Suitability of the Eggs of Two Species of Eucalyptus Longhorned Borers (Phoracantha Recurva and P

Biological Control 19, 95–104 (2000) doi:10.1006/bcon.2000.0853, available online at http://www.idealibrary.com on

Suitability of the Eggs of Two Species of Eucalyptus Longhorned Borers (Phoracantha recurva and P. semipunctata) as Hosts for the Encyrtid Parasitoid Avetianella longoi

K. A. Luhring, T. D. Paine,1 J. G. Millar, and L. M. Hanks2 Department of Entomology, University of California, Riverside, California 92521

Received August 20, 1999; accepted June 12, 2000

INTRODUCTION Eggs of the eucalyptus longhorned borer (Phoracan- tha semipunctata F.) are the typical host of the en- The eucalyptus longhorned borer, Phoracantha cyrtid egg parasitoid Avetianella longoi Siscaro. Both semipunctata (F.) (Coleoptera: Cerambycidae), native P. semipunctata and Phoracantha recurva Newman to Australia (Wang, 1995) and a serious pest of Euca- have become established in California, but A. longoi lyptus spp. in exotic plantings around the world, was appears to differentially parasitize these congeners. first discovered infesting eucalypt trees in southern Eggs of the two beetle species were tested for oviposi- California in 1984 (Scriven et al., 1986). P. semipunc- tional and developmental suitability for A. longoi at tata has a significant economic and environmental im- 0.5, 1.5, 2.5, 3.5, and 4.5 days after deposition in both pact in California and in other areas of the world where no-choice and choice bioassays. In no-choice experi- eucalypts are planted as ornamental or plantation ments, the parasitoid oviposition rate was similar in trees (e.g., Drinkwater, 1975; Ivory, 1977; Powell, the two host species when the host eggs were 0.5 days 1978; Mendel, 1987). Heavy larval infestations can old, but higher in P. semipunctata eggs when the host quickly kill trees, and even a few borer larvae can eggs were >1.5 days old. Females allocated signifi- girdle and kill a tree. Seasonal water stress in many cantly more eggs per host egg to P. semipunctata eggs areas where eucalypts have been planted (Hanks et al., than to P. recurva eggs when the eggs were 0.5–3.5 1991) and a lack of effective natural enemies (Paine et days old. Survival of the parasitoids was significantly lower in P. recurva eggs of all age classes than in P. al., 1993) contribute to the borer’s success in attacking semipunctata eggs. In choice experiments, female eucalypts outside of Australia. parasitoid oviposition rate was significantly higher in A congener of P. semipunctata, Phoracantha recurva P. semipunctata eggs regardless of host egg age. Para- Newman, recently appeared in southern California sitoid survival was also significantly higher in P. semi- (Hanks et al., 1998b), and has become the predominant punctata eggs. In both choice and no-choice experi- borer attacking eucalypts in many areas of southern ments, some P. recurva embryos survived the parasi- California (T.D.P. and J.G.M., unpublished data). P. toid attack and emerged as neonate larvae from recurva is also native to Australia where it is geograph- parasitized eggs, whereas P. semipunctata embryos ically sympatric and shares host species with P. semi- never successfully developed in parasitized eggs. punctata (Wang, 1995; Hanks et al., 1998b). These results have important implications for biolog- In southern California, the two congeneric beetles ical control of the two beetle species by A. longoi. © 2000 are generally similar in oviposition behavior. The noc- Academic Press turnally active females of both Phoracantha species lay Key Words: egg parasitoid; host egg age; parasitism; their egg masses under loose bark of eucalypts (Hanks Cerambycidae; Encyrtidae; Phoracantha semipunc- et al., 1993). Adult P. recurva have been found as early tata; Phoracantha recurva; Avetianella longoi; Cali- as February. In contrast, adult P. semipunctata are fornia. normally present from May to November (Hanks et al., 1993) and complete two overlapping generations per year. However, Phoracantha spp. adult emergence is

1 not tightly synchronized, and adults can live for over a To whom correspondence should be addressed. Fax: (909) 787- month under ﬁeld conditions. Therefore, Phoracantha 3086. E-mail: [email protected]. 2 Current address: Department of Entomology, University of Illi- eggs are present for most of the year in southern Cal- nois, 505 South Goodwin Avenue, Urbana, IL 61801. ifornia.

In 1993, the egg parasitoid Avetianella longoi Siscaro METHODS (Hymenoptera: Encyrtidae) was successfully introduced into California as one component of a biological Field parasitism of beetle eggs by A. longoi. Be- control program against P. semipunctata (Paine et al., tween August and October 1997, we assessed the host- 1993). Since its introduction into California, A. longoi finding abilities of A. longoi at two field sites (Rancho has been established throughout most areas of the Santa Fe and La Jolla, San Diego Co., CA) that had state where eucalypts occur. In California as well as in well-established populations of A. longoi. To obtain egg Europe, A. longoi has proven to be an effective natural masses of known age of both P. recurva and P. semi- enemy of P. semipunctata (Longo et al., 1993; Hanks et punctata under conditions that approximated natural al., 1996). Parasitism rates of all eggs in all naturally conditions, we created an oviposition substrate by sta- Ϸ ϫ laid egg masses of P. semipunctata sampled in 1995 pling two pieces ( 5 5 cm) of eucalyptus bark to- exceeded 90% (Hanks et al., 1996). The small size (1.2– gether and then placed these “bark sandwiches” into P. 1.6 mm) and dorso–ventral flattening of A. longoi fe- recurva and P. semipunctata colony cages overnight to males allow the adults to move under the loose bark allow female beetles to oviposit between the two bark and reach the eggs of the borers. sheets. The following morning, we collected the bark sandwiches from the colony cages, transported them to One of the major challenges facing egg parasitoids is the field sites, and stapled the sandwiches to freshly to locate and attack an ephemeral host resource felled logs. The bark sandwiches were collected after (Strand, 1986). In the field, A. longoi has a short win- 48 h and assessed for the success of female parasitoids dow of opportunity to locate and attack host eggs. The at finding host egg masses (percentage of masses par- larvae of P. recurva and P. semipunctata eclose 5–6 asitized) and for the acceptance of host eggs (percent- days after egg deposition at 22–25°C (Hanks et al., age parasitism in each egg mass). Parasitized host eggs 1993; unpublished observations). However, daily sum- can be identified readily without dissection because, in mer temperatures in southern California frequently common with other encyrtid wasp eggs (Clausen, exceed 35°C, and this can significantly shorten host 1962), the eggs of A. longoi have pedicels that protrude development time. Also, because A. longoi is diurnally from the outer surface of host eggs and anchor the active and Phoracantha spp. females are crepuscular parasitoid’s eggs to the chorion of the host egg. Our (unpublished laboratory and field observations, Hanks assessment of successful host finding may have under- et al., 1998a), the earliest that the female parasitoids estimated the number of egg masses that were located can locate and attack eggs of Phoracantha spp. is the because some egg masses visited by A. longoi females morning after they have been laid (4–12 h after ovipo- may not have stimulated oviposition. sition). Data were arcsine square-root transformed (Sokal In southern California, A. longoi females are faced and Rohlf, 1981) and one-way analysis of variance with increasing numbers of P. recurva eggs to use as (ANOVA) tests were performed using PROC GLM in hosts, as populations of P. semipunctata have been PC-SAS for Windows version 6.12 (SAS Institute, reduced to very low levels. Some areas of northern 1996) with the data blocked according to site. Statisti- California have not been colonized by P. recurva but cal significance was established at ␣ Յ 0.05. sustain large populations of P. semipunctata. Other General laboratory methods. Colonies of P. recurva areas of the state have a mixture of the two species, and P. semipunctata were maintained as described by particularly along the expanding edge of the geo- Hanks et al. (1993, 1995). Briefly, adult beetles were graphic distribution of P. recurva. Therefore, we con- caged in cylindrical hardware cloth cages with 14.5-cm- ducted field experiments and both choice and no-choice diameter plastic petri dishes covering the top and bot- laboratory bioassays to evaluate the reproductive re- tom of the cages. The beetles were fed 10% honey– sponse of the parasitoid to eggs of one or both beetle water solution and eucalypt pollen (Wando and Jarrah species. The specific objectives of the present study Tree Pollen, Great Health Co., Brea, CA). The bottom were (1) to determine whether A. longoi females locate of each cage was lined with filter paper. An oviposition eggs of each Phoracantha species at the same rate in substrate was prepared by covering the bottom of a the field, (2) to determine whether A. longoi recognizes 9-cm-diameter petri dish with filter paper and placing and accepts eggs of both species equally, as measured this dish on the filter paper lining the bottom of the by oviposition rates, in field tests, (3) to determine cage. Adult females of both Phoracantha species lay whether P. semipunctata and P. recurva eggs are egg masses of Ͼ10 eggs (Hanks et al., 1993; unpub- equally acceptable to ovipositing parasitoids in the lab- lished data) between the two layers of filter paper. For oratory, (4) to determine whether P. semipunctata and parasitism trials, freshly laid beetle eggs were col- P. recurva are equally suitable hosts for the develop- lected from colonies daily between 0700 and 0830 h. ment of A. longoi, and (5) to determine the effects of Therefore, the youngest Phoracantha eggs (0.5 day old) host egg age on parasitoid oviposition rates and devel- were Յ12 h old when used in parasitism bioassays. Egg opment and survival of immature parasitoids. masses were divided to achieve an egg mass size of HOST SUITABILITY FOR Avetianella longoi 97

10–20 eggs and held at 22°C and L14:D10 in covered associated with designing parasitoid behavioral assays petri dishes until used in the bioassays. To simulate (van Alphen and Jervis, 1996). After a 3-h oviposition the range of host egg ages that parasitoid females period (0900–1200 h), the egg masses were removed, would encounter in the field, we presented eggs that labeled, and held together in a 10-cm-diameter petri were 0.5, 1.5, 2.5, 3.5, and 4.5 days old to female A. dish at 22°C. Parasitoid oviposition and emergence longoi. (measured as described above), eclosion of beetle lar- A. longoi colonies were reared as described by Hanks vae, and egg death were recorded for each trial. Beetle et al. (1995). Briefly, A. longoi colonies were held in neonates were removed from the egg masses. For all 10-cm-diameter, cylindrical plastic cages with small, trials, there were four potential fates for the parasit- screen-covered holes on the top for ventilation. Adult ized beetle eggs: (1) beetle larvae eclosed, (2) parasi- parasitoids were fed honey applied to the surface of one toids emerged, (3) parasitoids developed but then dia- screened hole. Parasitoids were added to the colonies paused (eggs were held until the parasites emerged to by placing filter paper slips containing A. longoi pupae confirm wasp adult viability), or (4) both host and in the colony cage 1–2 days before adult eclosion. Adult parasitoids died. For the initial run of bioassays (n ϭ 5 parasitoids emerged and mated, and females began trials for each host egg age), egg masses were not held ovipositing Ϸ1 day after eclosion. To continue wasp individually through parasitoid emergence, thus, it rearing, filter paper slips containing freshly laid beetle was not possible to determine sex ratio of the emerged ϭ egg masses were placed in the A. longoi colony cages for parasitoids. For the remaining trials (n 15 trials for 1.5–2 days, removed, and held in covered petri dishes each host egg age), the parasitized egg masses were until the parasitoids darkened. Phoracantha spp. neo- held individually after beetle eclosion until adult para- nates were removed from the dish containing the par- sitoids emerged and the sex ratio of emerged parasi- asitized eggs. The parasitoids developed from egg to toids was calculated for each egg mass. adult in Ϸ14 days at 22°C. For all parasitism trials, A. Choice tests. For each trial (n ϭ 15), four naı¨ve longoi females were reared from P. semipunctata eggs. gravid A. longoi females were placed in a 5-cm-diame- Adult female parasitoids were Ͻ5 days old, had no ter petri dish and presented with two egg masses (one prior experience with Phoracantha eggs when they mass of P. recurva eggs and one mass of P. semipunc- were used in the bioassays, and were used only once. tata eggs; 10–20 eggs per mass; host egg density per All parasitism trials took place in 5-cm-diameter petri parasitoid female maintained constant in both choice dishes with locking lids that prevented parasitoid es- and no-choice experiments). A. longoi females were cape. Preliminary investigations demonstrated that held with the host eggs for 3 h (0900–1200 h), and then the egg pedicel does not deteriorate over the time pe- the eggs were removed, labeled, and held together in a riod of the experiments. Therefore, by externally exam- 9-cm-diameter petri dish until the remaining unpara- ining the parasitized host eggs under magnification, sitized eggs hatched. Parasitoid oviposition and emer- we obtained a paired data set for oviposition and de- gence, beetle larval eclosion, egg death, and sex ratio of velopment in the two Phoracantha spp. eggs. A. longoi emerging parasitoids were assessed as described egg pedicels protruding through the chorion of the host above. Egg masses from all trials were held individually after beetle eclosion until parasitoids emerged. For egg were counted 1–2 days postparasitism to deter- all trials, the fate of the parasitized eggs was determine parasitism rates. Survival rates of parasitoids mined as described above. were determined from the number of emerging adults. Statistics. Differences among treatment means No-choice oviposition bioassays. For each parasit- were tested by analysis of variance (Sokal and Rohlf, ism trial, two naı¨ve gravid A. longoi females were 1981) using PROC GLM and least squares means sep- placed in a 5-cm-diameter petri dish and presented aration and an ␣ Ͻ 0.05 (SAS Institute, 1996). Repli- with an egg mass (10–20 eggs) of either P. recurva or P. cates were blocked in each experiment to control for semipunctata on a filter paper slip. Based on previous overall changes in levels of parasitism at different studies (Hanks et al., 1995), the number of eggs re- times. Proportional data (oviposition rate, parasitoid flected the average daily oviposition rate of young fe- survival, beetle survival, and egg death) were arcsine male parasitoids rather than the average size of a transformed prior to analysis (Sokal and Rohlf, 1981). beetle egg mass to maximize use of available eggs as Count data (number of parasitoid eggs per host egg) well as restrict levels of “over stinging” of host eggs. In were square-root (Y ϩ 0.5) transformed prior to anal- addition, behavioral observations of ovipositing fe- ysis (Sokal and Rohlf, 1981). males have demonstrated that host feeding is an infre- quent event, but when it does occur, females will feed RESULTS on fluids from the egg in which they have just oviposited (unpublished data). Consequently, the number of Field parasitism of beetle eggs by A. longoi. P. re- host eggs provided to the females in each replicate was curva egg masses laid in the bark sandwiches had determined to be adequate to avoid potential problems significantly more eggs per mass (36.4 Ϯ 3.4 eggs, 98 LUHRING ET AL.

TABLE 1 Percentage of Phoracantha spp. Eggs per Egg Mass Parasitized by A. longoi, Number of Parasitoid Eggs Deposited per Host Egg, Survival of Parasitoid Eggs, and Sex Ratio of Emerged Parasitoids for the No-Choice Experiments

Host egg age (days after deposition) Host egg species 0.5 1.5 2.5 3.5 4.5

% Phoracantha spp. eggs P. recurva 52.6 ϩ 7.01,a 21.5 ϩ 7.3b 9.5 ϩ 4.0b 10.6 ϩ 2.4b 3.4 ϩ 2.0b parasitized per mass P. semipunctata 64.0 ϩ 0.07a 51.1 ϩ 8.5a 65.5 ϩ 6.6a 51.5 ϩ 7.2a 59.5 ϩ 7.8a **** No. A. longoi eggs P. recurva 1.34 ϩ 0.06a 1.34 ϩ 0.10a 1.04 ϩ 0.03b 1.02 ϩ 0.02b NA3 20 oviposited per host egg P. semipunctata 1.71 ϩ 0.05a 1.77 ϩ 0.14a 1.33 ϩ 0.06b 1.21 ϩ 0.04b 1.29 ϩ 0.07b *2 *** % A. longoi to survive P. recurva 34.3 ϩ 7.6a 9.4 ϩ 4.1a 40.9 ϩ 12.0a 37.7 ϩ 11.7a NA from egg to adult P. semipunctata 87.7 ϩ 4.6a 90.3 ϩ 2.4a 85.8 ϩ 3.4a 62.6 ϩ 8.9b 61.2 ϩ 8.9b **** Sex ratio of emerged A. P. recurva 54.5 ϩ 11.9a NA NA 58.3 ϩ 15.4a NA longoi (proportion females) P. semipunctata 70.7 ϩ 7.6a,b 85.4 ϩ 4.2a 78.6 ϩ 2.7a,b 72.4 ϩ 4.4a,b 55.4 ϩ 10.9b

1 All values are untransformed means ϩ SE. Statistics are based on the transformed data. Lower case letters indicate significant differences within host species (P Ͻ 0.05; least squares means comparisons). 2 Asterisks (*) indicate significant differences between the two host species at each egg age (P Ͻ 0.05; least squares means comparisons). 3 NA indicates that statistics were not calculated because sample sizes were too small (n Ͻ 5). mean Ϯ SD) than P. semipunctata egg masses (26.2 Ϯ A. longoi females allocated significantly more eggs 3.6 eggs) (t ϭ 2.22, df ϭ 145, P Ͻ 0.03), but this did per P. semipunctata host egg than per P. recurva host not appear to affect host finding by the parasitoid. Egg egg (Table 1; Species effect: F ϭ 44.3, df ϭ 1, P ϭ masses of either host species placed at the two field 0.0001, Time effect: F ϭ 17.55, df ϭ 4, P ϭ 0.0001). sites in bark sandwiches were located readily by A. Egg allocation rate was not compared between the two longoi, with no significant difference in the mean pro- species for 4.5-day-old eggs because A. longoi females portions of P. recurva (49.9% Ϯ 13.7, mean Ϯ SE) and oviposited in eggs of only 26% (n ϭ 4) of these P. P. semipunctata (74.6% Ϯ 13.4) egg masses located recurva egg masses. For P. semipunctata eggs, female (F ϭ 2.8, df ϭ 1, P Ͻ 0.128). However, A. longoi parasitoids allocated significantly more eggs per host parasitized significantly fewer P. recurva eggs per egg when the host eggs were 0.5 and 1.5 days old than mass (11.4% Ϯ 2.6, mean Ϯ SE) than P. semipunctata when the host eggs were older. Similarly, female para- eggs per mass (30.3% Ϯ 6.9) (F ϭ 13.04, df ϭ 1, P Ͻ sitoids oviposited more in younger P. recurva eggs than 0.0006). in older P. recurva eggs. No-choice tests. When P. semipunctata and P. re- Parasitoid survival was significantly higher in the curva eggs were 0.5 day old, the percentage of beetle eggs of P. semipunctata than in those of P. recurva eggs parasitized per mass was not significantly differ- when the host eggs were Ͻ4.5 days old at parasitism. ent between the two species, but for all other egg age A. longoi survival was Ϸ90% in P. semipunctata eggs classes (1.5, 2.5, 3.5, and 4.5 days) significantly more P. Ͻ3.5 days old compared to Ϸ40% or less in P. recurva semipunctata eggs per mass were parasitized (Table 1; eggs (Table 1; Species ϫ Time effect: F ϭ 3.2, df ϭ 3, Species effect: F ϭ 83.7, df ϭ 1, P ϭ 0.0001; Time P ϭ 0.03). Survival rate decreased significantly to effect: F ϭ 5.87, df ϭ 4, P ϭ 0.0002). Comparing egg Ϸ60% in older P. semipunctata eggs. Survival of para- age classes within beetle species, the percentage of sitoids in P. recurva eggs did not change with host egg parasitized P. recurva eggs per mass decreased dra- age. As previously described, differences in survival matically after the 1st day and then was not signifi- rates between parasitoids in 4.5-day-old eggs were not cantly different for the 1.5-, 2.5-, 3.5-, and 4.5-day-old compared statistically because of the low rates of par- eggs (Table 1). Conversely, there was no decrease in asitism in 4.5-day-old P. recurva egg masses. the percentage of P. semipunctata eggs parasitized per We were unable to test for species differences in the egg mass for the five age classes (oviposition ranged sex ratio of A. longoi progeny because very few P. between 52 and 65% for all five ages of host eggs; Table recurva eggs produced viable adults (Table 1). The 1). emerging parasitoid sex ratio from P. semipunctata HOST SUITABILITY FOR Avetianella longoi 99 eggs was female biased (70–85% female) for eggs Ͻ4.5 very similar to that for the no-choice experiments. Fe- days old. For the 4.5-day-old eggs, the parasitoid sex male parasitoids allocated significantly more eggs ratio was 55% female and significantly lower than that (Ϸ1.7 eggs per host egg) to younger P. semipunctata for the 1.5-day-old eggs. eggs. There was no significant trend in the rate of Neonate beetle larvae eclosed only from parasitized parasite egg allocation for P. recurva (Table 2), and eggs of P. recurva (Fig. 1). After the neonate P. recurva small sample sizes resulted in large standard errors eclosed, dead parasitoids were visible as brown 1st around the means. instar larvae attached to the host chorion. A higher The pattern of parasitoid survival in the choice ex- percentage of parasitized P. recurva eggs than parasit- periment was very similar to that in the no-choice ized P. semipunctata eggs died producing neither bee- experiment. Significantly more parasitoids survived in tles nor parasitoids when the host eggs were Ͻ4.5 days P. semipunctata eggs than in P. recurva eggs (Table 2; old at parasitism (Fig. 1; Species effect: F ϭ 21.1, df ϭ Species effect: F ϭ 216.74, df ϭ 1, P ϭ 0.0001; Time 1, P ϭ 0.0001; Time effect: F ϭ 3.89, df ϭ 4, P ϭ effect: F ϭ 3.91, df ϭ 4, P ϭ 0.006). These data 0.005). The percentage of P. semipunctata eggs dying suggest that P. semipunctata eggs are more suitable increased with egg age (Fig. 1). For P. recurva eggs, egg hosts than P. recurva eggs for development of A. longoi. death did not show any overall trends except that No comparison was possible for the survival of parasi- significantly fewer eggs died when they were parasit- toids in the 3.5- or 4.5-day-old P. recurva eggs because ized at 1.5 days old compared to that at 3.5 days old of the low parasitism rate and the fact that all para- (t ϭ 2.1, P ϭ 0.04; least squares means comparison). sitized eggs in the 4.5-day-old egg masses died without Parasitoids diapaused only in 1.5-day-old P. recurva producing either parasitoids or neonate larvae. eggs and in 1.5- to 3.5-day-old P. semipunctata eggs We were again unable to test for species differences (Fig. 1; Time effect: F ϭ 2.71, df ϭ 4, P ϭ 0.03, in the sex ratio of A. longoi progeny because very few P. Species ϫ Time effect: F ϭ 4.13, df ϭ 3, P ϭ 0.008). recurva eggs produced viable adults (Table 2). In P. semipunctata eggs, the emerging parasitoid sex ratio Choice tests. Across all host egg age classes, when was only slightly female biased (56–71% female) and female A. longoi were presented with both P. semipunc- not significantly different across all egg age classes. tata and P. recurva eggs of the same age, they prefer- For all age classes, the vast majority of parasitized P. entially chose P. semipunctata eggs as hosts (Table 2; semipunctata eggs produced adult parasitoids (Fig. 2). Species effect: F ϭ 252.8, df ϭ 1, P ϭ 0.0001, Time In contrast, for P. recurva eggs older than 0.5 day old, effect: F ϭ 3.61, df ϭ 4, P ϭ 0.008). When P. recurva Ͼ Ͼ few adult parasitoids were produced, and 40% of the eggs were 2.5 days old, less than 10% of the eggs per parasitized P. recurva eggs died, producing neither egg mass were parasitized (Table 2). When P. semi- parasitoids nor beetle larvae (Fig. 2; Species effect: F ϭ punctata eggs were 2.5 days old, the percentage of eggs 191.5, df ϭ 1, P ϭ 0.0001; Time effect: F ϭ 3.33, parasitized per egg mass was significantly lower than df ϭ 4, P ϭ 0.0001). As in the no-choice experiments, that at any other age class. At all other egg age classes, neonate beetle larvae eclosed only from parasitized P. approximately 70–83% of P. semipunctata eggs per egg recurva eggs. mass were parasitized. There was higher overall parasitism of P. semipunctata eggs in the choice experi- DISCUSSION ments than in the no-choice experiments, possibly a result of the increased number of female parasitoids Parasitoid foraging behavior has been divided into per dish during the choice experiments. The choice five steps: host habitat location, host finding, host ac- experiments were designed to have the same parasi- ceptance, host suitability, and host regulation (Doutt, toid/host ratio as that in the no-choice experiment, but 1959; Vinson, 1975). However, a more functional or may have resulted in a greater “effective” parasitoid natural selection-based approach to parasitoid behav- ratio, because one of the two available egg masses (i.e., ior would attempt to address the evolutionary conse- P. recurva) was of poor host quality. quences of specific behaviors and the selection pres- For host eggs 0.5–2.5 days old when parasitized, A. sures that shape them (van Alphen and Jervis, 1996). longoi females allocated significantly more eggs per P. Our studies addressed two fundamental questions: (1) semipunctata host egg than per P. recurva host egg whether there is a difference in the allocation of eggs (Table 2; Species effect: F ϭ 48.91, df ϭ 1, P ϭ by the parasitoid when provided these two congeneric 0.0001; Time effect: F ϭ 5.86, df ϭ 4, P ϭ 0.0003). beetle hosts and (2) whether there is a fitness conse- We did not test for interspecific differences in the num- quence to the host selection behavior. We have demon- ber of parasitoid eggs per host egg for 3.5- to 4.5-day- strated that parasitoids find egg masses of the two old eggs because wasps oviposited in only 33% (n ϭ 5) beetle species at the same rate in the field, but that and 13% (n ϭ 2) of the 3.5- and 4.5-day-old P. recurva parasitism levels in field-collected P. semipunctata egg masses, respectively. For P. semipunctata eggs, the eggs were significantly higher than those in field-col- pattern of egg allocation across the egg age classes was lected P. recurva eggs. This discrimination between 100 LUHRING ET AL.

FIG. 1. Fate of parasitized (A) P. recurva and (B) P. semipunctata eggs in the no-choice experiment. Bars show the percentage of parasitized eggs that developed into beetles ■, adult parasites ᮀ, diapause p, or eggs that died prior to either parasite emergence or beetle eclosion s. HOST SUITABILITY FOR Avetianella longoi 101

TABLE 2 Percentage of Phoracantha spp. Eggs per Egg Mass Parasitized by A. longoi, Number of Parasitoid Eggs Deposited per Host Egg, Survival of Parasitoid Eggs, and Sex Ratio of Emerged Parasitoids for the Choice Experiments

Host egg age (days after deposition) Host egg species 0.5 1.5 2.5 3.5 4.5

% Phoracantha spp. eggs P. recurva 23.4 ϩ 7.31,a 16.7 ϩ 5.8a 9.5 ϩ 4.4a,b 4.0 ϩ 1.9b 1.0 ϩ 0.7b parasitized per mass P. semipunctata 83.3 ϩ 5.5a 70.5 ϩ 8.9a 54.9 ϩ 6.1b 80.5 ϩ 4.6a 68.4 ϩ 7.6a,b *2 **** No. A. longoi eggs P. recurva 1.20 ϩ 0.06a 1.06 ϩ 0.13a 1.02 ϩ 0.02a NA3 NA oviposited per host egg P. semipunctata 1.79 ϩ 0.13a 1.73 ϩ 0.10a 1.35 ϩ 0.07b 1.45 ϩ 0.06b 1.33 ϩ 0.09b *** % A. longoi to survive P. recurva 34.8 ϩ 11.9a 9.4 ϩ 4.1b 15.5 ϩ 8.4b NA NA from egg to adult P. semipunctata 97.1 ϩ 1.2a 90.8 ϩ 2.9a,b 93.6 ϩ 1.9a,b 86.1 ϩ 4.9a,b 83.7 ϩ 3.9b **** Sex ratio of emerged A. P. recurva 77.7 ϩ 8.2 NA NA NA NA longoi (proportion females) P. semipunctata 66.3 ϩ 3.0a 56.4 ϩ 6.8a 67.4 ϩ 4.9a 71.5 ϩ 4.6a 63.9 ϩ 6.7a *

FIG. 2. Fate of parasitized (A) P. recurva and (B) P. semipunctata eggs in the choice experiment. Bars show the percentage of parasitized eggs that developed into beetles ■, adult parasites ᮀ, diapause p, or eggs that died prior to either parasite emergence or beetle eclosion s. HOST SUITABILITY FOR Avetianella longoi 103 semipunctata eggs. The vast majority of parasitized P. more fundamental interest, if A. longoi is much more semipunctata eggs produced a parasitoid, with only a effective against P. semipunctata than against P. re- small proportion of host eggs being killed and produc- curva, then the parasitoid may actually be a principal ing neither beetle nor wasp. In contrast, relatively few selective force in the ecological replacement of P. semi- P. recurva eggs produced adult parasitoids and a sig- punctata by P. recurva. That is, if the parasitoid is nificant proportion of P. recurva eggs died, producing preferentially parasitizing P. semipunctata, resulting neither a beetle larva nor an adult parasitoid. Further- in a reduction of populations of that species, then P. more, another large proportion of P. recurva eggs sur- recurva may face reduced interspecific competition for vived parasitism. Differences between the two beetle the same host material. species in their suitability as hosts for the parasitoid A few studies have implicated parasitoids as driving may be related to physiological incompatibility be- the success or failure of an introduced herbivore spe- tween host and parasitoid. Parasitoids may inject bio- cies (see Lawton, 1986). For example, Settle and Wil- chemical substances during oviposition that halt devel- son (1990) reported that preference of a parasitoid for a opment of the host (Strand et al., 1980). Incompatibil- native grape leafhopper reduced these leafhopper pop- ity of such substances with P. recurva eggs could ulations to low levels, thus allowing another leafhop- explain why a significant proportion of P. recurva em- per to become the dominant species in vineyards. Our bryos in parasitized eggs continued to develop and data strongly suggest that differential parasitism may produced beetle neonates. be one top-down factor driving the replacement of P. In some parasitized P. recurva eggs, the parasitoid semipunctata by P. recurva. The combined effects upon embryos turned brown and subsequently died. We ob- P. semipunctata of being chosen for parasitism more served these brown parasitoid larvae inside both dead frequently and parasitoid attack causing 100% mortal- P. recurva eggs and live P. recurva eggs that produced ity to the eggs strongly indicates that the effect of the viable beetle larvae, but never inside P. semipunctata parasitoid is much greater on P. semipunctata than on eggs. In addition, egg dissections demonstrated that P. recurva. In fact, parasitoid survival rates from P. this suspected melanization of the parasitoid larvae recurva eggs were so low that they might actually act was initiated while the larvae were alive (unpublished as a sink for A. longoi eggs. observation). Melanization is a common defense reac- Other factors may exacerbate the effects of differen- tion to parasitism by host larvae and is known to occur tial parasitism in the replacement of one cerambycid for several encyrtid parasitoids inside their larval species for another in this novel environment. For ex- hosts (Bartlett and Ball, 1966; Blumberg et al., 1993). ample, P. recurva appears to have both a shorter de- Salt (1938) stated that cellular responses of eggs to velopmental time and a longer annual activity period parasitism were not possible, and we have not found than P. semipunctata (unpublished data) which, in any published reports of melanization reactions occur- combination, lead to inherently faster rates of increase ring within eggs in response to egg parasitoids. for P. recurva. There also may be more subtle effects of When the biological control program for P. semipunc- P. recurva adults emerging earlier in the spring. This tata began in southern California, it was the only in- reproductive head start could allow larvae to exploit troduced longhorned borer species attacking eucalypts the shared host resources sooner and more effectively, (Paine et al., 1993). Introduction of natural enemies in particularly as the larvae of both species are cannibal- conjunction with urban forest management strategies istic, with larger, older larvae frequently killing appear to have significantly decreased P. semipunctata smaller larvae in densely populated host logs. populations and rates of host tree mortality. The de- With the decreased competitive pressure exerted by cline in tree mortality may be attributed, at least in P. semipunctata, P. recurva populations should con- part, to the high rates of parasitism of P. semipunctata tinue to increase. The integrated management pro- eggs by A. longoi (Hanks et al., 1996, 1998a). However, gram developed to reduce risk to individual trees from the appearance of P. recurva in southern California attack by both species of beetles must be maintained to and its increasing abundance relative to the decreasing reduce tree mortality throughout California. In south- size of P. semipunctata populations suggest that mor- ern California, however, the implication from A. longoi tality rates of eucalypts are likely to increase. having a distinct preference for P. semipunctata and The ability of A. longoi to attack P. recurva has never reduced survival in P. recurva eggs is that introduction been assessed. It is critically important to understand of a more effective parasitoid for P. recurva is required the fundamental behavioral ecology of the interactions for effective biological control of this beetle complex. between the beetles and the parasitoid to assess the comprehensive impact on the biological control pro- ACKNOWLEDGMENTS gram being conducted in California against both Pho- racantha spp. If A. longoi is not effective against P. We thank Chris Campbell, Jose Arredondo, and the Eucalyptus recurva, then other biological control agents must be longhorned borer lab at University of California, Riverside for their found to reduce the mortality of eucalyptus trees. Of expertise in rearing the beetle and wasp colonies. Dr. Stuart Reitz 104 LUHRING ET AL. provided many comments on the statistics and experimental design. Pak, G. A. 1986. Behavioural variations among strains of Tri- We also thank the College of Natural and Agricultural Sciences at chogramma spp.: A review of the literature on host age selection. University of California, Riverside for financial assistance. J. Appl. Entomol. 101, 55–64. Powell, W. 1978. Colonization of twelve species of Eucalyptus by REFERENCES Phoracantha semipunctata (F.) (Coleoptera: Cerambycidae) in Malawi. Bull. Entomol. Res. 68, 621–626. Bartlett, B. R., and Ball, J. C. 1966. The evolution of host suitability Ruberson, J. R., Tauber, M. J., and Tauber, C. A. 1987. Biotypes of in a polyphagous parasite with special reference to the role of Edovum puttleri (Hymenoptera: Eulophidae): Responses to devel- parasite egg encapsulation. Ann. Entomol. Soc. Amer. 59, 42–45. oping eggs of the Colorado potato beetle (Coleoptera: Chrysomeli- Blumberg, D., Wysoki, M., and Hadar, D. 1993. Further studies of dae). Ann. Entomol. Soc. Am. 80, 451–455. encapsulation of eggs of Metaphycus spp. (Hym.: Encyrtidae) by Salt, G. 1935. Experimental studies in insect parasitism. III. Host the pyriform scale, Protopulvinaria pyriformis (Hom.: Coccidae). selection. Proc. R. Entomol. Soc. London 117, 414–435. Entomophaga. 38, 7–13. Salt, G. 1938. Experimental studies in insect parasitism. VI. Host Clausen, C. P. 1962. “Entomophagous Insects.” Hafner, New York. suitability. Bull. Entomol. Res. 29, 223–246. Doutt, R. L. 1959. The biology of parasitic Hymenoptera. Annu. Rev. SAS Institute. 1996. “SAS. User’s Guide for Statistics.” SAS Insti- Entomol. 4, 161–182. tute. Cary, NC. Drinkwater, T. W. 1975. The present pest status of eucalyptus borers Scriven, G. T., Reeves, E. L., and Luck, R. F. 1986. Beetle from Phoracantha spp. in South Africa. In “Proceedings, 1st Congress of Australia threatens Eucalyptus. Calif. Agric. 40, 4–6. the Entomological Society of Southern Africa 1975,” pp. 119–129. Entomological Society of Southern Africa, Pretoria. Settle, W. H., and Wilson, L. T. 1990. Invasion by the variegated leafhopper and biotic interactions: Parasitism, competition and Hanks, L. M., Paine, T. D., and Millar, J. G. 1991. Mechanisms of apparent competition. Ecology 71, 1461–1470. resistance in Eucalyptus against larvae of the Eucalyptus longhorned borer (Coleoptera: Cerambycidae). Ann. Entomol. Soc. Am. Sokal R. R., and Rohlf, F. J. 1981. “Biometry.” Freeman, New 20, 1583–1588. York. Hanks, L. M., Paine, T. D., and Millar, J. G. 1996. Tiny wasp helps Stevens, P. S. 1995. Host preferences of Trichogrammatoidae bactrae protect eucalypts from eucalyptus longhorned borer. Calif. Agric. fumata (Hym.: Trichogrammatidae) an egg parasitoid of leafrollers 50, 14–16. (Lep.: Tortricidae). Entomophaga 40, 379–385. Hanks, L. M., Millar, J. G., and Paine, T. D. 1998a. Aerial dispersal Strand, M. R. 1986. The physiological interactions of parasitoids of the eucalyptus longhorned borer (Coleoptera: Cerambycidae) in with their hosts and their influence on reproductive strategies. In an urban landscape. Environ. Entomol. 27, 1418–1424. “Insect Parasitoids” (J. Waage and D. Greathead, Eds.), pp. 97– Hanks, L. M., Gould, J. R., Paine, T. D., Millar, J. G., and Wang, Q. 136. Academic Press, New York. 1995. Biology and host relations of Avetianella longoi (Hymen- Strand, M. R., and Vinson, S. B. 1983. Factors affecting host recog- optera: Encyrtidae), an egg parasitoid of the eucalyptus long- nition and acceptance in the egg parasitoid Telenomus heliothidis horned borer (Coleoptera: Cerambycidae). Ann. Entomol. Soc. Am. (Hymenoptera: Scelionidae). Environ. Entomol. 12, 1114–1119. 88, 666–671. Strand, M. R., Ratner, S., and Vinson, S. B. 1980. Maternally in- Hanks, L. M., McElfresh, J. S., Millar, J. G., and Paine, T. D. 1993. duced host regulation by the egg parasitoid Telenomus heliothidis. Phoracantha semipunctata (Coleoptera: Cerambycidae), a serious Physiol. Entomol. 8, 469–475. pest of Eucalyptus in California: Biology and laboratory-rearing procedures. Ann. Entomol. Soc. Am. 86, 96–102. van Alphen, J. J. M., and Jervis, M. A. 1996. Foraging behaviour. In “Insect Natural Enemies. Practical Approaches to Their Study and Hanks, L. M., Paine, T. D., Millar, J. G., and Campbell, C. 1998b. Evaluation” (M. Jervis and M. Kidd, Eds.), pp. 1–62. Chapman & Another tree-killing pest of eucalyptus invades California. Calif. Hall, London Plant Pest Disease Rep. 16, 19–21. van Dijken, J. M., Kole, K., van Lenteren, J. C., and Brand, A. M. 1986. Ivory, M. H. 1977. Preliminary investigations of the pests of exotic forest trees in Zambia. Comm. For. Rev. 56, 47–56. Host preference studies with Trichogramma evanescens Westwood (Hymenoptera: Trichogrammatidae) for Mamestra brassica, Pieris Lawton, J. H. 1986. The effect of parasitoids on phytophagous insect brassicae and Pieris rapae. J. Appl. Entomol. 101, 64–85. communities. In “Insect Parasitoids” (J. Waage and D. Greathead, Eds.), pp. 265–287. Academic Press, New York. Vinson, S. B. 1975. Biochemical coevolution between parasitoids and their hosts. In “Evolutionary Strategies of Parasitic Insects and Laing, J. 1937. Host finding by insect parasites. I. Observations on finding of hosts by Alysia manducator, Mormoniella vitripenoris, Mites” (P. W. Price, Ed.), pp. 14–48. Plenum, New York. and Trichogramma evanescens. J. Anim. Ecol. 6, 298–317. Vinson, S. B. 1994. Physiological interactions between egg parasi- Longo, S., Palmeri, V., and Sommariva, D. 1993. Sull’attivita di toids and their hosts. In “Biological Control with Egg Parasitoids” Avetianella longoi ooparassitoide di Phoracantha semipunctata (E. Wajnberg and S. A. Hassan, Eds.), pp. 210–217. CAB Interna- nell’Italia meridionale. Redia 76, 223–239. tional, Wallingford. Mendel, Z. 1987. Major pests of man-made forests in Israel: Origin, Vinson, S. B., and Iwantsch, G. F. 1980. Host suitability for insect biology, damage and control. Phytoparasitica 15, 131–137. parasitoids. Annu. Rev. Entomol. 25, 397–419. Paine, T. D., Millar, J. G., Bellows, T. S., Hanks, L. M., and Gould, Wang, Q. 1995. A taxonomic revision of the Australian genus Pho- J. R. 1993. Integrating classical biological control with plant racantha Newman (Coleoptera, Cerambycidae). Invertebr. Taxon. health in the urban forest. J. Arboric. 19, 125–130. 9, 865–895