Accurate tools for measuring plant uniformity
Seed spacing may vary due to planter errors, poor planter maintenance, or seed germination problems. Average plant spacing and standard deviation of plant spacing are often used to determine plant spacing accuracy. Having only an average of plant to plant spacing is not a good measure of what is present in the field, because spacing between plants is not normally distributed (not uniform). The standard deviation is not a very good measurement either since it is based on squared deviations of the mean and is therefore influenced by a few very large spacings (skips or misses). Because of these problems, Kachman and Smith (1995) concluded that the mean and standard deviation are not appropriate for summarizing distributions of plant spacing. They compared four other measures that were based on theoretical spacing and found that they do a good job of summarizing distributions of plant spacing for single seed planters.
Theoretical spacing is the targeted distance between plants, assuming no skips, no multiples, and no variability in seed drop. Theoretical spacing is abbreviated as xref . The existing stand is assessed relative to what the plant to plant spacing should be. The theoretical spacing is used to divide the observed spacings into five divisions:
Division I = 0 to 0.5 xref. These are multiple seeds at the same spot or seed spacings that are closer than ½ the theoretical spacing.
Division II = 0.5 xref to 1.5 xref. These are single plant spacings that are close to the theoretical spacing.
Division III =1.5 xref to 2.5 xref. These are single skips.
Division IV = 2.5 to 3.5 xref. These are double skips.
Division V = over 3.5 xref. These are triple skips etc.
Four measures of plant-spacing accuracy are based on the frequency of spacings that occur in the five divisions. They are as follows:
Multiples index, D (doubles, triples, etc.), is a percent of spacings that are less than or equal half to the theoretical spacing, D = nI / N x 100 where: nI is the number of spacings in region I and N is total number of spacings. Smaller values of D indicate better performance than larger values.
Miss index, M (skips), is the percentage of spacings greater that 1.5 times the theoretical spacing: M = (nIII + nIV + nV ) / N x 100 where: nIII, nIV, and nV are the number of spacings in regions III, IV, and V and N is total number of spacings. These skips could be due to the failure of the planter to drop a seed or the failure of a seed to produce a seedling. Smaller values of M indicate better performance than larger values.
Quality of feed index, A, the percentage of spacings that are more than half but no more than 1.5 times the theoretical spacings: A = nII/N x 100 where: nII is the number of spacings in region II and N is total number of spacings. This is a measure of how close the spacings are to the theoretical spacing. It is another way to look at information in the other two indices since: 100 - (D + M) = A. Larger values of A indicate better performance than smaller values.
Precision, C, a measure of the variability in plant spacing after removing the variability due to skips and multiples. Precision is similar to a coefficient of variation for the spacings that are classified as singles (i.e. plants in region II): C = sII/xref where: sII is the standard deviation of the n2 observations in zone II and xref is the theoretical spacing. It is not affected by outliers, multiples or skips. A practical upper limit is 29%. Smaller values of C indicate better performance than larger values.
Kachman, S.D. and J.A. Smith. 1995. Alternative measures of accuracy in plant spacing for planters using singe seed metering. Transaction of the American Society of Agricultural Engineers. 38(2):379-387.
This text, written by Roger Elmore, is taken from a Crop Watch article (University of Nebraska extension newsletter) written May 17, 2002.