|
 Essay - nov/dec 2003
Appearance, Perceptions and Reality
James C. Delouche
Professor Emeritus Mississippi State University
Things are not always as they appear to be. Perhaps, this is the reason that so-called reality programs are among the most popular shows on U.S. television. The reality program genre includes the televising of hours of actual judicial proceedings, usually lurid murder trails, and more recently special programs designed to expose and record the interactions, competition and developing relationships of a small group of "ordinary" but carefully selected people placed with limited resources in a stressful environment, i.e., survival mode, such as a small tropical island, a jungle, a house, etc. Reality programs are claimed to present vignettes of "life" as they actually are and as they occur rather than in the usually stylized, idealized and fictional manner of drama. What one sees is real: appearance equals reality.
Scientific data and findings are properly considered as inerrant unless there is some obvious flaw in the experimental design or instrumentation. They are reality. Most laypersons and even some scientists also assume that if the data are inerrant, the implications or conclusions based on them are also inerrant. The implications and conclusions, however, deal with the appearance of the findings to the investigator(s) which may or may not correspond to reality. Differences between appearance and reality in science are not uncommon. The widely differing conclusions and recommendations presented in the media based on essentially the same databases are abundant evidence of how differently scientific findings are treated and interpreted by different scientists and related professionals. For example, the increasing concentration of carbon dioxide in the atmosphere is an indisputable fact, i.e., reality, but there is no agreement on what this means in terms of the rate and extent of global warming. The field of human nutrition, e.g., diets, is especially afflicted by widely different conclusions and recommendations based on essentially the same databases. A high-fat, high-protein, low-carbohydrate diet is recommended and followed by many, while others are convinced that a low-fat, moderate-protein and relatively high-carbohydrate diet is the ideal.
There seems to be no escaping the pitfall of crowning an inerrant data set with an errant conclusion. Seed science and technology and seed scientists and technologists, including the author, are as vulnerable to errancy as others. There are many causes of errancy but the most common are inadequate or inappropriate experimental controls, invalid assumptions, confusion of cause and effect, and faulty logic. Many interesting and
significant instances of the separation of appearance and reality can be cited but to avoid pointing the finger at others I will start by pointing at myself.
Inadequate Controls. As a graduate student at Iowa State University in the early 1950s I studied seed dormancy in the common lawn weed, Digitaria sanguinalis. One of my hypotheses was that a water-soluble inhibitor in the hulls surrounding the seed
(caryopsis) was involved in the dormant condition. To test this hypothesis I designed a simple experiment: freshly harvested, deeply dormant seeds were soaked in flowing water for periods up to 7 days (168 hours). The control, 0 hours soaking, germinated less than 5% while about 90% of the seeds germinated after a 168 hour soak in flowing water. I concluded that the soaking treatment leached a water-soluble inhibitor that caused dormancy in seeds of D. sanquinalis and published a small paper reporting the results and my conclusions. A few years after I joined the faculty at MSU in the late 1950s I decided to take up studies on seed dormancy in D. sanguinalis again using seeds collected in Mississippi. I repeated the soaking experiment but it had no effect on dormancy: the seeds germinated less than 5% at 0 hours and less than 5% even after a 192 hour soak. I concluded that the Mississippi seeds represented a different ecotype from the Iowa seeds used in the first study, made arrangements to obtain some seeds from Iowa and repeated the seed soaking experiment with them. Again, the soaking treatment had no effect. I was both puzzled and embarrassed, put the studies away and turned to other research. More than a year later I had the sudden idea that the differences in response of D. sanguinalis seeds to the water soaking treatments in Iowa and Mississippi must be related to some property of the water, perhaps its temperature. A colleague at Iowa State measured the temperature of flowing water from a water tap in the same lab and month as in the original experiment and found that it was about 12 degrees C. I measured the temperature of flowing water at Mississippi State and found that it was about 22 degrees C. Since so-called "pre-chill" treatments are widely used to break dormancy in various grass (Poaceae) species, I set up a simple experiment to determine if the 12 degree temperature of the Iowa soaking water served as a "pre-chill" treatment to release seed dormancy in D. sanguinalis. It did. An inhibitor was not involved. I have to admit I never published a correction to my original paper on the inhibitor and never studied dormancy in D. sanguinalis again.
The Hill variety of soybean was the most widely planted early maturing variety in Mississippi in the 1960s. The seeds reached harvest maturity in early September. During the same period the Bragg variety was perhaps the most widely planted late maturing variety. The seeds reached harvest maturity in late October. Hill soybean seeds had serious quality problems, e.g. germination, vigor, appearance, nearly every year and producers had great difficulty meeting even minimum germination standards. On the other hand, Bragg soybean seeds had few quality problems and seed producers were able to label them well above the minimum germination standard. I examined many batches of Hill seeds - no experiments, just observations - and began to think that the Hill variety was simply genetically inferior to Bragg in terms of seed quality. After relating my views at length and in detail at a regional meeting of soybean seed production, I was challenged by one of my graduate students. He conceded that there were substantial differences in seed quality between the two varieties but contended they were related primarily to environmental rather than genetic factors and suggested a study to establish which factor was dominant. Whole plants of the Hill variety were cut at the soil line when seed moisture content decreased below 25 % (early September) and placed under an open-sided shed. When the Bragg variety reached the same below 25% moisture stage in mid-
October plants were also cut at the soil line and repositioned in essentially the same row in a wire frame. On the same day the Hill plants were removed from the shed and repositioned in wire frames in rows adjacent to those of Bragg. One treatment consisted of simply exposing the plants to the field environment for periods up to 40 days while in the second treatment the plants were sprinkled with about 1 cm of water daily at 1800 hours simulating a daily rain. Seed pods were harvested every 4 days, dried, threshed and the seeds were tested for germination and vigor. There were no differences in responses of the Hill and Bragg seeds exposed to the ambient field environment or the artificial weathering treatments in terms of germination and vigor. Seeds of both varieties lost germinability and vigor at the same rates - slowly for the ambient conditions treatment and rapidly for the simulated rainfall treatment. Additional studies established that the field environmental conditions were much more severe in early September at the time of Hill maturity, than in late October at the time of Bragg maturity. The average temperature in late October was about 10 degrees C. cooler and the relative humidity about 15% lower than in early September. The solution to the seed quality problem of early maturing varieties such as Hill turned out to be a shift in seed production to the most northern portions of the area of adaptation rather than breeding a more weather resistant variety. We worked with breeders on the latter approach but without much success.
These are two examples of errant conclusions based on scientific findings or observations that were the result of inadequate and/or inappropriate controls. In both cases the data and observations were real. Dormany in seeds of D. sanguinalis was released by the soaking in flowing water treatment in Iowa and the seeds of the Hill variety of soybeans were afflicted by more frequent and severe quality problems than those of Bragg. The "appearance" of the findings to the investigator (me), however, led to erroneous conclusions which were challenged and required additional and better controlled experiments to move closer to the reality position. These cases were very important in my professional development. I became more skeptical of my scientific findings and those of others. I began to design experiments much more carefully and, most importantly, I became very cautious about "jumping" to the seemingly obvious conclusion or the facile interpretation. I learned to think and think and think again before I published, made recommendations or even voiced strong views at meetings.
In the next essay I will relate some additional and familiar errancies in conclusions and recommendations that resulted from the confusion of cause and effect, invalid assumptions and faulty logic. It is my hope that these expositions will promote the wider adoption of an open-minded but skeptical attitude regarding the claims and counter claims in the heated and nearly continuous debates on GM products, environmental issues, and many, many other technical issues that have a heavy emotional content.
|
|
............
|
