Experimental plot setup
The experiment was conducted in 2005 and 2006 at the Eastern South Dakota Soil and Water Conservation Research Farm near Brookings, SD (USDA, Agricultural Research Service, Northern Plains Area). Corn lines were planted in fields managed under a four-year rotation of corn, soybeans, oats, and spring wheat. Crop rotation was used to ensure that experimental plots were not contaminated by surrounding natural rootworm populations, which would increase variability within the study.
In order to determine appropriate fertilization levels, the soil was sampled on 4 April 2005 and 12 April 2006 from five locations throughout the field. At each location, five samples were taken from two depths (0–15 and 15–61 cm), and sent to a soil testing lab (South Dakota State University, Soil Testing Lab, Brookings, SD). On 29 April 2005, 157 kg/ha starter fertilizer (14-36-13) and 151 kg/ha urea (46-0-0) were surface applied, while on 4 May 2006, 177 kg/ha starter (14-36-13) and 105 kg/ha urea (46-0-0) were applied. Fertilizer was added to the field (19 m × 20 m) according to soil testing recommendations for 7500 kg/ha yield, and incorporated into the top 10.2 cm of soil via field cultivation.
There were 12 maize lines used in the study: 10 experimental inbred lines (SDG6, SDG7, SDG9, SDG10, SDG11, SDG12, SDG15, SDG17, SDG19, and SDG20) and two susceptible publicly available inbred lines (B73 and W64A). None of the experimental or public inbred lines had seed treatments. The original crosses made to establish experimental populations (Eubanks 1997, 2001, 2002) involved Tripsacum dactyloides, which is resistant to corn rootworms, crossed with Zea diploperennis, resulting in fertile intergeneric hybrids. All experimental lines were introgressed with genes from one or two Tripsacum-diploperennis recombinant lines (Tripsacorn and/or Sun Star) (Eubanks 2002, 2006). Tripsacorn has a tetraploid T. dactyloides as the seed parent (Source: Indiana University, Bloomington, IN; originally collected from Santa Claus, Spencer County, IN; 1949–1954) and Z. diploperennis as the pollen parent (Source: Upper las Joyas, Sierra de Manantlan, Jalisco, Mexico; Iltis, Nee & Guzman Acc. #1250; 1979). In contrast, Sun Star has Z. diploperennis as the seed parent (Source: Jalisco, Mexico; R. Guzman M. Acc. #777) and a diploid T. dactyloides as the pollen parent (Source: Manhattan, KS; K. Anderson). SDG6, SDG9, SDG15, and SDG19 have T. dactyloides cytoplasmic genes, meaning that T. dactyloides was the female parent of the bridging cross with Z. diploperennis. For SDG7, SDG10, SDG11, SDG12, and SDG17, Z. diploperennis was the female parent in the cross. SDG20 was derived from a three-way cross with parentage tracing to T. dactyloides as the female parent for two individuals and to Z. diploperennis as the female parent of the third. Reciprocal crosses were then performed between maize and the intergeneric hybrids (T. dactyloides × Z. diploperennis or Z. diploperennis ×T. dactyloides), and the resulting trigeneric hybrid plants backcrossed to maize or one of the intergeneric hybrids. The experimental inbred lines used in this study were derived from a recurrent selection breeding program using the maize inbreds B73 and W64A with a minimum of 14 generations of recurrent selection, backcrossing, and selfing. For more detailed information on recurrent selection methods, development of the experimental inbred lines, and genes involved in expression of rootworm tolerance see Eubanks (1998, 2002, 2003, 2006). The Sun Dance Genetics (SDG) lines in this study ranged from approximately 72% corn 28% exotic, to 97% corn 3% exotic.
The study consisted of two experiments conducted simultaneously: an evaluation of maize line susceptibility to larval corn rootworm feeding damage (root damage ratings, root fresh weight), and an evaluation of how rootworm infestation impacted grain yield. Experiments were conducted in adjacent blocks within the same field. We utilized randomized complete block designs with four replicates per experiment. Experimental maize inbred lines were planted within east-west single-row experimental plots with 0.76 m row spacing. Each experimental row was separated by a buffer row.
On 16 May 2005, buffer rows (DeKalb® 440) were sown at the label recommended density of 74 130 seeds/ha using an eight row vacuum planter (Max Merge 7200, John Deere, Moline, IL), while on 12 May 2006 buffer rows (DeKalb® 46-26) were sown at the label recommended density of 63 010 seeds/ha. Hybrid lines were used in buffer rows to protect inbred lines from adverse weather conditions, especially wind damage. While hybrid maize plants were approximately 2.0–2.5 m high and may have shaded some inbred maize lines, these effects were uniform across experimental rows. Maize inbred lines were hand-planted on 19 May 2005 and 16 May 2006 using jab planters (Easy-Plant Model 98; R.T. Adkins, Parsonsburg, MD) with eight seeds of one maize line per single-row plot, 5 cm seed depth, and 23 cm plant spacing. There were no formal buffer plants between experimental plots planted in the same row, however, the first and last plant in each experimental plot were not sampled. There was 0.76 m of buffer plants (2005, DeKalb® 440, 94 days relative maturity; 2006, DeKalb® 46-26, 96 days relative maturity) on the end of each row of experimental maize plants.
For evaluating maize line susceptibility to rootworms, all rows with experimental maize inbred lines were infested with rootworm eggs. In contrast, the second experiment evaluating grain yield was a split plot experiment with two treatments, (1) an agar-only control and (2) rootworm infested plots. Rootworm-treated rows were mechanically infested on 17 May 2005 and 15 May 2006 with 1000 viable western corn rootworm eggs per 30 cm suspended in a 0.15% agar solution (Palmer et al. 1977) at an approximate depth of 10 cm using Sutter and Branson’s (1980). Control plots in the second experiment were mechanically infested with only the 0.15% agar solution. Buffer plants were not infested with rootworm eggs.
Rootworm eggs were obtained from the primary diapausing colony maintained at the North Central Agricultural Research Laboratory in Brookings, SD. Hatch controls were performed prior to infestation to determine the percentage of viable eggs. Using a fine paintbrush, three batches of 100 eggs were placed on moistened filter paper in separate Petri dishes (100 mm × 15 mm), incubated at 25°C, and monitored for up to 4 weeks. In 2005, 87%± 2% of the eggs hatched, while in 2006, 86%± 4% of the eggs were viable.
In order to estimate western corn rootworm larval development and maximum root feeding damage, on 18 May 2005 and 16 May 2006 we placed the soil probe of a biophenometer (Model: BIO-51-TP03C; Omnidata® datapod, Logan, UT) to a depth of 10 cm into the soil and monitored soil temperature and growing degree days (GDD) with an upper threshold of 35°C and a lower threshold of 11°C (Fisher et al. 1990). Corn rootworm development is linked to temperature (Jackson and Elliott 1988; Woodson and Jackson 1996), and maximum rootworm damage occurs around the time adults begin emerging (Branson 1986).
Root damage and root fresh weight
On 13 July 2005 and 12 July 2006, four root systems per plot were sampled and rated for root damage. Root systems were removed from the soil when the majority of the insect population reached the pupal stage of development, which occurs approximately 600(base 11) GDD after egg infestation (Riedell and Evenson 1993). Plant shoots were cut above the lowest visible node and discarded, and the remaining stalk was labelled with a water proof (Tyvek) tag attached with a cable tie. The root systems were dug with a four pronged potato fork. Loose soil was removed from the root systems in the field by gentle tapping. Root systems were then soaked outside in mesh baskets suspended in tanks of water with water softener (1.9 l/tank; Calgon®, Reckitt Benckiser Inc., Wayne, NJ) to help disperse soil aggregates. In 2005, roots were soaked for 1–5 days, while in 2006, roots were only soaked for 1 day. After soaking, root systems were laid across wire mesh baskets and gently washed with high pressure sprayers to remove remaining soil without damaging the roots. Although some deterioration occurred while roots were soaking, it was primarily confined to the stem and did not interfere with root washing and processing. Root systems were then placed within doubled plastic garbage bags and stored in a 7°C cold room to retain moisture and prevent deterioration.
After 6–8 days in cold storage, shoots were cut at the seventh node, the top portion of the stem was discarded, and root system fresh weight was recorded. Root systems were then rated for rootworm larval feeding damage using the Iowa 1–6 scale (Hills and Peters 1971). This rating scale is based upon the following criteria: 1 = no root damage or a few feeding scars, 2 = feeding scars, but no roots pruned to 3.8 cm of the plant, 3 = several roots pruned to 3.81 cm, but an entire node of roots not pruned, 4 = one node of roots pruned, 5 = two nodes pruned, 6 = three or more nodes pruned.
To assess the impact of larval feeding on grain yield of the experimental maize lines, all ears were hand harvested from both non-infested control plots and rootworm infested plots from the second experiment. From 17 October 2005 to 21 October 2005 and from 23 October 2006 to 25 October 2006, ears were picked from all plants within the plots, excluding the first and last plant in each plot, for a maximum of six sampled plants. Ear damage from insect pests was negligible. Ears were placed in tightly-woven mesh bags, labelled, and air dried at approximately 21–26°C in a greenhouse for 7–10 days. Grain was removed from the cobs by feeding ears through a modified corn sheller (McCormick-Deering; International Harvester Co., Chicago, IL), and any missed kernels removed by hand. Grain was then cleaned using an Almaco grain cleaner (Allan Machine Co., Nevada, IA), which blows air over the sample, thus removing chaff and other light debris. The cleaned grain was weighed on a Mettler PC 4400 balance (Mettler Instrument Corp., Hightstown, NJ) and tested for percent moisture using a John Deere Moisture Chek Plus™ (John Deere, Moline, IL). These data were used to extrapolate grain yield at a 150 g/kg moisture basis.
Plant and primary ear node height were measured on three plants per plot in three replicates. Plant height was measured from the ground to the tip of the tassel using a digital reading measuring pole (Sokkia, Senshin Industry Co., Ltd, Osaka, Japan), while primary ear node height was measured from the ground to the uppermost ear node on the stem.
Root fresh weight, grain yield, plant height, and primary ear node height data were log(X + 1) transformed prior to analysis. All data were analysed using the GLIMMIX procedure (SAS® 2004, 2005) followed by Tukey’s Honest Significant Difference multiple comparison test (Ramsey and Schafer 1997). Bartlett’s test was used to assess the homogeneity of error variances. Because pollen availability can influence yield (Uribelarrea et al. 2002; Westgate et al. 2003), in yield analyses average days to anthesis was used as a covariate. Average days to anthesis values were calculated from 2 to 5 years of data collected from nurseries in North Carolina and Florida. Some experimental inbred lines had poor germination and ear development, thus, if plants did not germinate, yield data were reported as missing values. If plants germinated but did not produce ears with grain, yield was considered zero. Four lines had extremely poor germination and/or ear development (SDG10, SDG11, SDG15, SDG19), and while mean yield and days to anthesis data were reported, these lines were excluded from statistical analyses of yield data.