The use of high nitrogen (N) inputs is becoming restricted in many cou
ntries due to environmental and economic considerations and reliable i
nformation is needed about parameters pertinent for breeding N-use eff
icient triticales (xTriticosecale Wittmack). This study reports estima
tes of variances, heritabilities, correlations, and regression for 36
triticales, grown at three locations for 2 years with two N regimes (z
ero and normal N supply). Eleven traits were assessed. A 45% mean redu
ction in grain yield by low N input was mainly the result of a 25% low
er number of tiller-bearing spikes per m(2) and 17% fewer kernels per
spike, whereas 1000-kernel weight was little affected. Genotype-nitrog
en interaction was of importance only in some environments and for cer
tain traits. Heritability estimates at both N levels were similar and
moderate to high. Correlations between N levels were high for nearly a
ll traits. Regression analysis for grain yield revealed only few N-eff
icient triticales with lower than unity slope together with a high gra
in yield. Based on the genetic parameters, it is concluded that a spec
ial low input breeding programme does not appear to be necessary in tr
iticale. The breeder can select at either of the N levels with similar
response. But he should include tests in low productivity environment
s at some early stages in the breeding process to avoid loss of geneti
c variation for N-use efficiency.