Coccinellids Associated with the Cotton Aphid (Homoptera: Aphididae) in Northeast Arkansas Cotton

The diverse beneficial insect fauna found in cotton, Gossypium hirsutum L., is well described (Whitcomb and Bell 1964, van den Bosch and Hagen 1966, Wells et al. 2001). The natural enemies commonly credited with controlling the cotton aphid, Aphis gossypii Glover, are the predaceous arthropods including the coccinellids, chrysopids, and syrphids, the parasitoids (braconids), and the entomogenous fungus, Neozygites fresenii (Nowakowski) Batko (Chambers 1986, Frazer 1988, Rosenheim and Cisneros 1994, Steinkraus et al. 1991). Coccinellid predators are commonly credited in providing cotton aphid regulation in several locations (Hodek and Honek 1996, Kidd and Rummel 1997, Wells et al. 2001). Adult and larval Hippodamia convergens Guerin, Coleomegilla maculata (Degeer), Coccinella septempunctata L., Harmonia axyridis (Pallas), Cycloneda munda (Say), Scymnus spp., and Diomus spp. are commonly collected from southern cotton fields (Knutson and Ruberson 1996).

Sampling techniques that provide accurate estimates of predator densities are critical for evaluation studies on natural enemies and must provide precise density estimates while minimizing costs (Frazer 1988, Obrycki and Kring 1998). The drop cloth is one of the most efficient methods for sampling coccinellids (Wade et al. 2006), although in warm weather the adults quickly fly away after falling on the cloth, thus making identification difficult (Michels and Behle 1992, Hagerty et al. 2002). The beat-pan method is another effective sampling method that allows additional time for species identification because adult natural enemies often land on the screen cover before flying away. This method is a modification to the “thrips-catcher” (Newsom et al. 1953) used by research entomologists to sample insects in cotton. Insects are dislodged by striking the plants onto a wire mesh covering a light-colored pan (Elkassabany 1994).

The annual development of insect communities within agricultural crops is influenced by the species pool composition of potential colonists and by habitat development within the field (Liss et al. 1986). For coccinellids, the species pool of potential colonists is influenced by several factors, including availability of adequate overwintering habitat, the population densities of coccinellids entering diapause, and the extent of overwintering mortality (Honek 1986). Predation of aphids by these coccinellids is potentially important in maintaining cotton aphid densities below the treatment levels during early stages of cotton development (Kerns and Gaylor 1993, Knutson and Ruberson 1996, Wells et al. 2001, Conway et al. 2006).

Our goal in this study was to characterize the development of coccinellid communities in cotton both within and between the growing seasons. The objectives were to identify trends in the seasonal patterns of population density and occurrence of coccinellid adults and larvae inhabiting cotton fields, to establish the predictability of development of the coccinellid communities, and to quantify the relative abundance of individual species.

Materials and Methods

A 3-yr (1999 - 2001) field study was conducted at the University of Arkansas Delta Research Station at Clarkedale, AR. Sixteen plots, each ~0.3 ha, were planted with Gossypium hirsutum L. (NuCott 33B, Delta and Pine Land Company, Scott, MS 38772) on 4 May 1999, 2 May 2000, and 28 April 2001 in ~1.0-m wide rows. Cotton was grown under standard cultural practices including preplant herbicides, in-furrow aldicarb application (targeting thrips and nematodes), fertilization as required, cultivation ~3 - 4 times yearly, insecticide applications after reaching plant bug or boll weevil thresholds, in-furrow irrigation, monthly manual removal of in-field weeds, and biweekly mowing of field edges.

Data were obtained from a cotton aphid threshold experiment with 4 replicates applied to cotton fields using 3 treatments: (1) untreated, (2) conventional treatment threshold, and (3) experimental treatment threshold (Conway et al. 2006). Conventional plots reached aphid treatment thresholds and were treated by applications of 0.22 L / ha of imidacloprid on 18 and 28 June 1999, 28 June and 3 July 2000, and 7 and 12 July 2001. Experimental threshold plots reached aphid treatment thresholds and were treated with 0.22 L / ha of imidacloprid on 28 June 1999, 3 July 2000, and 19 July 2001. In 2001, there were treatments of malathion for the boll weevil eradication program applied to all test plots in Clarkedale, AR, on 5 and 15 June, and on 3, 11, 18, and 24 July.

Natural enemies were sampled 2× per wk using the beat-pan method whereby plants were struck onto a hardware mesh (4 cells / cm²) covering a white, plastic wash basin (35.5 × 28.5 × 13.5 cm). Population density levels of beneficial insects were estimated by collecting 8, 1 row-m samples per plot with 128 samples (64 from untreated control, 64 from the treated plots) taken on each sample date. In 2000 and 2001, larval coccinellids were collected from the beat-pan samples for later identification and placed into vials of ethanol marked with date, plot number and treatment. Sampling began each year in early June and continued until late July in 1999 and 2000, and early August in 2001. Total adult and larval coccinellids were enumerated each year, and Coccinellinae were identified to species in 2000 and 2001 (Gordon and Vandenberg 1991, Chapin and Brou 1991, Rees et al. 1994). Adult Scymninae were grouped together in counts, and larval Scymninae were identified as either Scymnus spp. or Diomus spp.

Principal components analysis was conducted on the mean yearly values for cotton aphids, adult and larval Scymninae, and adult and larval Coccinellinae from each of the 16 plots each year (48 total per factor) over the 3-yr period of the study (SAS Institute 2006). Multivariate analysis was conducted for Hippodamia convergens Guerin, Scymninae (Scymnus spp. and Diomus spp.), Coleomegilla maculata (Degeer), Harmonia axyridis (Pallas), and Coccinellinae larvae and adults per plot (16 plots) over years (2 years) with Kendall's tau coefficients of association for pairwise combinations (SAS Institute 2006).

Results and Discussion

A total of 3634, 3562, and 1644 beneficial insects were collected from the cotton fields in 1999, 2000, and 2001, respectively. The dominant predators of cotton aphids from these samples were coccinellid adults and larvae, which together composed 79%, 68%, and 53% of the captured beneficial insects in 1999, 2000, and 2001, respectively. Coccinellids also were the most abundant arthropod natural enemies identified from cotton in Georgia (Abney et al. 2008) and Mississippi (Weathersbee and Hardee 1994).

The coccinellid species observed in Arkansas during the 3-yr study (in descending order of abundance) were: H. convergens, Scymninae (Scymnus spp. and Diomus spp.), C. maculata, H. axyridis, and C. septempunctata L. (Table 1). Wells et al. (2001) reported a similar composition of coccinellid species in Georgia cotton during a 2-yr (1997 - 1998) field study, with H. convergens and Scymninae (Scymnus spp. and Diomus spp.) being the most abundant coccinellids collected. Hippodamia convergens and C. maculata were the most abundant coccinellids found in a study in Oklahoma cotton (French et al. 2001).

Table 1. Total number of adult and larval coccinellids by species or genera in cotton in Clarkedale, AR (1999 - 2001).

View Fullscreen

In our study, there were 3 - 4 times more adult and 9 times more larval H. convergens than any other coccinellid species in 2000 and 2001 (Table 1). An overall decrease in abundance of adult and larval Scymninae and of larval Coccinellinae occurred over the 3-yr study period, likely due to the yearly decrease in aphid densities (Fig. 1). The number of Coccinellinae adults fluctuated less predictably among years with 2000 having a higher abundance of adults than either 1999 or 2001 (Table 1).

View Figure

• Download as Powerpoint
• Download Figure

Adult and larval densities of each species fluctuated widely within the field and among the years. In 1999, beneficial insect sampling was conducted for 6 wks from 17 June to 19 July. During that time, Coccinellinae densities increased over time peaking the first full week of July at (mean ± SE, 1.01 ± 0.12) adults per row-m. Adult Scymninae densities increased through time, peaking at (1.30 ± 0.17) adults per row-m during the second week of July (Fig. 2). Similarly, Coccinellinae and Scymninae larval densities increased over time peaking on the first full week of July at (3.45 ± 0.41) and (1.05 ±0.19) larvae per row-m, respectively, before densities declined after an epizootic of N. fresenii in cotton aphids (Fig. 3).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

In 2000, beneficial insect sampling was conducted for 9 wks from 29 May to 24 July. Adult coccinellid densities fluctuated at low numbers during the first 5 wks until late June when Scymninae densities peaked at (0.19 ± 0.04) adults per row-m during the first full week of July. In the second week of July, H. axyridis adults peaked at (0.04 ± 0.02) adults per row-m, and H. convergens and C. maculata peaked at (1.76 ±0.14) and (0.14 ± 0.03) adults per row-m, respectively, in the third week of July (Fig. 2). No larval coccinellids were collected until the third week of sampling. Then, the densities increased steadily, peaking the first week of July for H. axyridis, C. maculate, Diomus spp., C. septempunctata, and Scymnus spp. (Fig. 3). Hippodamia convergens larvae were the dominant coccinellid species peaking at (1.81 ± 0.23) larvae per row-m during the second week of July then decreasing after an epizootic of N. fresenii (Fig. 3).

In 2001, beneficial insect sampling was conducted for 10 wks from 6 June to 6 August. Scymninae and H. convergens were the only adult coccinellids regularly collected during the entire sampling periods, although at much lower densities than the previous 2 seasons. Scymninae densities increased through time, peaking the first week of July at (0.19 ± 0.04) adults per row-m. Hippodamia convergens numbers fluctuated with 2 main peaks at (0.19 ± 0.05) in late June and (0.39 ± 0.06) adults per row-m in early August. Adult C. maculata were collected on 3 and 19 July and 1 H. axyridis was collected on 6 June (Fig. 2). No larval coccinellids were collected until the fourth week of sampling. Diomus spp. and H. convergens were the only larval species collected on a regular basis. Diomus spp. were collected in July through the beginning of August, peaking at (0.05 ± 0.02) larvae per row-m at the end of July (Fig. 3). Larval H. convergens densities fluctuated throughout the sampling period with 3 less-discernable peaks at 0.34, 0.29 and 0.46 larvae per row-m (Fig. 3). Larval C. maculata were collected on 3, 5, and 31 July and on 3 August, whereas 1 larval Scymnus spp. was collected on 3 July. The boll weevil eradication program applications of malathion in the area clearly contributed to the low aphid and predator numbers, as well as to the great fluctuations (multiple peaks) in predator densities (Conway et al. 2006).

Coccinellid densities were variable from year to year and between locations within the study field, similar to studies in Georgia cotton (Wells et al. 2001, Abney et al. 2008) and South Dakota small grains (Elliott and Kieckhefer 1990, Kieckhefer and Elliott 1990). As a group, coccinellids were present in the cotton field throughout most of each growing season, although we collected the coccinellid abundance data only during the early season when aphid outbreaks were likely to occur. Hippodamia convergens and Scymninae adults were frequently present at low but detectable levels during the early dicotyledon and 2-leaf stages (Fig. 2).

Two of the most important mortality factors in coccinellid populations are cannibalism and interspecific predation (Schellhorn and Andow 1999). Coccinellid larvae are often observed preying on conspecific and heterospecific eggs, larvae, prepupae and pupae. Older instars kill younger instars (Agarwala and Dixon 1993), and late instars prey upon prepupae and pupae (Hodek and Honek 1996). Each year, when aphid densities declined after an epizootic of N. fresenii, coccinellids mainly had the choice of N. fresenii-infected prey (Simelane et al. 2007) or other coccinellids. Intraspecific and interspecific predation was observed by coccinellid larvae on other larvae, prepuae, and pupae. In 1999 and 2000, there may have been more cannibalism and interspecific predation due to the lower overall aphid densities whereas coccinellid densities remained high (Fig. 1). Overall, interguild predation by coccinellids is an important ecological interaction that in early-season cotton agrosystems does not seem to disrupt cotton aphid biological control (Colfer and Rosenheim 2001). In fact, interguild predation likely allows remnant coccinellids populations to subsist in the field and later attack other prey or disperse to other habitats.

Densities of each coccinellid species or group within a plot appeared to fluctuate independently of one another. This observation was tested by Kendall's tau coefficients of association for pairwise combinations of species in each plot for H. convergens, Scymninae (Scymnus spp. and Diomus spp.), C. maculata, H. axyridis, Coccinellinae larvae, and Coccinellinae adults per plot (16 plots) over years (2 yrs). Significant differences (P < 0.01) occurred in 14 of 15 pairwise comparisons, indicating that species fluctuated independently of each other except Scymnus spp. with H. axyridis adults (Kendall's tau = 0.269, dF = 32, P < 0.06). This result was similar to observations of Elliott and Kieckhefer (1990) observations in small grains.

When a prey item is the primary food source for a predator, a density-dependent relationship will often develop (Krebs 1985). In Georgia, density fluctuations of the dominant coccinellid species H. convergens and Scymnus spp. are closely associated with fluctuations in the cotton aphid density, where coccinellid densities increased in response to an increase in cotton aphid density (Wells et al. 2001). In our study, density fluctuations of the most abundant coccinellid species, as well as fluctuations in the total coccinellid densities, appeared closely associated with the fluctuations in cotton aphid density (Fig. 1). As cotton aphid densities increased, a delayed density-dependent relationship occurred with an increase in larval coccinellid densities. Adult coccinellid densities subsequently increased as larvae pupated and new adults emerged.

As depicted, the first two principal components provided statistically significant information, accounting for 95.35% of the multivariate correlations (Table 2). Principal component 1 represented 78.59% of the multivariate information and indicated strong correlations among cotton aphid densities, Scymninae densities, and Coccinellinae larvae densities (Table 2). Pairwise correlations indicated 92.6% correlation between aphid densities and Coccinellinae larval densities, 92.2% correlation of aphid densities with larval Scymninae densities, and 90.0% correlation of aphid densities with adult Scymninae densities. Principal component 2 represented 16.76% of the multivariate information and indicated a correlation of Coccinellinae adult with larval densities (Table 2). Pairwise correlations indicated 55.2% correlation of adult with larval Coccinellinae.

Table 2. First two principal components derived from multivariate analysis of 4 variables describing correlations of population density with temporal patterns of occurrence of coccinellids and cotton aphid in cotton at Clarkedale, AR.

View Fullscreen

Based on population densities, timing of colonization and consistent delayed density-dependent relationship with aphid densities, H. convergens and Scymninae (Scymnus spp. and Diomus spp.) were the most important cotton aphid predators in northeast Arkansas cotton fields. These coccinellids play an important role in maintaining cotton aphid densities at low levels early in the growing season when the cotton plants are more susceptible to damage. Coccinellid predators have a significant impact on aphids in general and often play a major role in cotton aphid regulation (Hodek and Honek 1996, Kidd and Rummel 1997, Wells et al. 2001). Following this research, AR enacted a new treatment threshold for the cotton aphid explicitly incorporating density estimates of adult and larval Coccinellinae in the decision-making process (Greene 2006).