Role of Emerald Ash Borer (Coleoptera: Buprestidae) Larval Vibrations in Host-Quality Assessment by Tetrastichus Planipennisi (Hymenoptera: Eulophidae)

ECOLOGY AND BEHAVIOR Role of Emerald Ash Borer (Coleoptera: Buprestidae) Larval Vibrations in Host-Quality Assessment by Tetrastichus planipennisi (Hymenoptera: Eulophidae)

MICHAEL D. ULYSHEN,1,2 RICHARD W. MANKIN,3 YIGEN CHEN,1 JIAN J. DUAN,4 1,5 1,5 THERESE M. POLAND, AND LEAH S. BAUER

J. Econ. Entomol. 104(1): 81Ð86 (2011); DOI: 10.1603/EC10283 ABSTRACT The biological control agent Tetrastichus planipennisi Yang (Hymenoptera: Eulophi- dae) is a gregarious larval endoparasitoid of the emerald ash borer, Agrilus planipennis Fairmaire (Coleoptera: Buprestidae), an invasive cambium-feeding species responsible for recent, widespread mortality of ash (Fraxinus spp.) in North America. T. planipennisi is known to prefer late-instar emerald ash borer, but the cues used to assess host size by this species and most other parasitoids of concealed hosts remain unknown. We sought to test whether vibrations produced by feeding emerald ash borer vary with larval size and whether there are any correlations between these cues and T. planipennisi progeny number (i.e., brood size) and sex ratio. The amplitudes and rates of 3Ð30-ms vibrational impulses produced by emerald ash borer larvae of various sizes were measured in the laboratory before presenting the larvae to T. planipennisi. Impulse-rate did not vary with emerald ash borer size, but vibration amplitude was signiÞcantly higher for large larvae than for small larvae. T. planipennisi produced a signiÞcantly higher proportion of female offspring from large hosts than small hosts and was shown in previous work to produce more offspring overall from large hosts. There were no signiÞcant correlations, however, between the T. planipennisi progeny data and the emerald ash borer sound data. Because vibration amplitude varied signiÞcantly with host size, however, we are unable to entirely reject the hypothesis that T. planipennisi and possibly other parasitoids of concealed hosts use vibrational cues to assess host quality, particularly given the low explanatory potential of other external cues. Internal chemical cues also may be important.

KEY WORDS Agrilus planipennis, auditory cues, cryptic, vibrokinesis, vibrotaxis

The emerald ash borer, Agrilus planipennis Fairmaire reared and released are Oobius agrili Zhang & Huang (Coleoptera: Buprestidae), a cambium-feeding bu- (Hymenoptera: Encyrtidae), an egg parasitoid prestid beetle native to Asia, was Þrst detected near (Zhang et al. 2005); Tetrastichus planipennisi Yang Detroit, MI, and Windsor, ON, Canada, in 2002 (Haack (Hymenoptera: Eulophidae), a gregarious larval en- et al. 2002). Since then, the species has killed millions doparasitoid (Yang et al. 2006); and Spathius agrili of ash (Fraxinus spp.) trees in northeastern North Yang (Hymenoptera: Braconidae), a gregarious larval America and is expected to kill many millions more ectoparasitoid (Yang et al. 2005). over the next decade (Kovacs et al. 2010). Conse- Emerald ash borers spend most of their lives as quently, controlling emerald ash borer is currently larvae concealed beneath the bark, passing through one of the most pressing challenges facing forest en- four instars within the phloem layer. T. planipennisi tomologists on the continent. A classical biological and S. agrili have been shown to allocate more eggs control program involving three hymenopteran para- (both species) and a higher proportion of females (S. sitoid species associated with emerald ash borer in agrili) to larger (i.e., higher quality) emerald ash borer Asia is currently under way in the northeastern United larvae and may parasitize late instars at a higher rate States (USDAÐAPHIS 2007). The three species being (T. planipennisi) than early instars (Liu et al. 2007, Wang et al. 2008, Ulyshen et al. 2010b). These Þndings 1 Department of Entomology, Michigan State University, East Lan- are not surprising given the selective advantage of host sing, MI 48824. quality discrimination in parasitoids (Charnov et al. 2 Current address: USDAÐForest Service, Southern Research Sta- tion, Starkville, MS 39759 (e-mail: [email protected]). 1981, Charnov and Skinner 1984, Waage 1986). 3 USDAÐARS, Center for Medical, Agricultural and Veterinary En- Although parasitoids of exposed hosts can directly tomology, Gainesville, FL 32608. discern host size (Dijkstra 1986), T. planipennisi, S. 4 USDAÐARS, BeneÞcial Insects Introduction Research Unit, New- agrili, and other parasitoids of concealed hosts cannot. ark, DE 19713. 5 USDAÐForest Service, Northern Research Station, East Lansing, How parasitoids of concealed hosts assess host quality MI 48823. remains almost entirely unknown even though they 82 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 1 have long been known to do so (e.g., Purrington and width), and inserted into 10-cm-long large-diameter Uleman 1972). External cues that vary with host qual- ash sticks (3.8 Ϯ 0.05 cm; range, 2.6Ð5.1 [note that this ity may be important in some cases. For example, is within the range of diameters used by T. planipennisi Steidle and Fischer (2000) showed the amount of in the Þeld]) and held in an incubator (25ЊC, Ϸ75% feces associated with wheat, Triticum aestivum L., humidity, and a photoperiod of 16:8 [L:D] h) for grains infested by granary weevils, Sitophilus granarius 24Ð36 h before recordings were made. This length of (L.), increased with weevil size and strongly inßu- time was chosen based on the previous observation enced the amount of interest exhibited by a parasitoid that almost all emerald ash borer larvae feed actively of the species. Internal chemical cues also may be within the Þrst day after insertion into sticks (M.D.U., important. For example, the hemolymph of greater personal observations). To reduce fungal growth, the wax moth, Galleria mellonella (L.), pupae has been sticks were Þrst gently scrubbed under running tap shown to contain chemicals that induce oviposition water, sealed at both ends with parafÞn, held in a 0.05% (Hegdekar and Arthur 1973), but whether the con- bleach bath for Ϸ5 min, and rinsed with running tap centrations of such chemicals vary with host size is not water for 15 min. One larva was inserted into a narrow known. groove chiseled beneath a small ßap of bark peeled Because T. planipennisi (Ulyshen et al. 2010a) and from one end of each stick. The bark ßaps were then S. agrili (Wang et al. 2010) do not parasitize inactive held closed over the inserted larvae with thin strips of emerald ash borer larvae, it seems likely that both ParaÞlm. The heads of the inserted larvae faced away species, like other parasitoids of concealed insects from the nearest ends of the sticks (always pointing (Meyho¨fer and Casas 1999, and references therein), downward) to encourage feeding in the direction with rely on vibrations produced during movements or the greatest available host resource. feeding to locate hosts. If such vibrations are impor- Immediately before recording, a drill was used to tant in locating hosts, they also may be used to assess insert a 1.59-mm-diameter titanium bit into the end of host quality. There is some evidence that vibrational each stick near the bark ßap. A steel ßange was at- cues vary with host size in wood-boring beetles. For tached to the bit with a screw, and an accelerometer example, in a comparison of sounds produced by dif- was connected to the ßange by a magnetic attachment ferent Anoplophora glabripennis (Motschulsky) larval (Mankin et al. 2008). The stick was placed on a piece stages, Mankin et al. (2008) found Þrst and second of foam insulation on a table in a quiet ofÞce with air instars produce less vibrational energy on average conditioning turned off to reduce external vibrations. than later instars. To our knowledge, however, no Signals of larval movement and feeding vibrations de- previous efforts have been made to determine tected by the accelerometer were passed through a whether such differences are used by parasitoids to charge ampliÞer and monitored with headphones con- distinguish host size. nected to a dual-channel, digital audio recorder sam- Here, we present the results from a laboratory study pling at 44.1 kHz (24 bits). For each larva tested, an designed to test this idea by exposing T. planipennisi to interval of Ϸ180 s, relatively free of interfering back- emerald ash borer larvae of a wide range of sizes from ground noise, was stored on a computer with digital which accelerometer recordings had been made of signal processing software for further analysis. their movements and feeding activity just before ex- Digital Signal Processing and Classiﬁcation. To fa- posure. We hypothesized that 1) T. planipennisi prog- cilitate identiÞcation of individual sound impulses eny number (i.e., brood size) or the proportion of produced during larval activity (Mankin et al. 2008), female progeny would increase with increasing initial the stored signals were band-pass Þltered between 0.2 host weight, 2) the amplitude or rate of vibrational and 5 kHz by using Raven 1.3 software (Charif et al. impulses produced by feeding emerald ash borer lar- 2008). The impulses were identiÞed and counted us- vae would vary with larval weight, and 3) there would ing customized Digitize, Analyze and Visualize Insect be signiÞcant correlations between T. planipennisi Sounds (DAVIS) software (Mankin et al. 2000). The progeny data and emerald ash borer vibration data. times of impulses and their amplitudes (vibration lev- els measured in dB between 0.2 and four kHz, see Mankin and Benshemesh 2006) were saved in a Materials and Methods spreadsheet. Mean rates of impulses and mean vibra- Parasitoids. A laboratory colony of T. planipennisi, tion amplitudes of impulses were calculated for each originally collected in 2008 from Liaoning province of recording. China, was used in this study. Only naõ¨ve wasps that Parasitoid Exposures. Each stick from which vibra- had not been presented with hosts were used in the tions were detected (n ϭ 94) was placed in a 710-ml experiment. They were 3Ð6 wk old at the time of use clear plastic drinking cup (GFS.com), and Þve female and were presumed to have mated as mating activities T. planipennisi were added. The sticks were held up- were observed almost immediately after female emer- right using tacks punched through the bottoms of the gence and both sexes were held together before being cups. The cup openings were covered by Þne screen exposed to hosts. held in place by lids in which 5.8-cm-diameter circular Emerald Ash Borer Larval Bioassays. In total, 102 openings had been made. The cups were held in an Þeld-collected emerald ash borer larvae at various incubator (25ЊC, Ϸ75% humidity, and a photoperiod stages of development (i.e., secondÐfourth instars) of 16:8 [L:D] h) with drops of honey (i.e., food for the were individually weighed, measured (i.e., head wasps) added to the tops of the screens (Ulyshen et al. February 2011 ULYSHEN ET AL.: HOST VIBRATIONS AND T. planipennisi 83

2010a). After 2 d, all T. planipennisi were removed mine how T. planipennisi progeny number and sex from the cups. If some were in the act of parasitizing ratio varied with emerald ash borer head width, at that time, they were given additional time to Þnish weight, stick diameter, and vibration impulse ampli- before being removed, never taking Ͼ2 h. Based on tude and rate, the same analyses as described above previous experience,2dissufÞcient time to get high were performed on a data set consisting of the 43 parasitism rates at a 5:1 parasitoid:host ratio (M.D.U., emerald ash borer larvae from which T. planipennisi personal observations). We avoided longer exposure progeny data were collected. The larvae were once periods to be more conÞdent about any observed again evenly assigned to size classes (“small” [Ͻ0.031 relationship between initial host weight and wasp g; n ϭ 14], “medium” [0.031Ð0.0534 g; n ϭ 14] and progeny number. Because a group of Þve females “large” [Ͼ0.0534 g; n ϭ 15]) for analyses of variance. instead of a single individual was added to each cup, it is possible that more than one female contributed to Results each parasitized host, but we consider this unlikely considering the short exposure period. Although a 1:1 Vibration Characteristics. Larvae of all tested sizes parasitoid:host ratio would have been preferable, pre- produced 3Ð30-ms impulses of various amplitudes and vious experience has shown that parasitism by a single rates, similar to those in previous studies of insect T. planipennisi female is very unreliable, especially larval vibrations in wood (e.g., Mankin et al. 2008). within a period of only two days (M.D.U., personal Examples of an 80-s period of signals from a 9.5-mg observations). Consequently, using small groups of larva and a 93-mg larva are seen in Fig. 1. Several females and a 2-d exposure period was a necessary closely spaced impulse bursts in the signal trace from compromise given the objectives of the study and the the 93-mg larva are indicated by arrows, where bursts realities of working with T. planipennisi. Two days are deÞned as series of impulses separated by Ͻ0.25 s, after removing the wasps, the sticks were peeled to Mankin et al. (2008). A 60-ms expansion of the Þrst recover the larvae. Each emerald ash borer larva was marked burst is seen in the inset of Fig. 1. The range placed in an individual Falcon petri dish (50 by 9 mm of amplitudes and the intervals between the three with tight-Þt lid) lined with moistened Þlter paper and impulses in the inset is typical of the range observed observed for parasitism. Forty-six (i.e., 49%) of the for other larvae as well. Both of these larvae served as larvae were parasitized, with developing wasp larvae successful hosts for T. planipennisi, with 29 progeny clearly visible beneath the cuticle. After the T. pla- emerging from the 9.5-mg larva and 32 emerging from nipennisi progeny exited the host, they were counted. the 93-mg larva. The frequency spectra of the vibra- Three broods failed to exit the host and were too tion impulses varied according to the density, elastic- decomposed upon dissection to count. The sex ratio ity, and hardness of the ash sticks, as well as their was calculated for 33 broods (i.e., accurate counts diameters, as noted in Mankin et al. (2008). could not be made for all 43 broods due to larval Vibration Impulse Amplitude and Rate in Relation mortality caused primarily by fungus) once the de- to Emerald Ash Borer Size and Stick Diameter. The veloping progeny reached the pupal stage. three parameters of the regression model (i.e., head Data Analysis. T. planipennisi sex ratio, emerald ash width, larval weight and stick diameter) accounted for borer head width and weight, stick diameter, and only 19.3 and 8.7% of the variance in vibration impulse sound impulse rate were log(x), (x2.45 Ϫ 1)/2.45, rate (F ϭ 7.2; df ϭ 3, 90; P Ͻ 0.01) and amplitude (F ϭ square root(x), log(x ϩ 1), and log(x) transformed for 2.9; df ϭ 3, 90; P ϭ 0.04), respectively. There were no normality, respectively (Sokal and Rohlf 1995). The signiÞcant correlations between either vibration vari- transformation for head width was determined using able and head width or larval weight (data not shown), the BoxÐCox power transformation method (Sokal but vibration impulse rate was signiÞcantly correlated and Rohlf 1995). These transformed data were used in with stick diameter (t ϭϪ4.13, P Ͻ 0.01). According all subsequent analyses. All statistical analyses were to analysis of variance (ANOVA), there were no sig- conducted using SAS 9.2 (SAS Institute, Cary, NC). To niÞcant differences in vibration impulse rate among determine how emerald ash borer vibration amplitude larval weight classes (F ϭ 1.1; df ϭ 2, 91; P ϭ 0.4), but and impulse rate varied with head width, weight and there were differences in vibration amplitude among stick diameter, separate multiple linear regression larval weight classes (F ϭ 4.4; df ϭ 2, 91; P ϭ 0.02), analyses were performed on a data set consisting of the being signiÞcantly higher from large hosts than small 94 emerald ash borer larvae that produced detectable hosts (Fig. 2). vibrations. When signiÞcant correlations were not de- Parasitoid Progeny Number and Sex Ratio in Rela- tected, analyses of variance were conducted after tion to Emerald Ash Borer Size, Vibrations, and Stick evenly assigning the larvae to three size classes (i.e., Diameter. The Þve parameters of the regression instars could not be determined with certainty) based model (i.e., head width, larval weight, stick diameter, on weight (“small” [Ͻ0.0236 g; n ϭ 31], “medium” vibration amplitude, and vibration impulse rate) ac- [0.0236Ð0.051 g; n ϭ 31], and “large” [Ͼ0.051 g; n ϭ counted for 26.8 and 27.0% of the variance in progeny 32]) to test the null hypothesis that there were no number (F ϭ 2.71; df ϭ 5, 37; P ϭ 0.03) and sex ratio signiÞcant differences in vibration impulse amplitude (F ϭ 2.00; df ϭ 5, 27; P ϭ 0.11), respectively. There was or rate among the three size classes. If the null hy- a signiÞcant positive correlation between progeny potheses were rejected, means were further separated number and stick diameter (t ϭ 2.59, P ϭ 0.01), and a using the TukeyÐKramer method. Similarly, to deter- strong, although not signiÞcant, positive correlation 84 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 1

Fig. 1. Signal traces of 80-s intervals recorded from a 9.5- and a 93-mg emerald ash borer larva, with three impulse bursts marked by arrows, and a 60-ms inset showing three individual impulses in the Þrst marked burst. between female:male sex ratio and host weight (t ϭ weight. Our second hypothesis that the amplitude or 1.82, P ϭ 0.08). According to ANOVA, there were no rate of vibrational impulses produced by feeding em- signiÞcant differences in progeny number among lar- erald ash borer larvae would vary with larval weight val weight classes (F ϭ 1.65; df ϭ 2, 40; P ϭ 0.2), even also was supported as the amplitude of vibrations pro- though large hosts produced almost 20 more progeny duced by feeding emerald ash borer varied signiÞ- on average than small hosts (Fig. 3). There were dif- cantly between large and small larvae. Our third hy- ferences in sex ratio among larval weight classes (F ϭ pothesis, however, was not supported as there were no 3.24; df ϭ 2, 30; P ϭ 0.05), with a greater proportion signiÞcant correlations between the T. planipennisi of females emerging from large hosts than small hosts progeny data and the emerald ash borer vibration data. (Fig. 3). Consequently, the mechanism by which T. planipennisi assesses host quality remains unclear. We are hesitant, however, to entirely reject the Discussion notion that T. planipennisi and other parasitoids of The results from this and previous studies indicate that T. planipennisi, like many other parasitoid species (Godfray 1994), produces more offspring (Liu et al. 2007, Ulyshen et al. 2010b) and a higher proportion of female offspring from large hosts than small hosts, thereby supporting our Þrst hypothesis that T. planipennisi progeny number or the proportion of female progeny would increase with increasing initial host

Fig. 3. Mean Ϯ SE T. planipennis progeny number (top) Fig. 2. Vibration amplitudes (mean Ϯ SE) produced by and female:male sex ratio (bottom) from emerald ash borer feeding emerald ash borer larvae assigned to three weight larvae belonging to three weight classes. Bars with different classes. Bars with different letters are signiÞcantly different letters are signiÞcantly different based on the TukeyÐKramer based on the TukeyÐKramer mean separation test. mean separation test. February 2011 ULYSHEN ET AL.: HOST VIBRATIONS AND T. planipennisi 85 concealed hosts use sound cues to assess host quality against emerald ash borer. We provide, for example, given that we found sound vibration amplitude to vary the Þrst evidence that the proportion of female prog- with host size. Several aspects of our study may have eny produced by T. planipennisi increases with host limited our ability to detect this relationship, if it exists. size. This Þnding and the previous observation that For example, because accelerometer recordings were signiÞcantly more T. planipennisi progeny are pro- made soon after drilling metal drill bits into the sticks, duced from large hosts than small hosts (Ulyshen et al. the emerald ash borer larvae may not have fed nor- 2010b) underscore the importance of using late instar mally for a period after this handling, particularly emerald ash borer in rearing operations. In addition, considering that disturbances are known to elicit be- consistent with previous work (Ulyshen et al. 2010a), havioral changes in other concealed insects (e.g., Dje- our results show that the number of progeny produced mai et al. 2001). Furthermore, given the high variabil- per larva increases with stick diameter. ity in vibration amplitude observed over the course of In conclusion, this study represents the Þrst docu- the recordings (data not shown), the vibration am- mented effort to determine whether parasitoids of plitude at the time of host acceptance by T. planipen- concealed hosts use variations in larval feeding sounds nisi may have differed considerably from the mean to assess host quality. Because vibration amplitude vibration amplitudes recorded. Furthermore, because varied signiÞcantly with host size, we are hesitant to we used a 5:1 parasitoid:host ratio in our exposures, reject the notion that T. planipennisi and other para- eggs may have been deposited by more than one sitoids of concealed hosts use sound cues to assess host female in some cases, possibly obscuring any relation- quality, particularly given the low explanatory poten- ship between host weight and clutch size. Finally, our tial of other external cues. More research will be sample size may have been too small considering the needed before deÞnitive conclusions can be reached. imprecision of host assessment and egg delivery by If external cues prove to be unimportant to T. pla- parasitoids (Godfray 1994). nipennisi and other parasitoid species in assessing host Because T. planipennisi will only parasitize actively size, internal cues may be used instead (Hegdekar and feeding larvae (Ulyshen et al. 2010a), we suspect that Arthur 1973), assuming they vary with host size. movement and feeding vibrations are the main cues used in host location, if not also for distinguishing between large and small hosts. Vibration is just one of Acknowledgments many external cues parasitoids use to locate concealed We thank Roger Fuester (USDAÐARS), Scott Horn (US- hosts that also may be used to assess host quality, DAÐFS), and two anonymous reviewers for comments that however. Others include olfactory semiochemicals greatly improved the manuscript. (e.g., hosts, host frass, microorganisms within host galleries), visual signals from plant tissue damaged by References Cited host feeding, and infrared radiation (Wang and Yang 2008). In contrast to the explanatory potential of host Broad, G. R., and D.L.J. Quicke. 2000. The adaptive signif- vibration, however, it seems unlikely that any of these icance of host location by vibrational sounding in para- other cues would be very useful to T. planipennisi in sitoid wasps. Proc. R. Soc. B 121: 63Ð84. distinguishing between active and inactive hosts or in Charif, R. A., A. M. Waack, and L. M. Strickman. 2008. assessing host quality. Assuming they can be detected Raven Pro 1.3 userÕs manual. Cornell Laboratory of Or- nithology, Ithaca, NY. by T. planipennisi through intact bark, for example, the Charnov, E. L., R. L. Los-den Hartogh, W. T. Jones, and J. van chemical signals associated with emerald ash borer den Assem. 1981. Sex ratio evolution in a variable envi- frass would not be expected to give a reliable indica- ronment. Nature 289: 27Ð33. tion of host size because emerald ash borer larvae are Charnov, E. L., and S. W. Skinner. 1984. Evolution of host often surrounded by the frass-Þlled galleries of other selection and clutch size in parasitic wasps. Fla. Entomol. larvae feeding nearby. Although frass quantity was 67: 5Ð21. shown to be important to parasitoids of granary wee- Dijkstra, L. J. 1986. Optimal selection and exploitation of vils in assessing host size (Steidle and Fischer 2000), hosts in the parasitic wasp Colpoclypeus ﬂorus (Hym., in that system the weevils were conÞned to individual Eulophidae). Neth. J. Zool. 36: 177Ð301. Djemai, I., J. Casas, and C. Magal. 2001. Matching host re- grains of wheat and the frass was detectable outside actions to parasitoid wasp vibrations. Proc. R. Soc. B 268: the grains due to small holes created by the developing 2403Ð2408. larvae. Godfray, H.C.J. 1994. Parasitoids: behavioral and evolution- Some parasitoid species are known to locate hosts ary ecology. Princeton University Press, Princeton, NJ. through vibrational sounding, whereby wasps tap the Haack, R. A., E. Jendek, H.-P. Liu, K. R. Marchant, T. R. substrate with modiÞed antennal segments and then Petrice, T. M. Poland, and H. Ye. 2002. The emerald ash detect echoes with enlarged subgenual organs in their borer: a new exotic pest in North America. Newsl. Mich. tibia (Broad and Quicke 2000, Vilhelmsen et al. 2001). Entomol. Soc. 47: 1Ð5. The antennae and tibia of T. planipennisi are not ob- Hegdekar, B. M., and A. P. Arthur. 1973. Host hemolymph chemicals that induce oviposition in the parasite Itoplectis viously modiÞed for this purpose (Yang et al. 2006), conquisitor. Can. Entomol. 105: 787Ð793. however, and evidence that vibrational sounding can Kovacs, K. F., R. G. Haight, D. G. McCullough, R. J. Mer- be used to assess host size is lacking. cader, N. W. Siegert, and A. M. Liebhold. 2010. Cost of Several results from this study have important im- potential emerald ash borer damage in U.S. communities, plications for mass rearing T. planipennisi for release 2009Ð2019. Ecol. Econ. 69: 569Ð578. 86 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 1

Liu, H.-P., L. S. Bauer, D. L. Miller, T.-H. Zhao, R.-T. Gao, Ulyshen, M. D., J. J. Duan, L. S. Bauer, and I. Fraser. 2010b. L. Song, Q. Luan, R. Jin, and C. Gao. 2007. Seasonal Suitability and accessibility of immature Agrilus planipen- abundance of Agrilus planipennis (Coleoptera: Bupresti- nis (Coleoptera: Buprestidae) stages to Tetrastichus pla- dae) and its natural enemies Oobius agrili (Hymenoptera: nipennisi (Hymenoptera: Eulophidae). J. Econ. Entomol. Encyrtidae) and Tetrastichus planipennisi (Hymenoptera: 103: 1080Ð1085. Eulophidae) in China. Biol. Control 42: 61Ð71. Vilhelmsen, L., N. Isidoro, R. Romani, H. H. Basibuyuk, and Mankin, R. W., and J. Benshemesh. 2006. Geophone detec- D.L.J. Quicke. 2001. Host location and oviposition in a tion of subterranean termite and ant activity. J. Econ. basal group of parasitic wasps: the subgenual organ, ovi- Entomol. 99: 244Ð250. positor apparatus and associated structures in the Orus- Mankin, R. W., J. Brandhorst-Hubbard, K. L. Flanders, M. sidae (Hymenoptera, Insecta). Zoomorphology 121: 63Ð Zhang, R. L. Crocker, S. L. Lapointe, C. W. McCoy, J. R. 84. Fisher, and D. K. Weaver. 2000. Eavesdropping on in- Waage, J. K. 1986. Family planning in parasitoids: adaptive sects hidden in soil and interior structures of plants. J. patterns of progeny and sex Allocation, pp. 63Ð95. In J. Econ. Entomol. 93: 1173Ð1182. Waage and D. Greathead [eds.], Insect parasitoids. Ac- Mankin, R. W., M. T. Smith, J. M. Tropp, E. B. Atkinson, and ademic, Orlando, FL. D. Y. Jong. 2008. Detection of Anoplophora glabripennis Wang, X., and Z. Yang. 2008. Behavioral mechanisms of par- (Coleoptera: Cerambycidae) larvae in different host asitic wasps for searching concealed insect hosts. Acta trees and tissues by automated analyses of sound-impulse frequency and temporal patterns. J. Econ. Entomol. 101: Ecol. Sin. 28: 1257Ð1269. 836Ð849. Wang, X.-Y., Z.-Q. Yang, H. Wu, and J. R. Gould. 2008. Meyho¨fer, R., and J. Casas. 1999. Vibratory stimuli in host Effects of host size on the sex ratio, clutch size, and size location by parasitic wasps. J. Insect Physiol. 45: 967Ð971. of adult Spathius agrili, an ectoparasitoid of emerald ash Purrington, F. F., and J. S. Uleman. 1972. Brood size of the borer. Biol. Control 44: 7Ð12. parasitic wasp Hyssopus thymus (Hymenoptera: Eulophi- Wang, X-Y., Z.-Q. Yang, J. R. Gould, H. Wu, and J.-H. Ma. dae): functional correlation with the mass of a cryptic 2010. Host-seeking behavior and parasitism by Spathius host. Ann. Entomol. Soc. Am. 65: 280Ð281. agrili Yang (Hymenoptera: Braconidae), a parasitoid of Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman the emerald ash borer. Biol. Control 52: 24Ð29. and Company, New York. Yang, Z.-Q., J. S. Srazanac, P. M. Marsh, C. Van Achterberg, Steidle, J.L.M., and A. Fischer. 2000. Quantity does matter: and W.-Y. Choi. 2005. First recorded parasitoid from how feces are used for host stage selection by granary China of Agrilus planipennis: a new species of Spathius weevil parasitoid Lariophagus distinguendus. J. Chem. (Hymenoptera: Braconidae: Doryctinae). Ann. Entomol. Ecol. 26: 2657Ð2664. Soc. Am. 98: 636Ð642. [USDA–APHIS] U.S. Department of Agriculture–Animal Yang, Z.-Q., J. S. Srazanac, Y.-X. Yao, and X.-Y. Wang. 2006. and Plant Health Inspection Service. 2007. The pro- A new species of emerald ash borer parasitoid from China posed release of three parasitoids for the biological con- belonging to the genus Tetrastichus Haliday (Hymenop- trol of the emerald ash borer (Agrilus planipennis)inthe tera: Eulophidae). Proc. Entomol. Soc. Wash. 108: 550Ð continental United States: environmental assessment. 558. Federal Register 72: 28947Ð28948. Docket APHIS-2007- Zhang, Y.-Z., D.-W. Huang, T.-H. Zhao, H.-P. Liu, and L. S. 0060. Bauer. 2005. Two new species of egg parasitoids (Hy- Ulyshen, M. D., J. J. Duan, and L. S. Bauer. 2010a. Interac- menoptera: Encyrtidae) of wood-boring beetle pests tions between Spathius agrili (Hymenoptera: Bra- from China. Phytoparasitica 33: 253Ð260. conidae) and Tetrastichus planipennisi (Hymenoptera: Eulophidae), larval parasitoids of Agrilus planipennis (Coleoptera: Buprestidae). Biol. Control 52: 188Ð193. Received 5 August 2010; accepted 15 November 2010.