|
 Essay - sept/oct 2007
Statistics, approximations and close truths IV
James C. Delouche
Professor Emeritus Mississippi State University
Botanists define a seed in a very precise, developmental and morphological sense as a developed ovule. They are usually surprised and sometimes shocked at the broad, rather imprecise and functional agricultural definition of a seed. Agriculturalists consider a seed or rather a seed unit as the structure or organ(s) of a plant that is planted to produce a crop. For some agricultural crops the seed unit corresponds rather closely to the dispersal unit naturally separated from plants for reproduction and dissemination, while in other crops the natural dispersal unit is substantially modified in harvesting and processing to facilitate planting. Some seed units such as those of of Phaseolus spp. and Zea mays consist of one functional seed , others such as Cenchrus ciliaris and some other species of Poaceae consist of one or several functional seeds plus accessory structures, and still others consist of structures with multiple functional seeds such as multi-germ Beta vulgaris. The accessory material that is part of the dispersal unit and continues to be associated with the seed unit contributes greatly to the approximations and close truths in seed analysis and seed lot specifications related to the pure seed.
The Pure Seed
An important measure of the planting value of a seed lot is the percentage of pure seed. The definition of pure seed, however, is as complex and perplexing as the definition of the seed unit. The seed units of the desired species can appear to be shriveled and immature, diseased, insect and mechanically damaged to the extent that common sense would consider them as having no planting value. Common sense, however, does not seem to prevail in the purity analysis of seeds, especially to novices and beginners, since these apparently valueless seed units are classified as pure seed. I vividly recall my consternation as a beginning seed analyst when I was required to classify fragments of broken wheat seed larger that one-half of a whole seed without an embryo as pure seed and fragments less than one-half in size but containing the embryo as inert material. I have previously cited my difficulties in determining pure seed in Poa pratensis and similar species, i.e., how to distinguish between a floret with an immature caryopsis (pure seed) and one with dried anthers (inert material). After some time and a lot of indoctrination, however, I began to appreciate that compromises of these sorts with common sense were necessary to reduce subjectivity in the analysis of seeds and the considerable and uncontrolled variability associated with it. There is sound logic behind the compromise with common sense for pure seed because the planting value of the questionable pure seed units can be determined more objectively and consistently in the germination test than by subjective judgments based on visual clues.
Compromise with the truest values have been accepted in some cases to relieve analysts of especially tedious procedures such as in the purity analysis of species that have multiple units, e.g., Festuca rubra, Dactylsis glomerata. The multiple units variously consist of both fertile and sterile florets and other inert structures. Analysts used to have to separate the fertile units as pure seed from the sterile units as inert matter, a very tedious and fatiguing procedure. A sound, simpler and research based procedure was adopted with consists of determining the single units. and multiple units separately and applying a factor to determine the single unit equivalency for the multiple units. This convenience compromise obviously cannot produce the truest value, but experience has shown that in approximates the true value at least as close as the average results that were obtained by fatigued analysts using the old and tedious procedure.
Weed Seed and Crop Seed
In the determination of weed seed in a purity analysis, the compromise is in the direction of common sense: immature, undersize, broken, insect damaged and otherwise defective weed seeds which by visual examination fall into specified categories and hence appear to be incapable of producing a plant are classified as inert matter. The logic behind allowing this level of subjectivity in the purity analysis is that weed seeds, unlike the pure seeds, are not tested for germination. Interestingly, however, the determination of seeds of other crop species that occur as contaminants in the purity analysis follows the specifications for pure seed although other crop seeds like weed seeds are not subject to a germination test.
Germination
The methods and procedures for determining the germination of crop seeds are specified in great detail in the various rules for testing seeds There are, however, many inconsistencies in the prescriptions for dealing with certain kinds of seeds and/or seed conditions. The methods for dealing with dormant seeds, especially, exhibit inconsistencies that are difficult to rationalize except in terms of inadequate knowledge. One example that I am very familiar with is Paspalum notatum which has several forage cultivars. All of the cultivars are dormant, often deeply dormant, but for the Pensacola cultivar only light is specified for the germination test while for the other cultivars the glumes and other structures are removed with a "sharp scalpel" and the surface of the caryopsis is "lightly" scratched. The philosophy of dealing with seed dormancy has generally been that dormancy breaking treatments should be of the types to which seeds are naturally subjected or exposed. Thus, pre-chill, low-temperature stratification, alternating temperature, potassium nitrate, and light treatments have been the favored treatments for overcoming dormancy because they mimicked natural phenomena. The hulling and caryopsis scratching treatments for the Paspalum notatum cultivars other than Pensacola, however, surely do not mimic natural phenomena. The germination methods prescribed for the very dormant Stipa viridula, two methods one involving a soak in sulfuric acid, treatment with gibberellic acid and a fungicide, are also not very natural. And, they are many other examples of difficult to rationalize procedures for germination testing in the methods prescribed for crop, vegetable and tree seeds. The problem is that seed dormancy is a very complex and the conditions for its release under cropping conditions are not very well known or understood.
The determination of germination percentage, i.e. normal seedlings, has long been recognized as frequently drifting rather far from truthful values. The concept of "germination energy" arose early in seed testing in recognition of germination test deficiencies. Likewise, the vigor tests developed since the middle of the last century represent a serious attempt to provide information on the plant producing value of seeds that is closer to the truth than just germination percentage.
Tolerances
The tolerances in seed analysis are rather strong evidence of the recognition and acceptance that it involves many approximations and close truths. They are the statistically based variation allowed between results obtained in a first and second test, or as is the usual case, between the results specified on the seed label (presumably based on a first test) and those obtained from an inspection sample by a control laboratory. The tolerances take into account seed conditions that influence variability such as how free flowing are the seeds. Two types and applications of tolerances have always bothered me: first, the allowed tolerance of 2 for a specified 0 noxious weed seeds; and, second, the one way application of tolerances for germination; i.e., attention usually given only to germination percentages lower than that specified and not to those higher than specified which could have as much influence on planting rate as a lower than specified value. The solution to the first situation is to specify "no tolerance applied" in the seed law for prohibited noxious weed seeds. The second situation is disappearing with the increasing value and costs of high-tech seeds.
Some Final Thoughts
It is not difficult to identify convenience compromises, inconsistencies, and general deficiencies in the standardized methods for determining the quality and value of seeds. I sincerely wish, however, that no reader has been offended by any of the content of the four parts of Statistics, Approximations and Close Truth because that was not my intention. I still claim to be a seed analyst and what I have done or tried to do is to take some looks into the structure of seed analysis.
Considering the number of species that are cultivated and in the seed trade the overall structure of seed testing and analysis must be rated as remarkable, or more "truthfully", as extraordinary. It is a structure under the constant scrutiny of many dedicated seed analysts and scientists from the commercial and government sectors that is continually being updated and improved. There is no question about the great service and crucial importance of traditional seed analysis and the new analyses for genetic traits and events in the improvements in crop production and production efficiency underway in the industrialized countries and, increasingly, in the lesser developed areas of our world. The rapid transition to a global agriculture that is occurring and its biotechnological base provide both many opportunities for even greater service and important challenges that will be met in the objective, but cautious and thoughtful ways that have always characterized the discipline of seed analysis.
|
|
............
|
