Herbicide strategies for reducing nutgrass (Cyperus
rotundus L.) density in cotton (Gossypium hirsutum L.).

Introduction

Cyperus spp. were identified by Charles (1991) as the second most important
weeds of Australian cotton production, infesting 79% of New South Wales
cotton properties and 15% of the cotton area (21 000 ha).  This weed problem
is primarily attributable to Cyperus rotundus L. (nutgrass), which has
been reported at field densities of 2000-4000 tubers/m2, with corresponding
reductions of cotton lint yield of up to 92% (Charles 1995).
Australian cotton growers have traditionally attempted to control nutgrass
using MSMA and interrow cultivation in-crop, and cultivation in-fallow.
 However, Charles (1995) found this strategy relatively ineffective in
controlling nutgrass, but repeated applications of glyphosate gave effective
control both in-crop and in-fallow.  Charles (1995) suggested an infallow
strategy for controlling nutgrass of repeated applications of glyphosate,
and an in-crop strategy of MSMA applied early season, followed by glyphosate
applied mid or late season.  Applications should occur at intervals of
no longer than 4 weeks, as new tubers are produced by 4 weeks after shoot
emergence (Hauser 1962).  Charles' results also indicated that the inclusion
of norflurazon in the strategy may further enhance weed suppression.
Other researchers have reported successful control of nutgrass with a
range of alternative herbicides.  Hauser (1963) controlled nutgrass with
amitrole, EPTC and
2,4-D, but noted that asynchronous tuber emergence made nutgrass difficult
to eradicate.  Holt et al. (1962) also reported control with EPTC, while
Chernicky and Watkins (1992) found EPTC gave only short-term suppression
of nutgrass and was phytotoxic to cotton.  Johnson (1975) observed good
control with bentazon and perfluidone, and Grichar et al. (1992) with
imazethapyr, while Anderson and Dunford (1969) reported limited control
of nutgrass with bensulide.  W. K. Vencill (pers. comm.) found some degree
of control with a range of herbicides including norflurazon, methazole,
fomesafen, lactofen, fluometuron and diuron, with best results generally
coming from combinations of herbicides.  Keeley et al. (1973), Costa and
Appleby (1976), and Obrigawitch et al. (1980) also observed control of
C. esculentus with a range of herbicides including prometryn, methazole,
atrazine and metolachlor.
Halosulfuron-methyl is a newer herbicide being evaluated for registration
in Australia and has shown good activity on nutgrass (Somervaille 1995).
 Some cotton growers also report encouraging results with glyphosate-trimesium
and a combination of glyphosate + ethephon.
In this study, experiments were conducted to test and further develop
the strategy for controlling nutgrass in cotton, using norflurazon, MSMA
and glyphosate, proposed by Charles (1995).  Treatment combinations with
halosulfuron-methyl and a range of other herbicides were assessed (by
the density of nutgrass tubers) to
determine if a more effective strategy can be developed
for the control of nutgrass.

Materials and methods

Details of sites

Experiment 1 overlays the experiment at 'Wilona', Warren (31 48'S, 147 59'E),
on an alluvial soil, pH 6.7 (1 : 5 water) at the soil surface, to 7.4
at 0.7 m (McKenzie 1992), described by Charles (1995).  Experiment 2 overlays
the experiment at 'Norwood', Moree (291 26'S, 149 47'E), on a grey, cracking
clay, pH 7.3 (1:5 water), also described by Charles (1995).
Treatments for experiments 1 and 2 were allocated in a pseudo-random design,
so that at the start of each experiment, each treatment had the same average
density of nutgrass and included both heavily infested and lightly infested
plots, based on visual estimates in autumn 1992.  Treatments which included
norflurazon [a residual herbicide with a half-life of up to 180 days (Humburg
et al. 1989)1, overlayed plots previously treated with this herbicide.
Experiments 3, 4 and 5 were positioned beside experiment 2 and had the
same farming history.
Experiment 6 was conducted at 'Paloma', Premer (31 26'S, 149 55'E), on
a grey, cracking clay, pH 7.4 (0.01 mol CaCl2/L).
Experiments 3, 4, 5 and 6 were randomised complete block designs.

Herbicide application and sampling (experiments 1-5)

Experiments were situated in commercial, flood irrigated, cotton fields.
 Irrigation scheduling was determined using a neutron probe.  Insect pressure
was assessed twice weekly and insecticides applied as necessary.  Cotton
beds were maintained throughout the experiments, with minimal disturbance.
At Wilona (experiment 1), trifluralin (1.2 kg a.i./ha), diuron (0.5 kg
a.i./ha) and fluometuron (0.5 kg a.i./ha) were applied on 26 August 1992
and 18 August 1993.  Fluometuron (0.9 kg a.i./ha in a 0.5 m band) was
applied on the hill at sowing in 1992 and 1993.  Nitrogen was applied
as anhydrous ammonia at 118 kg N/ha in July 1992 and 104 kg N/ha in August
1993, and as water-run urea at 50 kg N/ha in December 1992 and 74 kg N/ha
in December 1993.  Cotton cv.  CS 50 was planted at 13 kg/ha in 1992,
and cv.  Sicala VI at 16 kg/ha in 1993.
At Norwood (experiments 2, 3, 4 and 5), trifluralin (1.4 kg a.i./ha) and
diuron (1.2 kg a.i./ha) were applied on 18 June 1992, and fluometuron
(1.5 kg a.i./ha in a 0.3 m band) was applied on the hill on 7 October
1992.  TrifIuralin (1.1 kg a.i.lha) and diuron (1.2 kg a.i./ha) were applied
on 27 August 1993, and fluometuron (1.5 kg a.i./ha) and pendimethalin
(0.3 kg a.i./ha) were applied in a 0.4 m band on the hill on 3 October
1993.  Nitrogen was applied as anhydrous ammonia in June 1992 and April
1993, at 160 and 250 kg N/ha,
respectively.  Cotton cv.  CS 189 was planted at 12 kg/ha in 1992 and
cv.  Siokra L23 at 10 kg/ha in 1993.
Nutgrass tuber densities were assessed annually in spring, from 14 soil
cores per plot, taken from the top of the centre row, at 1 m intervals.
 Cores were 70 mm diameter by 150 mm deep, and tubers were counted after
washing.  Plots were 20 m long by 4 m (4 rows) wide, except in experiments
3 and 4, where plots were 10 m long.
Herbicides were applied using a hand-held boom of plot width, delivering
100 L/ha, at 200 kPa nozzle pressure.  Preplanting herbicides were incorporated
to 50 mm depth with a Lilliston rolling finger cultivator.  In-crop herbicides
were applied using 2 directed nozzles per row, positioned to spray the
inter-row area; November and December treatments were applied through
a tractor-mounted, curtained sprayer.  Agral 60, a non-ionic surfactant,
was used with halosulfuron-methyl, imazethapyr and sulfometuron at 0.5,
0.2 and 0.25% of the spray volume, respectively.
At harvest, the central 2 rows of each plot were picked with a single
row plot picker and the yield of seed cotton recorded.  Subsamples were
taken from each row and processed in a single saw gin to assess ginning
percentage and lint yield.

Nutgrass control strategies based on glyphosate, MSMA and norflurazon
Experiments 1 and 2 test the strategy for the control of nutgrass proposed
by Charles (1995).  Halosulfuron-methyl treatment combinations were included
as a comparison.
Experiment 1. Seventeen treatments (Table 1) were allocated to 2 blocks
of 9 plots with 5 replicates.
Cores to assess the density of nutgrass tubers were taken on 28 October
1992, 3 November 1993 and 23 August 1994.  In 1994, tuber viability was
assessed in a glasshouse from a subsample of 20 tubers from each plot.
 Tubers were placed in a 3:1 sand/peat moss potting mix, in seedling trays
35 mm deep and shoot emergence was recorded for 120 days.
Experiment 2. Seven treatments (Table 2) were allocated to 2 blocks of
7 plots with 4 replicates; each treatment was included twice per replicate,
at both high and low initial densities of nutgrass, to assess the impact
of density on treatment efficacy.

Nutgrass control strategies based on alternative herbicides Experiments
3-5 test the efficacy of a range of alternative herbicides and herbicide
combinations for reducing nutgrass density.  Combinations of herbicides
tested in experiment 1 were included for comparison.  In experiment 3,
which primarily compared herbicides applied pre-emergence, 21 treatments
(Table 3) were applied in 1992-93, with 5 replicates.  In experiment 4,
which compared herbicides applied post-emergence, 15 treatments (Table
4) were applied with 5 replicates.

Based on the results of experiments 3 and 4, experiment 5 comprised an
additional 6 treatments which were assessed in 1993-94 (Table 5), with
4 replicates.

Strategies to reduce nutgrass density in afallow Experiment 6 was designed
to test the efficacy of 2,4-D and 2,4-D combinations for reducing nutgrass
populations in a fallow.  Norflurazon and atrazine combinations were added
in 1991 for comparison.  Herbicides were applied using the methods described
for experiments 1-5.

The paddock had been used for grazing and was sown
to oats on a number of occasions.  It was in fallow for 12 months before
the experiment and was regularly cultivated.  The experiment was established
in November 1988 and continued until October 1993; the site was lightly
grazed during this period.
Twelve treatments (Table 6) were applied with 3 replicates.  Plots were
10 m long by 3.5 m wide.  To assess the density of nutgrass tubers, 3
soil cores (70 mm diameter by 150 mm deep) per plot were taken at 2-m
intervals on 19 September 1989, 19 September 1990, 30 August 1991, 25
September 1992 and 28 October 1993.  Tuber density was not assessed at
the start of the experiment.

Statistical analyses

Nutgrass density data were analysed using the poisson model in GENSTAT
5. Two dimensional trends were removed as row and column effects, with
soil core weight (loge) as an offset.  Tuber densities from the spring
prior to the initial treatments were used as
covariates (loge) in the analysis and results were converted to tubers/m2.
 Standard errors are presented.  Single degree of freedom contrasts were
used to test specific comparisons between treatments.  Ginning percentage
and lint yield data were analysed using the REML model.  The maximum standard
error of the
difference between means (s.e.d.) is presented.  Statistical
significance was determined at the 5% level.

Results

Nutgrass control strategies based on glyphosate, MSMA and norflurazon

Experiment 1. Rainfall during the 1992-93 and 1993-94 cotton seasons was
350 and 262 mm, respectively (Fig. 1).
The density of tubers on untreated plots rose 278% over the 1992-93 and
1993-94 seasons, increasing from 1903 tubers/m2 at the start of the experiment,
to 4389 tubers/m2 in spring 1993 and 7194 tubers/m2 in 1994 (Table 1).
 In contrast, the nutgrass density on the benfuresate + glyphosate treatment,
fell from 1931 tubers/m2 in spring 1992, to 73 tubers/m2 in spring 1994,
representing a 96% drop.  This treatment also had the highest average
lint yield of 1488 kg/ha.
In-crop glyphosate treatments gave consistent
reductions in tuber density, when compared with untreated plots, with
further reductions with each additional application.  Glyphosate applied
in December and January reduced tuber density more than applications in
November and January or December and May.  An additional post-harvest
glyphosate application did not consistently reduce tuber density (treatments
6 and 7 v. treatments 5 and 2).
Norflurazon + glyphosate (treatments 12 and 13) did not reduce the nutgrass
density more than the comparable glyphosate treatments in 1992-93 (treatments
2 and 3), but gave a large additional decrease in 1993-94.  The MSMA +
glyphosate combination (treatment 8) did not reduce tuber density more
than the comparable glyphosate treatment (treatment 3), although the addition
of MSMA reduced tuber density on the halosulfuron-methyl treatment (treatment
16 v. treatment 14). The combination of MSMA + 2 glyphosate applications
(treatment 8) reduced tuber density more than the combination of 2 MSMA
applications with 1 glyphosate application (treatment 10).  The glyphosate
treatments reduced tuber density more than the halosulfuron-methyl treatments.
There were no significant treatment effects on cotton ginning percentages,
which averaged 41.8 +- 0.4 and 41.7 +- 0.3% in the 1992-93 and 1993-94 seasons,
respectively (data not shown).
The density of nutgrass tubers had a large effect on cotton yield (Fig.
2).  Contrasts showed a 63% improvement in lint yield from 3 glyphosate
applications (treatment 5) compared with 1 application (treatment 2),
and a 21% increase from 4 applications (treatment 6) compared with 3 applications
(treatment 5).
Experiment 2. Rainfall during the 1992-93 and 1993-94 cotton seasons was
317 and 315 mm, respectively (Fig. 3).
Analysis showed significant treatment, density and year effects, with
significant 2-way interactions between the terms.

Averaged over treatments, the density of nutgrass tubers on the low initial
density plots fell from 556 tubers/m2 in spring 1992, to 311 tubers/m2
in spring 1994, and on the high initial density plots, from 2250 to 367
tubers/m2 in spring 1994 (Table 2).
Averaged over years, the density of tubers on treatments which initially
had lower densities remained lower, except on the halosulfuron-methyl
plots.  There were large differences between the initial density of
tubers on some of the treatments, in particular the norflurazon + glyphosate
treatment, but treatment trends were relatively consistent with time (except
the halosulfuron-methyl treatment).

Averaged over the initial density, the density of nutgrass tubers on the
untreated plots rose from 815 tubers/m2 at the start of the experiment,
to 1029 tubers/m2 in spring 1994 (Table 2).  The lowest densities in spring
1994 were on the glyphosate, norflurazon + glyphosate and norflurazon
+ MSMA treatments (treatments 2, 3, 5 and 6), which had 87, 96, 96 and
86% reductions, respectively.  The glyphosate treatment (treatment 3),
also had average or aboveaverage lint yields in both seasons.
The glyphosate treatments (treatments 2 and 3) reduced tuber density more
than the halosulfuron-methyl treatment (treatment 7), and the postharvest
glyphosate application also reduced tuber density (treatment 3 v. treatment
2).  Tuber density on treatments receiving 1 MSMA and 2 glyphosate applications
(treatment 4) was comparable with the density with 3 glyphosate applications
(treatment 2) in the 1992-93 season, but the extra glyphosate gave a greater
reduction in tubers in 1993-94.
There were no significant treatment effects on cotton ginning percentage,
which averaged 35.0 +- 0.9 and 44.1 +- 0.2% in the 1992-93 and 1993-94 seasons,
respectively (data not shown).
Analysis showed significant treatment, density and year effects on cotton
lint yield, with interactions
between density and year, and treatment and year.  Lint yield was highest
on plots which initially had fewer nutgrass tubers.  Averaging over treatments,
high initial tuber density reduced cotton yield by 12 (1992-93) and 30%
(1993-94) when compared with low initial density.

Strategies based on alternative herbicides

Experiment 3. The density of nutgrass tubers on untreated plots declined
from 1342 tuberS/M2 at the start of the experiment, to 310 tubers/m2 in
spring 1993 (Table 3).  The lowest final densities were on the benfuresate
+ MSMA, benfuresate, benfuresate + glyphosate, EPTC + glyphosate and fluometuron
+ glyphosate treatments (treatments 7, 5, 6, 12 and 19), which had 95,
91, 92, 87 and 84% reductions, respectively.
Compared over treatments, tuber density was reduced most by benfuresate
(89% reduction) and fluometuron (80% reduction) combinations.  MSMA did
not significantly reduce tuber density, but treatment combinations which
included glyphosate reduced tuber density by 86% overall.
There were no significant treatment effects on lint yield or ginning percentage,
which averaged 37.8 +- 0.4% (data not shown).

Experiment 4. The density of nutgrass tubers on untreated plots declined
from 582 tubers/m2 at the start of the experiment, to 47 tubers/m2 in
spring 1993 (Table 4).  The lowest final densities were on the glyphosate,
glyphosate + ethephon, methazole, dimethenamid, atrazine and atrazine
+ glyphosate treatments (treatments 2, 3, 6, 8, 11 and 12).
Lint yield was only significantly reduced by the suifometuron treatment.
 Ginning percentage was not affected by treatment and averaged 37.4 +- ±
0.6% (data not shown).

Experiment 5. The density of nutgrass tubers on untreated plots rose from
434 tubers/m2 in spring 1993, to 1029 tubers/m2 in spring 1994.  The lowest
final density was on the halosulfuron-methyl + glyphosate treatment (treatment
6), where the second glyphosate application significantly lowered the
density (treatment 6 v. treatment 5).  There were no other significant
differences between treatments.
There were no significant treatment effects on lint yield or ginning percentage,
which averaged 43.8 +- 0.3% (data not shown).

Strategies to reduce nutgrass density in a fallow Experiment 6. Total weekly
rainfall for 1989-July 1993 is shown in Figure 4.

The density of nutgrass tubers progressively declined during the experiment
(Table 6), with the lowest final tuber densities on the 2,4-D ester +
glyphosate treatments (treatments 7 and 8).  Comparison over treatments
showed a significant reduction in density due to glyphosate.

Discussion

Nutgrass tuber density

Many of the herbicides and herbicide combinations examined, reduced nutgrass
densities to levels significantly below the densities of untreated plots.
 Multiple in-crop glyphosate applications consistently reduced nutgrass
densities, with further reductions from successive applications (experiments
1 and 2).  These results support the findings of Zandstra et al. (1974),
Doll and Piedrahita (1982) and Charles (1995) that nutgrass can be controlled
with glyphosate.
However, although visual observations indicated glyphosate killed all
vegetative shoots, new nutgrass shoots quickly emerged and many plots
were again heavily infested within 4 weeks of a glyphosate application
(data not shown).  These new shoots were produced either from tubers which
were not attached to vegetative shoots at the time of glyphosate application
and consequently were not exposed to the herbicide, or
from tubers which received a sublethal dose of glyphosate, possibly due
to inadequate herbicide translocation.

Incomplete translocation of glyphosate from vegetative shoots to attached
tubers was shown by Chase and Appleby (1979) to account for large differences
in glyphosate efficacy.  They observed a 5-fold decrease in translocation,
when comparing unstressed plants in high humidity, with moisture-stressed
plants in lower humidity.  Keeley et aL (1985) also showed translocation
decreased with increasing plant age.  Anecdotal evidence indicates giyphosate
translocation is reduced by cold stress, such that glyphosate is relatively
ineffective in controlling nutgrass when applied to young plants in early
spring.  Consequently, opportunities for applying glyphosate to nutgrass
in cotton are limited to an 8 week period from mid November (about 4 weeks
after cotton emergence).

As a result, long-term control of a nutgrass population will be difficult
to achieve in commercial cotton with a glyphosate-based program, as irrigations,
adverse weather conditions and other demands make it difficult to apply
the herbicide at the optimum time.  A mid season delay of 4 weeks due
to adverse conditions will prevent glyphosate being applied before cotton
canopy closure, allowing nutgrass to grow unchecked through the remainder
of a season.
The problem of glyphosate timing could be overcome by using a residual
herbicide.  In experiment 1, treatment efficacy was improved by combining
glyphosate with norflurazon, although multiple glyphosate applications
were required for best results (treatment 13 v. treatment 12).  However,
there was no additional benefit from norflurazon in experiment 2 (treatment
5 v. treatment 2).  The norflurazon + MSMA combination also reduced the
nutgrass density, with above-average cotton yields in experiment 2, though
not in experiment 3. Two factors contributed to these apparently inconsistent
results.

First, norflurazon is only active at high soil moisture levels, with activity
generally declining 2-3 days after an irrigation or rainfall event.  As
a result, norflurazon is active on an irregular basis, and is most effective
in wet seasons.  The 2 seasons covered by these experiments were relatively
dry, with below-average rainfall (Figs 1 and 3).  Additionally, the Norwood
crops were not fully irrigated, as insufficient water was available (Fig.
3).  Consequently, a reduction in nutgrass density due to norflurazon
was less likely to be observed than would have been the case in a wetter
season with full irrigation.
Second, field experience in Australia has shown norflurazon is most effective
when applied in consecutive years, with efficacy improving with time.
 Norflurazon treatments in experiments 1 and 2 overlayed plots which had
received norflurazon in each of the
2 previous years, whereas the area used for experiment 3 had no history
of norflurazon and a reduction in nutgrass density due to norflurazon
was less likely to be observed.  Nevertheless, even though the seasons
were relatively dry, there was a significant reduction in the density
of nutgrass on norflurazon-treated plots in experiment 1, which received
full irrigation.
Several cotton growers reported improved nutgrass control from adding
ethephon to glyphosate, but these reports were not supported by the results
of experiment 1 (treatment 9).  However, during application it was observed
that the concentrates reacted and came out of suspension.  Treatment efficacy
and cotton yield were improved in experiment 4 (treatment 3), where the
concentrates were not mixed.
The benfuresate + glyphosate combination also reduced nutgrass density
and gave excellent cotton yields in experiment 1, but was not so outstanding
in experiment 3. However, this herbicide will not undergo further evaluation
as it is not registered in Australia and has been dropped from commercial
development.
The glyphosate + fluometuron combination gave amongst the best results
in experiment 3, and was included in experiment 5 (treatments 2 and 3),
where it gave similar results to the halosulfuron + glyphosate combination
(treatment 6).
The EPTC + glyphosate combination was not further evaluated, as EPTC is
phytotoxic to cotton (Chernicky and Watkins 1992), and the result in experiment
3 was not superior to the result for the glyphosate + fluometuron combination
(treatment 12 v. treatment 19).
Methazole showed good efficacy on nutgrass in experiment 4.(treatment
6), and is safe on cotton, but is not registered for use in cotton in
Australia.  Dimethenamid also showed good efficacy, but again is not registered
for use in cotton.  Atrazine reduced the nutgrass density (treatments
1 1 and 12), but is phytotoxic to cotton, although it remains an option
in other cropping systems.

Neither imazapyr nor imazethapyr reduced the nutgrass population at the
rates used in experiment 4, even though they have shown good activity
in other situations at higher rates (Grichar et al. 1992; Somervaille
1995).  At rates of 1.0-1.5 L a.i./ha, imazapyr acts as,a soil sterilant
and should give much better results.  Its use at these rates to eliminate
isolated nutgrass patches from an otherwise clean field may be practical
in some situations.
Halosulfuron-methyl also gave poor results in experiment 1 (treatments
14-17), experiment 2 (treatment 7) and experiment 3 (treatments 9 and
10).  Halosulfuron-methyl did reduce the nutgrass density at the higher
rate (0.056 compared with 0.038 kg a.i./ha) in experiment 5 (treatments
4-6), but still required multiple applications.

Herbicide applications did not effectively control nutgrass in a rain-fed
fallow, largely due to the limited number of opportunities to apply herbicides
in a series of dry summers (Fig. 4).  Time was by far the largest influence
on nutgrass density, which fell from an average of 4454 tubers/m2 in 1989,
to 657 tubers/m2 in 1993.

Cotton lint yield

There was a strong negative relationship between lint yield and nutgrass
density at the end of the season, in experiment 1 (Fig. 2).  This relationship
was apparent in both seasons, even though, at the start of the first season,
all treatments had about the same weed density.  Also, many of the treatments
were not applied until relatively late in the season, with the second
(December), and third (January) herbicides applied 11 and 15 weeks after
planting.  Nevertheless, contrasts showed that treatment in January resulted
in a yield increase.  The relationship between density and yield should
become more pronounced over time, as nutgrass densities decline on the
better treatments.

The relationship between lint yield and nutgrass density was not so apparent
in experiment 2, although the level of the initial nutgrass population
had a highly significant (P<0.001) effect on yield.  Lack of yield response
to treatment can be explained by 2 factors.  First, the level of competition
in experiment 2 was probably lower due to a lighter nutgrass infestation
than occurred in experiment 1. Consequently, by the end of the second
season, the nutgrass density on untreated plots was 1123 tubersIM2 in
experiment 2 compared with 7194 tubers/m2 in experiment 1. Second, cotton
seedling vigour and growth were more rapid in experiment 2, which was
250 km north of experiment 1, and temperatures were warmer by an average
1.30C in 1992 and 2.90C in 1993 (average of daily maximum and minimum
temperatures, 1 October-31 December).  Cotton plants in experiment 2 were
taller and more rapidly shaded the nutgrass.

Conclusions

Of the herbicides examined, only benfuresate was superior to glyphosate
in reducing nutgrass densities, although methazole, dimethenamid and atrazine
may be useful in some cropping systems.  Of the remaining herbicides,
combinations of glyphosate + norflurazon, glyphosate + fluometuron and
MSMA + norflurazon gave large reductions in the nutgrass population in
some experiments.

>From these results, a strategy for reducing nutgrass densities in cotton
can be developed using norflurazon preplanting, and multiple applications
of glyphosate and/or MSMA in-crop, as suggested by Charles (1995).  The
inclusion in this program of norflurazon, a residual herbicide, reduces
difficulties with crop sensitivity, application timing and adverse conditions
encountered
using glyphosate alone.  Norflurazon can be applied presowing, is active
throughout the cotton season, and is most effective in wet conditions,
when nutgrass is actively growing.  MSMA was relatively ineffective by
itself (Charles 1995), but had much better efficacy when used in combination
with norflurazon.  MSMA can be used earlier in the season than glyphosate
and has greater crop safety.  Glyphosate has a narrower application window
than MSMA and less crop safety, but has superior efficacy and consistency
for reducing the nutgrass density.  Also, its efficacy is improved when
combined with norflurazon.
An ideal strategy for reducing a nutgrass population, will use norflurazon
(applied preplanting), and in-crop applications of MSMA (early season)
and glyphosate (mid and late season).