Objectives In order to understand the responses of yield and phosphorus efficiency of winter wheat cultivar improvement to various soil fertility levels, we investigated grain yield, yield components and phosphorus physiological efficiency of winter wheat cultivars released in 1980s, 1990s and recent years under different soil fertility levels, which have been built up in a long-term fertilization experiment on manural loess soil in Guanzhong Plain, Shaanxi province, China.

Methods The long-term experiment included seven treatments, control (CK, no nutrient input), two doses of manure amendment (M1 and M2), the integrated application of manure and chemical fertilizers of nitrogen (N) and phosphorus (P), and both at two doses (M1N1P1, M1N2P2, M2N1P1 and M2N2P2). The tested winter wheat cultivars were named as Xiaoyan 6 (extensively grown cultivar in 1980s), Xiaoyan 22 (extensively grown cultivar in 1990s) and Xinong 979 (cultivar in recent years).

Results The wheat yields generally showed a tendency of progressive increase with the improvement of cultivar, and were significantly enhanced with the increase of soil fertility regardless of cultivars. Nevertheless, the pronounced yield increases due to cultivar improvement were only observed on soils with high fertility such as the MNP treatments, while the yield benefit was minor on low fertility soils (e.g. CK, M1 and M2 treatments). During the cultivar development process, under the lower soil fertility condition spike number per unit area of cultivar Xiaoyan 6 was higher than that of cultivar Xiaoyan 22, and the current cultivar Xinong 979 yielded the highest spike number. However, under the high soil fertility, the spike number per unit area tended to be increased with the cultivar improvement. In addition, the grain number per spike showed an inconsistent trend with the cultivar development under the low soil fertility levels, but the general increasing trend was detected under high soil fertility levels. The 1000 grains weight showed an increasing trend with the improvement of cultivars at all tested fertilizer treatments. Similar to the grain yield, the increases in soil fertility and/or nutrients input also enhanced the spike number per unit area, grain number per spike and the weight of 1000 grains, respectively, regardless of cultivars. The phosphorus requirement for producing 100 kg wheat grain decreased with the cultivar improvement whatever the soil fertility level is, with an order of Xiaoyan 6 > Xinong 979 > Xiaoyan 22, and the phosphorus requirement under high fertility soils was lower than that under low fertility soils. The P physiological efficiency improved with the cultivar improvement irrespective of soil fertility, and the P physiological efficiency under high fertility soils was higher than that under low fertility soils.

Conclusions Wheat yield gradually increased with the cultivar improvement in Guanzhong Plain, Shaanxi province, which manifested by the increases in the weight of 1000 grains and spike number per unit area for cultivar Xinong 979, and in grain number per spike and weight of 1000 grains for cultivar Xiaoyan 22. However, the synergistic increases in the yield and P efficiency as a result of the cultivar improvement can be made full use only on high fertility soils. Hence, it is necessary to improve soil fertility to match the feature of current wheat cultivar on one hand, meanwhile wheat breeding should take the adaptation of new cultivars to the low soil fertility into consideration to fully explore yield potential of a given cultivar on the other hand in the region.