INTRODUCTION

Tobacco has a terminal meristem and suppresses growth of axillary shoots through hormonal activity. In most tobacco types, the terminal flower is removed in the button stage 6 to prevent the development of the seed head and to allow transfer of energy to the leaves to increase leaf size, weight, body, nicotine, and other chemical constituents 20. Hormones that inhibit axillary shoot growth are produced in the terminal bud, so removal of the terminal bud also removes apical dominance 2. Axillary shoots (more commonly referred to as suckers in tobacco), primarily those in the top three to four leaf axils, will then begin to grow vigorously 2. The suckers in the uppermost leaf axils will re-establish partial apical dominance and suppress growth of suckers in lower leaf axils 14. There are three potential suckers in each leaf axil; the well differentiated primary sucker attached to the main stem, the less differentiated secondary sucker attached to the adaxial surface of the leaf petiole, and a meristematic area outward on the leaf petiole 15. If suckers are not removed and/or if growth is not controlled, suckers will form a flower and seed head 22, resulting in lower cured-leaf yield and reduced physical and chemical quality of the tobacco leaf 19. Tests conducted by Collins and Hawks found that tobacco plants that were topped but not suckered contained less total alkaloids and reducing sugars than tobacco that was topped and suckered 2. Suckers can also interfere with harvest, especially if the tobacco is mechanically harvested 3.

For many years, growers used hand labor to remove suckers, which was a difficult and time-consuming practice 2. Approximately 60 worker hours were required to hand sucker an acre of flue-cured tobacco if suckers were removed when they were 10 to 20 cm long 2. The use of chemicals to control the growth of tobacco suckers in the United States began in the 1940's when mineral oil was used to desiccate small suckers 18 and 11, but it was not until the introduction of maleic hydrazide (MH) in the 1950's that a significant reduction in labor required to remove suckers was achieved 12. MH is a true systemic inhibitor of plant cell division 12 and 13. It is a uracil anti-metabolite 2 and results in inhibition of mitosis and the loss of the layered organization in the tunica-corpus of the apical shoot meristem 1. In addition, it has been reported to inhibit DNA and RNA synthesis 20, uptake and assimilation of nitrate 5, respiration, and photosynthesis 4.

Once applied to actively growing tobacco plants, MH is rapidly absorbed and translocated in both the xylem and phloem to meristems, where it inhibits cell division but not cell elongation 12. Working with radio-labeled MH, Tso found that twenty-eight days after treatment, 30–40% of the absorbed (¹⁴C)-MH was translocated to the roots and released into the nutrient solution, 12–22% remained in the plant, 14–18% was extracted as methanol-soluble metabolites, and 25–35% remained in the roots and other tissues as a methanol-insoluble residue 20. Because MH is translocated to meristems, application results in control of small suckers as well as slightly larger suckers that develop at a slower rate 2. Collins and Hawks 2 noted that under normal conditions, MH will provide sucker control for about six weeks after application.

The range of registered rates for MH in North Carolina is 2.5 to 3.4 kg ai ha⁻¹ 6. Many factors influence MH residue levels, including chemical characteristics, application rate, duration of time between application and harvest, stalk position, and environment 20. MH is resistant to decomposition from UV radiation and high temperature, as well as loss to volatilization. It also is not readily metabolized once in the plant 2. MH is very water-soluble and rainfall is the single most important factor that affects MH residue levels after application. Residues levels are typically lower in seasons with above average rainfall and higher with below average rainfall 12. When 2.0 cm of irrigation water was applied 12 hours after the MH application, sucker control was unaffected, but MH residues were reduced from 62 to 30 ppm of MH compared to a non-irrigated treatment 16. Further research showed that producers can have significantly lower MH residues if the first harvest can be delayed until the plants receive rainfall ranging from 0.01 to 1 cm 8.

Because of its high residues in tobacco, MH has been scrutinized by political and governmental agencies throughout the world. For example, in 1978 the Federal Republic of Germany enacted a new food law, which stipulated that all pesticides must be approved for use or meet tolerance levels established for foods 21. Tobacco was considered a food in this new law 7, and a tolerance of 80 ppm for MH residues was set for tobacco products sold in Germany 23.

There have been considerable research and grower education efforts to reduce MH residues in US tobacco. The North Carolina Cooperative Extension Service recommends several management practices that reduce MH residues, including use of a reasonable nitrogen rate to reduce sucker pressure, using contact fatty alcohols and flumetralin (not relying solely on MH for sucker control), using only registered rates of MH, and allowing at least one week between application and harvest 6.

The experimental compound diflufenzopyr has shown the ability to suppress sucker growthin tobacco. Diflufenzopyr is a member of the semicarbazone herbicide family that inhibits polar auxin transport 9. According to Lym et. al. 10 an auxin transport inhibitor suppresses the transport of naturally occurring indole acetic acid and synthetic auxin-like compounds in plants. In general, diflufenzopyr interferes with the auxin balance needed for plant growth. Diflufenzopyr was first developed as a herbicide and has shown activity on velvetleaf, field bindweed, and mesquite. 9. In addition, diflufenzopyr has been evaluated in combination with dicamba and other auxin herbicides to enhance their activity for perennial weed control 10.

The objectives of this research were to evaluate the efficacy for sucker control and tolerance of tobacco to diflufenzopyr and to evaluate the ability of diflufenzopyr to reduce the need for MH and therefore reduce MH residues.

METHODS AND MATERIALS

Research was conducted in 2003 and 2004 at the Central Crops Research Station (CCRS) near Clayton, NC and at the Border Belt Tobacco Research Station (BBTRS) near Whiteville, NC, to evaluate the use of diflufenzopyr for the control of axillary bud (sucker) growth in flue-cured tobacco (Nicotiana tabacum L.). Soils were a Plinthic Paleudult and a Typic Paleudult at CCRS and BBTRS in both years, respectively. Flue-cured cultivar ‘K326’was transplanted on May 9^th in 2003 and April 26^th in 2004 at CCRS and ‘K346’ was transplanted on April 30^th in 2003 and April 22^nd in 2004 at BBTRS. Tobacco was produced using normal production practices for each research station and according to extension recommendations 17, except for treatments imposed.

Experimental design was a randomized complete block with a factorial treatment arrangement and four replications. Plots were two rows wide, and were 13.7 m in length with 1.20 m between rows at BBTRS and 12.2 m in length with 1.14 m between rows at CCRS. All treatments were applied with a CO₂ powered backpack sprayer with a delivery volume of 467 L ha⁻¹ at 138 kPa. Treatments were applied using three nozzles per row, with 26 cm spacing between nozzles. The outside nozzles were a TG 3 (TEEJET: Spraying Systems Co. Wheaton IL 60189) and the center nozzle was a TG 5. Treatments consisted of maleic hydrazide (MH) at four rates (0, 0.6, 1.3, and 2.5 kg ai ha⁻¹), flumetralin at two rates (0 and 0.7 kg ai ha⁻¹), and diflufenzopyr at two rates (0 and 0.017 kg ha⁻¹). All treatments were applied approximately seven days after the second contact fatty alcohol application (equivalent to ten days after the removal of the terminal flower). Timing of diflufenzopyr application was based on extension recommendations for the proper timing of MH and flumetralin application 17.

Cured leaf yield and quality, and sucker control data were collected from one row of the two-row plot to allow for a common border row between plots. Suckers were removed by hand from ten plants in each plot, counted and weighed. A control treatment with maximum sucker expression was established by removing the terminal flower from each plant and allowing unrestricted sucker growth (topped, but not suckered). Sucker weights in each treated plot were then compared to the maximum sucker expression data collected in the control to determine percent sucker control. The Tobacco Chemistry Lab at North Carolina State University performed analysis for total alkaloids and reducing sugars on a 50 g cured leaf sample, composited by weight over primings, for each plot. No diflufenzopyr related treatment differences were observed with total alkaloids or reducing sugars, therefore data are not reported. MH residue analyses were conducted on a composite cured leaf sample from each plot by Southern Testing, Wilson, NC. All data were subjected to a factorial analysis of variance (ANOVA) (Table 2) and treatment means were separated using a least significant difference value (LSD) at P≤0.05.

RESULTS

Data are reported for the appropriate main effects or interactions based on significance from ANOVA (Table 1). Sucker fresh weight per plant and yield data had a significant location by MH by flumetralin by diflufenzopyr interaction. Therefore, these data were analyzed by location and appropriate interactions and main effects are reported based on the ANOVA for each location. Means were separated using Fisher's F-protected LSD at P≤0.05.

Table 1 Analyses of variance (p-values) for sucker fresh weight per plant, number of suckers per plant, fresh weight per sucker, percent sucker control, yield, and grade index.

View Fullscreen

Percent sucker control

The Loc*MH*FLM*DIF and Loc*MH*FLM interactions were not significant for percent sucker control. However, the MH*FLM*DIF, Loc*MH*DIF and Loc*FLM*DIF interactions were significant at p values of 0.0002, 0.0310 and 0.0162 respectively (Table 1). Therefore, means were reported by location when appropriate, and data were averaged over the missing factor.

Sucker control with MH alone ranged from 62 to 85% (Table 2). No statistical differences were seen in percent sucker control between the 0.6 and 1.3 kg rates of MH, but 2.5 kg of MH used alone improved sucker control compared to the lower MH rates. Flumetralin and diflufenzopyr used alone resulted in 73 and 64% sucker control, respectively and provided sucker control equivalent to that from the 0.6 and 1.3 kg of MH rates alone. Neither flumetralin nor diflufenzopyr used alone was as effective for sucker control as the 2.5 kg of MH alone. When MH and flumetralin were used together, percent sucker control ranged from 79 to 93% and was not affected by the increase of MH rates. Combinations of MH and diflufenzopyr resulted in sucker control ranging from 70 to 85%, which increased as the rate of MH increased from 0.6 to 2.5 kg.

Table 2 Percent sucker control with maleic hydrazide, flumetralin, and diflufenzopyr averaged over locations.

View Fullscreen

When 1.3 kg of MH was used with diflufenzopyr, sucker control was equivalent to MH at 2.5 kg alone and 0.6 kg of MH combined with flumetralin, and was equivalent to the highest percent sucker control rating in the experiment. Diflufenzopyr and flumetralin used together resulted in 83% sucker-control, which was as effective as any other treatment. Three-way combinations of MH, flumetralin, and diflufenzopyr resulted in 84 to 87% sucker control, which was not any better than when flumetralin and diflufenzopyr were used with 1.3 or 2.5 kg of MH.

The Loc*MH*DIF interaction (Table 3) was significant at p = 0.0310, therefore means are reported for each location and data were averaged over FLM (flumetralin) applications. The interaction reported in Table 3 indicates that percent sucker control was poor when neither MH nor diflufenzopyr were applied. When both compounds were used percent sucker control was not improved when MH rates were increased from 0.6 to 2.5 kg or whether diflufenzopyr was used or not.

Table 3 Percent sucker control at four locations with maleic hydrazide and diflufenzopyr averaged over flumetralin rates.

View Fullscreen

The Loc*FLM*DIF interaction (Table 4) was significant at p = 0.0162. Applications of flumetralin and diflufenzopyr, whether used separately or together, increased percent sucker control at three of the four locations. The combination of flumetralin and diflufenzopyr did not consistently increase percent sucker control compared to using each compound alone at three of the four locations (Table 4).

Table 4 Percent sucker control at four locations for flumetralin and diflufenzopyr averaged over maleic hydrazide rates.

View Fullscreen

Number of suckers per plant

The four-way interaction of Loc*MH*FLM*DIF was not significant with p =  0.9301 for sucker number per plant. The three-way interaction of Loc*MH*FLM was significant with a p-value of 0.0007. Two-way interactions Loc*DIF and FLM*DIF, were also significant with p values of 0.0173 and 0.0052, respectively. All significant interactions were averaged over the missing factor (Table 1).

There were no differences in sucker number per plant when means were averaged over diflufenzopyr rates at CCRS in 2003 (Table 5). At the BBTRS in 2003 and 2004, the use of MH alone, flumetralin alone, or MH and flumetralin together reduced sucker number per plant more than diflufenzopyr alone when means were averaged over diflufenzopyr rates. At both locations the use of MH at increasing rates showed only a small reduction in sucker number per plant. At CCRS in 2004 combinations of MH and flumetralin significantly reduced sucker number per plant compared to the use of MH alone (Table 5).

Table 5 Number of suckers per plant at four locations with maleic hydrazide and flumetralin averaged over diflufenzopyr rates.

View Fullscreen

The Loc*DIF interaction (Table 6) suggests that there were no differences in sucker number per plant at two of four locations. At the BBTRS in 2003 and 2004 applications of diflufenzopyr reduced sucker number by more than 50%. The two-way interaction, FLM*DIF (Table 7), shows a reduction in sucker number per plant, when either flumetralin or diflufenzopyr were used. However, the combination of flumetralin and diflufenzopyr reduced sucker number, statistically the same as when flumetralin was used alone.

Table 6 Number of suckers per plant at four locations with diflufenzopyr averaged over flumetralin and maleic hydrazide rates.

View Fullscreen

Table 7 Number of suckers per plant with flumetralin and diflufenzopyr averaged over maleic hydrazide rates and locations.

View Fullscreen

Fresh weight per sucker

The Loc*MH*FLM*DIF interaction and all three-way interactions were not significant (Table 1). The Loc*DIF and Loc*MH were significant at 0.0125 and 0.0001, respectively. Also the main effect of FLM was significant at a p-value of 0.0157 (Table 1). The Loc*DIF interaction reported in Table 8 indicates that diflufenzopyr had no effect on fresh weight per sucker at two of the four locations. Fresh weight per sucker was only reduced at the BBTRS location in 2003 and 2004, by 34 to 48%. For the Loc*MH interaction at the CCRS 2003 location, MH had no effect on fresh weight per sucker. However, MH did significantly reduce fresh weight per sucker with the 0.6 kg rate or higher at BBTRS in 2003, with the 2.5 kg rate at CCRS in 2004, and with the 1.3 kg rate or more at the BBTRS in 2004 (Table 9). The main effect of FLM showed consistent reduction of fresh weight per sucker from 88 to 70 grams when averaged over locations, MH, and diflufenzopyr applications (Table 10).

Table 8 Fresh weight per sucker at four locations with diflufenzopyr averaged over maleic hydrazide and flumetralin rates.

View Fullscreen

Table 9 Fresh weight per sucker at four locations with maleic hydrazide averaged over diflufenzopyr and flumetralin rates.

View Fullscreen

Table 10 Main effect of flumetralin on fresh weight per sucker averaged over locations.

View Fullscreen

Sucker fresh weight per plant

The four-way interaction of Loc*MH*FLM*DIF was significant at a p-value of 0.0037 (Table 1). Therefore, each location was reported separately, including an analysis of variance for the main effects and all two and three-way interactions (Tables 11–14).

Table 11 Sucker fresh weight per plant at CCRS¹ in 2003 with maleic hydrazide and diflufenzopyr, averaged over flumetralin rates.

View Fullscreen

Table 12 Sucker fresh weight per plant at BBTRS¹ in 2003.

View Fullscreen

Table 13 Sucker fresh weight per plant at BBTRS¹ in 2004.

View Fullscreen

Table 14 Sucker fresh weight per plant at CCRS¹ in 2004 with maleic hydrazide and flumetralin, averaged over diflufenzopyr rates.

View Fullscreen

For the CCRS 2003 location the MH*DIF interaction was significant at a p-value of 0.0422 (Table 11). Data, averaged over flumetralin rates, indicated that diflufenzopyr applied alone significantly reduced sucker fresh weight per plant from 25 to 14 grams. Reductions in sucker fresh weight obtained by diflufenzopyr were statistically similar to reductions caused by applications of MH alone. The combination of MH and diflufenzopyr did not reduce sucker fresh weight more than when MH was applied alone.

Both the BBTRS 2003 and 2004 locations had a significant MH*FLM*DIF interaction at p-values of 0.0001 (Tables 12 and 13). At both locations all chemical treatments significantly reduced sucker fresh weight compared to the control. Combinations of MH and diflufenzopyr and MH and flumetralin reduced sucker fresh weight with no difference between the two combinations, at both locations. Three-way combinations, two-way combinations at the 1.3 and 2.5 kg rates of MH, and the diflufenzopyr and flumetralin combination without MH, all resulted in a consistent reduction of fresh weight per sucker per plant at both locations.

At the CCRS in 2004 the MH*FLM interaction was significant at a p-value of 0.0150 (Table 14). Flumetralin applications statistically reduced sucker fresh weight per plant. At this location, applications of flumetralin alone, averaged over diflufenzopyr rates was more effective in reducing sucker fresh weight than any MH treatment applied alone.

Yield

The Loc*MH*FLM*DIF interaction was significant at a p-value of 0.0141 (Table 1). Locations are reported separately with analysis of variance performed on the main effects, two-way, and three-way interactions.

There were no differences in yield at the CCRS 2003 and 2004 locations and data are not shown. At BBTRS in 2003 and 2004 the three-way interaction was significant (MH*FLM*DIF) with p-values of 0.0297 and 0.0031, respectively (Tables 15 and 16). All chemical treatments at the BBTRS 2003 and 2004 locations yielded higher than the non-treated control with unrestricted sucker growth. At BBTRS 2003 (Table 15), the highest yielding treatments were also the treatments that had the highest percent sucker control (Tables 3 and 4). At the BBTRS 2004 location (Table 16), diflufenzopyr treated tobacco yielded significantly less than tobacco treated with a combination of MH and flumetralin or either compound applied alone.

Table 15 Yield at BBTRS¹ in 2003.

View Fullscreen

Table 16 Yield at BBTRS¹ in 2004.

View Fullscreen

Quality

The four-way interaction and all three-way interactions were not significant for quality. The Loc*MH and MH*FLM interactions were significant with p-values of 0.0430 and 0.0255, respectively (Tables 17 and 18).

Table 17 Grade index at four locations with maleic hydrazide averaged over diflufenzopyr and flumetralin rates.

View Fullscreen

Table 18 Grade index with maleic hydrazide and flumetralin rates averaged over diflufenzopyr rates and locations.

View Fullscreen

The use of MH improved grade indices at BBTRS in 2003, and with 1.3 kg of MH at CCRS in 2003, but did not significantly affect grade index in 2004 (Table 17). Due to a significant MH*FLM interaction data were averaged over diflufenzopyr rates and locations. The lowest grade index was observed when no MH or flumetralin was used. Other differences with respect to grade index were inconsistent across treatments and locations (Table 18). No treatment related differences in reducing sugars and total alkaloids were observed (data not shown).

Residues of MH and diflufenzopyr

MH residue data for selected treatments were collected at BBTRS in 2003 and BBTRS and CCRS in 2004 (Table 19). MH residues ranged from below the detection limit of 10 ppm up to 31 ppm with 0.6 kg, from 12 to 84 ppm with 1.3 kg, and from 25 to 152 ppm with 2.5 kg of MH. At all locations, MH residues increased with increasing MH rates. MH residues were not affected by the combination of diflufenzopyr and MH. In previous experiments diflufenzopyr residues were below the detection limit of 0.05 ppm, even with rates substantially higher than those used is these experiments (data not shown).

Table 19 Maleic hydrazide residue levels at the BBTRS¹ and CCRS².

View Fullscreen

DISCUSSION

Control of sucker growth with all treatments that included compounds currently registered for sucker control in tobacco in the United States (MH and flumetralin) was similar to previous research 20 (Tables 2,3 and 4). Diflufenzopyr alone resulted in 64% sucker control, and control was similar to flumetralin alone and 0.6 and 1.3 kg of MH alone (Table 2). Percent sucker control from diflufenzopyr alone, however, was less than the standard treatment of a tank mixture of MH at 2.5 kg and flumetralin at 0.7 kg and is therefore below acceptable levels. Diflufenzopyr and MH at 1.3 kg (one-half the registered rate of MH) controlled sucker growth as well as any treatment in the experiment and was equivalent to the standard treatment. In addition, the three-way combination of diflufenzopyr, MH, and flumetralin allowed a further reduction in MH rate, down to 0.6 kg, without reducing sucker control compared to the standard treatment. The tank mixture of diflufenzopyr and flumetralin also resulted in sucker control equivalent to the standard treatment discussed above. However, sucker control with flumetralin in this experiment was better than would be expected when applied with typical grower equipment 20. Flumetralin requires contact with the leaf axils to control sucker growth. Treatments were applied with a backpack sprayer, which resulted in greater precision of application than is typically achieved with high clearance or tractor mounted sprayers used by growers.

Diflufenzopyr also reduced sucker number and size of suckers. Fresh weight per sucker was reduced by diflufenzopyr at two of four locations (Table 8). Sucker number was reduced by approximately 32% by diflufenzopyr when averaged over MH rates. However, diflufenzopyr was not as effective as flumetralin, which reduced sucker number by 75%, and the combination of diflufenzopyr and flumetralin did not reduce sucker number more than flumetralin alone (Table 7).

There were no yield differences at CCRS in either year. Yield at BBTRS in 2003 was improved with all chemical treatments and generally increased with increasing levels of sucker control. At BBTRS in 2004, however, seven of the eight treatments that included diflufenzopyr yielded less than chemical treatments that did not include diflufenzopyr. Yield losses were not related to poor sucker control and no visual injury from diflufenzopyr was observed; therefore there is no apparent reason for the yield reduction. Chemical treatments did not consistently affect quality or average price and differences in value per hectare were related to yield.

Acceptable levels of sucker control can be achieved with the tank mixture of diflufenzopyr and one-half the registered rate of MH (1.3 kg) or from diflufenzopyr, flumetralin, and one-fourth the registered rate of MH (0.6 kg). Diflufenzopyr would therefore allow MH residues to be substantially reduced without reducing overall sucker control (Table 2 and 19).