|[Image: Brent Theunissen harvesting at the research station]|
|[Image: Photo of variety testing]|
|[Image: Photo of rice field]|
|[Image: Photo of rice name after Caffey]|
|[Image: Table 1.]|
Breeders measure selection efficiency in Louisiana
Sterling Brooks Blanche and Xueyan Sha
The art and science of plant breeding has evolved much since it was first practiced during the early stages of crop domestication. And yet, some things remain the same. Historically, one of the greatest challenges plant breeders faced was to effectively differentiate between performance due to the plant and that of the environment in which the plant was grown. At some point in history, somebody probably tried to introduce the mango tree into Louisiana, and it likely didn’t do well in our sub-tropical climate. Were poor growth and yield caused by an inferior variety (genetics) or by an unsuitable climate (environment)? In fact, both genetic and environmental factors can affect what plant breeders see and, ultimately, the selections that they make.
The term environment here encompasses any and all factors that affect yield and performance at a single location. These factors are numerous and difficult to characterize, but they include climactic (rainfall amount and timing, daytime and nighttime temperatures, humidity, the ratio of sunny to cloudy days, etc.), biological (insect, disease, and weed presence) and cultural (soil texture, fertility and organic matter) factors. Human-made factors, such as planting date, row spacing, plant population and pesticide choices also constitute the environment.
For example, assume a breeder is working with rice varieties at two locations – one of which normally produces high yields and the other produces lower yields because of high levels of disease. If only the high-yielding material was selected for consideration, then all selections would come from a single location. Not all of the varieties at the lowyielding location are necessarily inferior, however, and some might exhibit valuable levels of disease resistance. In either case, it is the breeder’s job to sift through and use all of this information.
Progress in a breeding program depends on the ability of breeders to accurately quantify the variations among breeding lines and varieties. This critical ability encompassing both the art and science of plant breeding is paramount in the breeding and selection phase of variety development. All breeding programs are affected by environmental factors, and this is where the art of plant breeding is critical. Hundreds of factors interact to control what breeders view in the field, and it is impossible to scientifically model the environment so all factors are controlled or even understood. So, breeders use known standards to evaluate material on their merits in each situation.
No mathematical formula exists for selecting high-yielding plants; rather, it is an art of simultaneously considering a large number of yield-related factors or characteristics that contribute to high yields.
In 2007 and 2008, LSU AgCenter rice breeders evaluated the influence of different environments on their ability to estimate rice grain yield as part of the annual commercial-advanced trials. Trial locations included the Rice Research Station and on-farm test locations in Vermilion, Evangeline, Acadia and Jefferson Davis parishes. Two rice breeders visually evaluated 60 rice varieties for yield potential.
Each breeder’s criteria are unique, but they generally include important rice characteristics that include the strength/ thickness and height of the rice stalks (affects how rice can withstand high winds), the size of the rice panicle (total number of grains per panicle), the number of rice flowers that matured into whole rice grains on each panicle, the total number of panicles per square yard, seedling vigor (how aggressively the rice grows in cool, early-season conditions) and disease resistance.
The similarities between the actual yield and the estimates for each breeder were evaluated. A higher correlation indicates that the visual estimates were more predictive of actual yield than those with a lower correlation. The suitability of the environments for breeding purposes was based on the average yields of all varieties and the degree of differences among varieties at each location. Sheath blight disease ratings were recorded under significant disease pressure at the Vermilion and Jefferson Davis parish locations.
Rice yields varied between 5,114 and 9,266 pounds per acre across the seven environments. There were also significant differences in how much the environments separated the varieties. Some of the important environmental factors included a late planting date, high sheath blight disease pressure, excessive lodging (plants falling down), nitrogen deficiency, low tillering (production of multiple stems) and optimum yield conditions.
Table 1 shows the breeders’ visual ratings in 2007 and 2008. A higher number suggests the breeders were more accurate in determining the yield potential of a variety. Alternatively, a low number indicates that the true yield potential of a variety was not easily estimated. The major factor in the accuracy of the visual ratings was the year effect. There were no differences between the two breeders in either year; however, their ratings were more accurate in 2007 than in 2008. This is not surprising considering the tremendous variation in Louisiana weather patterns, both week-to-week and year-to-year.
Many environmental factors affect the way plant breeders perceive yield potential. In 2007, two locations were heavily affected by sheath blight disease, which provided the breeders a good visual clue to the yield potential of the varieties. While the ratings were more accurate where sheath blight was present, plant breeders also must consider that the best varieties at any location might not have the highest genetic potential in the absence of sheath blight.
Across all years and locations, the breeders’ yield estimates were similar to the actual yields, indicating that selections were highly effective. Research in other crops and locations has shown much lower predictive ability than in the current study. In some crops, researchers have not been able to demonstrate that visual selection was any more effective than randomly selecting plants regardless of their appearance. The variability in yield potential that the breeders perceived also was estimated. Not surprisingly, both breeders were more accurate in their selection of high yield potential when they perceived greater differences between varieties.
In agriculture, the overall challenge is to consistently produce a crop in spite of numerous variables. Each year presents different combinations of seasonal weather conditions that greatly affect the size and quality of the harvest. Variety development is a complex process that occurs over many years and in many different environments. Plant breeders must exploit rather than suffer from environmental differences and use them to their advantage. By the time an experimental line is considered for release, it has been exposed to many years of testing in many different locations. The hope is that the product of these efforts will be a time-tested variety that is well-adapted to the growing conditions and weather variables likely in Louisiana.
Sterling Brooks Blanche, Assistant Professor, Dean Lee Research Station, and Xueyan Sha, Associate Professor, Rice Research Station.
(This article was published in the winter 2011 issue of Louisiana Agriculture magazine.)