Aprostocetus nitens (Hymenoptera: Eulophidae), an Ectoparasitoid Proposed for Biological Control of the Destructive Erythrina Gall Wasp, Quadrastichus erythrinae, in Hawaiʻi

Simple Summary

Aprostocetus nitens, an eulophid parasitic wasp native to eastern Africa, is being evaluated in Hawaiʻi to control the destructive erythrina gall wasp (EGW), Quadrastichus erythrinae, which has devastated the native wiliwili trees, Erythrina sandwicensis. Host range assessment of six non-target gall-forming insects using choice and no-choice tests revealed that A. nitens, highly specific to EGW, targets its immature stages on leaves, flowers, and pods. It is expected to complement the existing biological control agent, the eurytomid wasp, Eurytoma erythrinae, to further reduce EGW populations and protect wiliwili trees on the Hawaiian Islands. We also report on the diverse tri-trophic associations of Erythrina host plants, gall wasps, and their parasitoid guilds in the native African regions.

Abstract

Aprostocetus nitens Prinsloo & Kelly (Hymenoptera: Eulophidae) was identified as one of four hymenopteran ectoparasitoids utilizing three erythrina gall wasps, Quadrastichus bardus, Q. erythrinae, and Q. gallicola) (Hymenoptera: Eulophidae) in the native eastern Africa. In Hawaiʻi, the eurytomid wasp, Eurytoma erythrinae Gates & Delvare (Hymenoptera: Eurytomidae), was introduced and approved for statewide release in 2008 to control the erythrina gall wasp (EGW) Q. erythrinae Kim. EGW has devastated the wiliwili trees, Erythrina sandwicensis Degener (Fabaceae), an ecologically and culturally important native Hawaiian tree species. However, the parasitoid’s impact on the galled inflorescences and shoots was not adequate to ensure adequate seed set and maturation for successful tree recruitment. Aprostocetus nitens was thus evaluated as a prospective natural enemy to enhance the biological control of EGW to further protect the wiliwili trees in Hawaiʻi. Both choice and no-choice host specificity tests were conducted on seven non-target gall formers in the Hawaii Department of Agriculture, Insect Containment Facility, and showed that the parasitoid was extremely specific to EGW. The potential for competition between this parasitoid and the established E. erythrinae was also investigated, showing that the release of a second parasitoid will potentially complement the success of the eurytomid wasp for control of EGW. Unlike what was found in the native region, the Hawaiian laboratory colony is thelytokous, producing only female offspring. The life cycle took 20.1 ± 0.28 days under the laboratory conditions. Non-ovipositing female survived for 102.5 ± 2.9 days when fed honey and laid eggs for 25.1 ± 2.3 days with average fecundity of 156.7 ± 22.3 offspring/female. This value is 3.9-fold higher than offspring produced by E. erythrinae. Aprostocetus nitens, host specificity, competition with E. erythrinae, and its tri-trophic association with 15 Erythrina host plants and 5 gall wasp assemblages in the native African regions were defined. Implications to reduce frequent galls on the native Erythrina plants and likely domination over E. erythrinae, are discussed.

1. Introduction

In April 2005, the Erythrina gall-forming wasp (EGW), Quadrastichus erythrinae Kim (Hymenoptera: Eulophidae), was detected on Erythrina trees (Fabaceae) on the island of Oahu (21.4389° N, 158.0001° W). Africa was confirmed as the area of origin of Q. erythrinae, and Tanzania is the putative source of the populations that invaded Hawaiʻi. Once the wasp was established, it was dispersed via wind and through human activities such as hiking and shipping. EGW rapidly spread and devastated Erythrina trees throughout the Hawaiian Islands. Thousands of Erythrina variegata L. trees used primarily for landscaping and windbreaks were all decimated throughout the state (Figure 1). The native plant, Erythrina sandwicensis O.Deg., native Hawaiian name “wiliwili”, is a highly valued dominant species of Hawaii’s lowland dry forests and was severely damaged with more than 40% tree mortality recorded in some areas [1,2,3,4,5,6,7,8].

EGW is a minute wasp, ranging from 1–1.5 mm in length. The life cycle of EGW in Hawaiʻi is nearly 20 days, which includes the egg stage (4 d), the larval stage (10 d), and the pupal stage (6 d) (HDOA Plant Pest and Control Branch 2008 records). The female is synovigenic and emerges with about 60 mature eggs in its ovaries. A female wasp can lay up to 350 eggs by inserting them into the young tissues of the plant. The larvae and pupae develop within the tissues, resulting in the swelling and formation of galls on young leaves and petioles (Figure 1C). Each gall harbors one larva in one chamber [4]. Heavily galled leaves and stems result in loss of growth and vigor as photosynthesis is reduced and plant health declines. Severe infestations eventually lead to tree mortality (Figure 1B) [3].

Because EGW feeds within plant tissues, it was thought that systemic pesticides could be used to control infestations of this species. However, widespread use of this method is unreasonable, and frequent, long-term use of insecticides in natural areas is neither feasible nor advisable. Attempts in Honolulu to control EGW on iconic landscaping trees (Erythrina variegata) using systemic pesticides were unsuccessful, and eventually resulted in the plants being cut down, after thousands of dollars were spent on chemicals. Insecticidal control was discontinued. The wasp spread quickly to several species of Erythrina in the state, including the Brazilian coral tree, Erythrina crista-galli L., Indian coral tree, E. variegata, and wiliwili, E. sandwicensis. Wiliwili, a native Hawaiian species, showed greater than 40% mortality in some affected populations [9,10]. Severe infestations have also devastated native and introduced trees of the genus Erythrina in the Western Pacific and Hawaii (China, Guam, Hawaii, Japan, Malaysia, Philippines, Samoa, Singapore, Taiwan, and Thailand). Sanitation through pruning and chemical control measures has failed to contain this pest, and biological control is thought to be the only long-term solution [3,5,11,12,13,14,15].

Failed attempts to control EGW with conventional methods prompted the Hawaiʻi Department of Agriculture (HDOA), along with the University of Hawaii and other government agencies, to initiate a biological control project to examine species that would be appropriate biocontrol against EGW. Shortly thereafter, the survey team of five exploratory entomologists collected infested Erythrina leaves and stem samples from Tanzania, east Africa, and west Africa and shipped them to the HDOA Insect Quarantine Facility for parasitoid emergence [16]. Erythrina trees in Tanzania were very healthy and had low levels of gall insects (Figure 2). Gall abundance, sizes, and shapes vary on different Erythrina trees (Figure 2). Four main parasitoids known to attack EGW were collected during the exploratory surveys in East Africa in 2005–2007; they parasitized most of the wasp immature stages in the galls. These were ectoparasitoids including three eulophid parasitoids (Aprostocetus exertus LaSalle, A. nitens Prinsloo and Kelly, and A. tritus Prinsloo and Kelly) and the eurytomid parasitoid, Eurytoma erythrinae Gates and Delvare. Eurytoma erythrinae and A. nitens were amenable to laboratory rearing and showed potential as biocontrol agents. Aprostocetus exertus was not suitable for captive propagation owing to its male-biased sex ratio after four generations of rearing. Aprostocetus tritus was not reared [16,17,18,19].

After extensive evaluations at the HDOA Insect Containment Facility for host-specificity to confirm that E. erythrinae prefers to feed exclusively on EGW, field releases took place in November 2008. About 4000 individuals were released at various sites on the islands of Hawaiʻi, Kauaʻi, Maui, Molokaʻi, and Oʻahu. Within the span of a few months, the wiliwili trees began to show signs of recovery, with healthy new leaves and vigorous overall growth. By the second year after the release of E. erythrinae, more than 60% of young shoots were predominantly free of damage by EGW, tree canopies recovered each year after the release. By 2011, 90% of the sample trees had full canopy coverage (Figure 1A) [20,21,22,23].

The progressive increase of flower and seed production indicated that the overall tree health of E. sandwicensis also improved [21]. Depending on the location, recent weather, and time of year, parasitism rates by E. erythrinae of EGW larvae inside galls range from 20–100%. Flowering and fruiting were restored, and the number of flowers increased each year post-release. However, 54% of the sampled inflorescences failed to form mature seeds due to gall wasp damage. Since the flowering season coincides with warm summer months (May–October), which is favorable for EGW development, the damage to wiliwili flowers continues. Most wiliwili seedlings are still unable to survive the damage of the pest. This adverse impact on flowers and seed pods production is not only an ecological concern, but also a cultural issue because Native Hawaiians value the scarlet wiliwili seeds for lei making (Figure 3E). Low production of viable seeds and seedling mortality remain threats to the future survival of E. sandwicensis. A potentially viable option is to seek another biocontrol agent that can augment the current impacts of E. eythrinae, and reduce the damage to reproductive structures and seedlings (Figure 3) [21,22].

Thus, despite the success of the release of E. erythrinae, it has been proposed that there was a need for a second species, A. nitens, to be released to enhance biocontrol and suppress EGW populations. Because E. erythrinae acts more as a predator than a parasitoid, with its larvae feeding on EGW galls and tunneling to feed on other larvae in adjacent galls, larger galls with many individual EGW are preferred for oviposition (Figure 4A,B). This feeding behavior excludes galled young Erythrina seedlings, as well as flowers and seed pods, where smaller galls tend to form at lower density, and effectively escape parasitism by E. erythrinae [22]. As a complementary biocontrol agent to E. erythrinae, HDOA thus proposed to release A. nitens from containment into the natural environment of the state of Hawaiʻi. Unlike E. erythrinae, A. nitens utilizes only one host individual to complete its development, and therefore, it can live on much smaller galls, such as those found in flowers, pods, and young seedlings (Figure 3). It is anticipated that the release of A. nitens will increase the likelihood of tree recruitment, and improve the overall survival of wiliwili.

Here, we report on foreign exploration for EGW parasitoids in Africa, the biology and host specificity testing of A. nitens, as well as the potential interaction between A. nitens and E. erythrinae. The tri-trophic interactions of the diverse Erythrina hosts, gall formers, and their associated parasitoids in Africa are also described.

2. Materials and Methods

2.1. Explorations and Origin of the Parasitoid Colony

Eight African countries were surveyed during this project by the HDOA and University of Hawaiʻi survey teams. East Africa (Kenya, Madagascar, Mozambique, South Africa, Tanzania) were surveyed in December 2005, January–June 2006, January–April 2007, and West Africa (Benin, Ghana, and Togo) were surveyed in May–June 2006 (see detailed collection information on Table S1). Samples of galled Erythrina leaves were shipped to the HDOA Insect Containment Facility (ICF). There, leaves were placed as layers on wire screens and paper towel in screened cages (30 × 30 × 60 cm, 70 mesh) for wasp emergence at (mean ± SEM, 21.8 ± 0.12 °C, 70.2 ± 2.4% RH, 12:12, D: L photoperiod). Specimens of emerging wasps were preserved in 70% alcohol for later identification. We used the Prinsloo and Kelly (2009) manuscript to identify the gall formers and parasitoids [19]. Voucher specimens were sent to La Salle (hymenopteran taxonomist, CSIRO, Pullenvale, QLD, Australia) and Prinsloo (Pretoria, South Africa) for confirmation of species identity. The characters of color and submarginal setae separated Quadrastichs from Aprostocetus parasitoids according to the key ([19], Figures S1 and S2). Diversity and distribution of the Hymenopterous gall inducers of native Erythrina were determined with the recovered parasitoids.

Voucher specimens of Aprostocetus nitens were deposited in the insect reference collection of the HDOA, the Bernice P. Bishop Museum, and UH insect collection, Honolulu, Oʻahu.

2.2. Propagation of Host Plants and EGW

Erythrina variegata was used as the host plant for EGW rearing in our tests due to its availability and rapid growth. Seeds were scarified using a grinding machine. Two seeds were planted in 10 cm square pots and kept in outdoor screened cages (76 × 88 × 127 cm, 70 mesh) to avoid contamination with any pest scales, mites, aphids, or thrips. Seedlings were watered every other day. In 6–8 weeks, seedlings were ready to be used in propagating gall insects; the ideal plants are ±45 cm tall, 2–3 branches (about 2 cm long), with new leaves. We set up each rearing cage for EGW with two uninfested E. variegata plants in saucers. Fifteen pairs of EGW were placed in the cage for 7 days until galls had formed on young leaves and petioles. Plants were then transferred to cages (42 × 42 × 62 cm, 70 mesh) and held in HDOA-ICF in the conditions described above and were watered every other day. Galled E. variegata seedlings were ready to be used for parasitoid propagation 14 days after exposure to EGW.

2.3. Rearing of A. nitens

Twenty adults of A. nitens were placed in each cage, containing four plants that were previously exposed to EGW for 14 days as described above. Only female adults were needed for inoculation since this parasitoid is a thelytokous strain producing only females under HDOA-ICF laboratory conditions. Female A. nitens lays eggs up to 30–60 days. Adults usually emerged within 20 days after exposure. If no emergence occurred after 28 days, the plants were inoculated again with new female A. nitens. Newly infested Erythrina plants with EGW galls were added every 2–3 weeks to ensure plants in the cage had a continuous supply of new galls available.

2.4. Longevity and Life-Time Fecundity of A. nitens

Longevity of non-ovipositing females was determined by collecting newly emerged adults of A. nitens and placing 5–10 females in clear plastic vials (27 mm Ø, 55 mm height). The vials were covered with a piece of muslin cloth material and secured with a ventilated plastic cap. Honey (SUE BEE^® SPUN^® (https://siouxhoney.com/sue-bee-spun-honey), Sioux City, IA, USA, accessed on 1 January 2025) was dotted on muslin cloth material for feeding the females. Females without access to honey were also measured for their longevity. The number of dying wasps was recorded daily until all perished. Tests were replicated ten times. The mean survivorship of 100 and 58 females was determined for honey-fed and starved treatments, respectively.

Fecundity tests were conducted to determine the number of progeny a female could produce in her lifetime. One E. variegata plant with 14-d-old galls was placed in a (30 × 30 × 60 cm, 70 mesh) aluminum cage with a newly emerged female A. nitens. The plant was exposed to a female parasitoid for 6–7 days, then the plant was removed and replaced with a new plant with new EGW galls until the female died. Exposed plants were held in a cage, and the emerged parasitoid offspring were counted and recorded daily. After all parasitoids emerged, galls were dissected to count the non-emerging wasps. This test was repeated seven times.

2.5. Host Specificity Testing of A. nitens

All host specificity tests for A. nitens were conducted in the HDOA-ICF. Both choice and no-choice tests were conducted to determine whether A. nitens would feed on non-target hosts in the presence or absence of its target host. Seven non-target gall-forming insects were tested (same species as used in trials of E. erythrinae, Nagamine et al. [23]): Hamakua pamakani gall fly, Procecidochares alani Steyskal, a biocontrol agent of Ageratina riparia; Maui pamakani gall fly, Procecidochares utilis Stone, a biocontrol agent of Ageratina adenophora; lantana gall fly, Eutreta xanthochaeta Aldrich, a biocontrol agent of Lantana camara (all Diptera: Tephritidae); banyan gall wasp, Josephiella microcarpae Beardsley & Rasplus (Hymenoptera: Agaonidae), pest of Ficus microcarpa, Chinese banyan; a native psyllid, (Pariaconus sp., Hemiptera: Psyllidae), on ʻōhiʻa lehua Metrosideros polymorpha; a eulophid wasp, Ophelimus sp. (Hymenoptera: Eulophidae), a pest of Eucalyptus sp.; and a scale insect, Tectococcus ovatus Hempel (Hemiptera: Eriococcidae), a biocontrol agent of strawberry guava, Psidium cattleianum. Whole plants or cuttings were used in the experiments for choice or no-choice tests.

In the choice tests, a plant or cutting harboring one of the seven non-target insects and an EGW-infested Erythrina plant were placed side-by-side in a screened aluminum cage (46 × 46 × 76 cm, 70 mesh). Ten newly emerged A. nitens females from the laboratory colony were placed in the cage with honey provided as a source of food for the adult parasitoids. Behavioral observations were conducted to record the number of visits by the 10 A. nitens to each plant or cutting and whether they were resting or attempting to oviposit. Six daily counts were made at successive hourly intervals for two consecutive days (8:00 a.m.–14:00 p.m.). Parasitoids were then removed after two weeks, and the test and control plants were placed in separate cages to await parasitoid emergence. After one month, galls from each test plant were dissected and examined under a microscope to determine whether parasitism had taken place. Procedures were similar for the no-choice test, except that the parasitoids were given only one plant or cutting infested with one of the seven non-target gall-formers. As a control, Erythrina plants infested with EGW were exposed to the parasitoids in separate cages. There were three replicates of each insect tested in both choice and no-choice tests.

2.6. Intersepcific Competition Between A. nitens and E. erythrinae

Competition trials were performed in the HDOA-ICF. Because A. nitens is proposed to be released to complement the beneficial impacts made by the introduction of E. erythrinae, multiple parasitism trials were performed to determine (1) the level of Erythrina gall wasp parasitism by the two parasitoids when either is used alone or when both are used concurrently; (2) if the sequence of Erythrina gall wasp exposure to the two parasitoids will have differential effects on their progeny; and (3) if interspecific competition would reduce the efficiency by each or both parasitoids. The experimental treatments were as follows:

• Control: Plants with Erythrina gall wasp only, no parasitoids;
• Galled Erythrina were exposed to 10 females of E. erythrinae only, for 72 h;
• Galled Erythrina were exposed to 10 A. nitens only, for 72 h;
• Galled Erythrina were exposed to 10 females of E. erythrinae and 10 females of A. nitens concurrently for 72 h;
• Galled Erythrina were exposed to 10 females of E. erythrinae for 72 h. Then, after 4 days, the same plants were exposed to 10 A. nitens for 72 h;
• Galled Erythrina were exposed to 10 females of A. nitens for 72 h. Then, after 4 days, the same plants were exposed to 10 females of E. erythrinae for 72 h.

For each of the treatments 2–6, gravid females of E. erythrinae (7 d-old) and/or A. nitens (3 d-old) were introduced to galled E. variegata. Before the introduction of the two parasitoids, the level of galling was rated to ensure that both were exposed to similar levels of galling. After this exposure time, the plants were removed from the cages and held for 4 weeks to rear out the subsequently emerging parasitoids or EGW. All emerging adults were captured and counted. Fourteen days after exposure to the parasitoids, a subsample of the galls was removed from each plant and dissected to determine the number of parasitoids developing within the galls or to determine the parasitism rate of each parasitoid. The experiment was repeated five times.

Treatments 2–4 comprise an additive-series design where the number of each parasitoid species was the same in the single- and two-species treatments at the same host density, so that possible intraspecific interactions remain constant across different treatments. This experimental design is suitable for examining the outcome of interspecific competition on the overall impacts on host suppression. If interspecific interactions between parasitoid species have no effect on the host population (i.e., two different parasitoids act independently), host mortality should follow a multiplicative risk model [24,25]

(P_exp = [P₁ + P₂] − [P₁ × P₂])

where _exp is the expected host mortality from both parasitoids together, P₁ is the observed host mortality by parasitoid species 1 alone, while P₂ is the observed host mortality by parasitoid species 2 alone. If the observed (O) and expected (E) parasitism is no different, there is no negative impact), if O > E, the impact on host mortality is additive, and if O < E, the impact on host mortality is negative.

2.7. Statistical Analysis

An analysis of variance was used to assess the potential significance of differences in the number of parasitoids produced by parasitism on the different hosts exposed to the wasps. Frequency of visits and parasitism were statistically analyzed using one-way ANOVA. Prior to the ANOVA, data were checked for normality and homoscedasticity, and percentage data were arcsine square-root transformed before analysis as needed to normalize the variance. Means were separated by Tukey’s standardized range honestly significant difference test and t-tests at the α = 0.05 level. The observed and expected levels of host mortality were compared across replicates using t-tests. All analyses were conducted using SAS JMP Version 11 [26].

3. Results

3.1. Explorations and Acquiring Parasitoids

Erythrina trees in Africa were healthy with sporadic galls on leaves during the surveys. Aprostocetus nitens was first described by Prinsloo and Kelly (2009) [19] as a parasitoid soon after it was discovered in Tanzania and South Africa as a potential biocontrol agent of Quarastichus species. Aprostocetus nitens is shiny black in color with a dark metallic green tinge and yellow gaster, antennae, and legs (see Prinsloo and Kelly 2009 for full description [19]), Figure 5 and Figure 6.

Aprostocetus exertus is known from Tanzania as a primary parasitoid of Q. erythrinae and Q. ingens in galls of E. latissima in South Africa [18,19]. Aprostocetus nitens and A. tritus are parasitoids of the Quadrastichus gall wasps and have been reared in association with Q. bardus, Q. erythrinae, and Q. gallicola. Aprostocetus nitens head and body black with a dark metallic green tinge (Figure 6). Aprostocetus tritus head and body without a metallic luster, blackish brown except base of gaster yellowish, and submarginal vein of the forewing with four dorsal setae (not reared).

3.2. Tri-Trophical Associations of Native Erythrina, Gall Wasps, and Their Parasitoids

Fifteen species of native African Erythrina were examined for galls (Table S1). Five species of gall inducers and four endemic natural enemies were recognized (Table S1, Figure 7A). Erythrina abyssinica from Tanzania and E. lysistemon from South Africa were mostly galled by Quadrastichus bardus, and Q. gallicola. Quadrastichus ingens was found only in association with Erythrina latissima from South Africa (Figure 8C–F). Initial infestations by Quadrastichus erythrinae were detected in Arusha, Bwawani, Iringa, and Morogoro (Tanzania). These were confirmed by La Salle (CSIRO, Pullenvale, QLD, Australia) based on morphological characters (Figure 8A,B). The third natural enemy, A. exertus, was reared for four generations under Hawaii ICF conditions and died owing to extensive male progeny produced in the colony (Figure 4C).

A tri-trophic association with the 15 surveyed Erythrina species, and five different gall formers were linked to four parasitoids in Figure 7B. Aprostocetus nitens was the dominant parasitoid emerging from Q. bardus, Q. erythrinae, and Q. gallicola from eastern Africa only. Followed by the Eurytoma erythrinae as the second major parasitoid from eastern and western Africa (Figure 4D). Aprostocetus exertus was present in Benin, South Africa, and Tanzania (Figure 4C). In South Africa, Q. ingens is being parasitized by A. exertus and E. erythrinae. From the food chain figure, Q. erythrinae was emerging from samples of E. abyssinica and E. latissima collected only from the Tanzanian leaf samples from Bwawani, Chalinze, Gweta, and Rubungo villages (Figure 8A,B). Those samples were heavily parasitized by the three common parasitoids and mostly by A. nitens. Aprostocetus exertus emerged from Benin samples along with the regular parasitoid Eurytoma erythrinae in association with Q. bardus and Qudrastichus sp. from galls of E. indica, and E. vogelii (Table S1).

3.3. Longevity and Life-Time Fecundity of A. nitens

Biological studies of A. nitens were performed at the HDOA-ICF. This species reproduces parthenogenetically (via thelytoky parthenogenesis, the absence of mating and subsequent production of all female diploid offspring) under laboratory conditions. However, the first generation of field collections of this species from South Africa and Tanzania produced both sexes.

Eggs are laid on the host larvae or pupae of EGW after stinging (Figure 5A). It took an average of three days for eggs to hatch. Larvae matured after 11 days (Figure 5B) confined in galls feeding on EGW larvae. The pupal stage took up to 6 days (Figure 5C,D). Adult females start to lay eggs after three preoviposition days (Figure 5E,F). The entire life cycle for this species, from egg to adult, took 20.1 ± 0.28 days, and newly hatched female offspring contain one or two mature eggs in their ovaries. Females are synovigenic, producing an average of 156.7 ± 22.3 offspring throughout their lifespan. This species can survive 4.0 ± 3.0 days without food, and non-ovipositing females lived significantly longer than ovipositing females for 102.5 ± 2.3 days when fed honey (F_2,199 = 349.04, p < 0.0001). Ovipositing females lived for 46.9 ± 3.5 days (Figure 9) and continued to lay eggs for 25.1 ± 2.3 days, with a rate of 5.9 ± 0.45 offspring per day (

Table 1. Reproductive attributes of thelytokous Aprostocetus nitens females.

Reproductive Parameter	n	Mean ± SEM	Range
Survival of non-ovipositing females, fed honey (days)	100	102.5 ± 2.3 a	1–152
Survival of ovipositing females, fed honey (days)	44	46.9 ± 3.5 b	22–97
Survival of starved females (days)	58	4.0 ± 3.0 c	2–5
Oviposition period (days)	7	25.1 ± 2.3	17–34
Realized fecundity	7	156.7 ± 22.3	107–282
Daily progeny	216	5.9 ± 0.45	0–53
Female lifecycle (days)	11	20.1 ± 0.28	19–22

Ovipositing females were provided EGW-infested plants once every 7 days until females died. Means followed by different letters are significantly different (F_2,199 = 349.04, p < 0.0001, Tukey’s HSD test).

).

3.4. Host Specificity of A. nitens

In both choice and no-choice host specificity tests, not a single A. nitens emerged from any of the seven tested non-target gall-forming species (Figure 10A,B). In contrast, 101.6 ± 15.9 ($\overline{\mathrm{x}}$ ± SEM) and 29.08 ± 2.6 wasps emerged from EGW in the choice and no-choice tests, respectively. These results show statistically significant differences between the parasitism of the target host and non-target candidates in all cases (choice test: F_6,28 = 8.1, p < 0.0001; no choice test: F_6,28 = 8.1, p < 0.0001). Although the parasitoid visited or examined some non-target host galls, the frequency of landing on the non-target host galls (0.83 ± 0.28) was significantly lower than on the target host galls (29.08 ± 2.66) (t₆₈ = 10.6, p < 0.0001). There was no attempt made by the wasps to parasitize the non-target galls. These results suggest that none of these non-target species was recognized as a suitable host substrate by A. nitens (Figure 10C).

3.5. Interspecific Competition Between A. nitens and E. erythrinae

Overall, more E. erythrinae emerged than A. nitens in both single and two parasitoid species exposures (Figure 11A). There was no significant reduction in the number of emerged wasps for both parasitoids when they were exposed alone or together to the hosts. Consequently, when the two parasitoids were released concurrently, the total levels of parasitism (84.4 ± 9.2%) would be higher than E. erythrinae alone (75.4 ± 7.4%) or Aprostocetus alone (33.6 ± 8.8) (F_7,32 = 3.89, p = 0.004). The observed and expected host mortality by both parasitoid species was not different (t₈ = 1.35, p = 0.233), i.e., no negative impact of the interspecific interaction on host suppression. Although there was much lower parasitism by A. nitens than E. erythrinae, the number of both parasitoid species dissected from leaves and stems during the competition tests was similar, suggesting that a slight preference by A. nitens for galls on stems and petioles and by E. erythrinae for galls on foliage (Figure 11B).

Field observations in Mozambique, South Africa, and Tanzania indicated that A. nitens emerged only from galls produced on several Erythrina species, rather than any other host plants considered during field work. Such galls may contain the Erythrina gall wasp, Q. erythrinae, or other African Quadrastichus species. This putative limited host range was confirmed by these experiments performed at the HDOA ICF.

4. Discussion

Aprostocetus is one of the largest of all chalcid genera of the subfamily Tetrastichinae (Hymenoptera: Eulophidae) with about 813 described species globally. They act as parasitoids and gall formers of many species. Many of them are important biocontrol agents of insect pests [27,28,29]. In this study, for the first time, we document A. nitens as an important parasitoid for biological control of EGW. We described its ecological interactions with host insects and host trees in its native range and evaluated its host specificity and potential interaction and outcomes with the established E. erythrinae for biological control of EGW. Our quarantine screening showed host specificity on EGW. Seven gall inducer non-target species were tested with apparently no attempted or successful attack or development. The target host was included in all replicates of non-target testing to demonstrate A. nitens competence, i.e., to show that wasps used in each test were healthy and the environmental conditions were suitable for successful parasitism. Thus, both field observations in its native range and quarantine evaluations in Hawaii showed that A. nitens is highly specific to EGW, and strongly indicate that the proposed release of this parasitoid for the biocontrol of EGW will unlikely cause undesirable, negative, non-target effects in the natural environment of the Hawaiian Islands.

Understanding and predicting potential competitive outcomes is important in the design of classical, augmentative, and conservation biological control programs when considering multiple species introductions [30]. Based upon the experimental data obtained in this study, the proposed release and establishment of A. nitens is expected to supplement the positive impacts of E. erythrinae in suppressing infestations of EGW. Interspecific competition between these two parasitoids may still occur if both attack limited hosts in old leaves and stems, but they seem to have slightly different preferences for host feeding niches, which can facilitate their co-existence and improve overall impacts on EGW. Releasing A. nitens would likely increasingly affect immature EGW stages in these young leaves, flowers, and seed pods. This should result in increased seed set and seed maturation and positively improve the opportunity for recruitment of the endemic Erythrina.

The higher fecundity rate of A. nitens compared to E. erythrinae may also confer an advantage upon the former species over the latter. Indeed, the dominance of A. nitens over the established parasitoid in the field, as seen from the African surveys, may be related to its high fecundity and likely plasticity of body growth and development on differently sized hosts. This phenomenon may be observed in several parasitoids in the field in Hawaiʻi. Psyttalia humilis (Silvestri) (Hymenoptera: Braconidae), a South African parasitoid of fruit flies (Diptera: Tephritidae), became rare after being released in Hawaii with the Australian parasitoid, Diachasmomorpha tryoni (Cameron), for biocontrol of Ceratitis capitata (Wiedemann). While D. tryoni contributed to the reduction of C. capitata populations, P. humilis died out [31,32,33]. This was attributed to the superior fecundity of D. tryoni [34,35]. The invasion of Bactrocera dorsalis (Hendel) in the 1950s led to the introduction of several opiine parasitoid species from Asia to Hawaiʻi [31,36,37]. Among them, Diachasmimorpha longicaudata (Ashmead) became widely established, but thereafter was largely replaced by Fopius vandenboschi (Fullaway) [38,39]. Subsequently, the abundance of these two species declined following the introduction and successful establishment of Fopius arisanus (Sonan) [38]. Fopius arisanus reproductive superiority makes it the dominant opiine on the islands, resulting in 74% of the total parasitism of C. capitata [32]. The abundance of D. tryoni also sharply declined (to <4.0% parasitism) after F. arisanus became established in C. capitata populations. Observations during the 1950s suggested that dominance and displacement of parasitoids were influenced by the superior competitive ability of species with high intrinsic rates of population increase [32,38,40] as well as competitive superiority by the early acting egg parasitoid [41]. We expect Aprostocetus to coexist with E. erythrinae, likely filling slightly separate feeding niches.

We do not know the causes of the thylotokous development of this parasitoid in our quarantine rearing, which warrants further study. However, environmental conditions may stimulate the wasps to revert to sexual reproduction after the release. For example, thelytokous reproduction in Trichogramma spp. (Hymenoptera: Trichogrammatidae) can change to arrhenotoky under exposure to antibiotics or high temperatures (>30 °C) under experimental rearing conditions. This results in a significant increase in the percentage of male offspring and a decrease in the percentage of female offspring in wild populations [30].

Other than Hawaiʻi, only Taiwan and Japan have implemented biocontrol of EGW. Japan imported E. erythrinae from Hawaiʻi, but no recent reports on the outcome of the releases have been published. In Taiwan, resident parasitoids were found to attack EGW, one of which was Aprostocetus felix LaSalle, Yang & Lin, a new species to science [14]. Aprostocetus felix, a solitary parasitoid collected from mature larvae and pupae, dominated the other local parasitoids of Q. erythrinae. Population levels of this species remain high in cooler times of the year when other parasitoid population levels have declined. The parasitoid populations have built up through the years in Taiwan and may serve as a biological control agent of EGW there [3].

5. Conclusions

To examine the community-wide effects of A. nitens in Africa, we constructed a food web chart showing interactions among 15 Erythrina species, five Quadrastichus spp. gall formers, and their four major parasitoids, three Aprostocetus spp. (Eulophidae) and Eurytoma erythrinae (Eurytomidae).

Field observations in Tanzania and host specificity tests in HDOA ICF provide strong support for the hypothesis that A. nitens will not attack non-target gall-forming species, and the parasitoid should be approved for field release in Hawaiʻi. Competition studies showed A. nitens will be unlikely to negatively affect E. erythrinae but is likely to contribute to the overall suppression of EGW owing to their different preference for host feeding niches. The high fecundity and longevity of A. nitens are also important attributes for a biological control agent for EGW. We predict that A. nitens will be a valuable addition to EGW control programs in Hawaiʻi and potentially other places.