Arthropod Predators of Velvetbean Caterpillar., Anticarsia gemmatalis Hiibner (Lepidoptera: Noctuidae)., Eggs and Larvae KRIS E. GODFREY,' WILLARD H. WHITCOMB, AND JERRY L. STIMAC2 Department of Entomology and Nematology, University of Florida, Gainesville, Florida 32611 Environ. Entomol. 18(1): 118-123 (1989) ABSTRACT Arthropod predators of velvetbean caterpillar, Anticarsia gemmatalis Hiib- ner, eggs and larvae were determined using a direct observation technique. Observations of predation events were made on soybean foliage and the ground below the foliage. Predation on eggs and three size classes of larvae is presented as proportional seasonal consumption over 14 to 16 h by each arthropod predator species. On the foliage, Spanogonicus albofas- ciatus (Reuter) (Hemiptera: Miridae) and Geocoris punctipes (Say) (Hemiptera: Lygaeidae) were observed eating velvetbean caterpillar eggs in 1981. For small velvetbean caterpillar larvae, Tropiconabis capsiformis (Germar) (Hemiptera: Nabidae) was the major consumer in 1981, whereas in 1982, G. punctipes and Calleida decora (F.) (Coleoptera: Carabidae) Downloaded from were the major consumers. For medium and large velvetbean caterpillar larvae in 1981, tettigoniids (Orthoptera) consumed the largest proportion. On the ground, the ant Pheidole morrisi Forel (Hymenoptera: Formicidae) was the major consumer of all three size classes of velvetbean caterpillar larvae in 1981. In 1982, the earwig Lab/dura riparia (Pallas) (Der- maptera: Labiduridae) ate the greatest proportion of small velvetbean caterpillar larvae on the ground. http://ee.oxfordjournals.org/ KEY WORDS Insecta, Arachnida, predation PREDICTING THE EFFECT of arthropod predators be used to partition estimates of predator-induced on pest populations is a prerequisite to effective mortality for each predator species (Elvin et al. use of arthropod predators in pest management 1983). Only arthropod predator species that are (An- programs. Such prediction assumes that rates of responsible for killing velvetbean caterpillar by guest on June 7, 2016 arthropod predation can be estimated quantita- ticarsia gemmatalis Hubner) eggs and larvae in tively and integrated with other aspects of pest soybeans in 1981 and 1982 are presented in this population dynamics. Population models can serve paper. as the integrative tools (Stimac 1981). The initial step in the construction of a predator- Materials and Methods prey model for multiple species of predators re- Arthropod predation was observed on velvet- quires estimates of the stage-specific rates of pre- bean caterpillar (VBC) eggs and small (first and dation mortality and identification of the predator second instar; length <12.5 mm), medium (third species responsible for the mortality (Stimac & and fourth instar; length, 12.5-25 mm), and large O'Neil 1983). To obtain this information, Elvin et (fifth and sixth instar; length, >25 mm) larvae in al. (1983) used a two-step method that combined 1981. Arthropod predation was observed on small mark-release-recapture of prey and direct obser- VBC larvae in 1982. The observations during both vation techniques. We used the mark-release-re- years were made in a I-ha plot of 'Bragg' variety capture of prey technique to estimate stage-specific soybeans, Glycine max (L.) Merrill, located at Green mortality. This technique provided the informa- Acres Agronomy Farm near Gainesville, Fla. Soy- tion required to partition estimates of total mor- beans were planted on a 76 cm row spacing on 12 tality into predator-induced and non-predator-in- June 1981 and 9 June 1982. Observations were duced components. The direct observation made once a week from 7 July through 22 Septem- technique is used to identify the arthropod pred- ber in 1981 and once a week from 6 July through ator species responsible for predation mortality. 14 September in 1982. The time during the day When sufficient observations are made, the obser- the observations were made will be discussed later. vation of predation events, defined as the successful On each day that observations were made in 1981 capture of.one prey by an arthropod predator, can and 1982, upper, middle, and lower canopy tem- peratures, and soil temperatures (about 12.5 cm , Current address: 8011 Hickory Lane, Lincoln, Nebr. 68510. below the soil surface) were measured. , Address reprint requests to J. L. Stimac. Observations of predation events in both years 0046-225x/89j01l8-0123$02.00jO © 1989 Entomological Society of America February 1989 GODFREYET AL.: ARTHROPODPREDATORSOF VBC 119 were made using a technique modified from that tions was made from 0600 to 2000 hours. The short- proposed by Whitcomb (1967). Observations were er time interval in 1982 was chosen based on the made of predation on the foliage and on the ground results of the observation studies conducted in 1980 beneath the foliage. Predation was observed at 20 (Elvin et al. 1983) and 1981. foliage stations and four ground stations arranged The results are presented as seasonal proportions in a circular configuration across the field. of predation by each predator species over the 14- Foliage observation stations consisted of three 16-h observation period. The proportions were plants located in 91 cm of row. Known densities computed by taking the ratio of the number of of VBC eggs or larvae were released on three leaves, times a given predator species (or genus or family) one leaf on each plant, which was marked with was observed consuming a prey during all obser- flagging tape tied around its rachis. Only one life vation periods within the season to the total number stage of VBC was placed at a station. An observer of predation events by all predator species over the moved from station to station and recorded all ob- entire season. served predation events. Predators consuming prey were identified to the lowest appropriate taxon in Results and Discussion the field, or the predator was collected and iden- The direct observation technique used on the tified in the laboratory. Before leaving a station, foliage seemed to work well for small and medium an observer would replace any missing eggs or lar- larvae, but we had problems observing predation vae to maintain the number placed at a station. on eggs and large larvae. Of the 35 total predation The density of VBC eggs and larvae placed at records on the foliage, 30 were of predators preying Downloaded from a station in 1981 was varied over the season in an upon small and medium larvae (Table 1). The large attempt to provide more realistic conditions for number of records for these two stages was prob- observing predation. The number of eggs, small ably due to the amount of time necessary for a larvae, and medium larvae placed at a station was predator to capture and consume a prey. The pred- five or ten individuals, and for the large larvae, ators of eggs and large larvae can either quickly three or five individuals. The minimum density of consume their prey or they capture and carry the http://ee.oxfordjournals.org/ VBC eggs and larvae was placed at foliage stations prey to another location to eat it. Some of the 011 7 July and 14 July. On 21 July and 28 July, ten predators of small and medium larvae, on the other of the foliage stations were run at maximum den- hand, remain at the site of capture for a long time sity and ten foliage stations were run at minimum to eat a prey. This increases the probability of ob- density. All foliage stations were run at maximum serving a capture on small and medium VBC lar- density from 4 August through 22 September. In vae. 1982, the density of larvae placed at a foliage sta- The small number of observation records for tion did not vary over the season, but remained large larvae also may be due to the limited number by guest on June 7, 2016 constant at ten individuals per station. of predator species that can capture large larvae, Ground observation stations for both years con- because of their size and their violent, twisting, and sisted of the area between two rows and extended jumping defensive behavior. Small and medium 91 cm down the row. Upon arrival at a ground larvae also use this defensive behavior, but because station, an observer placed three VBC larvae on of their smaller size, they can be more easily cap- the ground. In 1981, one small, one medium, and tured. one large larva were placed on the ground- at a The total number of predation events on the station and observed for 10 min during the daytime foliage for eggs in 1981 was four with the largest observation period (0800 to 1800 hours EDST) and proportion of mortality attributable to unknown for 5 min during the night observation period (1800 causes (Table 1). Two predators, the mirid Spano- to 0700 hours). In 1982, three small larvae were gonicus albofasciatus (Reuter), and the Iygaeid placed on the ground at a station and observed for Geocoris punctipes (Say), were observed preying 10 min. During daylight observations, all larvae upon eggs (Table 1). These predators have been were placed in the shade of the soybean foliage in reported previously as predators of the eggs of VBC an attempt to reduce heat-induced mortality. All and other noctuid species in soybeans (Neal et al. predation events were recorded. A ground pre- 1972, Buschman et al. 1977, McCarty et at. 1980). dation event was defined as the successful capture The total numbers of observations of predation of one prey by one or more predators. Unsuccessful events on the foliage for small VBC larvae were predation attempts (i.e., a larva was attacked, but 11 in 1981 and 32 in 1982 (Table 1). In 1981, not captured) also were recorded. Predators prey- Tropiconabis capsiformis (Germar) (a nabid) adults ing upon the larvae were identified in the same consumed the largest proportion of small larvae, manner as for the observations of predations on whereas in 1982, G.