. Thirdly, there was inadequate apparent retranslocation of pre-anthesis reserves to compensate for the lack of post-anthesis assimilation.

Biofumigation refers to the suppression of soil-borne pathogens and pests by biocidal compounds released by Brassica crops when glucosinolates (GSL) in their residues decay in soil. We conducted field studies at 2 sites to investigate the hypothesis that biofumigation by Brassica break crops would reduce inoculum of the take-all fungus Gaeumannomyces graminis var. tritici (Ggt) to lower levels than non-Brassica break crops, and thereby reduce Ggt infection and associated yield loss in subsequent wheat crops. High and uniform levels of Ggt were established at the sites in the first year of the experiments by sowing wheat with sterilised ryegrass seed infested with Ggt. Ggt inoculum declined more rapidly under Brassica crops than under linola and this reduction coincided with the period of root decay and reduced root glucosinolate concentrations around crop maturity. There was no consistent difference in inoculum reduction between canola (Brassica napus) and Indian mustard (Brassica juncea), nor between cultivars with high and low root GSL within each species. Despite significant inoculum reduction attributable to biofumigation, there were no differences in the expression of disease and associated impacts on the yield of subsequent wheat crops across the sites. Seasonal conditions, in particular the distribution of rainfall in both the summer–autumn fallow following the break crops and during the subsequent wheat crop, influenced inoculum survival and subsequent disease development. In wet summers, inoculum declined to low levels following all break crops and no extra benefit from biofumigation was evident. In dry summers the lower inoculum levels following brassicas persisted until the following wheat crops were sown but subsequent development of the disease was influenced more by seasonal conditions than by initial inoculum levels. Significant extra benefits of biofumigation were observed in one experiment where wheat was sown within the break crops to simulate grass weed hosts of Ggt. Under these circumstances there was greater reduction in Ggt inoculum under canola than linseed and an associated decrease in disease development. For host-dependent pathogens such as Ggt, we hypothesise that the benefits of biofumigation to subsequent wheat crops will therefore be restricted to specific circumstances in which inoculum is preserved during and after the break crops (i.e. dry conditions, grass hosts present) and where conditions in the following wheat crop lead to significant disease development (early sowing, wet autumn and spring, dry periods during grain filling).

The term dual-purpose canola describes the use of a canola crop for forage before seed production. It could potentially provide a profitable and flexible break-crop option for mixed farms, but there have been no studies to test the concept in Australia. We investigated the feasibility of using canola in this way in field experiments near Canberra, Australia, from 2004 to 2006, using European winter and mid–late maturing Australian spring canola varieties. Winter varieties sown from early March to mid-April produced 2.5–5.0 t/ha of biomass providing 0.3–3.5 t/ha of high-quality forage grazed by sheep in winter. The spring varieties produced similar amounts of vegetative biomass from April sowing but were unsuited to the earlier March sowing as they flowered in early winter and did not recover from grazing. The canola forage was readily eaten by sheep; alkane-based estimates of diet composition indicated that >85% of the organic matter intake consisted of canola. Canola forage was also highly digestible (86–88%) and Merino hoggets grew at 210 g/day from a dry matter intake of 1530 g DM/day. The canola generally recovered well when grazed in winter before bud elongation. Delays in flowering associated with heavy grazing ranged from 0 to 4 days when grazed before buds were visible, to 28 days if the crop had commenced flowering. Significant delays in flowering (>14 days) associated with winter grazing did not reduce seed yield or oil content when favourable spring conditions allowed compensatory growth. Yield loss was observed when winter and spring conditions were unfavourable for compensatory growth, or if grazing continued too late into spring (late September) irrespective of seasonal conditions. The yield loss was more than offset by the value of the grazed forage and the mean gross margin for dual-purpose canola over the four experiments was $240 to $500 higher than for grain-only canola depending on the value assumed for the forage. The study indicates there is considerable scope to capture value from grazing early-sown canola crops during winter without significant, uneconomic trade-offs with seed yield. Further investigations in other medium to high rainfall environments in southern Australia are warranted.