Damage to Cotton Fruiting Structures by the Fall Armyworm, Spodopterafrugiperda (Lepidoptera: Noctuidae)

Cotton at three stages of crop phenological development was infested with eggs and third instar larvae of the fall armyworm, Spodoptera frugiperda (J.E. Smith), to determine the effect of larval feeding on fruit damage and yield. Regression analyses indicated that numbers of damaged squares and bolls were significantly (P @ 0.05) influenced by the number of egg masses and third instars placed on the plants. More damage resulted with infestations of third instars than with eggs. Fruiting structures that were penetrated tended to have lower probabilities of survival to harvest than did non-penetrated or undamaged fruit. In a subsequent study with individual larvae confined for 48 h by cotton bags on fruiting structures, third instars damaged 0.63 squares, 0.72 small bolls, and 0.40 large bolls; fourth instars damaged 0.71 squares, 0.76 small bolls, and 0.63 large bolls; and fifth instars damaged 0.83 squares, 0.81 small bolls, and 0.66 large bolls. Damage to squares by all instars resulted in a significant (P @ 0.05) reduction in survival of fruit to harvest. Feeding on small bolls by fourth and fifth instar larvae, but not third instar larvae, resulted in significant (P @ 0.05) reductions in probability of harvest. Feeding on large bolls did not reduce the probability of survival of fruit to harvest. The fall armyworm prefer species in the grass family as hosts (Luginbill, 1928), but damaging populations can occur on other economically important plants, such as cotton and soybean (Bass, 1978; Pitre, 1979; Young, 1979; Pitre and Hogg, 1983). Studies have shown that one fall armyworm strain—the corn-cotton strain—prefers cotton as a host, while another strain prefers grasses (Pashley and Martin, 1987; Pashley et al., 1992). The fall armyworm is a sporadic pest that does not overwinter in most areas of the USA (Luginbill, 1950). Population densities vary tremendously from year to year and place to place, but damaging populations infrequently develop on cotton in the Midsouth. As corn acreage increases in traditional cottonproduction regions such as the Midsouth, there is concern that the fall armyworm will become a pest of increasing importance on cotton. Another concern, as it relates to fall armyworm, is the growing popularity of transgenic cottons, which express endotoxin proteins of Bacillus thuringiensis kurstaki. The fall armyworm is one of the least susceptible lepidopterans to the endotoxin proteins expressed in cotton (MacIntosh et al., 1990; Wan, 1994). Damaging densities of fall armyworm on these transgenic cottons will likely need to be treated with conventional insecticides. R. G. Luttrell, Department of Entomology and Plant Pathology, Mississippi Agricultural and Forestry Experiment Station, Box 9775 Mississippi State University, Mississippi State, Mississippi 39762; and J. S. Mink, Zeneca Ag. Products, Rt. 1, Box 65, Leland, MS 38756. Received 2 July 1998. Correspondence K. Knighten, Department of Entomology and Plant Pathology, Box 9775, Mississippi State, MS 39762, ([email protected]). 36 R. G. LUTTRELL AND J. S. MINK: DAMAGE TO COTTON FRUITING STRUCTURES The fall armyworm does not routinely infest cotton in the Midsouth, and little research has been conducted to quantitatively describe feeding and damage of this pest on cotton. Quantifiable economic thresholds are not available and control recommendations are based largely on subjective opinions and personal experiences. Several papers on the biology and ecology of fall armyworm on cotton have been published during the past decade. They included studies on the distribution of fall armyworm egg masses on cotton (Ali et al., 1989), survival of fall armyworm immatures on cotton (Ali and Luttrell, 1990), distribution of fall armyworm larvae within the cotton canopy (Ali et al., 1990a), and developmental rates of fall armyworm feeding on cotton (Ali et al., 1990b). Mink and Luttrell (1989) and Smith et al. (1993) have reported on commonly used insecticides for control of fall armyworm. Chandler (1995) reported on the potential use of insect growth regulators for Spodoptera spp. Earlier literature contains numerous reports of fall armyworm damaging cotton (Dew, 1913; Walton and Luginbill, 1916; Luginbill, 1928; Vickery, 1929; Pitre, 1979; Clower, 1984; Smith, 1985), but no quantitative information exists on the amount and type of damage caused by fall armyworm feeding on this crop. This type of information is available for fall armyworm feeding on corn and sorghum (Henderson et al., 1966; Morrill and Greene, 1974; Cruz and Turpin, 1983). A considerable amount of information can be found in the literature describing damage to cotton by other lepidopteran pests, particularly bollworm (Helicoverpa zea Boddie) and tobacco budworm (Heliothis virescens F.) (Adkisson et al., 1964; Kincade et al., 1967; Hartstack et al., 1978). These studies were conducted to measure the effects of feeding of fall armyworm larvae on damage and retention of cotton fruiting structures. The results should be useful to those interested in the development of economic thresholds (Stern et al., 1959) for this pest on cotton. MATERIALS AND METHODS Fall armyworm larvae were confined on individual cotton fruiting structures to measure fruit retention from different levels of feeding damage. All studies were conducted in a field of 'Stoneville 506' cotton planted during early May 1986 and 1987 and maintained under normal growing conditions on the Plant Science Research Farm at Mississippi State Univ., Mississippi State, MS. Presquaring cotton was thinned to 10 plants m plant density (74,100–98,800 plants ha). Additional thinning to 5 plants m (49,500 plants ha) was done 1 wk prior to insect infestation in order to facilitate plant handling during artificial infestations. Applications of the insecticides cypermethrin (0.32 kg [AI] ha) and azinphosmethyl (1.35 kg [AI] ha) were made before and after infestations when necessary to protect the cotton against bollworm, tobacco budworm, and boll weevils (Anthonomous grandis grandis Boheman). The study plots were sprayed with a non-persistent insecticide (mevinphos 1.35 kg [AI] ha) 1 d prior to infestation to reduce the number of predators present. Egg masses were obtained from a fall armyworm colony maintained at the USDA-ARS Crop Science Research Laboratory, Mississippi State, MS. This colony was initiated with insects collected from corn. Subsequent studies, that involved infesting cotton with fall armyworm larvae, were conducted with insects originally collected from cotton. This was to ensure that the corn-cotton strain of fall armyworm (Pashley and Martin, 1987; Pashley et al., 1992) was used in the studies. For all studies, insects were reared by the procedures described by Davis et al. (1985) and fed a wheat germ-casein diet (Davis, 1989). Infestation of Cotton with Egg Masses In 1986 egg masses as uniform in size as possible (45 ± 9 eggs mass) were obtained by placing a wire screen (0.32 × 0.32 cm) onto wax paper before oviposition and visually selecting the appropriate masses. Four egg mass densities (0 per plant, 3 per 10 plants, and 1 and 5 per plant) were applied by pinning the egg mass through the border of the wax paper to the midribs on the abaxial leaf surface in the middle portion of the plant. Cotton was infested at three stages of crop phenological development (squaring, flowering, and boll maturation). During the squaring stage, the fruit load on the plant averaged 12.8 squares, <1 bloom, and <1 boll per plant. At the flowering stage, the plants had an average of 14.6 squares, 1.9 blooms, 37 JOURNAL OF COTTON SCIENCE, Volume 3, Issue 2, 1999 and 9.1 bolls per plant. Fruit on plants at the boll maturation stage were mainly bolls (average of 1.9 squares, <1 bloom, and 16.0 bolls per plant). To allow the fall armyworm larvae to disperse and behave as if an entire field was infested at uniform densities, we also infested 10 plants on each side of the plots with egg masses. Plots contained five plants, were at least 1 m long, and were replicated 10 times. The plots were monitored weekly after infestation by making whole plant observations until no fall armyworm larvae were found. Information was recorded on fall armyworm density and sizes (estimated larval instar), plant damage, and fruit production. Plants were hand harvested during the last week of September, and the harvested seed cotton was weighed to determine yield differences among treatments. The experiment was conducted in a split plot design with whole plots and sub-plots arranged in a randomized complete block design. Our whole plots were the stages of plant phenological development; egg mass densities were the sub-plots. Data were analyzed using analysis of variance and regression analysis (MSTAT, 1986) at P @ 0.05. Infestation of Cotton with Larvae High mortality of egg masses from unknown causes occurred in our 1986 studies. Therefore, additional studies were conducted during 1987 with plants infested with third instar fall armyworm. During three phenological stages of plant development (squaring, flowering, and boll maturation), four densities of larvae (0, 2, 6, and 18 per plant) were placed on cotton-plant fruiting structures using low-tension forceps. Each phenological stage was treated as a separate study. The fruiting structures on the plants during the squaring stage averaged 27.4 squares, <1 bloom, and <1 boll per plant. During the flowering stage, an average of 19.2 squares, 2.9 flowers, and 12.9 bolls per plant were observed in the plots. Fruit on the plants at the boll maturation stage of crop development consisted mainly of bolls (averaging 1.9 squares, <1 bloom, and 16.0 bolls per plant). If a plant did not contain adequate fruiting structures in which to place the larvae (1 larva per fruiting structure), larvae were placed on leaves in a randomized complete block design. Each plot was 3 rows wide and 6 m long. All plants within each plot were infested, but observations were restricted to 10 consecutive plants in the center row. Infestation during the squaring stage was replicated seven times in natural uncaged environments. During the flowering stage, four replicates of natural uncaged environments and four replicates of caged environments using 6.1 × 6.1 × 2.1 m field cages (32-mesh screen) were infested. During the boll maturation stage, only four replications in caged environments were infested. Plants were mapped by recording the main stem and branch node position of all fruit prior to infestation, at 3 and 7 d post-infestation, and at harvest. Data were also recorded for each selected fruit class (small square = square with bract width <0.5 cm, medium square = square with bract width 0.6 cm to 1.2 cm, large square = square with bract width >1.2 cm, flowers, small bolls = boll diameter 1.5 cm to 2.5 cm and large bolls = boll diameter >2.5 cm) including the total number of fruit, damage to the fruit, fall armyworm larvae density, and type of fall armyworm injury. The types of damage caused by fall armyworm are described in detail by Mink (1988) and include “calyx feeding,” “penetration” (feeding had resulted in a hole in the corolla of squares or carpel wall of bolls, “bract plus calyx feeding,” “bract feeding plus penetration,” “calyx feeding plus penetration,” “bract plus calyx feeding plus penetration,” and “total destruction” (extensive internal feeding and hollowing out of fruiting structure). Mapping the plants prior to infestation provided sufficient information to follow all fruiting structures from infestation until harvest. Fruit survival to harvest also was recorded for each damage class. The plots were hand harvested on 1 and 8 October when about 75 and 95% of the total bolls were open in the uninfested plots, respectively. The data were analyzed using analysis of variance (ANOVA) and regression techniques (MSTAT, 1986). To determine the impact of the different types of fall armyworm injury on each cotton fruiting structure, the types of damage within fruiting structure classes were grouped across larval density treatments in each experiment, and the probability of harvest was determined. The types of damage were categorized into two groups, “vegetative” 38 R. G. LUTTRELL AND J. S. MINK: DAMAGE TO COTTON FRUITING STRUCTURES (feeding restricted to bract and/or calyx) and “penetrative” (feeding had resulted in a hole in the corolla of squares or carpel wall of bolls). The undamaged fruit within each age class were used as a standard with which to compare among the damage categories. An additional category, total damage (total number of fall armyworm damaged fruit within a fruit class), was included in the analysis for comparison purposes. These individual comparisons were performed using a K-sample binomial test for equal proportions (Marascuilo and McSweeney, 1975). This pair-wise comparison procedure adjusts alpha levels for each comparison to control experiment-wise Type I error (P @ 0.05). Caging Larvae on Individual Fruit Third, fourth, and fifth instar larvae were caged individually on cotton fruiting structures in 5.5 x 7.0 cm cotton cloth bags. Four separate groups of larvae were used in order to avoid removing a bag more than once in 24h. The observation periods for the four groups were 6 and 30, 12 and 36, 18 and 42, and 48h. Damage to young fruiting structures late in the growing season is generally considered to have little effect on yield. Thus, the age classes of fruit included in the present experiments were large squares (1 to 1.5 cm bract width), small bolls (ca. 1.5 to 2.5 cm in diameter), and large bolls (>2.5 cm in diameter). Larvae were placed on undamaged fruiting structures by transferring them from their artificial diet with low-tension forceps. Preliminary studies indicated that fall armyworm larvae transferred from artificial diet survived and damaged cotton equally as when they had been allowed to feed on cotton fruiting structures 48 h prior to infestation (Mink, 1988). After larvae were placed on the fruit, the cotton bags were placed around the fruiting structures and closed with a drawstring. A cloth laundry tag (2 x 4 cm) was attached around the stem of the fruiting structure to aid in post treatment observations. All infestations were made either early in the morning or late in the afternoon to reduce the effects of high temperature on survival of larvae. Separate experiments were conducted for each age class of fruit. All of the experiments were conducted in a randomized complete block design with three treatments (larval instar) and four replications. Each treatment included 25 individually caged larvae on plants in a 20-m section of row ( 200 plants per plot). A single plant generally had only one caged insect. All data were analyzed by ANOVA and means were separated by Duncan's (1955) multiple range test. Means were significantly different at P @ 0.05. The experiment with squares was conducted in July (14 and 25) when the plants were in the squaring stage and the fruit on the plant consisted of mainly squares and a few blooms. The experiment on bolls was conducted during the first week of September when the plants contained a few squares and flowers but mainly small and large bolls. Bags were removed at either 24 or 48 h to determine the status of the insect and to note the types of damage to the fruiting structure. Damage was categorized as vegetative, penetration, and total destruction. Another category, no damage, was also included. Observations were again made at harvest (13 and 14 October) by removing and counting the tags that were associated with harvestable cotton bolls when ca. 90% of the fruit in the untreated check (bagged fruit without an insect) were mature.