on Hemlock Woolly Adelgid

BIOLOGICAL CONTROLÑPARASITOIDS AND PREDATORS

Propagation, Synchrony, and Impact of Introduced and Native Laricobius spp. (Coleoptera: Derodontidae) on Hemlock Woolly Adelgid in Virginia D. L. MAUSEL,1 S. M. SALOM, L. T. KOK, AND J. G. FIDGEN2 Virginia Polytechnic Institute and State University (Virginia Tech), Department of Entomology, Price Hall MC 0319, Blacksburg, VA 24061

Environ. Entomol. 37(6): 1498Ð1507 (2008)

ABSTRACT Synchrony and impact of the predators Laricobius nigrinus Fender and Laricobius rubidus LeConte, on hemlock woolly adelgid, Adelges tsugae Annand, were studied in an eastern hemlock Þeld insectary in Virginia. First, a Þeld insectary for propagation of the introduced L. nigrinus was established by planting hemlocks in 2001, infesting them with hemlock woolly adelgid in 2002 and 2003, followed by releasing 258 L. nigrinus in 2003. Initial sampling showed that the native L. rubidus was present in the area. Hemlock woolly adelgid and both Laricobius species populations increased annually, from which 305 F3 L. nigrinus adults were collected and redistributed to forests in 2007. Second, the phenology of hemlock woolly adelgid and Laricobius spp. life cycles were monitored in 2005 and 2006. Adult L. nigrinus (F2) and L. rubidus were active on hemlock from fall through mid-spring and overlapped with second-instar sistentes nymphs through progredientes eggs. The predatorsÕ eggs were oviposited and larvae developed (i.e., F3 L. nigrinus) from late winter to mid-spring on progredientes eggs, indicating synchrony with suitable prey life stages. Third, a predator exclusion experiment was used to examine the relationships between the predators and prey in 2005 and 2006. When exposed to L. nigrinus (F2 adults and F3 larvae) and L. rubidus, hemlock woolly adelgid survival and ovisac density were lower and ovisac disturbance was higher than hemlock woolly adelgid protected in cages. The establishment and production of L. nigrinus at a Þeld insectary, synchronization with, and impacts on hemlock woolly adelgid after a small release 2 yr earlier makes it an important potential biological control agent of hemlock woolly adelgid. KEY WORDS classical biological control, Adelges tsugae, Þeld insectary, synchrony, exclusion cage

The hemlock woolly adelgid, Adelges tsugae Annand (Hemiptera: Adelgidae), is invasive in eastern North America on eastern [Tsuga canadensis (L.) Carrie`re] and Carolina hemlock (Tsuga caroliniana Engelmann). Infestation ranges from Maine to Georgia and inland to Kentucky, West Virginia, and upstate New York (USDA Forest Service 2007). Hemlock woolly adelgid uses piercing-sucking mouthparts to feed on a hemlockÕs energy reserves in xylem ray parenchyma cells (Young et al. 1995), which causes decline in growth and eventual death (Orwig and Foster 1998). Hemlock woolly adelgid is mostly sedentary and has two wingless asexual generations per year called the sistentes (summer through early spring generation) and progredientes (spring generation) (McClure 1989, Havill and Foottit 2007). There is also a winged sexual generation in the spring called the sexuparae that are 1 Corresponding author: University of Massachusetts, Department of Plant, Soil and Insect Sciences, Agricultural Engineering Bldg., 250 Natural Resources Rd., Amherst, MA 01003 (e-mail: dmausel@ psis.umass.edu). 2 Current address: Idaho Department of Lands, 3780 Industrial Ave. S., Coeur dÕAlene, ID 83815.

typically produced on deteriorating hemlocks (McClure 1991a). Hemlock woolly adelgid has been a focus of biological control efforts since the mid-1990s (Cheah et al. 2004). Several predators from the adelgids native range in Asia and the PaciÞc Northwest have been released into eastern North America, including the hemlock woolly adelgidÐspeciÞc beetle Laricobius nigrinus Fender (Coleoptera: Derodontidae). In the PaciÞc Northwest, Kohler et al. (2008) identiÞed many hemlock woolly adelgid predators on infested western hemlock [Tsuga heterophylla (Raf.) Sargent], of which L. nigrinus was the most widespread and abundant. L. nigrinus adults are small (2Ð3 mm), entirely black, and covered in Þne setae (Fender 1945, ZilahiBalogh et al. 2006). Its life cycle is univoltine and maintains phenological and numerical synchrony with hemlock woolly adelgid in the PaciÞc Northwest (Zilahi-Balogh et al. 2003a). Both L. nigrinus and hemlock woolly adelgid are active in the fall, winter, and spring, and aestivate in the summer. L. nigrinus eggs, larvae, and adults are cold hardy (Humble and Mavin 2005) and feed on hemlock woolly adelgid when other

0046-225X/08/1498Ð1507$04.00/0 䉷 2008 Entomological Society of America

December 2008

MAUSEL ET AL.: IMPACT OF Laricobius SPP. ON HEMLOCK WOOLLY ADELGID

predators are inactive. Adults emerge from the soil in the fall, disperse to hemlock branches, and feed on developing hemlock woolly adelgid sistentes nymphs, adults, and progredientes through early spring. Females synchronize oviposition with the presence of sistentes ovisacs (i.e., progredientes eggs) in the early spring and subsequently die. Larvae primarily eat eggs and by June develop through four instars, drop to the soil pupate, and aestivate (Zilahi-Balogh et al. 2003b). Laricobius nigrinus was Þrst imported from Victoria, British Columbia, Canada, into Virginia in 1997 for host-range testing and other studies (Zilahi-Balogh et al. 2002, 2003b, c) and then subsequently released from quarantine. In Þeld cages, adults survived in southwestern Virginia during winter and each fed on three to six sistentes nymphs per day (Lamb et al. 2005b). In a two-part study, Lamb et al. (2006) detected signiÞcant predator oviposition and impact on hemlock woolly adelgid density within Þeld cages and conducted the Þrst Þeld release in spring 2003, which resulted in the recovery of F2 adults on or near release trees in fall 2004. L. nigrinus is being mass-reared in the laboratory (Lamb et al. 2005a) and throughout the current range of hemlock woolly adelgid. To expedite production and release in natural forests, Þeld insectaries were advocated for rearing L. nigrinus (Kok and Salom 2002). Species in the genus Laricobius Rosenhauer prey exclusively on the Adelgidae, a family that only feeds on trees in the Pinaceae (Lawrence and Hlavac 1979, Havill and Foottit 2007, Zilahi-Balogh et al. 2007). The lone species native to eastern North America is Laricobius rubidus LeConte (Clark and Brown 1960, Lawrence 1989). The adults are also small (2Ð3 mm), but contrast with L. nigrinus by being red and black. The primary host of L. rubidus is the pine bark adelgid, Pineus strobi Hartig (Hemiptera: Adelgidae), to which it is phenologically synchronized (Clark and Brown 1960). L. rubidus has been collected from hemlock woolly adelgidÐinfested eastern hemlock (Montgomery and Lyon 1996, Wallace and Hain 2000), and laboratory studies have shown that it can also reproduce and complete development on hemlock woolly adelgid (Zilahi-Balogh et al. 2005). The life history of L. rubidus is similar to L. nigrinus, except that it has a hibernal diapause in addition to an aestival diapause (Clark and Brown 1960, Montgomery and Lyon 1996, Zilahi-Balogh et al. 2005). Adults are active in the fall and spring and larvae develop in the spring. This research consists of three separate studies with the following objectives: (1) Þeld insectary study to determine the effectiveness of Þeld planted hemlocks infested with hemlock woolly adelgid for propagating L. nigrinus by monitoring the predator and prey populations and collection of adult predators for release elsewhere; (2) synchrony study to determine the current synchrony of L. nigrinus, L. rubidus, and hemlock woolly adelgid by monitoring their phenology and numbers at the insectary; and (3) predator exclusion study (a) to determine the short-term impact of L. nigrinus and L. rubidus on hemlock woolly adelgid survival, ovisac density, and ovisac disturbance at the

1499

insectary by the use of exclusion cages and (b) to describe the relationship between L. nigrinus and L. rubidus density and hemlock woolly adelgid density. L. nigrinus can be effectively reared in a Þeld insectary, and its synchronization with and impacts on hemlock woolly adelgid make it an important potential biological control agent of hemlock woolly adelgid.

Materials and Methods Establishment of the Field Insectary. An eastern hemlock plantation was established at Virginia TechÕs Kentland Farm, Montgomery Co., VA, in October 2001 (Kok and Salom 2002). A Þeld that was left fallow with naturally occurring wild grass on a northeast-facing slope (10 Ð15%) was selected for the location of a 0.4-ha Þeld insectary. Adjacent to the Þeld insectary was a 15-yr-old white pine, Pinus strobus L., plantation that was naturally infested with pine bark adelgid. For the hemlock plantation, twelve 12 by 20-m blocks were spaced 5 m apart in a 4 by 3 block rectangle. Thirty trees were planted per block in six rows with Þve trees per row. Trees were spaced 2.4 m within and 3.7 m between rows. Three hundred 1.2- to 2.4-m tall hemlock trees with 0.6-m-diameter root balls wrapped in burlap were purchased from a local nursery. Ten of the 12 blocks were planted with these large hemlocks. The other two blocks were planted at the same spacing but with 60 potted 0.6-m-tall hemlocks obtained from a nursery in Pennsylvania, for a total of 360 trees. Augers (60 and 15 cm diameter) were used to make holes for the root balls of the large and small hemlocks, respectively. Bark mulch was put down 10 Ð15 cm deep, and 4 Ð 8 liters of water were poured to the drip line of each tree immediately after planting. They were watered again at 3 and 6 wk after planting. A survey in 2002 showed that 85 large hemlocks had suffered needle loss and another 91 trees had dead tops caused by prolonged drought conditions in 2001. A subsequent survey in 2003 showed that 75 hemlocks had died, with 60 of those being the potted hemlocks in two blocks. The remaining dead trees were scattered throughout most blocks. Between 2002 and 2005, dead tops were removed, and the dead trees were replaced with potted 0.6-m-tall hemlocks. Weeds in the plot were managed by mowing, clipping, hand pulling, and herbicide treatment (Roundup Pro 2% solution; Monsanto, St. Louis, MO) during the growing seasons starting in 2002. In summer 2002, 0.5 kg of 5Ð10-10 (N-P-K) fertilizer was placed around the drip line of the large hemlocks and 0.2 kg around the small hemlocks. In November 2004, the trees were fertilized again, and the tallest hemlocks had their leaders pruned to favor a low and spreading crown form. Hemlock woolly adelgid was released on the trees three times in March 2002 and three times in April 2003 (⬇1st, 15th, and 30th days). Two ⬇30-cm-long hemlock clippings infested with progredientes eggs from local forests were implanted in the mid-crown of each living tree.

1500

ENVIRONMENTAL ENTOMOLOGY

Laricobius nigrinus adults were collected in Victoria, British Columbia, Canada, in late winter 2003 and shipped to Virginia Tech. Starting in early spring 2003 using laboratory rearing methods (Lamb et al. 2005a), the predators reproduced, and larvae developed, pupated, and entered aestival diapause through early summer 2003. Adults emerging from soil aestivation containers in fall 2003 were placed in plastic containers with 50 conspeciÞcs on hemlock woolly adelgidÐ infested eastern hemlock. Containers were kept at 4 and 2⬚C (day and night, respectively) and given fresh hemlock woolly adelgid every 2 wk until Þeld release. We estimated the density of hemlock woolly adelgid on 193 trees ⬎1.0 m tall on 12 November 2003 to determine which trees had enough prey for a predator release. The number of new shoots infested with one or more sistentes hemlock woolly adelgid and total number of new shoots were counted on 30-cm segments at the middle third of a branchÕs length. One branch was measured at each cardinal point, and the percentage of infested shoots from the four branches was averaged for each tree. Predators were released on infested trees with the number per tree based on the estimated density of hemlock woolly adelgid as follows: six adult L. nigrinus were released on each of four heavily infested trees (ⱖ76% of new shoots with at least one hemlock woolly adelgid), four were released on 42 moderately infested trees (25 ⱕ x ⱕ75%), and one was released on 66 lightly infested trees (ⱕ25%). One, four, or six beetles were packaged in petri dishes and placed on an infested part of a tree with a paintbrush on 18 November 2003 from 1100 to 1330 hours under light rain, ambient 11.2⬚C, and 11.3 km/h wind conditions. This open release amounted to 258 L. nigrinus adults (unknown sex ratio) on 112 hemlocks that ranged in height from 1.8 to 3.0 m. Field Insectary Survey. Yearly population estimates of L. nigrinus and L. rubidus were conducted in winter of 2005 through 2007 on 137 hemlocks ⬎1.5 m tall using 71-cm2 canvas beat sheets (Bioquip, Rancho Dominguez, CA). We sampled the most heavily hemlock woolly adelgidÐinfested branches of a tree Þrst, followed by other branches. Sampling was conducted on days with no precipitation, ambient temperature ⬎10⬚C, and winds ⬍10 km/h. We spent ⬇1 min per tree hitting branches with a 1.0-m-long bamboo stick in the afternoon (1300 Ð1700 hours) and counting dislodged L. nigrinus and L. rubidus adults on the sheet. The species were differentiated by color and returned to the trees. In 2007, both species were collected from the beat sheets with an aspirator and transported to the insectary for sorting. L. nigrinus adults (F3) were packaged for release in hemlock woolly adelgidÐinfested natural forests. In winter of 2004 through 2007, hemlock woolly adelgid infestation level was categorized yearly as heavy, moderate, light, or uninfested (after ⬇1 min of searching) on the same 137 hemlocks. These estimates were based on the quantitative estimates taken on 12 November 2003, when the number of L. nigrinus to release per tree was initially determined (described previously). In 2007, tree decline for the hemlocks was

Vol. 37, no. 6

classiÞed into the following categories: healthy (⬍10% hemlock woolly adelgid damage symptoms), light (10 Ð25%), moderate (26 Ð50%), severe (51Ð100%), or dead (no green foliage). Damage symptoms included no new shoot growth, twig dieback, foliage discoloration, and needle loss (Young et al. 1995). Data for the total number of each predator collected and the total number of trees in each infestation category per year were presented graphically. Synchrony. To determine the phenology and numbers of L. nigrinus (F2) and L. rubidus adults, we selected 42 Þeld insectary trees for sampling at ⬇2-wk intervals with beat sheets for 1 yr (1 September 2005 to 15 August 2006). The selected trees were known to have each predator species present. Heavily infested branches were sampled in the afternoon, as described previously, and dislodged L. nigrinus and L. rubidus were counted. Fourth-instar L. nigrinus (F3) and L. rubidus larvae were also dislodged from beating and counted on the sheets. Because identiÞcation of the larvae to species was not possible, 80 were collected on 28 April and reared to adulthood in the laboratory to determine the relative numbers of L. nigrinus to L. rubidus. The presence of other predators on the beat sheet was noted. Hourly temperature data were summarized as the mean daily soil temperature at 10 cm depth and ambient temperature at a weather station 1.4 km from the Þeld insectary. Hourly wind speed (km/h) data from the weather station were averaged for each afternoon sampling period. To determine the phenology and numbers of each hemlock woolly adelgid stage, L. nigrinus (F3), and L. rubidus immatures (eggs and larvae), we clipped two heavily infested hemlock shoots from each of the 42 trees. Shoots dating back to the 2005 growing season were clipped from the north and south sides of each tree. Branch clipping occurred on the same dates as the beat sheet sampling, but from March through May 2006, shoots were clipped weekly to more precisely monitor oviposition by hemlock woolly adelgid and the two Laricobius species. Samples were frozen until inspection with a dissecting microscope (12Ð20⫻) in fall 2006. To examine hemlock woolly adelgid, shoots were randomly subsampled until a minimum of 200 live hemlock woolly adelgid were examined for each sample date. N1 through N4 instars (N ⫽ nymphs), adults, and adults with eggs (termed ovisacs) were counted by stage. Instars were determined by adding one to the number of molts by counting exuviae within the woolly ßocculence. In cases of missing or disturbed ßocculence, the body or antennal length was measured with an eyepiece micrometer to determine the instar (McClure 1989). With a bent teasing needle, probing of hemlock woolly adelgid that was frozen live produced purple colored hemolymph when punctured, rebounding to the touch of the needle, and appeared normal shaped (McClure 1991a, Palmer and Sheppard 2002). We ignored dead hemlock woolly adelgid, which had desiccated bodies or cloudy brown/black hemolymph. To examine Laricobius spp. eggs and larvae, which are oviposited and embedded in hemlock woolly adelgid woolly ßocculence (Zilahi-

December 2008


Balogh et al. 2003b), we counted them by stage during the hemlock woolly adelgid probing described above and from the remaining shoots collected per sample date. The Laricobius spp. larval instars L1ÐL4 (L ⫽ larvae) were determined by the use of head capsule measurements for L. nigrinus (Zilahi-Balogh et al. 2003b). The percentages of live hemlock woolly adelgid and Laricobius spp. in each life stage per sample date were calculated. The presence of other predators on hemlock woolly adelgid or hemlock in the samples was noted to the family level. Hemlock woolly adelgid phenology, numbers of L. nigrinus and L. rubidus collected with beat sheets, and ambient temperature over 1 yr are shown graphically. The percentages of immature Laricobius spp. and hemlock woolly adelgid populations in each life stage over 1 yr are also displayed graphically. In addition, overlap values (Cih) were calculated for pairwise comparisons of the predators and prey stages to measure phenological and numerical synchrony. Values range from 0 to 1, with 1 being perfect overlap (Colwell and Futuyma 1971): C ih ⫽ 1 ⫺ 1/2 兺兩p ij ⫺ p hj兩, where pij equals the proportion of a predatorsÕ stage on a sample date out of all the samples and phj equals the proportion of a hemlock woolly adelgid stage on a sample date out of all the samples. The overlap value incorporates both the phenology and numbers of Laricobius spp. and hemlock woolly adelgid stages. Predator Exclusion Experiment. Exclusion cages were used to estimate the effect of L. nigrinus (F2 adults and F3 larvae) and L. rubidus predation on hemlock woolly adelgid at the Þeld insectary. Fourteen trees with moderate to high hemlock woolly adelgid densities (25Ð100% of shoots infested) were selected for a randomized complete block design. Three treatments were used: (1) closed cage, where the cage was closed at both ends to exclude predators and monitor cage and other effects on hemlock woolly adelgid; (2) open cage, where the cage was open at both ends to monitor predator, cage, and other effects on hemlock woolly adelgid; and (3) no cage, where predator and other effects on hemlock woolly adelgid were monitored. Cages were made of sewn white chiffon fabric (polyester, 0.25-mm mesh size) 45 cm wide and 60 cm long. Identical cages similarly placed in the partial shade of lower hemlock canopies had no signiÞcant effect on interior temperatures compared with exterior temperatures in other studies (Lamb et al. 2005b, 2006). Three branches with similar hemlock woolly adelgid densities were tagged per tree, and each treatment was randomly assigned to a branch. The treatments were set up on 8 November 2005 after generalist predator activity had stopped. To set up the closed cage treatment, the branch was tapped 10 times along its length to dislodge any active predators. The cage was pulled over the branch and the open end was cinched with a cable-tie to the branch stem. For the open cage treatment, a cage with both ends open was pulled over the branch and fastened along the branch stem with

1501

thin wire poked through the fabric. Other than tagging, nothing was done to the no-cage treatment branches. The treatments were cut from the trees after 5 mo (11 and 14 April 2006), placed into plastic bags, and frozen. The timing of cage removal coincided with the resumption of Harmonia axyridis Pallas (Coleoptera: Coccinellidae) adult activity, approximate conclusion of hemlock woolly adelgid sistentes oviposition, and Laricobius spp. larval feeding. On thawing, the branch clippings were inspected with a dissecting microscope in fall 2006, as described previously. Shoots dating back to the 2005 growing season were cut off and randomly subsampled until a minimum of 200 hemlock woolly adelgid (second-instar nymphs through adults) per sample were probed and recorded as live or dead. Hemlock woolly adelgid remain Þrmly attached to branches by the stylet and woolly ßocculence after predation or other mortality factors. Laricobius spp. typically pierce the exoskeleton of nymphs with their mandibles, consume the hemolymph, and leave the carcass (Franz 1958). Each sistentes ovisac in a sample was inspected and recorded as disturbed or undisturbed. Disturbed ovisacs had tattered and loosened woolly ßocculence, characteristic evidence of late-instar L. nigrinus, L. rubidus, and other predators feeding on the hemlock woolly adelgid eggs embedded in the wool (D.L.M., personal observation). This evidence of predation on ovisacs has been observed in other Laricobius species (Franz 1958, Brown and Clark 1962). Frequently, disturbed ovisacs coalesced into indistinguishable masses of woolly ßocculence. In these cases, to determine the number of ovisacs that were disturbed, we counted the sistentes adults attached to the stem beneath the tattered wool. Each disturbed ovisac was probed to count Laricobius spp. and other predator eggs, larvae, or exuviae to determine the percentage that had direct evidence of a predator. The predator presence was classiÞed as (1) Laricobius spp.; (2) syrphids (the only other predator observed); or (3) “unidentiÞed” if only tattered and loosened woolly ßocculence was evident (i.e., circumstantial evidence of predation). Undisturbed ovisacs were neatly round and Þrmly attached to the stem. Each undisturbed ovisac was probed to count immature Laricobius spp. and other predators within the woolly ßocculence. The length of shoots per sample was measured to the nearest 0.1 cm to calculate the total ovisac and immature predator densities. Percent hemlock woolly adelgid survival was calculated as the total number of live N2 through adults and undisturbed ovisacs divided by live and dead N2 through adults and total ovisacs. Two-way analysis of variance (ANOVA) was used to determine whether a signiÞcant proportion of the variation in mean percent survival, sistentes ovisac density, and percent ovisac disturbance was attributable to the cage treatment using SPSS version 10.0, followed by a TukeyÕs honestly signiÞcant difference (HSD) test to separate signiÞcantly different means. Proportion data were arcsine square root transformed before analyses, and

1502

ENVIRONMENTAL ENTOMOLOGY 450

80 Total no. of trees

400

Total no. of adults

350 300 250

Vol. 37, no. 6

Uninfested Light HWA Moderate HWA Heavy HWA

60 40 20 0

F3

'03/'04

'04/'05

'05/'06

'06/'07

200 150 100

L. rubidus L. nigrinus

F2

50 F1

0

'03/'04

'04/'05

258 L. nigrinus released

'05/'06

'06/'07

Year

Fig. 1. Numbers of L. nigrinus and L. rubidus adults and hemlock woolly adelgid infestation levels over time (inset) at an eastern hemlock Þeld insectary in Montgomery County, VA. On 18 November 2003, 258 L. nigrinus adults were released.

ovisac density was 冑x ⫽ 3/8 transformed to stabilize variance (Zar 1999). A paired t-test was used to compare immature Laricobius spp. and syrphid density between the open and no cage treatments. The percentage of disturbed ovisacs caused by Laricobius spp., syrphids, and unidentiÞed causes were compared between the open and no cage treatments by paired t-tests. All averages are presented as means ⫾ SEM, and the 0.05 signiÞcance level was used. A simple linear regression was used to describe the relationship between hemlock woolly adelgid ovisac density and immature Laricobius spp. density (eggs and larvae). Results Field Insectary Survey. The numbers of L. rubidus increased over time, as did L. nigrinus from introduction to recovery of F3 adults (Fig. 1). Two pairs of L. nigrinus and L. rubidus were observed mating interspeciÞcally after being dislodged onto the beat sheet (21 February 2007). The Þeld-reared L. nigrinus were large (⬇3 mm length), and in 2007, 305 F3 L. nigrinus were collected and released in a hemlock woolly adelgidÐinfested forest in Pennsylvania (121) and Maryland (184). The total number of uninfested and lightly infested trees decreased over 4 yr after hemlock woolly adelgid release, and the number of moderately infested trees increased (Fig. 1, inset). The number of heavily infested trees increased rapidly from 2003/ 2004 to 2004/2005 and stabilized at 30 Ð 40 trees. In 2007, 89% of the trees were healthy and 6% had light, 3% had moderate, and 2% had severe decline symptoms. Synchrony. Beat sheet sampling recovered 193 F2 L. nigrinus and 383 L. rubidus adults from fall 2005 through mid-spring 2006 (Fig. 2B). The soil temperature had dropped from 19.1 to 11.2⬚C from 13 October to 3 November when the Þrst predators were recovered. No recoveries were made on 15 December when

the mean ambient temperature was ⫺2.1⬚C, and snow and sleet were accumulating. On 29 December, gusty winds (18 km/h) and light sleet were associated with reduced recovery. Similarly, L. nigrinus recovery was low on 10 February because of wind (19 km/h) and cold (mean temperature was ⫺1.1⬚C). Branch clipping sampling conÞrmed that aestivating hemlock woolly adelgid neosistentes resumed development in fall, developed through four instars and became adults in winter, and oviposited in late winter to mid-spring (Figs. 2A and 3A). The progredientes developed rapidly, matured, and oviposited in mid- to late spring. We did not observe sexuparae. Aestivating neosistentes were present in summer through early fall. Laricobius spp. eggs and larvae were collected from late winter to mid-spring (Fig. 3B). Beat sheet sampling collected fourth-instar larvae in early to midspring (Fig. 2B). In fall 2006, 47 adults were successfully reared in the laboratory from the 80 larvae collected with beat sheets on 28 April 2006. Of these, 48% were L. nigrinus and 52% were L. rubidus. H. axyridis adults were observed during beat sheet sampling on 11 March and were numerous on 10 April. Coccinellid and chrysopid larvae were observed on 28 April during beat sheet sampling, and coccinellid larvae were numerous on 8 May and 19 May. In branch clipping samples, we observed in low numbers syrphid eggs on 11 March to 21 April, syrphid larvae on 17 March to 21 April, chrysopid larvae on 10 April, and coccinellid larvae on 28 April and 8 May. Overlap values for both L. nigrinus and L. rubidus adults were the highest with sistentes nymphs and adults, lower with sistentes ovisacs, and the lowest with progredientes nymphs and adults (Table 1). Overlap values for Laricobius spp. eggs were the highest with sistentes ovisacs (i.e., progredientes eggs), lower with progredientes nymphs and adults, and the lowest with sistentes nymphs and adults. Overlap values for Laricobius spp. larvae were the highest with

December 2008

MAUSEL ET AL.: IMPACT OF Laricobius SPP. ON HEMLOCK WOOLLY ADELGID A)

1503

Aestivating neosistentes N2-adult sistentes Progredientes eggs N1-adult progredientes Sistentes eggs

90 80 70

30 L. nigrinus L. rubidus Laricobius Larvae spp. larvae Ambient Ambient temp. temp.

25 20

60 50 40

15

30 20

5 0

9-1 9-15 10-1 10-13 11-3 11-19 12-1 12-15 12-29 1-13 1-28 2-10 2-24 3-11 3-24 4-10 4-28 5-8 5-19 6-5 6-20 7-8 7-15 7-31 8-15

10 0

10

Temp. (ºC)

No. adults or larvae

B) 100

-5

Date

Fig. 2. Synchrony of hemlock woolly adelgid and Laricobius spp. adults on eastern hemlock during 2005 and 2006. (A) Phenology of hemlock woolly adelgid. (B) Numbers of L. nigrinus adults, L. rubidus adults, fourth-instar larvae (both species pooled), and mean daily ambient temperature.

sistentes ovisacs, lower with progredientes nymphs and adults, and the lowest with sistentes nymphs and adults. Overlap of Laricobius spp. eggs, larvae, or adult stages and aestivating neosistentes or progredientes ovisacs did not occur. Predator Exclusion Experiment. Hemlock woolly adelgid survival differed signiÞcantly among the three cage treatments (F ⫽ 7.2; df ⫽ 2,26; P ⫽ 0.003). Survival on branches inside the closed cages was higher (71.2 ⫾ 3.2%) than the open (58.8 ⫾ 2.6%) or no cage treatments (56.2 ⫾ 3.9%), where predators were allowed free access. Hemlock woolly adelgid sistentes ovisac density differed signiÞcantly among treatments (F ⫽ 3.3; df ⫽ 2,26; P ⫽ 0.05). Ovisac density in the closed cage was higher than the no cage treatment, but the open cage was not different from either (Fig. 4). The percentage of ovisacs that were disturbed also differed signiÞcantly among the treatments (F ⫽ 36.2; df ⫽ 2,26; P ⬍ 0.001). Ovisac disturbance was lowest in the closed cage, higher in the open cage, and highest in the no cage treatment (Fig. 4). No predators were found in the closed cages, and the disturbed ovisacs in this treatment (⬇1%) were entirely from unidentiÞed causes. In the pooled open and no cage treatments, direct evidence of Laricobius spp. was found in 19.6 ⫾ 6.1%, syrphids in 0.6 ⫾ 0.5%, and unidentiÞed causes (i.e., circumstantial evidence of predation) in 75.9 ⫾ 6.9% of the disturbed ovisacs. We pooled these data because the percentage of ovisacs disturbed by Laricobius spp., syrphids, or unidentiÞed causes between the open and no cage treatments were not signiÞcantly different (Laricobius spp.: t ⫽ 0.04, df ⫽ 13, P ⫽ 0.9; syrphids: t ⫽ 0.3, df ⫽ 13, P ⫽ 0.7; unidentiÞed: t ⫽ 0.8, df ⫽ 13, P ⫽ 0.5). No coccinellid or other predator eggs or larvae were observed. Laricobius spp. density (0.03 ⫾ 0.01/cm twig) was an order of magnitude higher than syrphid density

(0.002 ⫾ 0.0008/cm twig) in the pooled open cage and no cage treatments. We pooled these data because the density of Laricobius spp. or syrphid immatures between the open cage and no cage treatments were not signiÞcantly different (Laricobius, t ⫽ 0.3, df ⫽ 13, P ⫽ 0.8; syrphids, t ⫽ 0.5, df ⫽ 13, P ⫽ 0.6). Immature Laricobius spp. density was related positively to hemlock woolly adelgid ovisac density (F ⫽ 55.03; df ⫽ 1,25; r2 ⫽ 0.69; P ⬍ 0.0001, Fig. 5). Hemlock woolly adelgid ovisac density or Laricobius spp. density in the open and no cage treatments were not signiÞcantly different and pooled for the regression. Discussion Field Insectary Survey. Field insectaries are recognized as an effective way to rear several biological control agents of invasive weeds and insects (Stoyer and Kok 1986, Debach and Rosen 1991, Kok and Salom 2002). We are aware of only one case where a Þeld insectary was used to rear a natural enemy of a forest pest: Didea fasciata (Diptera: Syrphidae), a predator of the Þr bark aphid, Cinara piceae (Hemiptera: Lachnidae), in Europe (Stary´ 1976). L. nigrinus easily established at the Þeld insectary, and several factors incorporated into the design favored multiplication, collection, and limited dispersal. These were an abundance of hemlock woolly adelgid, perhaps aided by nitrogen fertilization (McClure 1991b); hundreds of open grown and healthy hemlock trees with little vegetative competition; easy access (low tree height and vehicle accessibility); and isolation from natural hemlock forests. Despite increasing hemlock woolly adelgid populations at the insectary, the majority of the trees appeared healthy, which may have been because of their young age and good growing conditions, impact of predators, or both. L. nigrinus, which has become

1504


Vol. 37, no. 6

% of population

A) 100 90 80 70 60 50 40 30 20 10 0

HWA Ae. sis. N1 Sis. N2 Sis. N3 Sis. N4 Sis. ad. Sis. ovi. Pro. N1 Pro. N2 Pro. N3 Pro. N4 Pro. ad. Pro. ovi.

% of ovisacs with predation

Laricobius spp.

5-8 5-19 6-5 6-20 7-8 7-15 7-31 8-15

4-14 4-21 4-28

Eggs L1 L2 L3 L4

11-3 11-19 12-1 12-15 12-29 1-13 1-28 2-10 2-24 3-3 3-11 3-17 3-24 3-31 4-10

100 90 80 70 60 50 40 30 20 10 0

9-1 9-15 10-1 10-13

% of population

B)

Date Fig. 3. Synchrony of hemlock woolly adelgid and Laricobius spp. immatures on eastern hemlock during 2005 and 2006. (A) Percentages of the hemlock woolly adelgid population in each life stage per sample date. (B) Percentages of the Laricobius spp. (pooled L. nigrinus and L. rubidus) population in each life stage per sample date. (Ae sis, aestivating neosistentes; Sis, sistentes; Ad, adults; Ovi, ovisacs; Pro, progredientes; L, larvae).

established and increased in numbers, and L. rubidus likely contributed to continuing tree health. The presence of L. rubidus was expected given the population in the adjacent pine bark adelgidÐinfested white pine stand, but the high numbers were surprising. Laboratory studies indicate that hemlock woolly adelgid is a physiologically suitable host for L. rubidus (ZilahiBalogh et al. 2005), and this study indicated it is an ecologically suitable host as well. The observed mating attempts between L. nigrinus and L. rubidus indicated that these species are not antagonistic toward each other and makes further study regarding their interaction necessary.

Based on experience, we recommend Þeld insectaries to enhance biological control of hemlock woolly adelgid. The L. nigrinus reared under Þeld conditions were acclimated to the environment, appeared larger and healthier, and may be more successful than laboratory-reared beetles. Zilahi-Balogh (2001) showed that there is a signiÞcant positive relationship between female size and fecundity. Alternatives to planting trees include utilization of heavily infested hedges, nurseries, and open grown trees on campuses, cemeteries, golf courses, and city parks that have not been nor are likely to be treated with insecticides. In fact, it is under these conditions that L. nigrinus is collected

Table 1. Pairwise overlap values for the phenological and numerical occurrence of the predators L. nigrinus, L. rubidus, and their prey, hemlock woolly adelgid, from Sept. 2005 through Aug. 2006 in Montgomery Co., VA Occurrence of

Neosistentes aestivating

Sistentes N2-adult

Sistentes ovisacs

Pro. N1adult

Pro. ovisacs

L. nigrinus adults L. rubidus adults Laricobius spp. eggs Laricobius spp. larvae

0.00 0.00 0.00 0.00

0.71 0.75 0.14 0.05

0.46 0.47 0.88 0.55

0.11 0.26 0.23 0.46

0.00 0.00 0.00 0.00

Overlap values range from 0 to 1, with 1 being a perfect match. N, nymphs; Pro., progredientes.

2.5

MAUSEL ET AL.: IMPACT OF Laricobius SPP. ON HEMLOCK WOOLLY ADELGID HWA ovisac density % ovisacs disturbed

A

C

0.18

40

0.16

35 30

1.5 25

1.0

AB

20

B B

10

0.5 A 0

15

% ovisacs disturbed

No. HWA ovisacs / cm shoot

2.0

45

Closed cage

Open cage Treatment

No cage

No. immature predators / cm shoot

December 2008

0.12 0.1 0.08 0.06 0.04 0.02

0

0

in large numbers in its native range. A population of L. nigrinus would likely be self-sustaining at a Þeld insectary and provide a supply of predators for release. A highly productive Þeld insectary will require maintaining an optimal predator: prey ratio that has not yet been determined. Synchrony. Phenological and numerical synchrony between a natural enemy and its host are prerequisites for successful classical biological control because asynchrony results in natural enemy starvation, establishment failure, or insigniÞcant impacts on the host (Van Driesche and Bellows 1996). Our results indicate synchrony between suitable hemlock woolly adelgid prey stages and L. nigrinus in Virginia. This synchrony may be because of a decline in soil temperatures that signaled termination of aestivation and emergence at the same time hemlock woolly adelgid broke aestivation. Laboratory-reared L. nigrinus adults emerged coincident with declines in soil temperatures (Lamb et al. 2007). Overlap of L. nigrinus oviposition and larval development with sistentes ovisacs was also critical because larvae require progredientes eggs to develop (Zilahi-Balogh et al. 2003b). Last, no overlap of L. nigrinus occurred with aestivating neosistentes and progredientes ovisacs, the nonsuitable host stages. Adults die when aestivating neosistentes are the only available prey, which can occur in the laboratory when adults emerge in the summer out of synchrony with hemlock woolly adelgid breaking aestivation in the fall (Lamb et al. 2007). The predators also do not feed on progredientes ovisacs in the Þeld because adults reach the end of their life span, and larvae complete development before they are present in late spring (Zilahi-Balogh et al. 2003a). The phenologies of hemlock woolly adelgid and L. nigrinus in this study were slightly different from previous descriptions. In this study, hemlock woolly adelgid terminated aestival diapause 1 mo later than those in Victoria, British Columbia, Canada (ZilahiBalogh et al. 2003a), Connecticut (McClure 1987), and elsewhere in Virginia (Gray and Salom 1996). The

y = 0.0603x - 0.0119 r2 = 0.69, P < 0.0001

0.14

5

Fig. 4. Laricobius nigrinus and L. rubidus impact on hemlock woolly adelgid ovisac density and the percentage of ovisacs that were disturbed (mean ⫾ SEM) in a predator exclusion experiment.

1505

0

0.5

1 1.5 2 No. HWA ovisacs / cm shoot

2.5

3

Fig. 5. Relationship between hemlock woolly adelgid ovisac density and immature Laricobius nigrinus and L. rubidus density (pooled).

unique conditions at the Þeld insectary may have caused this delay. Sistentes began ovipositing 1 mo later, concluded oviposition 1 mo earlier, and progredientes began ovipositing 1Ð2 wk earlier than those in Victoria. No sexuparae were observed probably because the trees were not stressed. Adult L. nigrinus activity began 1 mo later and ended 1 mo earlier than those in Victoria. The only occasion when adults were not collected (15 December), the temperature was below freezing, and the weather was adverse, as described previously (Zilahi-Balogh et al. 2003a). Laricobius spp. oviposition began 1 mo later, and larvae concluded development 1 mo earlier than those in Victoria. The shorter duration of sistentes oviposition in this study than at Victoria corresponded with the predatorsÕ oviposition and larval development. The differences between the marine west coast climate in Victoria and continental climate in Virginia likely explain the phenological differences. The activity of L. rubidus adults in the winter and synchrony with hemlock woolly adelgid was unexpected because it was described as having a hibernal diapause and synchrony with pine bark adelgid (Clark and Brown 1960, Montgomery and Lyon 1996, ZilahiBalogh et al. 2005). Other predators occurred at low densities and later in the spring on the hemlock woolly adelgidÐinfested hemlocks, consistent with previous studies (Montgomery and Lyon 1996, Wallace and Hain 2000). Predator Exclusion Experiment. A predator impact on hemlock woolly adelgid in this study occurred because of the L. nigrinus release 2 yr earlier and the high numbers of L. rubidus. The cages were removed ⬇3 wk before larval predation on sistentes ovisacs ended, and our results may be conservative. In another predator exclusion study aimed at assessing existing eastern natural enemies (including L. rubidus) before classical biological control attempts, Wallace and Hain (2000) did not detect a signiÞcant impact on progredientes. Laricobius spp. impact on hemlock woolly adelgid survival was not striking in this study. Survival was

1506


only 12Ð15% less when exposed to predation in the no cage and open cage treatments than when protected in the closed cage treatment. High survival was expected because small numbers of L. nigrinus were released 2 yr earlier, the population was still small, and each adult consumes only a few hemlock woolly adelgid per day (Lamb et al. 2005b). However, exposure to predation resulted in less than one ovisac per centimeter of twig. Most disturbed ovisacs were classiÞed as unidentiÞed, but the primary cause was likely L. nigrinus and L. rubidus larvae because other predators were inactive. Laricobius spp. are known to crawl along shoots and feed on numerous ovisacs without leaving direct evidence (i.e., their presence or exuviae) (Franz 1958, Clark and Brown 1960). Hemlock woolly adelgid ovisac disturbance caused by Laricobius spp. larval feeding was considerable because each larva needs ⬎300 hundred eggs to complete development (Zilahi-Balogh et al. 2003b). Last, the predators responded to high hemlock woolly adelgid density by increasing their oviposition, aggregation, or both, and this density-dependent numerical response could lead to hemlock woolly adelgid suppression in the future. Ovisac disturbance was signiÞcantly higher in the no cage than in the open cage treatment, indicating a cage effect. Perhaps the cage protected ovisacs from some type of disturbance, such as branch abrasion during high winds. Alternatively, the open cage could have somehow affected predator behavior, but immature L. nigrinus and L. rubidus densities in the open and no cage treatments were equal. The open cage could have reduced H. axyridis adult predation on ovisacs, but they became active near the very end of the experiment. Syrphids probably contributed little, if at all, to the unidentiÞed ovisac disturbance because they were quite uncommon during the sampling periods. Ovisac disturbance was uncommon in the closed cage and must have been from abrasion with the cage or the inclusion of some predators that our hemlock woolly adelgid dissections did not detect. The large L. rubidus population did not permit us to evaluate L. nigrinus impact on hemlock woolly adelgid in isolation. The Þeld insectary, however, was a typical situation because L. rubidus is relatively common in hemlock woolly adelgidÐinfested hemlock forests (Montgomery and Lyon 1996, Wallace and Hain 2000, Mausel 2007). Hemlock forests usually have pine bark adelgidÐinfested white pines in the area that support L. rubidus populations. Both predators likely played a synergistic role in affecting sistentes and progredientes hemlock woolly adelgid populations. Syrphids, H. axyridis, and other generalist predators had no impact on sistentes N2 through adults in this study because of their inactivity, but they may have some impact on the progredientes generation and sistentes eggs (Wallace and Hain 2000, Flowers et al. 2006). Acknowledgments We thank A. Lamb, M. Cornwell, and H. Benero for help with rearing L. nigrinus; D. Hamilton for laboratory assistance; and two anonymous reviewers for helpful comments.

Vol. 37, no. 6

Funding was provided by USDA Forest Service, Forest Health Protection Agreement 04-DG-11083150-030 and USDA National Biological Control Institute Agreement 20811-110A-007-333-1.

References Cited Brown, N. R., and R. C. Clark. 1962. Studies of predators of the balsam woolly adelgid, Adelges piceae (Ratz.) (Homoptera: Adelgidae) X. Field identiÞcation of Laricobius erichsonii Rosen. (Coleoptera: Derodontidae). Can. Entomol. 94: 191Ð193. Cheah, C., M. E. Montgomery, S. M. Salom, B. L. Parker, S. Costa, and M. Skinner. 2004. Biological control of hemlock woolly adelgid. U.S. Department of Agriculture. Forest Service, Washington, DC. Clark, R. C., and N. R. Brown. 1960. Studies of predators of the balsam woolly aphid, Adelges piceae (Ratz.) (Homoptera: Adelgidae) VII. Laricobius rubidus Lec. (Coleoptera: Derodontidae), a predator of Pineus strobi (Htg.) (Homoptera: Adelgidae). Can. Entomol. 92: 237Ð 240. Colwell, R. K., and D. J. Futuyma. 1971. On the measurement of niche breadth and overlap. Ecology 52: 567Ð576. Debach, P., and D. Rosen. 1991. Biological control by natural enemies. Cambridge University Press, Cambridge, United Kingdom. Fender, K. M. 1945. A new Laricobius from Oregon. PanPac. Entomol. 21: 152. Flowers, R. W., S. M. Salom, and L. T. Kok. 2006. Competitive interactions among two specialist predators and a generalist predator of hemlock woolly adelgid, Adelges tsugae (Hemiptera: Adelgidae), in south-western Virginia. Agric. For. Entomol. 8: 253Ð262. Franz, J. M. 1958. Studies on Laricobius erichsonii Rosenh. (Coleoptera: Derodontidae) a predator on Chermesids. Part I, distribution, life-history, and ecology. Entomophaga 3: 109 Ð164. Gray, D. R., and S. M. Salom. 1996. Biology of the hemlock woolly adelgid in the southern Appalachians, pp. 26 Ð35. In S. M. Salom, T. C. Tigner, and R. C. Reardon (eds.), Proceedings of the Þrst hemlock woolly adelgid review. U.S. Department of Agriculture Forest Service, Washington, DC. Havill, N. P., and R. G. Foottit. 2007. Biology and evolution of Adelgidae. Annu. Rev. Entomol. 52: 325Ð349. Humble, L. M., and L. Mavin. 2005. Preliminary assessment of the cold tolerance of Laricobius nigrinus, a winteractive predator of the hemlock woolly adelgid from western Canada, pp. 304. In B. Onken and R. C. Reardon (eds.), Third symposium on hemlock woolly adelgid in the Eastern United States. U.S. Department of Agriculture Forest Service, Washington, DC. Kohler, G. R., V. L. Stiefel, K. F. Wallin, and D. W. Ross. 2008. Predators associated with the hemlock woolly adelgid (Hemiptera: Adelgidae) in the PaciÞc Northwest. Environ. Entomol. 37: 494 Ð504. Kok, L. T., and S. M. Salom. 2002. Possible use of Þeld insectaries to rear hemlock woolly adelgid predators, pp. 225Ð232. In B. Onken, R. C. Reardon, and J. Lashcomb (eds.), Proceedings: hemlock woolly adelgid in the Eastern United States symposium. Rutgers New Jersey Agricultural Experimental Station, New Brunswick, NJ. Lamb, A. B., S. M. Salom, and L. T. Kok. 2005a. Guidelines for rearing Laricobius nigrinus Fender, pp. 309 Ð318. In B. Onken and R. C. Reardon (eds.), Third symposium on hemlock woolly adelgid in the Eastern United States. U.S.

December 2008


Department of Agriculture Forest Service, Washington, DC. Lamb, A. B., S. M. Salom, and L. T. Kok. 2005b. Survival and reproduction of Laricobius nigrinus Fender (Coleoptera: Derodontidae), a predator of hemlock woolly adelgid, Adelges tsugae Annand (Homoptera: Adelgidae), in Þeld cages. Biol. Control 32: 200 Ð207. Lamb, A. B., S. M. Salom, L. T. Kok, and D. L. Mausel. 2006. ConÞned Þeld release of Laricobius nigrinus (Coleoptera: Derodontidae), a predator of the hemlock woolly adelgid, Adelges tsugae (Hemiptera: Adelgidae), in Virginia. Can. J. For. Res. 36: 369 Ð375. Lamb, A. B., S. M. Salom, and L. T. Kok. 2007. Factors inßuencing aestivation in Laricobius nigrinus (Coleoptera: Derodontidae), a predator of Adelges tsugae (Hemiptera: Adelgidae). Can. Entomol. 139: 576 Ð586. Lawrence, J. F. 1989. A catalog of the Coleoptera of America north of Mexico, Family: Derodontidae. U.S. Department of Agriculture, Washington, DC. Lawrence, J. F., and T. F. Hlavac. 1979. Review of the Derodontidae (Coleoptera: Polyphaga) with new species from North America and Chile. Coleop. Bull. 33: 369 Ð 414. Mausel, D. L. 2007. Release and monitoring of Laricobius nigrinus (Coleoptera: Derodontidae) for biological control of the hemlock woolly adelgid in the eastern US. PhD dissertation, Virginia Polytechnic Institute and State University, Blacksburg, VA. McClure, M. S. 1987. Biology and control of hemlock woolly adelgid. Connecticut Agricultural Experiment Station, New Haven, CT. McClure, M. S. 1989. Evidence of a polymorphic life cycle in the hemlock woolly adelgid, Adelges tsugae (Homoptera: Adelgidae). Ann. Entomol. Soc. Am. 82: 50 Ð54. McClure, M. S. 1991a. Density-dependent feedback and population cycles in Adelges tsugae (Homoptera: Adelgidae) on Tsuga canadensis. Environ Entomol. 20:258 Ð264. McClure, M. S. 1991b. Nitrogen fertilization of hemlock increases susceptibility to hemlock woolly adelgid. J. Arboric. 17: 227Ð230. Montgomery, M. E., and S. M. Lyon. 1996. Natural enemies of adelgids in North America: their prospect for biological control of Adelges tsugae (Homoptera: Adelgidae), pp. 89 Ð102. In S. M. Salom, T. C. Tigner, and R. C. Reardon (eds.), Proceedings of the Þrst hemlock woolly adelgid review. U.S. Department of Agriculture Forest Service, Washington, DC. Orwig, D. A., and D. R. Foster. 1998. Forest response to the introduced hemlock woolly adelgid in southern New England, U.S.A. J. Torrey Bot. Soc. 125: 60 Ð73. Palmer, D. J., and J. L. Sheppard. 2002. Mass rearing Pseudoscymnus tsugae at the New Jersey Department of Agriculture: challenges and lessons, pp. 214 Ð220. In B. Onken, R. C. Reardon, and J. Lashcomb (eds.), Proceedings: hemlock woolly adelgid in the Eastern United States symposium. New Jersey Agricultural Experiment Station, New Brunswick, NJ. Stary´, P. 1976. Cinara piceae (Panz.) (Hom., Lachnidae), a pest of young European Þr trees (Abies alba Mill.) and its natural enemy complex in Czechoslovakia. Studia Entomol. For. 2: 171Ð180.

1507

Stoyer, T. L., and L. T. Kok. 1986. Field nurseries for propagating Trichosirocalus horridus (Coleoptera: Curculionidae), a biological control agent for Carduus thistles. J. Econ. Entomol. 79: 873Ð 876. U.S. Department of Agriculture Forest Service. 2007. Hemlock woolly adelgid distribution maps. (http://www.na. fs.fed.us/fhp/hwa/). Van Driesche, R. G., and T. S. Bellows. 1996. Biological control. Chapman & Hall, New York. Wallace, M. S., and F. P. Hain. 2000. Field surveys and evaluation of native and established predators of the hemlock woolly adelgid (Homoptera: Adelgidae) in the southeastern United States. Environ. Entomol. 29: 638 Ð 644. Young, R. F., K. S. Shields, and G. P. Berlyn. 1995. Hemlock woolly adelgid (Homoptera: Adelgidae): stylet bundle insertion and feeding sites. Ann. Entomol. Soc. Am. 88: 827Ð 835. Zar, J. H. 1999. Biostatistical analysis. Prentice Hall, Englewood Cliffs, NJ. Zilahi-Balogh, G.M.G. 2001. Biology of Laricobius nigrinus Fender (Coleoptera: Derodontidae) and its potential as a biological control agent of the hemlock woolly adelgid, Adelges tsugae Annand (Homoptera: Adelgidae) in the eastern United States. PhD dissertation, Virginia Polytechnic Institute and State University, Blacksburg, VA. Zilahi-Balogh, G.M.G., L. T. Kok, and S. M. Salom. 2002. Host speciÞcity of Laricobius nigrinus Fender (Coleoptera: Derodontidae), a potential biological control agent of the hemlock woolly adelgid, Adelges tsugae Annand (Homoptera: Adelgidae). Biol. Control 24: 192Ð198. Zilahi-Balogh, G.M.G., L. M. Humble, A. B. Lamb, S. M. Salom, and L. T. Kok. 2003a. Seasonal abundance and synchrony between Laricobius nigrinus (Coleoptera: Derodontidae) and its prey, the hemlock woolly adelgid (Hemiptera: Adelgidae). Can. Entomol. 135: 103Ð115. Zilahi-Balogh, G.M.G., S. M. Salom, and L. T. Kok. 2003b. Development and reproductive biology of Laricobius nigrinus, a potential biological control agent of Adelges tsugae. Biocontrol 48: 293Ð306. Zilahi-Balogh, G.M.G., S. M. Salom, and L. T. Kok. 2003c. Temperature-dependent development of the specialist predator Laricobius nigrinus (Coleoptera: Derodontidae). Environ. Entomol. 32: 1322Ð1328. Zilahi-Balogh, G.M.G., C. D. Broeckling, L. T. Kok, and S. M. Salom. 2005. Comparison between a native and exotic adelgid as hosts for Laricobius rubidus (Coleoptera: Derodontidae). Biocontrol Sci. Technol. 15: 165Ð171. Zilahi-Balogh, G.M.G., L. M. Humble, L. T. Kok, and S. M. Salom. 2006. Morphology of Laricobius nigrinus (Coleoptera: Derodontidae), a predator of the hemlock woolly adelgid. Can. Entomol. 138: 595Ð 601. Zilahi-Balogh, G.M.G., J. Jelı´nek, T. J. McAvoy, S. M. Salom, and L. T. Kok. 2007. Two new species of Laricobius (Coleoptera: Derodontidae) from China, and a key to Laricobius in the southeastern Palaearctic. Proc. Entomol. Soc. Wash. 109: 377Ð384. Received 3 March 2008; accepted 20 August 2008.