Foreign exploration for natural enemies of pink hibiscus mealybug, Maconellicoccus hirsutus, was conducted in Australia from 2000 to 2002. In Queensland, the predaceous beetle Cryptolaemus montrouzieri, the predaceous drosophilid fly, Cacoxenus perspicax and the encrytid parasitoid Gyranusoidea indica were recovered. In Western Australia and the Northern Territory a predatory noctuid, Mataeomera sp., an aphelinid parasitoid Coccophagus sp., and a probable encyrtid hyperparasitoid, Coccidoctonus sp. were reared from M. hirsutus on a native Hibiscus species. A field study was conducted from February 2000 to March 2002 in Sherwood, Queensland to document the seasonal phenology of M. hirsutus in its native habitat on its preferred host, Hibiscus rosa-sinensis. Populations of the mealybug stayed at or below detectable levels for most of the study with minor population peaks in the summer months.
The pink hibiscus mealybug, Maconellicoccus hirsutus (Green) (Hemiptera: Pseudococcidae) invaded Caribbean islands, during the mid 1990s, was detected in Belize, Mexico, and the Imperial Valley of California in 1999, and the Bahamas in 2000 (Meyerdirk 2000; Michaud & Evans 2000; Roltsch et al. 2001). In the absence of effective natural enemies, the highly polyphagous M. hirsutus can reach very high population levels (Stibick 1997). At outbreak densities, damage to agricultural crops and ornamental landscape plantings can be quite extensive, actually killing hibiscus and some trees such as soursop, Annona muricata L. and Samanea saman (Jacq.) Merr. (Meyerdirk 1999; Kairo et al. 2000). In the presence of natural enemies, Michaud & Evans (2000) report that the host range of the mealybug in the field is much reduced and that noticeable damage was restricted to its preferred host, hibiscus, Hibiscus rosa-sinensis L. The parasitoids Anagyrus kamali Moursi and Gyranusoidea indica Shafee, Alam & Agarwal (Hymenoptera: Encyrtidae), and predator Cyrptolaemus montrouzieri Mulsant (Coleoptera: Coccinellidae) were introduced into the Caribbean and are providing excellent control of M. hirsutus (Kairo et al. 2000; Meyerdirk, 1999, Meyerdirk, 2000; Michaud & Evans 2000).
The two parasitoid species A. kamali and G. indica were recently introduced into the Imperial Valley of California within one month after the initial detection of M. hirsutus. It is not known how well these natural enemies will perform in the hot, dry Imperial Valley climate, although preliminary results are positive (Roltsch et al. 2001). In anticipation of the need for new natural enemy species that are climatically adapted to the environment in the Imperial Valley, a search for new natural enemies was conducted in climatically similar parts of Australia. Maconellicoccus hirsutus is believed to be native to Australia and is distributed throughout the wet and dry subtropical and tropical areas of Queensland, Northern Territory, and Western Australia (Williams 1985).
In Australia, Maconellicoccus hirsutus is under excellent natural control by predators and parasitoids and rarely reaches pest status. It has been collected from Annona squamosa L., Citrus spp., Erthryina sp. Glycine max L., Gossypium spp., Sida acuta Burm. f., Parkinsonia sp. Parthenium hysterophorus L., and Hibiscus spp. (Williams 1985) Radunz & Allwood (1981) report that it is an occasional pest of ornamentals in the Northern Territory. Perhaps due to its minor pest status, very little is known about the seasonal phenology of M. hirsutus in its native range. A field study of M. hirsutus was conducted in subtropical southeast Queensland (Goolsby 2000) where it is native, on its preferred host—hibiscus. The study is designed to be used as a comparison for studies of M. hirsutus where it is invasive.
Materials and Methods
Garden sites with hibiscus, H. rosa-sinensis were selected in suburban Brisbane, Queensland to study the seasonal phenology of M. hirsutus. Monthly collections over a two-year period from April 2000 to March 2002 were made in Sherwood, Graceville, Chelmer and Indooroopilly, Queensland, all residential suburbs in the Brisbane metropolitan area. Each garden site had mature plantings of hibiscus. Corrugated cardboard bands were used to estimate the density of M. hirsutus. This technique has been used in several biological control programs to monitor population densities of mealybugs and evaluate the impact of their natural enemies (DeBach 1949, Browning 1959, Furness 1976, Berlinger 1977). Banding has the advantage of being the least destructive measure of pest density. Destructive sampling of plant terminals on individual hibiscus plants could have an effect on the mealybug population. Six bands per site (ea. 10 × 7 cm) were wrapped around the main trunk and limbs of the hibiscus plants and secured with wire garden tie. Bands were left on the plants for a period of one month. Counts of mealybugs in the bands were conducted in the laboratory. All stages of mealybugs on the surface of the bands were counted; however, the counts overwhelmingly represent mature ovipositing females. Bands were held in paper cartons streaked with honey for emergence of the natural enemies.
Arthropods collected by the Australian Biological Control Laboratory (ABCL) were assigned a specific site collection number. Each accession is unique prefaced by the acronym for the laboratory with the year collected and a serial number associated with the field collection (i.e., ABCL 2000809). If an organism is later exported to the U.S. for a biological control program, the number is used as an identifier in the ROBO (Releases of Beneficial Organisms) database that is maintained by USDA-Agricultural Research Service. Genbank accession numbers for the exported agents are also included as an identifier in ROBO (i.e., GENBANK-1786201). This method allows biological control workers to track the identity of the organism, including its DNA profile, from field collection through quarantine evaluation and release.
Natural enemies were collected from bands used in the M. hirsutus field study conducted in Brisbane, Queensland from Jan 2000 to March 2002. After field collection, the bands were held in paper cans streaked with honey for emergence of the natural enemies. In addition, two distinct collections were made in the Northern Territory and Western Australia in July 2000 and May 2001. Leaves, terminals and roots were sampled for M. hirsutus in these collections. Mealybugs of all stages were transferred to cotton stoppered vials with honey and held in a humiditron (Debach & Rose 1985) at 70% RH for emergence of parasitoids. Mealybugs were sent to the Queensland Department of Primary Industries for identification. Parasitoids and predators were sent to the ARS—Systematic Entomology Laboratory (SEL) in Beltsville, MD for identification. Gyranusoidea indica collected in Queensland was compared to material from rearing colonies in the California Dept. of Food and Agriculture in Brawley, CA. Molecular sequencing of the D2 expansion domain of the 28S rRNA gene was used to compare the two populations. The method was employed by Babcock & Heraty (2000) and De Barro et al. (2000) to distinguish Encarsia species of parasitic Hymenoptera.
Results and Discussion
Three of the six garden sites were originally selected because they had visible populations of mealybugs on the terminals of the hibiscus plants. The terminals of the plants were impacted and exhibiting bunchy growth, but no leaf loss. The population density in February 2000 at these three sites were the highest encountered in the study with 100, 92, & 102 M. hirsutus individuals collected in the bands respectively. The high population levels at these three sites were not sustained. For the remainder of the study, populations at all the sites remained at very low levels for most of the year with two small peaks occurring during the summer months (November-January) of 2001 (3.2 ± 0.8 M. hirsutus per band) and 2002 (3.3 ± 0.9 M. hirsutus per band) (Fig. 1). Populations were below detectable levels during the winter months (June-August) in 2001-02. Cryptolaemus montrouzieri was consistently recovered from the bands throughout the study when mealybugs were present. The characteristic larvae were also commonly observed on infested terminals. The parasitoid, G. indica was rarely recovered. It may have been underrepresented in our study due to predation by C. montrouzieri in the bands, or due to the tendency of the bands to harbor later instar mealybugs that are not preferred by the parasitoid. The adult female M. hirsutus often uses the bands for shelter during production of ovisacs. Our observations of the natural enemies suggest that C. montrouzierri plays the key role in regulating M. hirsutus in this environment.
Retrospectively, we compared population levels in our study to those in Puerto Rico post release of A. kamali and G. indica (Michaud & Evans 2000). We concluded that for most of the two-year study, hibiscus stands showed light or no damage with less than 5% of the plants with live M. hirsutus, which is far below levels detected post release of natural enemies in Puerto Rico.
The predators, parasitoids and hyperparasitoids reared from all the collections of M. hirsutus are listed in Table 1. In Queensland (QLD), C. montrouzieri was the most commonly encountered natural enemy. Populations of the beetle persisted year round, responding to even low densities of M. hirsutus. Cryptolaemus montrouzieri is highly polyphagous in its native habitat. Larvae and adults feed on crawlers and immature soft scales, armored scales, wax scales, fluted wax scales, and all stages of many mealybug species (Smith et al. 1997). However, its preference appears to be for mealybugs (Don Sands, CSIRO Entomology, pers. comm.). Cryptolaemus montrouzieri is extant in parts of California; therefore there are no plans to introduce this species. A predaceous drosophilid fly, Cacoxenus perspicax Knab was collected, but only at high M. hirsutus densities. The host range of C. perspicax and its impact on M. hirsutus is not known. No plans were made to introduce this species to the United States. The encrytid parasitoid G. indica was occasionally collected from M. hirsutus in Queensland. There were no differences in the D2 sequences between the G. indica populations from Queensland and those in culture in California; therefore we elected not to ship this population to U.S. quarantine facilities. Representative specimens of G. indica from Queensland were vouchered at SEL and the colony was terminated.
Collections of M. hirsutus in the Northern Territory (NT) and Western Australia (WA) revealed different species of natural enemies from those found in Queensland. Several potential wild malvaceous hosts were surveyed in addition to ornamental hibiscus including, Gossypium australe F. Muell, Gossypium sturtianum J. H. Willis, Gossypium robinsonii F. Muell., Hibiscus panduriformis N.L. Burman var. australis Hochr., Hibiscus sp. C and Abutilon lepidum (F. Mueller) A. Mitch. Maconellicoccus hirsutus was collected from ornamental hibiscus, at several locations in the NT and WA, but no parasitoids emerged. Collections made of G. australe in northwestern Western Australia near Derby and Broome revealed substantial numbers of M. hirsutus just below ground level feeding on the roots and crown of the plant. Populations of M. hirsutus on the roots were heavily defended by ants. Near Kununurra, WA, M. hirsutus was collected from the terminals of a wild malvaceous host identified using the Flora of the Kimberley (Wheeler et al. 1992) as Hibiscus sp. C. A predaceous noctuid larvae, Mataeomera sp., an aphelinid parasitoid Coccophagus sp., and a probable hyperparasitoid, Coccidoctonus sp. were reared from M. hirsutus on this native host plant (Table 1). Mataeomera sp. could have potential as a biological control agent as some species in this genus are believed to have a narrow host range and can be quite damaging to mealybug populations (Don Sands, CSIRO Entomology, pers. comm.). This species, as well as the primary parasitoid Coccophagus sp. might be considered for importation if additional natural enemies are needed for control of M. hirsutus in California.
In its native range, M. hirsutus is rarely a pest and under excellent natural control by indigenous natural enemies. The predator C. montrouzieri appears to be the key natural enemy in subtropical eastern Australia on ornamental stands of hibiscus. More intensive sampling is needed to determine the impact of parasitism by G. indica. In addition, the natural enemies Mataeomera sp. (Lepidoptera: Noctuidae), and Coccophagus sp., (Hymenoptera: Aphelinidae) from the hot, monsoonal environment of northern Australia and could be useful in biological control efforts in California if A. kamali and G. indica do not provide adequate control.
The authors would like to acknowledge Anne Bourne for statistical advice; three anonymous reviewers for their helpful comments and Ryan Zonneveld, CSIRO Entomology, for conducting the field and laboratory work. We would also like to thank the staff at the Queensland Department of Natural Resources at the Alan Fletcher Research Station and the homeowners in Chelmer, Graceville, Indooroopilly, and Sherwood, Queensland for allowing us to use their gardens as test sites. Alan Donaldson, Queensland Department of Primary Industries identified the mealybugs, Don Sands, CSIRO Entomology, identified the predatory Noctuidae, Mike Schauff, Alan Norboom, Michael Gates identified the Hymenoptera and Diptera recovered in the study. The United States Department of Agriculture; Agricultural Research Service—Office of International Research Programs and the Animal and Plant Health Inspection Service—National Biological Control Institute provided funding for the research.