Printer Friendly

Challenges in testing transgenic and nontransgenic cotton cultivars.

COTTON CULTIVARS transgenically enhanced to impart weed and/or insect pest management features have been widely adopted since commercial introduction in the mid-1990s. About 77% of the 2002 U.S. cotton hectarage was planted to transgenic cultivars containing genes conferring glyphosate herbicide resistance, Bt-mediated lepidopteran insect resistance, or both (glyphosate + Bt), plus smaller hectarage planted to bromoxynil-resistant cultivars (USDA-AMS, 2002). Thus, transgenic cultivars can now be considered to be conventional cultivars because they dominate plantings. Despite the dominance of transgenic cultivars in U.S. production, OCTs remain the primary vehicle for testing cultivars. Transgenic technology developers, growers, and some scientists have advocated that OCT test protocols evolve in tandem with transgenic technology, while also questioning the validity of OCT performance data.

A key role of cultivar testing is to provide unbiased agronomic performance in a defined environment while employing an accepted, common management system. The primary goals of cotton OCTs are to rank cultivars within and across locations for lint yield, lint fraction (ratio of lint to seed-cotton), and fiber properties. Before the advent of transgenic, pest-managing cultivars, imposition of common insect and weed management protocols on all cultivars in OCTs was an accepted practice. Differential effects of the common pest management system on certain cultivars in OCTs may have existed, but such bias was not considered of sufficient magnitude to invalidate OCT results. Cotton germplasm can possess host resistance traits such as gossypol glands, providing some protection from certain lepidopteran insects (Parrott et al., 1989; Calhoun et al., 1997; Jenkins et al., 1997) compared with non- or less-glanded cultivars. The elevated lepidopteran insect control afforded by gossypol can be further enhanced under more frequent insecticide application regimes appropriate for non-host-resistant germplasm; hence, densely glanded cultivars can have a yield advantage in OCTs (Lukefahr et al., 1966). Cotton cultivars can also vary in leaf trichome density (Lee, 1985) and leaf shape (Thomson et al., 1987), traits associated with pest insect oviposition nonpreference (Lukefahr et al., 1971; 1975), and reduced larval populations (Robinson et al., 1980). Thus, bias from common pest management protocols imposed by OCTs existed before inclusion of transgenic cultivars (Bourland et al., 2000). The high level of lepidopterous insect control conferred by the Bt technology (Jenkins et al., 1997), plus the capability of herbicide-resistant cultivars to be produced with fewer soil-applied herbicides and potentially less herbicide-induced yield loss (S.C. Culpepper, 2002, personal communication; May and Murdock, 2002), has led cultivar development firms, transgenic technology providers, and some producers to question the validity of testing transgenic cultivars in OCTs and fairly comparing them with nontransgenic cultivars. Our objectives were to assess the merits of OCTs to test transgenic and nontransgenic cultivars and to recommend alternatives.

The First Contingent of Transgenic Cotton Cultivars

A brief synopsis of the development of transgenic cultivars affords a broader understanding of their appeal to growers. Plus, it is instructive to describe the pest management features of transgenic cultivars to appreciate the complexities of conveying performance of the transgenic trait and the genetic background into which the gene(s) have been introduced.

Transgenic cotton cultivars became commercially available with the introduction of bromoxynil herbicide-resistant cultivars BXN 57 and BXN 58 in 1995 (Collins, 1996). These bromoxynil herbicide-resistant cultivars were the first transgenic cultivars grown on extensive hectarage in the USA. Both were derived from direct selections within 'Coker 312' germplasm transformed via an Agrobacterium vector. The BXN technology allows over-the-top application of bromoxynil to cotton up to 60 d before harvest combined with excellent crop tolerance (Culpepper and York, 1999). Bromoxynil controls several troublesome weed species encountered in cotton production, including morning-glory (Ipomoea spp.) and cocklebur (Xanthium strumarium L.), but is mostly ineffective at controlling grasses, pigweed (Amaranthus spp.), and sicklepod {Cassia obtusifolia L. [= Senna obtusifolia (L.) H.S. Irwin & Barneby]}.

Being a selective, contact herbicide, production of BXN cultivars typically includes use of the soil-applied herbicides (pendimethalin [N-(1-ethylpropyl)-3,4-dimethyl-2,6-dinitrobenzenamine] or trifluralin [2,6-dinitro-N,N-dipropyl-4-(trifluoromethyl) benzenamine], applied at or before planting, followed by over-the-top bromoxynil applications, then additional postdirected herbicides for weed species not controlled by bromoxynil (Culpepper and York, 1997, 1999). The cultivar BXN 47, derived from the popular cultivar Stoneville 474, has gained market share and has supplanted most of the hectarage planted to BXN 57 and BXN 58 (USDA-AMS, 2000, 2001).

The next class of transgenic cultivars commercialized in the USA was Bt lepidopterous larvae resistant (Bollgard) cultivars in 1996 (Table 1). Although several companies developed Bt genes conferring lepidopterous insect resistance, the Monsanto Bt gene [cryIA(c)] dominates commercial production (Benedict et al., 1996; Jenkins et al., 1997; Benedict and Altman, 2001, p. 169). Development of Bt transgenic cultivars was spurred through successful efforts by Monsanto to hyperexpress the CryIA(c) insecticidal protein from Bacillus thuringiensis spp. kurstaki (Perlak et al., 1991). Cultivars containing the Monsanto Bt gene are highly resistant to tobacco budworm [Heliothis virescens (Fabricius)] and pink bollworm [Pectinophora gossypiella (Saunders)] larvae, but are less resistant to the cotton bollworm [Helicoverpa zea (Boddie)] and the American bollworm [H. armigera (Hubner)] larvae, some of the most damaging insect species to cotton (Benedict et al., 1993; Jenkins et al., 1997; Fitt et al., 1998; Benedict and Altman, 2001). The Bt cultivars NUCOTN33B and NUCOTN 35B, backcross derived from recurrent parents 'Deltapine 5415' and 'Deltapine 5690', respectively, garnered immediate grower acceptance in the first year of commercial availability in 1996, as they were planted to 12% of the U.S. cotton hectarage (Table 1) despite lack of OCT testing before commercialization. All current U.S. Bt cultivars were derived through introgression of the Bt gene from one donor parent cultivar Coker 312 transformation event 531, chosen for its ability to protect against budworm and bollworms and lack of nontarget effects (Jenkins et al., 1997). Production of Bt cultivars still requires pest insect enumeration during squaring and fruiting, especially in regions infested with high populations of cotton bollworm (Mahaffey et al., 1995). As such, several states have established bollworm egg and larvae density thresholds to trigger insecticide application to Bt cultivars to prevent larvae from reaching growth stages beyond which the Bt toxin is ineffective (Roof, 2000). Depending on insect pest densities, the Bt technology can reduce cotton production costs (Carpenter and Gianessi, 2000; Gerloff, 2001), but the immediate gain with Bt cotton is the convenience of fewer insecticide sprays and the opportunity to exploit integrated pest management practices.

Glyphosate-resistant cultivars alone or stacked with Bt became commercially available in 1997 in the USA, and immediately gained grower acceptance (Table 1). Glyphosate resistant cotton was developed by Monsanto through transformation of Coker 312 with a glyphosate-tolerant version of the enzyme 5-enolpyruvylshikimate-3-phosphate synthase (Nida et al., 1996). Glyphosate-resistant cultivars can be sprayed twice over-the-top with up to 0.56 kg a.i. [ha.sup.-1] glyphosate through the fourth true leaf stage, allowing 10 d and two nodes of growth between applications (Kerby and Voth, 1998). Thereafter, postdirected applications minimizing leaf contact are necessary to prevent boll abortion. Glyphosate-resistant cultivars were derived through backcrossing, initially with one donor parent, a Coker 312 line containing transformation event no. 1445 (Nida et al., 1996; Sheets and Speed, 1997). The popularity of glyphosate-resistant cotton lies in the ease of weed management afforded by over-the-top applications and the broad-spectrum control of many annual and perennial grass and broadleaf weed species (Wilcut, 1996; Wilcut et al., 1996; Culpepper and York, 1999).

Before the availability of bromoxynil--and glyphosate-resistant cultivars, postemergence weed control in cotton relied on sometimes-unsuccessful establishment of height differences between weeds and cotton such that herbicides could be directed to contact weeds but minimize contact with cotton (Buchanan, 1992). The broad spectrum of weeds encountered in cotton production that are controlled by glyphosate makes total postemergence weed management possible (Culpepper and York, 1998), plus an added benefit can be to lessen chances of yield drag from soil-applied herbicides for certain soil types (Welch et al., 1997; S.C. Culpepper, 2002, personal communication; May and Murdock, 2002). Production of bromoxynil--and glyphosate-resistant cotton often still requires use of soil-applied herbicides, typically pendimethalin or trifluralin applied preplant incorporated, followed by postemergence-applied herbicides (Culpepper and York, 1999). Despite the capability to apply bromoxynil over-the-top of BXN cultivars later in crop development compared with glyphosate-resistant cultivars, glyphosate-resistant cultivars are more popular because glyphosate controls more weeds encountered in cotton production (Culpepper and York, 1999).

Except for BXN cultivars, the first wave of transgenic cultivars received little or no public testing before wide-scale grower adoption (May et al., 2000), while the particular transgenic technology itself was widely tested by entomologists and weed scientists (Benedict et al., 1996). Most of these technology trials employed transformed Coker germplasm, and not candidate cultivars, plus a relatively narrow range of weed and insect management systems (Dotray and Keeling, 1997; Jenkins et al., 1997; Benedict and Altman, 2001). Consequently, transgene x background interactions that might alter performance of the pest management trait, or position effects of transgene insertion on the popular genetic backgrounds were not extensively investigated before commercialization, nor were effects of pest management programs on performance of the transgenic technology and background genotype.

Overall, the dominance of transgenic cultivars in commercial production testifies to the efficacy and convenience of insect and weed management possible with transgenic cultivars. A minority of growers report fruit abortion and yield loss with glyphosate-resistant cultivars alone or stacked with Bt (Kerby and Voth, 1998), despite claims of adherence to Monsanto's glyphosate application protocol, while others have experienced less-than-desired insect control with the Bt technology (Benedict and Altman, 2001, p. 176). These issues have focused attention on the validity of OCT test data for transgenic cultivars and whether such trials fairly compare transgenic and nontransgenic cultivars.

Status of Cotton Cultivar Testing

Nearly all states in the U.S. cotton belt conduct OCTs, typically divided into preliminary strain trials to test candidate cultivars and advanced cultivar trials, where established cultivars are compared with newly commercialized cultivars. Several states separate strains or cultivars by relative maturity compared with a cultivar of known maturity into early and later maturity trials, with the strain or cultivar owner deciding appropriate maturity (Bowman, 1997). Conduct of separate strain and cultivar trials delineated by relative maturity has in part arisen to reduce the size of trials in the face of large increases in cultivars submitted for testing since development of transgenic cultivars (Bourland et al., 2000). Variation in production practices in OCTs across the Cotton Belt reflects differences in heat unit accumulation, prevalence of irrigated or rain-fed culture, insect pests, weed species and densities, and harvest practices such as stripping or number of spindle pickings (Bowman, 1997). Most cotton OCTs employ randomized complete block designs, relatively small plots to accommodate oftentimes large numbers of entries to keep the size of field blocks manageable, and weed and insect pest management appropriate for nontransgenic cultivars (Bowman, 1998; Creech et al., 1998; Caldwell et al., 1999; May et al., 1999; Benson et al., 2000; Glass et al., 2000; Day et al., 2001). The small plots in OCTs, the typically large number of trial entries, and the possibility of accidental insecticide and/or herbicide spray drift to adjacent plots that were not intended to receive such treatments practically mandates that all cultivars in the trial receive the same herbicides and insecticides. Consequently, insect control measures are imposed on all cultivars in OCTs on the basis of pest insect densities established for non-Bt cultivars, typically consisting of several insecticide applications during the squaring and fruiting cycles. And all cultivars in OCTs receive the same herbicide regime, but glyphosate--and bromoxynil-resistant cultivars do not receive applications of herbicides they respectively tolerate and may receive applications of herbicides that they were intended to avoid (May et al., 1999; Benson et al., 2000).

Transgenic, pest-management-enhanced cultivars can be viewed as dual-purpose products, intended to control certain insects, facilitate weed management through herbicide tolerance, or combine the two, and to also produce yield. Hence, performance data should verify if the transgene(s) impart the pest management capability as intended in the genetic background wherein the genes are inserted, and under the cultural, environmental, and pest management programs and challenges encountered in commercial cotton production. We must also test the yield and fiber quality to ensure that expression of native genes has not been adversely influenced by insertion of foreign gene(s). The ability of cotton to compensate throughout the growing season for fruit loss due to abiotic and biotic factors complicates interpretation of OCT data because failure of the transgenic technology to perform in the particular genetic background (e.g., protect against insect-induced fruit loss or fruit loss from intolerance to the applied herbicide purported to be possible with the transgenic technologies) can be masked by nontransgenic pest management regimes such as yields of Bt cultivars under non-Bt insecticide regimes employed in OCTs. Thus, efficacy of the transgenic technologies and cotton's innate ability to compensate for fruit loss can be confounded. These considerations lead to discussion of the merits of OCTs to provide performance data for transgenic, pest managing cultivars.

Can OCTs Adequately Test Herbicide-Resistant Cultivars?

This question was first raised after grower complaints of fruit shed from glyphosate-resistant cultivars in portions of the Cotton Belt in 1997 and 1998 (Kerby and Voth, 1998). Growers subsequently voiced concerns that performance of herbicide-resistant cultivars in OCTs may not be valid since the respective herbicide is not applied (Hargett, 2000), while also expressing concern about herbicide-resistant cultivar performance when treated with traditional herbicides (Coley, 2000).

The excellent crop tolerance to bromoxynil in BXN cultivars and its availability in only a few cultivars has lessened concerns that OCTs do not evaluate cultivar variation in bromoxynil resistance. Growers adopting BXN cultivars, particularly BXN 47, report satisfaction with both the weed management technology, except where high populations of sicklepod, pigweed, and certain grass species not controlled by bromoxynil are encountered in cotton production (Culpepper and York, 1999), and the performance of the BXN 47 cultivar. Initial complaints about the BXN technology were related to poor agronomic performance of BXN 57 and BXN 58, which were derived from selections within transformed Coker 312, itself agronomically obsolete compared with current cultivars (Isgett et al., 1996; Wilcut, 1996).

Only a few studies have addressed cultivar variation in resistance to glyphosate when applied according to Monsanto's directions. Most research with glyphosate resistant cultivars has focused on effects of applying glyphosate in ways Monsanto does not condone, such as topical applications past the four-leaf crop stage (Vargas et al., 1998; Jones and Snipes, 1999). Additionally, the glyphosate resistance technology was tested mostly with noncandidate cultivar germplasm lines (Nida et al., 1996), while glyphosate resistant cotton cultivars were first offered to growers in 1997 without public testing in intended production systems or OCTs (May et al., 2000).

May and Murdock (2002) conducted a 2-yr field study to evaluate glyphosate-resistant cultivar x herbicide system interactions for yield. They tested 14 glyphosate-resistant cultivars in 1998 and 10 in 1999, including many cultivars still popular in 2001. Herbicide treatments in their study included a regime of typical soil-applied herbicides as might be used in OCTs, but no glyphosate, and two regimes with glyphosate applied according to Monsanto's directions. One glyphosate regime included a preplant, soil-incorporated herbicide, plus a single over-the-top glyphosate application at the four-leaf stage of cotton followed by postemergence, precision-directed herbicides. The second regime consisted of the same four-leaf glyphosate application followed by a precision-directed glyphosate spray, but without pre- or postemergence soil-applied herbicides. On the basis of no cultivar X herbicide system interaction, the authors concluded that cultivar variation in crop tolerance to glyphosate was not evident as measured by yield. Thus, transgene X background effects, if present, were not manifested in yield, suggesting OCTs can discriminate among glyphosate-resistant cultivars for relative yields.

Research by S.C. Culpepper (2002, personal communication) and Phipps et al. (2002) affirm this conclusion. Both groups assessed glyphosate-resistant cultivar X herbicide system interactions in North Carolina and Arkansas, respectively, where herbicide systems consisted of either soil-applied and postemergence-applied herbicides but no glyphosate or a glyphosate-only regime. The Arkansas and North Carolina studies also found no cultivar X herbicide interactions for yield. In summary, these findings do not exclude the expression of glyphosate resistance gene X background interactions under specific environmental conditions, such as the anomalous performance of glyphosate-resistant cultivars during the historically cold weather in the spring and early summer in 1997 across much of the cotton belt (Kerby and Voth, 1998). However, the finding of no cultivar X herbicide interactions supports the general capability of OCTs to rank glyphosate-resistant cultivars for yield. Another issue in testing bromoxynil and glyphosate-resistant cultivars is the potential for yield reduction from effects of soil-applied herbicides on crop growth (Keeling and Abernathy, 1989). Welch et al. (1997) and Sanders et al. (2000) found under weed-free conditions that glyphosate-resistant cotton cultivars produced with soil-applied herbicides yielded less compared with glyphosate-only herbicide systems. Greenhouse and growth chamber studies report reduced plant or root growth when cotton was grown in the presence of dinitroaniline herbicides or fluometuron (Kappelman and Buchanan, 1968; Kappelman et al., 1971; Murray et al., 1979; Bailey and Bourland, 1986), possibly explaining yield reduction in field trials employing such herbicides. May and Murdock (2002) reported that glyphosate-resistant cultivars may suffer yield loss when high rates of soil-applied herbicides are applied to minimize weed competition effects, compared with herbicide regimes containing fewer or no soil-applied herbicides. The main effect of herbicide system in their studies revealed that glyphosate-resistant cultivars yielded 20 to 30% more under glyphosate-only herbicide regimes compared with regimes containing only soil-applied herbicides. S.C. Culpepper (2002, personal communication) also reported a 7% yield increase at one of six locations when glyphosate-resistant cultivars were produced in a glyphosate-only herbicide system compared with a herbicide regime containing soil-applied herbicides.

These studies suggest that the genetic yield potential of glyphosate-resistant cultivars can be affected by the herbicide system. Thus, OCTs applying high rates of soil-applied herbicides can impose a yield penalty on glyphosate and perhaps bromoxynil-resistant cultivars, potentially questioning whether yields of herbicide-resistant and nonresistant cultivars can be compared.

The Merits of OCTs For Evaluation of Bt Cultivars

Concerns about OCT data for evaluating Bt cultivars center around reporting their yields when produced with insecticide regimes appropriate for non-Bt cultivars and possible yield advantages Bt cultivars have compared with non-Bt cultivars. Additional considerations include possible cultivar variation in Bt efficacy; efficacy of the Bt insecticidal protein differs with target pest; effects of the background genotype can attenuate toxicity or availability of the Bt endotoxin to pest insects that would not be apparent under OCT insecticide regimes; and OCT cultural practices which affect cultivar vegetative and fruiting status.

Monsanto requires establishment of Bt functional gene equivalency between candidate Bt cultivars and the original Bt donor parent, Coker 312 transformation event 531, as a prerequisite to commercialization that should render cultivar variation in Bt efficacy unlikely (Greenplate et al., 2001). But efficacy of Bt against target pests could differ among cultivars if CryIA [delta]-endotoxin levels vary among plant parts where target pests oviposit and feed, followed by destruction of fruit considered essential for high yields (Jenkins et al., 1990). Such effects could be masked in OCTs under typically more intense insecticide regimes appropriate for non-Bt cultivars. Adamczyk et al. (2000) reported that among 17 Bt cultivars, NUCOTN33B and 451BR differed in larval survival in no-choice laboratory leaf feeding assays, with 451BR supporting significantly higher cotton bollworm larval growth. Furthermore, they found that seasonal mean CryIA [delta]-endotoxin levels were lower in bracts and bolls of 451BR compared with NUCOTN33B, but CryIA [delta]-endotoxin in squares and flowers did not differ. Greenplate et al. (2001) also reported that seasonal mean CryIA [delta]-endotoxin in squares did not differ among Bt cultivars, but that it could vary among plant terminals, prime oviposition, and feeding sites for initial instar lepidopterous insects. Also of significance was their finding of no genotype X location interactions for CryIA [delta]-endotoxin, but main effects of environment could influence CryIA [delta]-endotoxin levels (Greenplate, 1999; Greenplate et al., 2001). Despite the finding that CryIA [delta]-endotoxin can vary among plant parts and cultivars, Greenplate et al. (2001) point out that such variation has not been associated with cultivar variation in control of the most susceptible target pests (tobacco budworm, pink bollworm), but this finding may have relevance for control of less susceptible American bollworm or cotton bollworm where this pest is the predominant species attacking cotton.

Current Bt cotton cultivars essentially control attack from tobacco budworm, but are less resistant to cotton bollworm larvae (Benedict et al., 1993; Jenkins et al., 1997; Benedict and Altman, 2001), among other pests. Thus, concern exists that OCTs might boost Bt cultivar yields due to subthreshold pest insect control resulting from application of insecticides under a regime intended for non-Bt cultivars. Yield enhancements of 10% or more for Bt cultivars can occur when insecticides are applied to control high levels of bollworm, beet armyworm (Spodoptera exigua Hubner), or fall armyworm [Spodoptera frugiperda J.E. Smith] (Benedict et al., 1996; Jenkins et al., 1997; Bell et al., 1999). Additionally, the high level of tobacco budworm resistance afforded by the Bt technology can provide a yield advantage to Bt cultivars vs. non-Bt cultivars in OCTs conducted where high populations of tobacco budworm or insecticide-resistant tobacco budworm are present (Jenkins et al., 1997; Smith, 1997). In summary, yields of Bt cultivars in OCTs can be biased upward when certain lepidopterous insects are not completely controlled by the Bt technology.

Another consideration is that effects of the background genotype may alter efficacy of target pest control by the Bt technology. Cotton contains condensed tannin compounds that increase in concentration with plant age from emergence to anthesis (Zummo et al., 1984; Lege et al., 1992), but tannins may reduce efficacy of the CryIAc insecticidal protein against target pests. Olsen and Daly (2000) reported variation in quantity and toxicity or availability of the Bt toxin as cotton plants aged in laboratory bioassays challenging H. armigera in Australia. The CryIAc protein isolated from fruiting plants was less toxic to H. armigera compared with that from pre-square-stage plants. They attributed the lesser toxicity of the CryIAc protein from fruiting plants to not only lower quantity of toxin, but to also plant-toxin interactions. Their research implicated tannins or other plant compounds in reducing the toxicity and/or availability of CryIAc protein to H. armigera. In contrast, plant terpenoid compounds (Lukefahr et al., 1966) can increase Bt control of tobacco budworm and cotton bollworm. Thus, densely glanded Bt cultivars could have a yield advantage that is further enhanced under insecticide-sprayed conditions.

In summary, OCTs face the greatest difficulty to impart the performance of Bt cultivars and to fairly compare them with non-Bt cultivars. The genetic yield potential of such cultivars may be realized only when produced under a Bt insect-management protocol. Interaction of the Bt gene with the background genotype, insecticide regimes that can artificially enhance their yields, and lack of target pest insect pressure under non-Bt insecticide regimes largely obviate OCTs from determining the dual purpose performance of Bt cultivars.

Systems Trials Approach for Testing All Cultivar Types

The limitations outlined above for OCTs to convey Bt cultivar performance plus potential negative effects of OCT herbicide regimes on glyphosate--or bromoxynil-resistant cultivar yields are two reasons that OCT data alone may be inadequate to choose among cultivars and pest management protocols. An alternative is to abandon the traditional approach of separate agronomic and pest management efficacy testing by unifying efforts across disciplines. Rather than conducting separate trials, agronomists, entomologists, and weed scientists, as appropriate, should cooperate in the comprehensive evaluation of cultivar agronomic and pest management capabilities through systems trials imposing the respective pest management programs. Cultivar-specific pest management for the current generation of transgenic and nontransgenic cultivars would entail imposing the following treatments: non-Bt insect management and non-herbicide-resistant weed management, non-Bt insect management and BXN weed management, non-Bt insect management and glyphosate weed management, Bt insect management and non-herbicide-resistant weed management, Bt insect management and BXN weed management, and finally, Bt insect management and glyphosate weed management.

The logistics of imposing cultivar-specific pest management influences choice of experimental and treatment designs. Treatment designs restricting the allocation of cultivars to experimental units such as split plot, strip plot, or latin square (Gomez and Gomez, 1984, p. 108-115) to result in land areas of sufficient size or shape to apply insect and weed management treatments (considering practical limitations of pesticide application equipment) result in the nesting of cultivars within pest management systems. For example, if insect management systems Bt and non-Bt are imposed through blocking to achieve land areas large enough to accommodate both pesticide application equipment and to prevent spray drift from treated to untreated plots, Bt and non-Bt cultivars cannot be cross classified with the insect management regimes. The result is that the Bt and non-Bt cultivars are nested within the respective insect management regime, and thus yields of the cultivar groups cannot be compared by ranking all cultivars for yield. The consequence of nesting cultivars within a management system is that unless the main effect of the management system and the whole plot error are negligible (in which case they can be ignored and the experiment analyzed as a randomized, complete block design), cultivars cannot be ranked for yield across management systems.

An alternative to blocking by pest management system is to increase plot size to the minimum necessary to accommodate pesticide application equipment and then assign cultivars to experimental units as in randomized, complete block designs. In the simplest case, all cultivars would be produced with a version of their respective pest management programs. For example, all cultivars might receive a preplant incorporated application of a soil-applied herbicide, but preemergence and postemergence herbicide programs would be tailored to the particular cultivar type (Culpepper and York, 1999). Insect management by plot could be imposed based on locally established thresholds for pest densities and cotton growth stages to trigger insecticide applications. With this protocol, cultivars can be ranked for yield, but unlike OCTs, systems trial yields should reflect the background genotype, pest management system, and performance of any pest management trait(s) (Bryant et al., 1999, 2000; May et al., 2002). Still, limitations of this approach include only one of several possible pest management systems can be imposed on a cultivar type, thus the necessity of defining best management practices in advance. And unpredictable densities of pests, especially in small plots, can render the cultivar response to pest management system and pest pressure an unreliable predictor of response under the more intense challenge encountered in commercial cotton production.

Beyond these considerations, as new transgenic pest management traits are developed, it remains necessary to establish pest management options provided by the new technology. Entomologists would compare yields of new insect-protected cultivars under positive and negative insecticide regimes and insect-challenged conditions, while weed scientists would investigate yields of herbicide-resistant and nonresistant cultivars under weed-infested conditions and the requisite herbicide regimes to observe performance of the new technology. Lastly, bias imposed by nontransgenic pest management regimes in OCTs can be quantified through trials levying transgenic and OCT-type pest management treatments on transgenic cultivars (S.C. Culpepper, 2002, personal communication; May and Murdock, 2002).

SUMMARY

Cultivar yield ranks in OCTs can be influenced by nontransgenic, common pest management regimes imposed without regard to cultivar type. Yields of Bt cultivars in OCTs can be biased upward relative to non-Bt cultivars, while herbicide-resistant cultivars may be penalized. As such, yield ranks among cultivars in OCTs can be affected, questioning the capability of OCTs to discern the highest-yielding cultivars.

We do not, however, recommend that OCTs evolve to impose cultivar-specific pest-management regimes, despite the ever-expanding hectarage planted to transgenic cotton cultivars, and an impending new generation of transgenically enhanced insect- and herbicide-resistant cultivars. The practical limitations of plot size, number of trial entries, and consequent size of field blocks to conduct trials are reasons to leave OCTs as they are. Standard OCTs still have value for determining relative adaptation of nontransgenic and transgenic cultivars across locations. Instead, we advocate that OCTs be augmented with systems trials that evaluate transgenic cultivars with the local insect and weed management systems established for the particular transgenic technology, while nontransgenic cultivars are managed according to accepted protocols for their production. The size of the systems trial can be controlled by limiting entry to those cultivars that perform best in the OCT. The resulting yields from the systems trial should reflect the genetic potential of the cultivar to produce yield and any pest management capabilities whether transgenically imparted or endogenous. New transgenic traits such as glyphosate resistance past the current four-leaf growth stage and dual Bt genes will be available to growers in the next several years, affirming the need for agricultural economists, agronomists, entomologists, and weed scientists to cooperate in defining best management practices for the new technologies, highlighting any bias imposed on the new transgenic cultivars by OCTs, and determining the agronomic and pest-management virtues of the new cultivars.

Abbreviations: Bt, Bacillus thuringiensis var. kurstaki; OCTs, official cultivar trials.
Table 1. Proportion of U.S. cotton hectarage planted to nontransgenic
and transgenic cultivars since introduction of transgenic cultivars
in 1995. ([dagger])

                                                 Year

Cultivar Type ([double dagger])     1995     1996     1997    1998

                                             % Hectarage

Nontransgenic                      99.8     87.8     77.2     54.9
Bt                                  0.08    12.0     18.0     18.7
R                                   0        0        3.2     17.0
Bt + R                              0        0        0.45     3.6
BXN                                 0.07     0.16     1.2      5.8
BXN + Bt                            0        0        0        0

                                               Year

Cultivar Type ([double dagger])    1999    2000     2001    2002

                                            % Hectarage

Nontransgenic                      40.7    28      22.3     23.0
Bt                                 15.6    10.8     3.8      2.3
R                                  20.5    25.9    33.3     36.2
Bt + R                             15.4    28.1    37.1     35.7
BXN                                 7.8     7.2     3.6      1.6
BXN + Bt                            0       0       0.05     0.6

([dagger]) Data from USDA-AMS (1995-2002).

([double dagger]) Bt = cultivar contains gene from Bacillus
thuringiensis imparting resistance to certain lepidopteran insects; R,
cultivar contains gene for glyphosate resistance; BXN, cultivar
contains gene for bromoxynil resistance.


REFERENCES

Adamczyk, J.J., Jr., L.C. Adams, and D.D. Hardee. 2000. Quantification of CryIA(c) [delta] [delta]-endotoxin in transgenic Bt cotton: Correlating insect survival to different protein levels among plant parts and varieties, p. 929-932. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Counc. Am., Memphis, TN.

Bailey, B.A., and F.M. Bourland. 1986. The influence of seed quality on response of cotton seedlings to the pre-plant herbicide trifluralin. Field Crops Res. 13:375-382.

Bell, D.E., G. Slaughter, P.M. Roberts, C.E. Ellis, L. Willingham, and T. Cary. 1999. The effects of oversprays on Bt cotton, p. 956-958. In Proc. Beltwide Cotton Conf., Orlando, FL. 3-7 Jan. 1999. Natl. Cotton Counc. Am., Memphis, TN.

Benedict, J.H., and D.W. Altman. 2001. Commercialization of transgenic cotton expressing insecticidal crystal protein, p. 137-201. In J.N. Jenkins and S. Saha (ed.) Genetic improvement of cotton. Science Publ., Enfield, NH.

Benedict, J.H., E.S. Sachs, D.W. Altman, W.R. Deaton, R.J. Kohel, D.R. Ring, and S.A. Berberich. 1996. Field performance of cottons expressing transgenic CryIA insecticidal proteins for resistance to Heliiothis virescens and Helicoverpa zea (Lepidoptera: Noctuidae). J. Econ. Entomol. 89:230-238.

Benedict, J.H., E.S. Sachs, D.W. Altman, D.R. Ring, T.B. Stone, and S.R. Sims. 1993. Impact of b-endotoxin-producing transgenic cotton on insect-plant interactions with Heliothis virescens and Helicoverpa zea (Lepidoptera: Noctuidae). Environ. Entomol. 22:1-9.

Benson, N.R, F.M. Bourland, W.C. Robertson, J.M. Hornbeck, and F.E. Groves. 2000. Arkansas cotton variety tests 2000. Res. Rep. 481. Arkansas Agric. Exp. Stn., Fayetteville, AR.

Bourland, F.M., N.R. Benson, and W.C. Robertson. 2000. Inherent biases in the Arkansas cotton variety testing program, p. 547 549. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Counc. Am., Memphis, TN.

Bowman, D.T. 1997. Cotton variety testing recommendations, p. 490. In Proc. Beltwide Cotton Conf., Nashville, TN. 9-12 Jan. 1996. Natl. Cotton Counc. Am.. Memphis, TN.

Bowman, D.T. 1998. North Carolina measured crop performance. Soybean and cotton. Rep. 176. North Carolina Agric. Res. Service., Raleigh, NC.

Bryant, K.J., C.T. Allen, F.M. Bourland, and L.D. Earnest. 1999. Cost and return comparisons of Roundup Ready and Bollgard cotton varieties, p. 236-238. In Proc. Beltwide Cotton Conf., Orlando, FL. 3-7 Jan. 1999. Natl. Cotton Counc. Am., Memphis, TN.

Bryant, K.J., W.C. Robertson, G. Lorenz, C.T. Allen, F.M. Bourland, and L. Earnest. 2000. Economic evaluation of transgenic cotton systems in Arkansas. p. 3843. In Proc. of the 2000 Cotton Research Meeting. Spec. Rep. 198. Arkansas Agric. Exp. Stn., Fayetteville, AR.

Buchanan, G.A. 1992. Trends in weed control methods, p. 47-72. In C.G. McWhorter and J.R. Abernathy (ed.) Weeds of Cotton: Characterization and control. The Cotton Foundation, Memphis, TN.

Caldwell, W.D., R.C. Griffin, D.J. Boquet, E.M. Holman, S. Moore. J.I. Dickson, P.D. Colyer, G.O. Myers, R. Gable, J. Thomas, A.B. Coco, P.R. Vernon, M. Walcotte, M. Deloach, and E.P. Millhollon. 1999. 1999 Louisiana cotton variety trials. Res. Summary 117. Louisiana Agric. Exp. Stn., Baton Rouge, LA.

Calhoun, D.S., J.E. Jones, J.J. Dickson, W.D. Caldwell, E. Burris, B.R. Leonard, S.H. Moore, and W. Aguillard. 1997. Registration of 'H1244' cotton. Crop Sci. 37:1013-1014.

Carpenter, J.E., and L.P. Gianessi. 2000. Value of Bt and herbicide resistant cottons, p. 76-79. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Council Am., Memphis, TN.

Coley, C.B. 2000. Seed selection: A southeastern grower's perspective. p. 26. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Council Am., Memphis, TN.

Collins, J.R. 1996. BXN Cotton: Marketing plans and weed control programs utilizing buctril, p. 201. In Proc. Beltwide Cotton Conf., Nashville, TN. 9-12 Jan. 1996. Natl. Cotton Counc. Am., Memphis, TN.

Creech, J.B., T.P. Wallace, J.R. Johnson, D.M. Ingram, N.W. Buehring, and W.H. McCarty. 1998. 1998 Mississippi cotton variety trials. Inf. Bull. 352. Mississippi Agric. and Forestry Exp. Stn., Mississippi State, MS.

Culpepper, A.S., and A.C. York. 1997. Weed management in notillage bromoxynil-tolerant cotton (Gossypium hirsutum L). Weed Technol. 11:335-345.

Culpepper, A.S., and A.C. York. 1998. Weed management in glyphosate-tolerant cotton. J. Cotton Sci. 2:174-185.

Culpepper, A.S., and A.C. York. 1999. Weed management and net returns with transgenic, herbicide-resistant, and non-transgenic cotton (Gossypium hirsutum L.). Weed Technol. 13:411-420.

Day, J.L., A.E. Coy, W.D. Branch, O.L. May, S.S. LaHue, and L.G. Thompson. 2001. Georgia 2001 peanut, cotton, and tobacco performance tests. Rep. 677. Georgia Agric. Exp. Stn., Athens, GA.

Dotray, P.A., and J.W. Keeling. 1997. Roundup Ready cotton tolerance to Roundup Ultra applied at various growth stages, p. 778. In Proc. Beltwide Cotton Conf., New Orleans, LA. 7-10 Jan. 1997. Natl. Cotton Counc. of Am., Memphis, TN.

Fitt, G.P., J.C. Daly, C.L. Mares, and K. Olsen. 1998. Changing efficacy of transgenic Bt cotton--Patterns and consequences, p. 189-196. In M.P. Zalucki et al. (ed.) Pest management--Future consequences. Univ. Queensland Printery, Brisbane, Australia.

Gerloff, D. 2001. Cotton budgets for 2001 [Online]. [30 p.] Available at: http://www.cottoninc.com/ag_cotton budgets/Tennessee/ cottonbudget01%20TN.pdf [posted 1 Jan. 2001; verified 6 Mar. 2003]. Agric. Ext. Service Publ. AE 01-42. Univ. Tennessee Inst. of Agric., Knoxville, TN.

Glass, K.M., C.D. Monks, and C.H. Burmester. 2000. 2000 Alabama cotton variety report. Rep. 233. Alabama Agric. Exp. Stn., Auburn, AL.

Gomez, K.A., and A.A. Gomez. 1984. Statistical procedures for agricultural research. John Wiley and Sons, New York.

Greenplate, J.T. 1999. Quantification of Bacillus thuringiensis insect control protein Cry1Ac over time in Bollgard cotton fruit and terminals. J. Econ. Entomol. 92:1377-1383.

Greenplate, J.T., W. Mullins, S. Penn, and K. Embry. 2001. Cry1Ac levels in candidate commercial Bollgard[R] cultivars as influenced by environment, variety, and plant age: 1999 gene equivalency field studies, p. 790-793. In Proc. Beltwide Cotton Conf., Anaheim, CA. 9-13 Jan. 2001. Natl. Cotton Counc. Am., Memphis, TN.

Hargett, J. 2000. Seed selection: A mid-south grower's perspective. p. 26. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Council Am., Memphis, TN.

Isgett, T.D., E.C. Murdock, and A. Keeton. 1996. Weed control in BXN[TM] cotton: Performance in 1995 applied research and grower's fields, p. 1535. In Proc. Beltwide Proc. Beltwide Cotton Conf., Nashville, TN. 9-12 Jan. 1996. Natl. Cotton Counc. Am., Memphis, TN.

Jenkins, J.N., J.C. McCarty, Jr., R.E. Buehler, J. Kiser, C. Williams, and T. Wofford. 1997. Resistance of cotton with 3-endotoxin genes from Bacillus thuringiensis var. kurstaki on selected lepidopteran insects. Agron. J. 89:768-780.

Jenkins, J.N., J.C. McCarty, Jr., and W.L. Parrott. 1990. Effectiveness of fruiting sites in cotton yield. Crop Sci. 30:365-369.

Jones, M.A., and C.E. Snipes. 1999. Tolerance of transgenic cotton (Gossypium hirsutum L.) to topical applications of glyphosate. J. Cotton Sci. 3:19-26.

Kappelman, A.J., Jr., and G.A. Buchanan. 1968. Influence of fungicides, herbicides, and combinations on emergence and seedling growth of cotton. Agron. J. 60:660-663.

Kappelman, A.J., Jr., G.A. Buchanan, and Z.F. Lund. 1971. Effect of fungicides, herbicides, and combinations on root growth of cotton. Agron. J. 63:3-5.

Keeling, J.W., and J.R. Abernathy. 1989. Pre-emergence weed control in a conservation tillage cotton (Gossypium hirsutum) cropping system on sandy soils. Weed Technol. 3:182-185.

Kerby, T., and R. Voth. 1998. Roundup Ready[R]--Introduction experiences in 1997 as discussed in the beltwide cotton production conference. Weed management: Transgenics and new technologies panel. p. 26-29. In Proc. Beltwide Cotton Conf., San Diego, CA. 5-9 Jan. 1998. Natl. Cotton Counc. Am., Memphis, TN.

Lee, J.A. 1985. Revision of the genetics of the hairiness-smoothness system of Gossypium. J. Hered. 76:123-126.

Lege, K.E., C.W. Smith, and J.T. Cothren. 1992. Genotypic and cultural effects on condensed tannin concentration of cotton leaves. Crop Sci. 32:1024-1028.

Lukefahr, M.J., J.E. Houghtaling, and D.G. Cruhm. 1975. Suppression of Heliothis spp. with cottons containing combinations of resistant characters. J. Econ. Entomol. 68:743-746.

Lukefahr, M.J., J.E. Houghtaling, and H.H. Graham. 1971. Suppression of Heliothis populations with glabrous cotton strains. J. Econ. Entomol. 64:486-488.

Lukefahr, M.J., L.W. Noble, and J.E. Houghtaling. 1966. Growth and infestation of bollworms and other cotton insects on glanded and glandless strains of cotton. J. Econ. Entomol. 59:817-820.

Mahaffey, J.S., J.R. Bradley, and J.W. Van Duyn. 1995. Bt Cotton: Field performance in North Carolina under conditions of unusually high bollworm populations, p. 795-798. In Proc. Beltwide Cotton Conf., San Antonio, TX. 4-7 Jan. 1995. Natl. Cotton Counc. Am., Memphis, TN.

May, O.L., A.S. Culpepper, D.S. Shurley, and P.M. Roberts. 2002. Cultivar evaluation: Evolution of systems trials to compare transgenic and non-transgenic cultivars. In Proc. Beltwide Cotton Conf., Atlanta, GA. 8-12 Jan. 2002. Natl. Cotton Counc. Am., Memphis, TN.

May, O.L., and E.C. Murdock. 2002. Yield ranks of glyphosate-resistant cotton cultivars are unaffected by herbicide systems. Agron. J. 94:889-894.

May, O.L., B. Nichols, T. Kerby, S. Brown, and J. Silvertooth. 2000. Proposed guidelines for pre-commercial evaluation of transgenic and conventional cotton cultivars, p. 503-508. In Proc. Beltwide Cotton Res. Conf., San Antonio, TX. 4-8 Jan. 2000. Natl. Cotton Council, Memphis, TN.

May, O.L., M.J. Sullivan, D.K. Barefield, Jr., D.M. Robinson, and G.M. Veazy. 1999. Performance of field crops in South Carolina-Cotton. Agric. and Forestry Exp. Stn. Circ. 184. Clemson Univ., Clemson, SC.

Murray, D.S., J.E. Street, J.K. Soteres, and G.A. Buchanan. 1979. Growth inhibition of cotton (Gossypium hirsutum) and soybean (Glycine max) roots and shoots by three dinitroaniline herbicides. Weed Sci. 27:336-342.

Nida, D.L., K.H. Kolacz, R.E. Buehler, W.R. Deaton, W.R. Schuler, T.A. Armstrong, M.L. Taylor, C.C. Ebert, G.J. Rogan, S.R. Padgette, and R.L. Fuchs. 1996. Glyphosate-tolerant cotton: Genetic characterization and protein expression. J. Agric. Food Chem. 44:1960-1966.

Olsen, K.M., and J.C. Daly. 2000. Plant-toxin interactions in transgenic Bt cotton and their effect on mortality of Helicoverpa armigera (Lepidoptera: Noctuidae). J. Econ. Entomol. 93:1293-1299.

Parrott, W.L., J.N. Jenkins, J.E. Mulrooney, J.C. McCarty, Jr., and R.L. Shepherd. 1989. Relationship between gossypol gland density on cotton squares and resistance to tobacco budworm (Lepidoptera: Noctuidae) larvae. J. Econ. Entomol. 82:589-592.

Perlak, FJ., R.L. Fuchs, D.A. Dean, S.L. McPherson, and D.A. Fischhoff. 1991. Modification of the coding sequence enhances plant expression of insect control protein genes. Proc. Natl. Acad. Sci. USA 88:3324-3328.

Phipps, B.S., N.R. Benson, F.M. Bourland, and C. Tingle. 2002. Are Roundup applications necessary for evaluation of Roundup Ready cultivars? [CD-ROM computer file]. Proc. Beltwide Cotton Conf., Atlanta, GA. 8-14 Jan. 2002. Natl. Cotton Counc. Am., Memphis, TN.

Robinson, S.H., D.A. Wolfenbarger, and R.H. Dilday. 1980. Antixenosis of smooth leaf cotton to the ovipositional response of tobacco budworm. Crop Sci. 20:646-649.

Roof, M.E. 2000. Cotton insect control [Online]. [2 p.] http://cufan. clemson.edu/pmguide/documents/cotton%20insect %20ccotton%20 bollworm%20 and %20tobacco %20budworm % 201arvae.htm [verified 6 Mar. 2003]. Clemson Univ. Cooperative Ext. Service, Clemson, SC.

Sanders, D.E., J.W. Barnett, S.T. Kelly, R.D. Bagwell, M.A. Martin, and R.D. Neal. 2000. Large plot comparisons of pre. vs. no pre. herbicides in Roundup Ready[R] (glyphosate-resistant) cotton, p. 1471-1472. In Proc. Beltwide Cotton Conf., San Antonio, TX. 5-9 Jan. 2000. Natl. Cotton Counc. Am., Memphis, TN.

Sheets, R.H., and T. Speed. 1997. Paymaster Cottonseed cotton varieties PM 2200 RR and PM 2326 RR. p. 39-40. In Proc. Beltwide Cotton Conf., New Orleans, LA. 7-10 Jan. 1997. Natl. Cotton Counc. Am., Memphis, TN.

Smith, R.H. 1997. An extension entomologist's 1996 observations of Bollgard (Bt) technology, p. 856-858. In Proc. Beltwide Cotton Conf., New Orleans, LA. 7-10 Jan. 1997. Natl. Cotton Counc. Am., Memphis,

TN.

Thomson, N.J., P.E. Reid, and E.R. Williams. 1987. Effects of the okra leaf, nectariless, frego bract, and glabrous conditions on yield and quality of cotton lines. Euphytica 36:545-553.

USDA-AMS. 1995. Cotton varieties planted--1995 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 1996. Cotton varieties planted--1996 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 1997. Cotton varieties planted--1997 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 1998. Cotton varieties planted--1998 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 1999. Cotton varieties planted--1999 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 2000. Cotton varieties planted--2000 Crop. USDA-AMS, Memphis, TN.

USDA-AMS. 2001. Cotton varieties planted--2001 Crop. USDAAMS, Memphis, TN.

USDA-AMS. 2002. Cotton varieties planted--2002 Crop. USDA-AMS, Memphis, TN.

Vargas, R.N., S. Wright, and T.M. Martin-Duvall. 1998. Tolerance of Roundup Ready[R] cotton to Roundup Ultrar applied at various growth stages in the San Joaquin Valley of California. p. 847-848. In Proc. Beltwide Cotton Conf., San Diego, CA. 5-9 Jan. 1998. Natl. Cotton Counc. Am., Memphis, TN.

Welch, A.K., P.R. Hahn, R.D. Voth, J.A. Mills, and C.R. Shumway. 1997. Evaluation of preplant and preemergence herbicides in Roundup Ready[R] cotton, p. 784-785. In Proc. Beltwide Cotton Conf., New Orleans, LA. 7-10 Jan. 1997. Natl. Cotton Counc. Am., Memphis, TN.

Wilcut, J.W. 1996. New weed management technologies for cotton: A perspective for North Carolina and Georgia. p. 205-206. In Proc. Beltwide Cotton Conf., Nashville, TN. 9-12 Jan. 1996. Natl. Cotton Counc. Am., Memphis, TN.

Wilcut, J.W., H.D. Coble, A.C. York, and D.W. Monks. 1996. The niche for herbicide-resistant crops in U.S. agriculture, p. 213-230. In S.O. Duke (ed.) Herbicide-resistant crops: Agricultural, environmental, economic, regulatory, and technical aspects. CRC Press, Boca Raton, FL.

Zummo, G.R., J.G. Segers, and J.H. Benedict. 1984. Seasonal phenology of allelochemicals in cotton and resistance to bollworm (Lepidoptera: Noctuidae). Environ. Entomol. 13:1287-1290.

O. L. May, * F. M. Bourland, and R. L. Nichols

O.L. May, Dep. of Crop & Soil Science, Univ. of Georgia, Coastal Plain Exp. Sm., P.O. Box 748, Tifton, GA 31793-0748; F.M. Bourland, Univ. of Arkansas, Northeast Res. and Ext. Ctr., Keiser, AR 72351; and R.L. Nichols, Cotton Incorporated, 6399 Weston Parkway, Cary, NC 27513. Received 5 July 2002. * Corresponding author (lmay@tifton.uga.edu).
COPYRIGHT 2003 Crop Science Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2003 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:May, O.L.; Bourland, F.M.; Nichols, R.L.
Publication:Crop Science
Geographic Code:1USA
Date:Sep 1, 2003
Words:7454
Previous Article:Transgenic herbicide tolerant canola--the Canadian experience.
Next Article:Translocation breakpoints in soybean classical genetic linkage groups 6 and 8.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters