INTRODUCTION
Johnsongrass (Sorghum halepense (L.) Pers.) is among the
world`s worst weeds. It has been problem in many crops including cotton
(Holm et al., 1991). It is a Mediterranean plant where Turkey
is located (Holm et al., 1991).
Cotton is an important crop in Turkey, which is sown over 600 000 ha
area (FAO, 2006). Although total cotton sown area have not changed much
in Turkey; cotton sown areas shifted after initiation of Southeast Anatolia
Irrigation Project (GAP) where the cotton areas have expanded in parallel
to increasing irrigation possibilities and become main crop in the irrigated
fields. Johnsongrass is one of the worst weeds in cotton fields in the
GAP Region as well as other regions of Turkey (Uludag and Uremis, 2000).
The interference of johnsongrass with cotton has been already studied.
Keeley and Thullen (1981), found that 20 shoots of johnsongrass per square
meter could reduce seed cotton yield 20% and densities exceeding 75 shoots
m-2 caused over 75% loss. In addition, cotton stands and height
were reduced by johnsongrass competition. Bridges and Chandler (1987),
reported that 2 johnsongrass plants in 9.8 m row length of cotton was
critical observed density. The average loss was 1% when 1 johnsongrass
plant was established, it reached to 70% when density was 32 plants/9.8
m. Intraspecific competition was clearly evident at high johnsongrass
densities, which the total number of panicles and johnsongrass fresh weight
were the parameters measured. Cotton yield was differently affected by various johnsongrass densities
but cotton cultivars with different height had similar competitiveness
(Bridges and Chandler, 1988).
Although herbicides and mechanical control methods applied intensively,
johnsongrass stays as a problem in the cotton fields in Turkey (Uludag
and Uremis, 2000). There are examples of differential effect of weeds
on crops depending on regional difference (Zimdahl, 2004). The aim of
the current study was to quantify the effects of johnsongrass on cotton
in the Southeast Anatolia Region where johnsongrass is native.
MATERIALS AND METHODS
Four field experiments were carried out in three locations, which
were Diyarbakir, Sanliurfa and Kahramanmaras provinces in 1992, 1997 and
2002, respectively, in the Southeast Anatolia Region of Turkey with natural
johnsongrass densities in order to show the effect of johnsongrass densities
on cotton yield. Fields had different soil characteristics (Table
1).
Each year different cotton varieties, which were KAT 64, St 250/1 and
Maras 92 in 1992, 1997 and 2002, respectively, were sown before mid May.
| Table 1: |
Soil type and properties of experiments |
 |
The competitiveness of cotton
cultivars was not different in previous experiments (Bridges and Chandler,
1988). Before planting trifluralin at 2 kg ha-1 were applied
and incorporated into soil. Fertilizers were applied at planting and in
June. Two different types of experiments were set in the first week of
June. One was based on 4 m of cotton row length with 0, 3, 6, 8, 11, 14
and 22 johnsongrass densities in 1992 and 0, 3, 6, 8, 14 and 22 plant
densities in 1997. The other was with 0, 1, 2, 4, 8, 16 and 32 johnsongrass
densities per 8 m cotton row length, which were carried out in 1997 and
2002. Plots included 6 rows of cotton, which were 0.70 m apart from each
other. The length of plots was either 4 or 8 m depending on experiment
type. Undesired plants were removed by hand pulling and hand hoeing every
10 days until the end of the August. Inter row cultivation were applied
twice, before experiments were set and at the end of the June except in
1992, experiment. Cotton was irrigated as recommended for the region.
All experiments were conducted in a randomized complete block design
with four replications except the experiment in 1992 with three replications.
Cotton was harvested by hand in October and seed cotton was weighed. Data
were subjected to ANOVA and regression analysis.
RESULTS AND DISCUSSION
Observed mean cotton yield at weed free plots was 3281, 2506 and
3182 kg ha-1 in 1992, 1997 and 2002, respectively. Cotton yield
was significantly affected by johnsongrass densities. The effect of johnsongrass
depended on the effect of years. Due to the effect of years and differences
in plot sizes, each experiment again underwent to ANOVA and regression
analyses were run separately.
Data were fitted to equations Y = 1/(a+b*X8) in 1997 and 2002
for the number of johnsongrass in 8 m of row length and Y = 1/(a+ b*X4-1/2)
in 1992 and 1997 for the number of johnsongrass in 4 m of row length,
where Y is cotton yield (kg ha-1), a and b are constants and
X8 is the number of johnsongrass plants in 8 m cotton row and
X4 the number of johnsongrass in 4 m of row (Fig.
1 and 2). Parameter estimates were presented at Table
2. Ratkowsky (1990), stated that this model has been used in agricultural
research for yield-density studies and is known the reciprocal model.
Cotton yield was affected by all johnsongrass densities (Table
3). One johnsongrass plant in 8 m of cotton row decreased cotton yield
4.82 and 9.41% in 1997 and 2002, respectively. In our conditions crop
loss was higher in low densities.
| Table 2: |
Parameter estimates and regression coefficient for experiments |
 |
|
| Fig. 1: |
The effect of johnsongrass densities at 8 m row length
on cotton yield (solid line, predicted curve for 2002; dotted line,
predicted curve for 1997; ♦, measured cotton yields in 2002; ▲ , measured cotton yields in 1997) |
|
| Fig. 2: |
The effect of johnsongrass densities at 4 m row length
on cotton yield (solid line, predicted curve for 1992; dotted line,
predicted curve for 1997; ♦ , measured cotton yields in 1992; ▲ , measured cotton yields in 1997) |
Bridges and Chandler (1987) found 1% crop loss
when a johnsongrass plant competed in 9.6 m row length. However, similar
crop losses were obtained at higher densities with earlier studies. Estimated
yield loss for 32 johnsongrass shoots reached to 76.88 and 61.86% in 2002
and 1997, respectively.
| Table 3: |
Estimated crop losses due to johnsongrass densities,
which is calculated from predicted regression curves |
 |
The relative yield loss was higher in 1992 and 2002 comparing to 1997. In this region, over
90% of fields were infested by johnsongrass and johnsongrass mean density for
overall region after mechanical and herbicidal activities completed in fields
was still 2.91 plants m-2 (Uludag and Katkat, 1991). Cotton yield
loss corresponding to mean regional yield was estimated minimum 40%, which roughly
8 plants per 4 m or 16 plants per 8 m (Table 3). Recent surveys
in Kahramanmaras province had showed that 77% of fields were infested and overall
regional mean is 0. 65 plants m-2 (Gozcu and Uludag, 2006), where
16.86% yield loss would be in the region if there had not been control measures
for johnsongrass (corresponding density is in experiments is roughly 4 plants
per 8 m).
It is clear that johnsongrass causes huge crop losses. Current techniques
applied by farmers are not well enough to eliminate negative effect of
johnsongrass. In addition, world wide problem of herbicide resistance
requires new and more effective measures for johnsongrass management in
cotton.
Subscribe Today
