Internationally, sticky cotton is a major concern for the textile industry (Hector & Hodkinson 1989). Physiological plant sugars in immature fibres, contaminants from crushed seed and seed coat fragments, grease, oil and pesticide residues are all potential sources. However, all are insignificant compared with honeydew contamination from Bemisia tabaci and Aphis gossypii (Hector & Hodkinson 1989; Ellsworthy et al. 1999a; Hequet & Abidi 2002). The underlying reason is the distribution of the sugars along the fibre. Physiological sugars, grease and oil are usually distributed evenly over the fibre whereas sugars from honeydew tend to be in scattered concentrations (Bruno 1984). In the latter, these sugars lead to uneven yarn which is prone to breaking during weaving and knitting of fabrics (Hequet & Abidi 2002) as well as impeding fibre handling and causing in severe circumstances mill shutdown to clean equipment (Ellsworthy et al. 1999a). A reputation for stickiness has a negative impact on sales, exports and price for cotton from regions suspected of having stickiness. Reductions in the market value of lint due to stickiness are applied regionally and indiscriminately. In Arizona, perceptions regarding stickiness lead to a -5.63c/lb discount relative to Californian cotton (Ellsworthy et al. 1999a).

More than 20 different sugars are excreted in honeydew (Hendrix & Wei 1994) and most are insect rather than plant derived (Tarczynski et al. 1992; Salvucci et al. 1997). The major sugars excreted by A. gossypii are melezitose, sucrose, glucose and fructose while for B. tabaci, there is the additional sugar, trehalulose. Analysis by Hendrix et. al. (1992) of aphid and silverleaf whitefly honeydew from insects feeding on cotton indicated around 40% of total sugars present was melezitose in the aphid honeydew, while silverleaf whitefly honeydew exhibited about 40% trehalulose plus about 17% melezitose.

The two sugars that contribute most to cotton stickiness problems are trehalulose and melezitose (Henneberry et al. 1995, 1996, 1998a, 1998b; Gamble 2001) and both are produced as a result of transglycosylation reactions involving dietary sucrose (Wei et al. 1997). The lint content of both trehalulose and melezitose were quantitatively linked to insect numbers. The composition of honeydew collected directly from Bemisia was found to be virtually identical to that recovered from contaminated lint (Hendrix 1995) and poinsettia (Byrne & Miller 1990). For B. tabaci first and second instars produce less trehalulose than third and fourth instars while adults produce more than nymphs. In contrast, more melezitose was produced by nymphs than adults (Henneberry et al. 1999). While trehalulose and melezitose are significant contributors to sticky cotton the interaction between these, other sugars and cotton stickiness is poorly understood and complicated by the fact that other sugars such as sucrose glucose and fructose occur in both honeydew and cotton lint (Henneberry et al. 1998a).

Using the thermodetector method to measure cotton stickiness (Henneberry et al. 2000) increases in thermodetector sticky cotton counts were closely correlated with increasing numbers of whitefly nymphs and adults. The experience from the USA suggests that the threshold of concern for sticky cotton is indicated by thermodetector measurement ≥ 5 (Frydrych 1986; Brushwood & Perkins 1993). To reach this level, whitefly numbers needed to be ≥ 8.9 adults per leaf or ≥ 3.2 nymphs/cm2 (Henneberry et al. 1998a). This is well below the action thresholds for insect growth regulators (Ellsworthy et al. 1999b). Further, Yee et al. (1997) and Henneberry et al. (1998) indicated that insecticide applications at an average of 10 adults per leaf reduced honeydew production as effectively as applications at 5 adults per leaf. However, standardised sampling protocols, and the relationship between levels of stickiness in the field and problems arising in textile mills are still not well understood (Henneberry et al. 1998a). This is apparent if one plots the data from the USA against thermodetector readings (Henneberry et al. 1998a, 1998b, 2000) (Fig 1). The considerable scatter associated with the correlation underlines the lack of a clear cut decision point with regards to concentrations of trehalulose and melezitose and stickiness.