Cooperative Extension Service
College of Agricultural and Environmental Sciences
The University of Georgia

Volume 4/Number 6

June 2000

George Boyhan, Extension Horticulturist-Vegetables
East Georgia Extension Center
Rosenwald Building, Third Floor
PO Box 8112
Georgia Southern University
Statesboro, GA 30460
912-681-5639 | 912-681-0376, Fax | 912-682-3481, mobile

Results of Vidalia Onion Variety Trial

The table below has results of our variety trials with Vidalia onions from this past year. We had two trials - one at the Vidalia Onion and Vegetable Research Center and the other at an on-farm location in Tattnall County. We followed recommended practices for fertility, pest management, weed control, and water management in producing these crops.

The experimental design was a randomized complete block design with four replications. Each plot was 50 feet long and 6 feet wide. A 15-foot length of each plot from both trials was harvested when the onions from that variety were considered at optimum for harvest. Onions were pulled the day prior to clipping and bagging. Field weights were recorded at harvest. Onions were transported to the Vidalia Lab in Tifton, where they were graded by size and re-weighed. In addition, all plots were evaluated for seedstems, doubles and bacterial disease.

Overall, this was a very good season for onion production. The lack of rain, which has been the bane of growers for the past two years, has actually worked to the advantage of onion growers. Since they have to irrigate to produce the crop anyway, the lack of rain has helped in reducing disease pressure while allowing growers to partially cure onions in the field and to schedule their harvests without rain delays.

Looking at the data for disease incidence, the disease pressure was much higher at the Vidalia Farm compared to Bland's Farm. Overall, the results were consistent from the two locations. Some of the highest incidences of bacterial disease occurred with 'Pegasus' and 'Centaur.' 'Nikita,' which is a common Texas grano type, had the highest incidence of disease at the Vidalia Farm. Some of the lowest incidences of disease occurred with 'Georgia Pride,' 'WI-609,' and 'WI-3115,' which are Japanese overwintering onions.

Field weights were very good at both locations. Yields ranged from 1,164 bags/acre for 'DPS 1033' to 508 bags/acre for 'DPS 1058' at the Bland Farm. At the Vidalia Farm, yields ranged from 1,107 bags/acre for 'Savannah Sweet' to 554 bags/acre for 'SSC 6372.'

Probably the most telling bit of information from the trial is the dramatic increase in culls due to disease that occurred as the later varieties were harvested. The rate of culls was 100 percent for 'Centaur' at Bland's Farm and 98 percent at the Vidalia Farm. All of the later varieties had very high rates of culls.

Garlic Variety Trial Results

Georgia has several small elephant garlic producers but does not produce any regular garlic commercially. For the past couple of years, I have been running trials to evaluate garlic varieties in Georgia.

Below are the results of this past year's trials. One trial was held at the Vidalia Onion and Vegetable Research Center just outside Reidsville, and the other was held at an on-farm location in Screven County.

The design was a randomized complete block design with four replications at the Vidalia Farm and three replications at the on-farm location. Yields were taken in the field directly after harvest. The width of five bulbs from each plot at the Vidalia Farm also was measured. Plot size at the Vidalia Farm was 8 ft long by 6 ft wide. At the Screven County location, the plots were 10 ft long by 6 ft wide.

Yields at the Vidalia Farm ranged from 176-332 bags per acre and at the Screven County location, from 220-528 bags per acre. At the Vidalia Farm, the earlier harvested varieties performed better than the later harvested varieties. The California market opens in June, and we hope to beat this market. Representatives from Bejo Seed Company, while visiting the Vidalia Farm, inspected our garlic and indicated the material harvested on May 8 was immature. The material we harvested on June 5 and 6 appeared to be past its prime harvest date. We think that the ideal harvest date in southeast Georgia would probably be between these two dates, particularly for 'California Early' and similar types. Later maturing garlic such as 'California Late' do not appear to do very well in southeast Georgia. This may be due to increased heat during their optimum harvest window.

Several plots were killed early on by what appeared to be a soil-borne disease. Plots of 'California Late' and 'VE-2550' had this problem. The numbered lines from Basic Vegetables can be described as L-5150 - like California Late; VE-2550 - like California Early; LE-4050 - like California Early; and VE-1750 - like California Early. The commercial variety most suitable to southeast Georgia at this time is 'California Early.'

Interpreting Statistical Results

Since I'm supplying the results of both our onion and our garlic trials with this newsletter, I thought it might be a good time to briefly go over the statistics used with these results. There are three statistics listed at the bottom of some of the columns to help interpret the results of that column. The statistics are R2, which is the squared multiple correlation; CV, which is the coefficient of variation; and LSD, which is the least significant difference.

CV

CV is the easiest to compute. This value is the standard deviation divided by the mean and multiplied by 100. This value is actually much more useful when measuring items for sameness. For example, you may have test strips that measure pH, and you want to know how repeatable these strips are. So you make many measurements with the strips and compute the CV. A low CV (below 1 percent, for example) indicates a high degree of repeatability. Many researchers, particularly when working under laboratory conditions, want to see CVs below 10 percent. But remember, when dealing with treatments that have large effects on the experimental units, it is not uncommon to have very high CVs. In general, however, lower CVs are considered better. It should be noted also that the CVs can compute to more than 100 percent. The real usefulness of using a CV is that it is independent of the units of measure. This means you can use this value to compare one experiment to another, even if the units of measure are completely different.

R2

The R2 is a more complex value to compute and is best left up to the computer. This value will range from 0-1, with values closer to 1 considered better. This value gives an indication of how well the independent variables (variety, replication) predict the dependent variables (yield, disease incidence, bulb size, etc.). For example, a value of 0.75 percent would indicate that the varieties and replication account for 75 percent of the predicted values such as yield. The remaining 25 percent is a random effect or is not accounted for by the model and is thus unknown.

LSD

The LSD is the most useful of the three statistics. To give some background to this statistic, we need to discuss an analysis of variance (ANOVA). When evaluating a population, there are two types of measurements that can be made. One is measures of central tendency. These would include measurements like the mean and median. There are also measurements of dispersion. The most important of these is a variance. Variances can be pulled from a population based on a variety of criteria. The underlying assumption is that all variances pulled from a population should be the same. Differences among variances are an indicator that you are dealing with different populations. Differences are generally measured in terms of a probability. In most agricultural settings, a probability of either 1 percent or 5 percent is generally used. Occasionally 10 percent is used. In our case, we measure differences at the 5-percent level. Once you've established that different populations exist, then an LSD can be computed from data generated in an ANOVA to give you an idea of which populations differ. In our variety trials, the populations are the different varieties. To be absolutely correct in using an LSD, it should be used only to compare adjacent means that are in rank order. In practice, however, an LSD is often used to compare any two means.

Vidalia Onion Business Council Asks for Changes

There continues to be concern over the onion varieties that growers are using. Last year, several varieties were eliminated as Vidalia onions. These included 'Sugar Queen,' 'Spring Express' and 'Sweet Dixie.' This year, there has been concern over the variety 'Georgia Pride.' The Commissioner of Agriculture, Tommy Irvin, is considering eliminating this variety as a Vidalia onion variety.

In addition, the Vidalia Onion Business Council has requested that growers not grow '6712 (Centaur),' 'SSC 6371,' 'SSC 6372,' 'SSC 6436,' 'RCS 1919 (King Midas),' 'DPS 1058 (Sweet Advantage),' 'Southern Honey' and 'Southern Belle.' They have requested that the Commissioner disallow the growing of these varieties until more information is available on their performance.