Mortality Factors Affecting Whitefly Populations in Arizona Cotton Management Systems: Life Table Analysis

Direct-observation studies were conducted in replicated experimental plots to identify causes and estimate rates of mortality of whiteflies in cotton over the course of six generations from late June through late October. In plots receiving no whitefly or Lygus insecticides, predation and dislodgment were major sources of egg and nymphal mortality, and overall survival from egg to adult ranged from 0-18.2%. Similar patterns were observed in plots treated with the insect growth regulator (IGR) Knack. Applications of the IGR Applaud or a mixture of endosulfan and Ovasyn caused high levels of small nymph mortality and reduced rates of predation on nymphs during the generation immediately following single applications of these materials in early August. Whitefly populations declined to very low levels by mid-August in all plots, and few differences were observed in patterns of whitefly mortality among treated and control plots 4-6 weeks after application. The population crash was associated with an unknown nymphal mortality factor which reduced immature survivorship during this first posttreatment generation to zero. The application of insecticides for control of Lygus in subplots modified patterns of mortality in all whitefly treatments by generally reducing mortality from predation during generations observed from mid-July through August. Parasitism was a very minor source of mortality throughout and was unaffected by whitefly or Lygus insecticides. Introduction Many biotic and abiotic mortality factors impact the population dynamics of Bemisia tabaci (Biotype B) in agricultural ecosystems, yet we have a poor understanding of the rates of these mortality factors and how they may be involved in overall population regulation. For a multivoltine and multi-crop pest like B. tabaci estimating rates of mortality in the field is extremely complex and difficult. The effects of various conventional insecticides are generally well known; however, studying the effects of such factors as predation and parasitism, or even insect growth regulator insecticides, is much more difficult. This task is made even harder because of overlapping generations of whiteflies in the field and because pest management activities provide further sources of mortality that may enhance or disrupt natural enemies. Life table analysis (Deevey 1947) is the most direct and robust method for estimating the sources and rates of mortality affecting a population. Life table studies categorize the sources of mortality and provide a means to quantify rates of death from various factors over the course of a generation. Over the past three years we have been conducting studies to demonstrate and evaluate different strategies for whitefly management in Arizona cotton (Ellsworth et al. 1998, Ellsworth and Naranjo 1999, Naranjo and Hagler 1997, Naranjo et al. 1998a,b). These multi-component, commercial and quasi-commercial scale experiments have evaluated sampling methodologies, thresholds for insecticide application, methods of application, insecticide resistance, and natural enemy conservation. An integral part of these studies has also been the development of life tables which have allowed us to categorize and quantify sources and rates of whitefly mortality relative to different management systems. Our previous results suggest that the use of the insect growth regulators (IGR) Applaud_ (buprofezin) and Knack_ (pyriproxyfen) for whitefly control helps conserve native predators and whitefly parasitoids (Naranjo and Hagler 1997, Naranjo et al. 1998a, b). Here we present preliminary results of a second year of replicated life table studies. Materials and Methods Studies were conducted at the University of Arizona, Maricopa Agricultural Center in Maricopa, AZ. The study was conducted using NuCOTN 33B and contrasted four whitefly control regimes; Applaud used first, Knack used first, a rotation of conventional insecticides (1995-IRM), and an untreated control. The threshold for use of IGR treatment was 1 large nymph/disk plus 3-5 adults/leaf (Ellsworth et al. 1996). All conventional insecticide applications were made at 5 adults/leaf (Ellsworth et al. 1995). In addition to applications for whitefly, each treatment plot was split, and each half either received sprays for Lygus hesperus or were left untreated for Lygus. Insecticides for Lygus were applied as needed based on a threshold of 15-20 Lygus (adults + nymphs, with nymphs present) per 100 sweeps (Ellsworth 1998, Ellsworth and Diehl 1998). All applications were made by ground, and seasonal usage of insecticides for the studies described here is summarized in Table 1. Each insecticide regime was replicated 4 times using a randomized block, split-plot design in a total area of about 9 acres. Additional detail on the entire experiment is provided in Ellsworth and Naranjo (1999). Cohorts of newly laid eggs and settled 1st instar nymphs were established over the course of 2 pre-spray and 4 postspray generations from late June through late October in the replicated experimental plots described above. Cohorts consisted of at least 50 individuals of each stage in each plot. The location of each individual was marked on leaves with a non-toxic felt-tip pen. Each stage was then examined every 2-3 days directly in the field with the aid of a 15x hand lens. Mortality was recorded as due to insecticides, predators, parasitoids, inviability (eggs only), unknown and missing. The unknown category simply catalogs mortality that could not be attributed to one of the other causes of death and probably reflects physiological or natural mortality. The missing category represents mortality associated with weather (wind and/or rain) or chewing predation. In this preliminary report we present only apparent rates of mortality observed for each source. Future analyses will involve the estimation of marginal rates of mortality which corrects for the effects of contemporaneous mortality by multiple factors (e.g. predation, parasitism and insecticides) and allows for more robust statistical comparisons among treatments. The particular treatments in which life table studies were conducted varied over the season. The first cohorts were established in control plots prior to the use of any whitefly or Lygus insecticides when whitefly abundance was extremely low. The second, fourth and sixth sets of cohorts contrasted control plots that were sprayed or not sprayed with Lygus insecticides. The third cohort contrasted the effects of whitefly insecticides and their interaction with Lygus sprays, and the fifth cohort contrasted the effects of whitefly insecticides alone.