The value of Leucaena leucocephala (Lam.) de Wit (leucaena) in silvopastoral systems is well documented. However, damage by the psyllid insect pest (Heteropsylla cubana Crawford), constrains productivity and limits adoption of leucaena pastures in more humid regions of Queensland. The interspecific hybrid L. pallida (Britton & Rose) x L.leucocephala ssp. glabrata (Rose) S. Zarate, referred to as KX2 hereafter, has agronomic attributes that can overcome this problem. However, the commercial utilization ofKX2 is prohibited by the difficulty of Fi hybrid seed production. Vegetative propagation ofKX2 is possible, but expensive and not suited to large-scale broad-acre planting. A recurrent selection breeding program was undertaken to produce a genetically stable, advanced generation KX2 hybrid suited to commercial seed production.
Agronomic attributes, heritabilities and selection strategies to breed a genetically stable hybrid were investigated. Five elite KX2 F1 hybrids, identified in earlier work, were chosen as the foundation population of the breeding program. Two cycles of mass selection, conducted at different sites during the years 2001-2004, are reported in this work. In both cycles, alternating reference trees of cv. Tarramba and the KX2 F1 hybrid were used to benchmark psyllid damage and biomass yield, and were arranged in an alpha-lattice design. Dry matter (DM) yield was estimated from a non-destructive yield index (plant height x basal stem diameter2). Rapid assessment methods were used to measure psyllid damage as well as other important agronomic traits of tree form and leaf production. Spatial variation in environmental and competition effects across the plots were accounted for using a linear mixed model by residual maximum likelihood (REML) (ASReml, New South Wales Agriculture, Australia). In both cycles, elite KX2 trees were successfully selected by interrogating the data with Microsoft Access® and subsequent field validation.
A high degree of segregation for all agronomic attributes was observed in both cycles. The average DM yield of the Elite KX2 F2 (cycle 2) and Elite KX2 F3 (cycle 3) populations were 198% and 82% that of cv. Tarramba respectively. The yield index from the Elite KX2 F3 (82%) did not improve compared to the control cv. Tarramba (100%), and probable inbreeding depression for yield was identified from the low yield of the F3 population relative to cv. Tarramba, and the heavily skewed frequency distribution of yield in the KX2 F3 towards poor yield.
The psyllid damage rating (PDR) ranged from 1 (no damage) to 9 (stem death) in both cycles. Psyllid resistance was evident in the elite KX2 trees as they were clearly the most psyllid tolerant lines with PDR values of 1.3 (cycle 2) and 1.5 (cycle 3), compared to the psyllid-susceptible cv. Tarramba (PDR = 6, loss of 50% of young leaves).
Branchiness (1 = single stem, 5 = heavily branched) and leafiness (1 = few leafs, 5 = full of leafs) showed the same segregation from 1 to 5. Elite trees had excellent form for forage shrubs, with greater basal and secondary branching than cv. Tarramba. The average branchiness and leafiness ratings were 3.0 and 2.9 in cycle 2 and 3.2 and 2.8 in cycle 3 respectively. The cv. Tarramba had lower ratings with the exception of leafiness in cycle 3, which was higher than the Elite KX2 F3.
Narrow-sense heritability for each trait was estimated from inbreeding coefficients and variance components from ANOVA. The inbreeding coefficients for F2 and F3 populations were 0.117 and 0.271 respectively. Assuming no self fertilization occurred, equal pollen contribution from all F2 parents, no dominance effects and diploid inheritance estimated heritabilities of 0.31 ± 0.05 for yield, 0.28 ± 0.04 for psyllid resistance, 0.19 ± 0.03 branchiness and 0.26 ± 0.04 leafiness were observed in the breeding program.
Two alternative breeding strategies are proposed to overcome inbreeding depression for yield, whilst capitalizing on the genetic gain made in psyllid resistance. The first approach is to develop genetically stable inbred lines from self-compatible Elite KX2 F3 hybrids and discontinue selection for self-incompatible out-crossing lines. The second strategy involves backcrossing cv. Tarramba to these highly psyllid resistant Elite KX2 F3 hybrids to produce a psyllid-resistant cultivar, which may contain 75 or 87.5% of the genetic composition of cv. Tarramba.