4. Discussion
Predictions of DM yield suggest that this parameter
varied greatly across the ®eld, but particularly at cut 1
(Fig. 4(a)), when a 5.5 t haÿ1 dierence was predicted
between the lowest and the highest yielding points. The
close relationship between the predictions of DM and N
yield (Fig. 6), and the preponderance of negative (DRIS)
N indices in the lower-yielding parts of the ®eld (Fig. 7),
indicated that N availability had been the factor most
limiting to sward production at each cut. Total N yield
(at both cuts combined) varied between 144 and 290 kg
haÿ1 (Fig. 5(c)), i.e., between 76% and 153% of the total
fertiliser N input. Some of this variability may have
arisen because of errors associated with the prediction of
DM yield and hence N yield. However, the fact that
ÔrealÕ N-yield values for the mowed strips also varied
widely (particularly at cut 1), implies that the variability
in mineral N availability across the ®eld had indeed been
substantial.