Initially five diverse and broad-based populations were sampled for leaf sheath composition in order to form a population designated WFISILO that was used to evaluate the effect of fiber, lignin and silica on insect behavior. The initial selection was to develop germplasm with extremes in fiber composition. The five source populations were as follows:1) Mo-ECB2 (Barry and Zuber, 1984), 2) Caribbean flint-dent composite (PI451690), 3) Northern Temperate Region (NTR) composite developed by CIMMYT, 4) Composite ARG AC, also known as “Argentine Colorado x Caribe” developed by J.H. Lonnquist , and 5) Maize-teosinte composite also developed by J.H. Lonnquist.
In 1983, early-flowering, vigorous plants from the five populations were self-pollinated to produce 193, 54, 49, 59, and 39 S1 families from Mo-2 ECB, the Caribbean flint-dent composite, NTR, ARG AC, and the Maize-Teosinte composite, respectively. In 1984, the S1 families were evaluated for leaf sheath composition. Twenty S1 families were selected based on low levels of acid detergent fiber (ADF), lignin and silica. Selected families were chosen from each of the five source populations (six from Mo-2 ECB, two from the Caribbean flint-dent composite, seven from NTR, four from ARG AC, and Maize-teosinte composite), and these were intermated for two generations to form WFISILO C0.
Three cycles of S1 recurrent selection were used to decrease fiber and lignin concentrations of stover tissue in WFISILO. During the first cycle of selection, 100 S1 families from WFISILO C0 were evaluated for leaf sheath and stalk composition at anthesis. An additive selection index comprised of four characteristics; leaf sheath NDF + leaf sheath lignin + stalk NDF + stalk lignin, was calculated for each S1 family. Ten S1 families with low INDEX values were selected and intermated for one generation to form WFISILO C1. Selection for silica concentration was discontinued because of unreliability of existing laboratory procedures and expense of adapting better assays. During the second cycle of selection, 100 S1 families were evaluated for whole-plant NDF and lignin. A selection index based on NDF + lignin, was used for selection. Ten families with the lowest index values from WFISILO C1 were intermated one generation using the bulk entry method to create WFISILO C2. The same procedure was used for the third cycle of S1 selection to produce WFISILO C3.
Similar procedures were used to also create a high fiber and lignin population designated WFISIHI, and the results of several population evaluation experiments have been reported in the literature (Buendgen et al., 1990; Wolf et al., 1993a, 1993b; Ostrander and Coors, 1997). It was apparent that selection was effective at altering fiber amount and fiber composition, and three inbreds with improved nutritional value have been released from the WFISILO population.
At the completion of third cycle in WFISILO, the 10 selected S1 families were crossed to the single cross H99 x Mo17, and then all three-way crosses were intermated for two generations to create the initial cycle of WQS (WQS C0, previously designated as QCOMP). Inbreds H99 and Mo17 were chosen based on stover assessments performed from 1991 to 1993 under the auspices of the previous UW Silage Research Consortium. Both inbreds have excellent stover digestibility on both a dry matter and fiber basis, as well as low stover NDF. By including these inbreds in WQS, the combining ability pattern of the population was better established.
WQS Selection Protocol
Approximately 200 S1 families were produced from WQS C0 in the 1995/6 winter nursery. During the summer of 1996 these S1 families were visually evaluated in a three-replicate trial at high plant-density (~30,000 plants/acre) for plant health and general agronomic acceptability. Approximately one-half were discarded, and 110 S2 families from the remainder were then crossed to LH119 in the 1996 winter nursery. These testcrosses were evaluated at Madison and Arlington during 1997 for silage yield and nutritional value. Twenty superior S2 families were selected based on an heritability-based selection index consisting of total dry matter yield, low NDF, high in vitro dry matter and NDF digestibility, and high protein (see table). The selected S2 families were recombined during 1998 to form WQS C1, and they were also included in the inbred development nursery.
The second cycle of selection began in 1999 when 365 S1 families developed from WQS C1 during the 1998/9 winter nursery underwent visual screening in a three-replicate trial under high planting density at the West Madison station. About 70% of these families were discarded, and the remaining 92 families were testcrossed as S2 families to LH198 in the 1999/2000 winter nursery. These testcrosses were evaluated in 2000 for yield, NDF, in vitro dry matter and NDF digestibility, protein, and starch.
Instead of selecting families by using a heritability-based selection index as was done for WQS C1, milk/ton of forage and milk/acre were estimated based on the MILK2000 equations developed with help of Joe Lauer (UW Agronomy), Randy Shaver (UW Dairy Science), Pat Hoffman (UW Dairy Science), and Eric Schwab (UW Dairy Science). Twenty S2 families from WQS C1 were selected based on superior yield and milk production potential. (See field trials from 2000 – LH198 – Forage yield evaluation and LH198 – Nutritional evaluation.) The 20 S2 families were recombined in the 2000/2001 winter nursery to create WQS C2.
The third cycle began in 2001. We were not able to produce new S1 families in time for a visual screening in the summer of 2001. Instead, we produced 186 new S1 families during the summer, and then 462 S2 families (2-3 per S1 family) during the winter. These S2 families were visually screened and crossed to HC33 during the summer of 2002. In 2003, 226 HC33 testcrosses were evaluated for yield and quality. Selection was based on procedures similar to those used for Cycle 2. Twenty S2 families from WQS C2 were selected for superior yield and milk production potential. (See field trials from 2003: WQS – Forage yield evaluation and WQS – Nutritional evaluation.) The 20 S2 families were recombined in the 2003/2004 winter nursery to create WQS C3. Approximately 300 new S1 families were created during 2004 to begin cycle 4, and S2 families will be screened and crossed to HC33 or related tester during the summer of 2005.
The fourth cycle of selection began in 2004. We produced 300 new S1 families during the summer, and then 887 S2 families (2-3 per S1 family) during the winter. These S2 families were visually screened under stress conditions (early planting, high plant density and disease pressure), and they were also crossed to LH244 during the summer of 2005. In the summer of 2006, 101 S2-3 testcrosses were evaluated for yield and quality in trial WQS244. The best families (20) were recombined in the summer of 2007 to create WQS C4. (See field trials from 2006: WQS244 – Forage yield evaluation and WQS244 – Forage quality evaluation.)
The fifth cycle of the WQS population was started in 2008. During the winter nursery of 2007-08, 20 WQS half-sib families derived from the intermating process conducted in summer 2007 were selfed. At least 10 self-pollinations were conducted for each of the half-sib families. Between 4 and 12 S1 families were derived from each of the 20 half sib families. A total of 200 S1 families were planted in our summer nursery in 2008 and approximately half of them were eliminated based on agronomic appearance. Two ears were harvested from each of the selected S1 families producing a total of 200 S2 lines. Those lines were crossed by LH244 during 2009 summer nursery. A set of 188 WQS C4 derived S2 lines crosses were evaluated at two locations (West Madison and Arlington) in 2010 with two replications per location in trial WQS C4 S2. The best 20 S2 lines were identified from this evaluation and will be recombined in summer nursery 2011 to create WQS C5. (See field trials from 2010: WQS C4 S2 – Forage yield evaluation and WQS C4 S2 – Forage quality evaluation.)