MATERIALS AND METHODS

This study was conducted at three cotton field’s located Harran Plain in 2003-2004. Three delta pheromone traps (one trap for each field) were placed for determination of adult population development and traps were re-baited every three weeks. The counting of adults in traps was conducted weekly. In addition, 100 bolls of cotton were picked randomly from 100 cotton plants weekly. The collected bolls were split in the laboratory and examined whether they were infected with PBW.

The relation between variables was modeled with analyze of regression as referred by Draper and Smith (1998) using SPSS statistical software (SPSS, 1999). In addition, R² for goodness of fit was counted as:

(1)

Correlation co-efficient (r) for variables was obtained by Pearson correlation coefficient as:

(2)

RESULTS

Relationships between adult number and boll infestation were observed simultaneously. That is, when the adult number was increased, boll infestations were also increased. Pearson correlation co-efficient between adult number catches in trap and boll infestation to determine simultaneous variation was observed.

Correlation co-efficients in all three localities were very important in the level of p<0.01. R-values for three locations (Kulunce, Cavdarli and Huzurlu) were determined as 0.87, 0.88 and 0.88, respectively. A significant positive increase between adult number catches in traps and boll infestation was found.

Adult number catches in trap (x) is independent variable and boll infestation (y) is dependent variable. To model relation between these two variables, scattered diagram was applied. In order to determine the relationship between variables linear regression model (y = a+bx) was fitted. R²-values for locations (Kulunce, Cavdarli and Huzurlu) were obtained as 0.75, 0.78 and 0.78, respectively. The significance levels concerning regression variance analyzes were found to be very important for three localities (p<0.01).

Fig. 1:	The graphic of combined regression of all years and localities

The relationships adult number catches in trap and boll infestation were linear in all localities. So, the relation between variables to put forward with only one model, the data obtain from all localities were combined and regression graphic was given in below (Fig. 1).

As seen in the Fig. 1, linear relationship among variables exits. R² in combined data was 0.742 and the result was significant (p<0.01). The correlation co-efficient (r) among variables was 0.86 and very significant positive relation was revealed in the level of p<0.01.

DISCUSSION

PBW has been caused harmful on cotton in Harran Plain since 1997 (Ozpinar et al., 1998). It has still continued damage and importance in Harran Plain (Unlu et al., 2006; Unlu, 2006, 2007). The growers in the Plain do not apply insecticide for Pink Bollworm. In addition, they are not aware of the cultural practices such as using delinte seed and destruction of plant remains.

This study was conducted to determine the relationships adult number catches in trap and boll infestation of Pink Bollworm. In the study, significant correlation among variables was found. In all years and localities, both one by one and combined R² were also very significant. In Greece, there was a significant positive linear correlation between the number of P. gossypiella moth catches and infestation percentage from first and second stage larvae of PBW and the former was lower than the latter and were R = 0.694 and R = 0.739, respectively (Buchelos et al., 1999). Henneberry and Clayton (1982) also reported similar results between catches and larvae per boll (R = 0.82) and the percentage of boll infestation with larvae per boll (R = 0.91). Campion (1994) showed that, there was a positive correlation between the numbers of male adults caught within 3-4 days prior to the first cotton flower bud invasion by PBW and the numbers of larvae in bolls. A study was conducted in Amik Plain (Turkey) in 1996-1999, the damage level in 1997 was significantly related to the number of adults trapped (R = 0.799) (Mart et al., 2002).

Attique et al. (2001) reported that, chemical control of this pest is difficult, so emphasis is made on timely and effective cultural practices to reduce the over-wintering population. Unlu and Bilgic (2004) reported that, a 1% increase in infestation would reduce almost 2.5-6% of cotton yields. We recommend that, cotton growers must apply cultural practices and biotechnical control as to as control of Pink Bollworm, such as mating disruption and pheromone traps. In this way, they apply efficiently and economic methods to control of the pest.