Knowledge of how reduction in the rate of herbicide application or rotation of their mode of action influences weed growth will provide insight into how successful these practices will be in an integrated weed management program. Field experiments were conducted in 1996 and 1997 to quantify velvetleaf growth response to three postemergence herbicides, each with a different mode of action. A monoculture of velvetleaf was treated with halosulfuron, dicamba, and flumiclorac at 0, 0.10, 0.25, 0.50, 0.75, and 1.0 × the labeled rate for weed control in corn. Percent plant mortality increased with rate of application; the greatest mortality occurred in flumiclorac treatments in 1996 and in halosulfuron and flumiclorac treatments in 1997. Growth rate temporarily decreased as application rate increased. Maximum height decreased as rate of application increased, with the dicamba treatment resulting in the greatest (27%) reduction. Early-season leaf area index decreased with increasing rate of application, the greatest reduction occurring with halosulfuron (1997) and flumiclorac (1996 and 1997) treatments. The number of leaves produced per plant was temporarily reduced by all treatments, but treatment with dicamba later resulted in larger numbers of small leaves. The number of velvetleaf seed capsules produced per surviving plant was not reduced by any treatment, but the number of capsules per square meter was reduced by the 0.5 × rate of flumiclorac (1996) and the 0.5- and 1.0 × rates of halosulfuron (1997). Research is needed to evaluate whether the temporary suspension of velvetleaf growth after herbicide treatment is sufficient to prohibit crop yield reduction and velvetleaf capsule production.

Managing fertilizer in cropping systems may be an important component of integrated weed management programs. A field study was conducted to determine the effect of various application methods of the 15N-enriched nitrogen (N) fertilizer on N uptake in green foxtail, wild mustard, and spring wheat and on crop–weed competition. N application methods consisted of ammonium nitrate in solution applied broadcast on the soil surface, applied in pools on the soil surface at 20-cm intervals between every second wheat row, and point injected 10 cm deep at intervals similar to those of the surface pools. An unfertilized control treatment was also included. N uptake by green foxtail throughout the growing season was often greater from surface broadcast than from surface pools or point-injected N and was sometimes greater from surface pools than from point-injected N. In contrast, N uptake by wild mustard was rarely affected by the fertilizer placement method. In the presence of weeds, the ranking of N uptake by wheat usually was point injected > surface pools > surface broadcast. Weed biomass was often greater with surface broadcast than with either surface pools or point-injected N. In the absence of weeds, wheat yields were similar across the three N application methods. However, in the presence of green foxtail, wheat yields were greater with point-injected N than with surface broadcast N in two of the 3 yr and with surface pools of N in one of the 3 yr. In the presence of wild mustard, wheat yields were greater with surface pools and point-injected N compared with the unfertilized control in two of the 3 yr, whereas yields with broadcast N were never greater than the unfertilized control. Study findings suggest that point-injected N results in suppressed weed growth, not by reduced N uptake by weeds but instead by greater N uptake by wheat that increases its competitiveness with weeds. Information gained in this study will be utilized to develop a more integrated program for weed management in spring wheat.

Leafy spurge is a tenacious perennial weed of the Northern Plains. This plant maintains a perennial growth cycle by controlled production and growth of numerous underground adventitious buds. We are using molecular tools to identify signaling pathways that control underground adventitious bud growth and development in leafy spurge. Toward this end, we have used three techniques to identify genes that are differentially expressed concomitantly with the breaking of quiescence in underground buds of leafy spurge. These techniques include differential display of cDNAs, random cloning and sequencing of genes expressed in growing buds, and microarray technology. To date, we have identified more than 16 genes that are differentially expressed in underground buds of leafy spurge during dormancy break and growth initiation. A detailed expression analysis of these genes will allow them to be grouped by their responses to various signals known to play a role in control of underground bud growth. This information will be used to identify key cis-acting elements involved in the regulation of these genes. How such information on signal transduction processes may be used for developing new weed control strategies by the identification of novel target pathways and development of DNA-based herbicides is presented.

Studies were conducted to evaluate weed management programs in nontransgenic, bromoxynil-resistant, and glyphosate-resistant cotton in nontilled and tilled environments. Tillage did not affect weed control provided by herbicides. Early-season stunting in nontilled cotton was 3% regardless of the herbicide system and was no longer evident at midseason. Cotton yield was 10 to 15% greater, on an average, under tilled conditions than that under nontilled conditions. Excellent (> 90%) common lambsquarters, entireleaf morningglory, ivyleaf morningglory, jimsonweed, pitted morningglory, prickly sida, tall morningglory, and velvetleaf control was achieved with treatments containing pyrithiobac, bromoxynil, and glyphosate. Preemergence (PRE) or postemergence-directed (PD) herbicide inputs were necessary for adequate large crabgrass and goosegrass control. Bromoxynil and pyrithiobac postemergence did not control sicklepod unless supplemented with MSMA and followed by a late-postdirected treatment of cyanazine plus MSMA. Treatments that included glyphosate controlled sicklepod regardless of the late-PD treatment. Economic returns were at least $930 ha−1 and not different from the higher yielding programs in nontransgenic cotton when fluometuron applied PRE was included in the bromoxynil programs. Late-season weed control was usually greater than 90% from glyphosate programs, and net returns from glyphosate programs were as high or higher than the net returns from programs that used midseason treatments of bromoxynil, pyrithiobac, or fluometuron plus MSMA.

Managing weed infestations in a spatially precise manner requires accurate and cost-effective weed identification techniques. The goal of our research was to quantify the accuracy of continuous weed presence–absence maps and assess how management based on those maps may affect producer net returns. Each continuous sampled map covered the entire field and contained vector polygons labeled as either wild oat presence or wild oat absence. The accuracy of the continuous wild oat maps at each sampling time was determined from georeferenced quadrats of wild oat densities. The accuracy of the continuous wild oat seedling maps ranged from 48.3 to 87.1% among the six site-years. The accuracy of the wild oat seedling maps improved by at least 8% when a 10-m buffer was included around areas mapped as wild oat presence. The accuracy of continuous wild oat panicle maps from the combine at harvest ranged from 65.8 to 90.9% among the six site-years. The variation in accuracy for the wild oat seedling maps among sites was greater than the accuracy of the panicle maps. Net returns ($ ha−1) for four site-years were calculated and compared for four possible weed management approaches on each field. A site-specific herbicide application to areas mapped as wild oat presence always generated higher net returns than a herbicide application over the entire field for four sites. A site-specific herbicide application to areas mapped as wild oat presence plus a surrounding 10-m buffer area only resulted in a higher net return in one of the 12 site-years compared with a site-specific herbicide application without the 10-m buffer. This site had the lowest (48.3%) wild oat seedling map accuracy, and uncontrolled wild oat had a high-yield effect. This research indicates that using a continuous weed sampling method based on presence or absence for site-specific herbicide application can be profitable over a herbicide application to the entire field, even with the associated technology cost and seedling map errors.

We studied the molecular genetics of dormancy-related traits in an F2 family of poplar (Populus) hybrids derived from a cross between a northern genotype of black cottonwood and a southern genotype of eastern cottonwood by mapping quantitative trait loci (QTLs) and candidate genes. Dormancy-related traits included timing of vegetative bud set, fall frost damage, chilling response, timing of vegetative bud flush, and winter survival under field conditions, as well as photoperiodic responses (timing of bud set and number of new leaves) in a warm greenhouse under either a uniform 8-h photoperiod or a naturally shortening photoperiod in the fall. QTL analyses were conducted using a linkage map consisting of AFLP, microsatellite, and candidate gene markers. The candidate genes were chosen because of their potential roles in either photoperiodic perception (PHYB1, PHYB2) or abscisic acid signal transduction (ABI1B, ABI1D, ABI3). Significant QTLs were detected for all dormancy-related traits, except for winter survival, which had a relatively low heritability compared with the other traits. Interestingly, half of the field bud set QTLs did not map near photoperiodic QTLs. This is consistent with the moderate genetic correlation between these traits (0.53 to 0.60) and suggests that genetic differences in photoperiodic responses play only a modest role in explaining genetic differences in the timing of bud set under field conditions. Except for ABI1D, each of the candidate genes tested mapped near one or more of the dormancy-related QTLs. We conclude that molecular markers and QTL analyses can be used to study the genetics of dormancy-related traits, to design more effective breeding programs, and to provide new insights into tree physiology.

An experiment conducted at five locations in North Carolina during 1998 and 1999 evaluated weed management systems in cotton with CGA-362622 and pyrithiobac. Weed management systems evaluated different combinations with or without fluometuron preemergence (PRE) followed by (fb) CGA-362622 early postemergence (EPOST), postemergence (POST), or EPOST + POST; or pyrithiobac EPOST fb prometryn plus MSMA late postemergence directed (LAYBY) or no LAYBY treatment. The weed species evaluated include common ragweed, entireleaf morningglory, pitted morningglory, prickly sida, sicklepod, tall morningglory, and yellow nutsedge. Fluometuron PRE improved the control of all weed species by at least 17 percentage points and increased cotton lint yield compared with the systems that did not use fluometuron PRE. Prometryn plus MSMA LAYBY improved the control of all weed species and increased lint yield compared with the systems that did not use prometryn plus MSMA LAYBY when PRE or POST herbicides were used. Control with CGA-362622 at all application timings was greater than 70% for all weed species evaluated (common ragweed, entireleaf morningglory, pitted morningglory, sicklepod, tall morningglory, and yellow nutsedge), except prickly sida. Control of all three morningglory species and prickly sida was at least 70% with pyrithiobac, whereas control of common ragweed, sicklepod, and yellow nutsedge was lower. The only cotton that yielded over 800 kg ha−1 was treated with fluometuron PRE fb CGA-362622 EPOST, POST, or EPOST + POST fb prometryn plus MSMA LAYBY. Cotton treated with pyrithiobac EPOST gave yields that were similar to those given by cotton treated with CGA-362622 EPOST in systems with fluometuron PRE and less than those given by cotton treated with CGA-362622 EPOST in systems without fluometuron PRE. Early-season injury with CGA-362622 was greater than 60% at Clayton and Rocky Mount in 1998, whereas 12% or less injury was observed at the other locations. Pyrithiobac resulted in 25 to 45% injury at these two locations. No injury was observed 45 d after treatment.

Neither herbicides nor Aphthona spp. biological control agents provide satisfactory leafy spurge control in all environments where the weed is found. Open-release and caged field studies were conducted to determine the effect of herbicides used in combination with Aphthona spp. flea beetles on both the leafy spurge control and the biological control agent populations. Leafy spurge density was reduced more rapidly when A. nigriscutis or a mixture of A. czwalinae and A. lacertosa was combined with a fall-applied treatment of picloram plus 2,4-D than when either method was used alone. The combination treatment reduced leafy spurge density 3 to 5 yr earlier than did either the herbicides or the biocontrol agents alone and once reduced, the Aphthona flea beetles maintained acceptable control for at least 7 yr. Spring-applied herbicides eliminated the adult food source and negatively affected the flea beetle establishment, but fall-applied treatments did not reduce the flea beetle establishment or its reproduction. The combination treatment was most successful when used with established Aphthona spp. populations that were not increasing in size or reducing leafy spurge density. Land managers could realize a three- to fivefold cost savings with the combination treatment compared with herbicides alone.

Field experiments were conducted in 1996 and 1997 to evaluate the effect of EPTC (S-ethyl dipropyl carbamothioate) plus ethalfluralin at 2.4 plus 0.83 kg ai ha−1, rotary hoeing, in-row cultivation, rotary hoeing plus in-row cultivation, and dry bean canopy on weed seedling emergence. Cumulative weed emergence in 1996 and 1997 was similar in cropped and noncropped areas. Herbicides were more effective than mechanical cultivation in reducing weed emergence 91% in 1996 and 88% in 1997. Weed emergence was similar in both rotary hoed area and cultivated area in 1996 but weed emergence was 44% lower in rotary hoed plots than in cultivated plots in 1997. The Gompertz equation did not adequately predict weed seedling emergence in the untreated control and with in-row cultivation in 1996. Initial weed seedling emergence was observed at about 120 growing degree-days with 3 to 9% cumulative emergence among treatments. In 1997, the Gompertz equation adequately described weed seedling emergence in plots with or without disturbed soil. Weed emergence was first observed at 80 growing degree-days with 6 to 16% cumulative emergence among treatments. Predicted percent weed emergence closely approximated observed emergence in 1996 and 1997. Rotary hoeing plus in-row cultivation reduced maximum percent emergence rate 37% on an average. The greater maximum percent emergence rate obtained with in-row cultivation suggests that this treatment increased weed seedling emergence in 1997. On average, weed seedling emergence in the untreated check was lower in cropped areas than in noncropped areas, implying a competitive effect by the dry bean crop. Although weed seedling emergence occurred throughout the growing season, more weed seedlings emerged in June and early July than in late July and August.

Legume green manures have been used for millennia as sources of N for succeeding crops, but they are also sources of phytotoxic compounds that may selectively influence the performance of crop and weed species. To determine whether substitution of legume green manure for synthetic N fertilizer could enhance crop yield while suppressing weed growth, we conducted a field experiment in which common bean and wild mustard were sown in monocultures and mixtures. Three soil management treatments were employed: red clover residue, ammonium nitrate fertilizer (84 kg N ha−1), and a control that received neither red clover nor ammonium nitrate. In the absence of wild mustard, bean seed yield was equivalent in the red clover and ammonium nitrate treatments, where yields were 11 to 26% greater, respectively, than in the control. When bean grew in competition with wild mustard, its seed yield was as high (1995 and 1996) or higher (1994) with red clover residue than with ammonium nitrate. Averaged over bean competition treatments, wild mustard biomass production was 37% lower with red clover residue than with ammonium nitrate in 1994, 53% lower in 1995, and equivalent in these two treatments in 1996. Bean seed yield and wild mustard biomass production were strongly correlated with whole-plant N content and more weakly correlated with leaf area duration. Of the variation observed in wild mustard seed production, 89 to 93% was predicted by the variation in wild mustard biomass. The results of this study indicate that substitution of red clover green manure for ammonium nitrate fertilizer can be compatible with bean production goals and can contribute to the management of wild mustard.

Downy brome, an obligately selfing winter annual, has invaded a variety of habitats in western North America. Seeds are at least conditionally dormant at dispersal in early summer and lose dormancy through dry after-ripening. In the field, patterns of germination response at dispersal vary among populations and across years within populations. Degree of dormancy at summer temperatures in recently harvested seeds, as well as rate of dormancy loss during dry storage, can be related to the risk of premature summer germination in different habitats. Patterns of dormancy loss are predictable and can be modeled using hydrothermal time concepts. To assess the relative contribution of genotype and maturation environment, multiple parental lines from contrasting populations were grown for three generations under manipulated greenhouse conditions. Significant germination response differences among populations were observed, as well as major differences among full-sib families. Among-population variation accounted for over 90% of the variance in germination traits, whereas within-family variance accounted for 1% or less. Populations from predictable extreme environments (subalpine meadow and warm desert margin) showed significantly less variation among families than did populations from less predictable environments (cold desert, foothill, and plains). Environmental conditions that shortened the seed ripening period (water stress and high temperature) resulted in reduced seed dormancy level at maturation, but there were strong inbred line–environment interactions. For fully after-ripened seeds, inbred line and environmental effects were no longer evident, indicating that differences in genotype and maturation environment function mainly to regulate dormancy and dormancy loss rather than to mediate response patterns of nondormant seeds.

Productivity of dry bean is constrained by the competition with weeds for scarce nutrients and water in eastern Africa. Trials were conducted at Cornell University in 1996 and in central Uganda during the two seasons of 1997 to test the hypothesis that bean crop nutrition can be improved while increasing the relative competitiveness of bean with annual weed species. Soil levels of available N, P, and K were varied in the main plots. Subplots consisted of bean and two weed species in pure stands and bean mixed with each of the weed species. The weed species were black nightshade and smallflower galinsoga at Cornell and smallflower galinsoga and hairy beggarticks in Uganda. Bean yield was the most suppressed by hairy beggarticks with a mean reduction of 48%. Bean nutrient uptake and growth decreased relative to the weed nutrient uptake and growth when N and P were applied, but the relative competitiveness of bean increased with K application. The K effect on bean yield was greater than the P effect in two out of three trials. Alternative practices for the supply of N and P need to be evaluated for increasing bean yields while reducing the relative benefit to weeds.

Experiments were conducted in Lewiston, NC, in 1999 and 2000 and Rocky Mount, NC, in 1999 to evaluate weed management systems in strip- and conventional-tillage peanut. The peanut cultivars grown were ‘NC 10C’, ‘NC 12C’, and ‘NC 7’, respectively. Weed management systems consisted of different combinations of preemergence (PRE) herbicides including diclosulam and flumioxazin plus commercial postemergence (POST) herbicide systems. Dimethenamid plus diclosulam or flumioxazin PRE controlled common lambsquarters, eclipta, and prickly sida at least 91%. Diclosulam and flumioxazin provided variable control of three Ipomoea species (59 to 91%) and bentazon plus acifluorfen POST provided > 90% control. Only diclosulam systems controlled yellow nutsedge 90% late season. Annual grass control required clethodim late POST, regardless of tillage system. Dimethenamid plus diclosulam or flumioxazin PRE produced equivalent yields and net returns with no significant differences between the two PRE options. Both systems produced higher yields and net returns than dimethenamid regardless of the POST herbicide option. The tillage production system did not influence weed control of eight weeds, peanut yields, or net returns. The addition of diclosulam or flumioxazin to dimethenamid PRE improved weed control compared with dimethenamid PRE alone.

An open grassland at Sugarloaf Ridge State Park, Sonoma County, CA, was burned during three consecutive summers (1993–1995) to control yellow starthistle. By 1996, the yellow starthistle seedbank, seedling density, and mature vegetative cover were reduced by 99, 99, and 91%, respectively, and the plant community had greater diversity and species richness, particularly of native forbs. After the cessation of the prescribed burning after 1995, the community was monitored for 4 yr to determine if the reduced yellow starthistle population represented a stable state or if the population would quickly recover. The yellow starthistle seedbank rose dramatically over 4 yr. Seedling counts and summer vegetative cover also rose, though less rapidly. Total forb cover, particularly native species, total plant cover, and plant diversity decreased significantly after cessation of the burning. Grass cover did not show any strong trends, and year-to-year variation in the grass cover appeared to be more important than the treatment effects. In the absence of some overall changes in management, e.g., periodic prescribed burning, herbicide treatments, or revegetation, it may not be possible to establish and maintain a stable state with a low population of yellow starthistle in annual grasslands in California.

A laboratory experiment was conducted to compare the adsorption and desorption of fluometuron between two soils, one collected from an eastern gamagrass filter strip and the other from a cropped field. Fluometuron adsorption to soil collected from the filter strip was higher than to soil collected from the cropped field. Kd values for fluometuron ranged from 1.9 to 3.6 for soil from a cropped area, compared with 2.9 to 5.3 for soil from the filter strip, indicating a weak to moderate binding affinity for fluometuron. The total fluometuron desorbed ranged from 48 to 79% of that adsorbed, most of which (50 to 59% of the total amount desorbed) occurred during the first desorption cycle with both soils. Approximately 11% less fluometuron desorbed when a filter strip was present. Results indicate that eastern gamagrass filter strips can influence adsorption–desorption processes between fluometuron and a Brooksville silty clay soil by altering the soil properties, specifically soil organic matter.

Sprayable and granular formulations of isoxaben and trifluralin were applied to container plant nursery beds to determine formulation effects on herbicide runoff and weed control. In 1998 herbicide application was followed by 0.8 cm of irrigation delivered in 60 min, and runoff water samples were collected on the day of application (DOA). The highest concentrations of isoxaben and trifluralin detected in runoff water were 0.50 and 0.15 μg ml−1, respectively. Total isoxaben in runoff water was greater from the granular than from the sprayable formulation, but no differences attributable to formulation were detected for trifluralin. In 1999 herbicide application was followed by irrigation applied in three pulse cycles of 30 min each (0.8 cm per cycle). Runoff water samples were collected daily through 2 d after herbicide application (DAA). The highest herbicide concentrations were found in the first runoff sample on the DOA. Isoxaben was detected at 1.6 μg ml−1 from both formulations, and trifluralin concentrations were 0.8 and 0.2 μg ml−1 from the sprayable and granular formulations, respectively. Isoxaben concentrations were similar between formulations on the DOA, higher from the granular formulation 1 DAA, and higher in five of the nine runoff samples 2 DAA. Trifluralin concentrations were higher from the sprayable formulation for the first pulse cycle on the DOA and higher from the granular formulation in three of the nine samples 1 DAA and in four samples 2 DAA. The total amount of isoxaben recovered was 9.1% of applied from the granular formulation and 7.3% of applied from the spray formulation. The total amounts of trifluralin found were similar among formulations (0.5% of applied). Weed control was effective for both formulations in both years.

Knowledge and prediction of seasonal weed seedling emergence patterns is useful in weed management programs. Seed dormancy is a major factor influencing the timing of seedling emergence, and once dormancy is broken, environmental conditions determine the rate of germination and seedling emergence. Seed dormancy is a population-based phenomenon, because individual seeds are independently sensing their environment and responding physiologically to the signals they perceive. Mathematical models based on characterizing the variation that occurs in germination times among individual seeds in a population can describe and quantify environmental and after-ripening effects on seed dormancy. In particular, the hydrothermal time model can describe and quantify the effects of temperature and water potential on seed germination. This model states that the time to germination of a given seed fraction is inversely proportional to the amount by which a given germination factor (e.g., temperature or water potential) exceeds a threshold level for that factor. The hydrothermal time model provides a robust method for understanding how environmental factors interact to result in the germination phenotype (i.e., germination pattern over time) of a seed population. In addition, other factors that influence seed dormancy and germination act by causing the water potential thresholds of the seed population to shift to higher or lower values. This relatively simple model can describe and quantify the germination behavior of seeds across a wide array of environmental conditions and dormancy states, and can be used as an input to more general models of seed germination and seedling emergence in the field.

Metabolites of the fungus Fusarium solani (Sud 96) inhibited Striga hermonthica germination induced by the germination stimulant GR24. The active principles were identified as trichothecenes acuminatin, neosolaniol, 8-acetylneosolaniol, and tetraacetoxy T-2 tetraol (neosolaniol diacetate) on the basis of their chromatographic behavior and nuclear magnetic resonance and mass spectra. Inhibitory activity of the four trichothecenes against Striga germination increased with acetylation of the hydroxyl moieties. The most abundant inhibitor produced by the fungus, 8-acetylneosolaniol, completely inhibited Striga germination at 24 μM. The fungal toxin did not affect the germination of sorghum, a host crop, but retarded root and shoot elongation of the seedlings by 60 and 30%, respectively, at the same concentration.

The effects of MT-101 and its herbicidally active form, NOP, on the germination and seedling growth of hemp sesbania and rice were investigated. MT-101 decreased the germination of hemp sesbania by 57 and 90% at 0.05 and 0.5 mM, respectively, 1 d after treatment (DAT) in petri dishes. The germination, however, recovered such that there was no significant difference between treatments 4 to 6 DAT. NOP completely inhibited the germination of hemp sesbania at both 0.05 and 0.5 mM 1 DAT. However, germination also similarly recovered, and there was no difference between treatments 4 to 6 DAT. Neither MT-101 nor NOP decreased the germination of rice 3 to 6 DAT. In greenhouse trials preemergence (PRE) application of MT-101 at 2.25 kg ai ha−1 decreased the density (number of plants pot−1), plant height, and dry weight of hemp sesbania by 85, 67, and 91%, respectively. When applied postemergence (POST), MT-101 at 2.25 kg ha−1 decreased the density, plant height, and dry weight by a maximum of 58, 61, and 82%, respectively, indicating that MT-101 may have greater activity when applied PRE. NOP had greater activity than MT-101 on hemp sesbania. NOP at 2.25 kg ai ha−1 decreased the density, plant height, and dry weight of hemp sesbania 99, 78, and 97%, respectively, with PRE application. A POST application of NOP at 2.25 kg ha−1 decreased the dry weight of hemp sesbania 91 to 94%. A PRE application of NOP at 2.25 kg ha−1 decreased the dry weight of rice by 58%. Rice was not affected by POST applications of MT-101 but was affected slightly by NOP. These results suggest that MT-101 is a possible weed control agent in rice.

In many cases, the logical presentation of scientific results masks the actual unfolding of experimental progress. Our work to identify the mechanisms by which chemicals break red rice seed dormancy is presented as an example. Doing the simplest experiments first, allowing the data to drive subsequent work, frequent and aggressive study of published literature, and examination of potentially key findings using multiple, interlocking strategies have been the keys to defining the relationship between the structure and action of dormancy-breaking chemicals. Our experiments show that chemical activity is dependent upon the lipophilicity as well as the nature and position of functional groups. However, the chemical that is applied to a dormant seed is not necessarily the chemical that actually breaks dormancy. A combination of approaches using structure–activity studies, metabolic inhibitors, and physiological measurements indicate that alcohols must first be metabolized to carboxylic acids to elicit germination of red rice. This report summarizes these findings as a personal account, illustrating the importance of logic, luck, and imagination in scientific discovery, as well as to highlight some important lessons about conducting research.