Sampling Buprestidae ( Coleoptera ) in Washington state with Cerceris californica Cresson ( Hymenoptera , Crabronidae )

The beetle-hunting habits of ground nesting wasps in the genus Cerceris Latreille have been recently exploited as a survey technique for exotic and native Buprestidae, particularly Agrilus planipennis Fairmaire (the emerald ash-borer). While such methods have been developed for the wide-ranging eastern Cerceris fumipennis Say, the survey potential of western buprestid-hunting Cerceris spp. has not been explored. Cerceris californica Cresson is the most well-studied of the western buprestid feeders, and the only one known to occur in Washington state. Here we report the results of surveys conducted in Washington in 2012–2013 for C. californica colonies, and numbers of buprestid beetles collected from monitored colonies. Eight C. californica colonies were found through visual search of 228 baseball fields and sandy clearings, but only four were large enough to monitor. Fifty-four beetles were recovered from the four colonies, comprising five native species. Four of these are new prey records for C. californica, and one (Chrysobothris quadriimpressa Gory & Laporte) is newly recorded from Washington. Cerceris californica colonies do not appear to be large or common enough in Washington to be a significant exotic buprestid survey strategy. However, even the limited monitoring resulted in more buprestid captures than nearby purple sticky traps, and monitoring C. californica nests may be a locally useful supplement for general buprestid surveys.


Introduction
The recent spread of two exotic metallic woodboring beetle species (Coleoptera, Buprestidae) in parts of North America has resulted in widespread tree mortality and looming long-term changes in ecological communities.Both are in the speciose genus Agrilus, which includes species capable of feeding upon and potentially killing live, healthy trees.The emerald ash borer (EAB), A. planipennis Fairmaire, 1888, was first detected in North America in 2002 (Cappaert et al. 2005, Poland andMcCullough 2006), although it may have been present for many years prior (Kovacs et al. 2009).To date, the beetle has been detected in 23 states and two provinces in the northeastern United States and Canada (http://tinyurl.com/lyaka3y, accessed April 18, 2014).The congener goldspotted oak borer, A. auroguttatus Schaeffer, 1905, (in some past literature treated under A. coxalis Waterhouse, 1889, a Mexican and Central American species) has been introduced into southern California from its native range, probably from southeastern Arizona or northern Mexico, although in Mexico it has been recorded only from Baja California Sur.It has caused extensive mortality among oak species and has been steadily expanding its range northward since 2002 (Coleman andSeybold 2008, Coleman et al. 2012).
There is considerable need to effectively monitor for these species and other exotic woodborers to maximize the possibility of early detection and eradication of newly established populations.Methods for detecting and monitoring Agrilus planipennis and A. auroguttatus include visual survey for impacted trees, use of trap trees, twig sampling, sticky ash leaves, and large sticky traps baited with plant volatiles or beetle decoys (Francese et al. 2008, Poland et al. 2010, McCollough et al. 2010, McCollough et al. 2011, Ryall et al. 2011, Crook et al. 2012, Domingue et al. 2012).Developing lures and traps that are effective at detecting A. planipennis at low densities has been an ongoing challenge, and an ideal trap design remains elusive.The current USDA-APHIS EAB trapping protocols employ large prismatic purple sticky traps and a sampling map derived from a risk-based model emphasizing high-risk locations.Traps are armed with volatile lures and deployed in the canopy of appropriately sized Fraxinus trees, ideally within the model-generated trapping cells (USDA-APHIS 2013).
An alternative buprestid survey and monitoring method developed in the eastern states and provinces exploits the biology of a ground-nesting wasp in the family Crabronidae, Cerceris fumipennis Say, 1837.The genus Cerceris includes hundreds of species worldwide, adults of which hunt and collect various beetle groups as a larval food source (Bohart and Menke 1976).The majority of species for which prey data are known target Curculionidae, with fewer species preying upon Tenebrionidae, Bruchidae, Chrysomelidae, and Buprestidae.In North America, Buprestidae are the primary prey items for the five species in Scullen's (1965) Group II.Cerceris fumipennis is the most thoroughly studied of this group, occurring from Ontario to Florida and west into Texas and Wyoming.More than 100 species of buprestids have been recorded as prey items, including adventitious species with which it shares no evolutionary history (Scullen and Wold 1969, Rutledge et al. 2011, Swink et al. 2013, Hellman and Fierke 2014).
In 2005, new provincial distribution records for several buprestid species collected from foraging Cerceris fumipennis in Ontario launched a research program to develop a buprestid detection and monitoring tool using the wasp, sometimes dubbed "biosurveillance" (Careless et al. 2014).The beetle-hunting habits of the wasp were already well known, but it had not previously been used to investigate the occurrence, distribution, or spread of buprestid species.Careless et al. (2014) developed a thorough methodology for using C. fumipennis to detect A. planipennis, which are readily captured by the wasps, and several eastern states and provinces have successfully implemented wasp-based monitoring programs.The first A. planipennis recorded in Connecticut was collected from a foraging wasp in 2012, concurrent with a capture on a purple sticky trap (Rutledge et al. 2013).
The other four North American Cerceris species known or presumed to prey upon Buprestidae occur in the western states and provinces, and northern Mexico (Scullen 1965, Bohart andGrissell 1975).Of these, C. californica Cresson, 1865, ranges from British Columbia south into Mexico, and east to Texas (where it occurs with C. fumipennis).Females are active from at least mid-June through August, when they construct simple nests in compacted sandy soil (Fig. 1) and provision them with paralyzed buprestid beetles.Although less studied than C. fumipennis, it remains one of the better understood Cerceris species in North America, with 29 recorded prey species (Linsley and MacSwain 1956, Evans and Rubink 1978, Nelson and Westcott 1991, Davidson 2003).By comparison, few prey records are documented for the other three species attacking Buprestidae (Davidson 2003, Looney and Westcott unpublished data).
This project explored adapting the eastern sampling methodology to Cerceris californica in Washington to examine the potential utility of this survey method.We report on survey results for C. californica nest sites in Washington and new buprestid prey records from monitored nests.Data are limited, but we also compare the location of monitored nests and beetles recovered from C. californica with the location and results of 2013 EAB trapping with purple sticky traps.The potential contribution of C. californica to exotic buprestid monitoring is evaluated in the context of the C. fumipennis program in the east, using criteria described in Careless et al. (2014).

Methods
There are no published locations of Cerceris californica nest sites in Washington.Potential wasp range was inferred from Scullen (1965) and specimens housed at Washington State University, the University of Idaho, the University of California, Riverside, and Oregon State University.In Washington, wasps have been collected only east of the Cascade Range, from localities such as White Swan, Prosser, Moses Lake, Buena, Wawawai, and Lake Paha.All of these specimens were collected from flowers or at large, not from nesting sites.Surveys for C. californica nesting aggregations were made in Washington between July 7-25 2012, and May through July 2013 (Fig. 2).Possible C. californica burrows can be identified by an evenly distributed ring of excavated soil, or tumulus (Linsley and MacSwain 1956;Fig. 1).
Following Careless et al. (2014) and Nalepa et al. (2012), nest search concentrated primarily on baseball diamonds in hope of maximizing opportunities to find large colonies that would lend themselves to buprestid monitoring.Baseball diamonds are often occupied by Cerceris colonies, and offer several benefits for Cerceris survey.They represent large areas of consistent habitat for wasps and are easily located using online satellite imagery (Nalepa et al. 2012).Baseball diamonds may not be the only occupied habitat in a region, but when wasps are locally present in an area they are more easily found in ball fields than other habitat (e.g.dirt roads/trails, exposed earth) (Careless et al. 2014, Hellman andFierke 2014).By focusing on easily-located baseball diamonds, towns may be relatively quickly surveyed for wasps.Baseball diamonds are typically located on public land (schools, parks) and are therefore usually accessible to government and citizen surveyors.Since they are located in parks, there are often numerous tree species planted relatively close, which provide good habitat for buprestid beetles.Drawbacks to focusing on baseball fields include potentially high levels of maintenance and disturbance, which may deter Cerceris nesting, and unfavorable soil formulations (Nalepa et al. 2012).Some campgrounds, dirt roads, and vacant lots were also included in the survey.
Beetle collection followed the methodology described in Careless et al. (2014) for Cerceris fumipennis.First, aggregations were censused and the location of all Cerceris nests was mapped.The threshold for monitoring was set at ~15 burrows (Table 1).Any beetles found on the ground while surveying were collected."Careless collars", plastic tabs with holes, were placed over each nest to mark location and impede entry by wasps returning with beetle prey (Careless et al. 2014).Frequency of wasps returning with and without beetle prey was recorded.Beetles were collected from wasps returning to the nest primarily by capturing them with a net.Wasps occasionally dropped the beetle they were carrying, and it was retrieved at that time.All buprestid specimens were retained.Beetles collected in 2012 were weighed and their length measured on the collection day, or immediately frozen and weighed soon after.All buprestid specimens were identified by one of us (RLW).The locations of confirmed colonies were compared with the 2013 USDA-APHIS EAB trapping cells and purple prism trap locations in 2013 (Fig. 3).Beetles retrieved in 2013 were also compared with beetles captured on four purple prism traps nearby, which were deployed from June 28 until 25 September, 2013.Three traps were located in Yakima, and one near Ellensburg (Fig. 3).

Results
In total, 228 baseball diamonds or sandy clearings were inspected for Cerceris californica in 2012 and 2013 (Fig. 2).Survey during 2012 detected eight C. californica colonies.Of these, only four (located in the cities of Yakima, Wenatchee, and Walla Walla) were robust enough to support monitoring in 2012 (Fig 2).One of these had only seven burrows, but was included due to the general paucity of colonies (Table 1).
Only the two large colonies located in Yakima were monitored in 2013.Observed nest density ranged from 1-24 nests in 2012, and from 1-34 in 2013 (Table 2).No colonies Cerceris californica activity varied widely between sites and days, from a low of zero wasps observed to a returning wasp observed every 4 minutes.Beetles collected from foraging wasps varied similarly, from 0-7 beetles recovered per hour of monitoring.In 2012, 36 beetles were collected during 28 hours of monitoring; 7 dropped beetles and 29 taken from wasps.In 2013, 18 beetles were collected over 19 hours of monitoring; 2 dropped beetles and 16 taken from wasps.No Buprestidae were captured on purple prism traps in the Yakima or Wenatchee area in 2013.
Beetle weights in 2012 ranged from 5.2 mg to 52.3 mg, and length between 5.21 mm and 10.4 mm.The lightest beetle was a dropped specimen, which appeared to be desiccated.Excluding this specimen, average weight was 19.0 mg, and the average length was 7.80 mm.Over both years five buprestid species were collected in Washington, four of which are new prey records for Cerceris californica (Table 2).One of these, Chrysobothris quadriimpressa Gory & Laporte, 1837, is newly recorded from Washington.Hellman, 1 specimen.This is probably the most widespread species of the genus in North America, likely occurring in every state and province (Nelson et al. 2008).Its primary hosts are Salix and Acer (Nelson et al. 2008).This species is also taken by C. fumipennis (Swink et al. 2013).Nelson et al. 2008), although it has also been reared from Larix occidentalis (Nelson and Westcott 1976).Nelson et al. (2008) erroneously listed it from "NB" instead of NE (c.f.Nelson and Westcott 1976).

Discussion
The percentage of search sites occupied by Cerceris californica observed in Washington appears to be much lower than for C. fumipennis in several eastern states.1).Colony size was also smaller than generally reported for C. fumipennis.
The largest colony located and monitored in this study included 33 burrows.In contrast, C. fumipennis colonies in the northeastern states and provinces could contain hundreds of burrows (Nalepa et al. 2012, Careless et al. 2014).It was expected but still disheartening to find no C. californica colonies west of the Cascade Range, where native ash and oak species are found in Washington.Activity at wasp colonies was rare by August in 2012, suggesting wasps may have been primarily foraging earlier in the season than indicated by museum records and published studies.In 2013, visits to the two largest colonies (in Yakima) began in May.Wasp activity was not observed until mid-June, and appeared to peak in early July; by August almost no wasp activity observed.This phenology is shorter than that observed for Cerceris fumipennis, which displays active enough foraging to enable monitoring through at least mid-August, and sometimes into September (Careless et al. 2014, Hellman andFierke 2014).
The number of wasps observed and the number of beetles collected per colony were both much smaller than anticipated based on work with Cerceris fumipennis.Careless et al. (2014), Hellman andFierke (2014), andSwink et al. (2013) recovered hundreds of beetles per year, from multiple colonies, compared with the 54 beetles collected in this study.The comparatively abbreviated collecting period in this study may account for part of this difference.Even so, wasp colonies were consistently smaller and less frequent, limiting the potential gains from C. californica monitoring.Dropped beetles were seldom found, which presents a challenge to locating colonies since the presence of dropped beetles is a useful indicator (Careless et al. 2014).It also limits the amount of survey data gleaned from each colony, since even damaged beetles collected from the ground can be informative (e.g.Grossbeck 1912, Swink et al. 2013).It is not clear whether the lack of dropped beetles reflects differences in biology or behavior between Cerceris congeners, or if it is characteristic of the environment near monitored colonies-low numbers of dropped and captured beetles could simply be a function of low buprestid density near the colonies monitored in this study.
The fairly small pool of prey species observed in this study may similarly derive from relatively poor buprestid habitat located near the study sites.Only five beetle species were retrieved from Cerceris californica, although existing literature indicates they prey upon many more species of appropriately sized buprestids given the opportunity.This species-depauperate catch also must be a product of lower abundance and diversity of Buprestidae in WA generally (Barr 1971), particularly compared to the Southwest where most of the C. californica prey records originate.For instance, the majority of the recorded prey (60%) of C. californica are of Acmaeodera spp., which is species-poor in WA and much more diverse in the southwestern USA.
The beetle species captured were not dissimilar in weight and size to Agrilus planipennis and A. auroguttatus.Since buprestid prey selection appears to be a function of availability and size (Hellman and Fierke 2014), Cerceris californica foraging habits observed here and described in other reports indicate that the species has the potential to be an effective monitoring tool if large enough colonies can be located.It is notable that, although the beetle diversity and numbers retrieved from Cerceris were low in each year, the purple prism traps placed in Yakima in 2013-which ran for approximately three months in the same general location-captured no buprestids.In contrast, even the limited C. californica monitoring in 2013 resulted in the capture of 18 beetles in four species.Although EAB purple prism traps are hung in ash tree canopies, potentially biasing the available species pool towards ash-feeding buprestids, other species are captured.For example, Buprestis aurulenta L., 1767, and Phaenops drummondi (Kirby, 1837), were recovered from purple prism traps in western Washington in 2012 and 2013 (J.Cena, personal communication), and detections of other adventive Agrilus spp.have come from purple sticky trap captures in other regions (Westcott 2007, Jendek andGrebennikov 2009).
Numerous other wasps and bees were common at baseball diamonds, both occupied and un-occupied by Cerceris.Common Hymenoptera genera observed during the survey included other Cerceris, Eucerceris, Philanthus, Bembix, Halictus, and Polistes (Fig. 4).Some of these species are similar in outward appearance to C. californica, and confusing them could lead to lost time or other complications.Polistes dominulus (Christ), 1791, has somewhat similar coloring and size to C. californica and flies relatively slowly.Mistaking these two species could lead to a nasty sting.Discriminating between nests of ground-dwelling species may also be confusing for the novice.Halictus farinosus Smith, 1853 (Halictidae) was collected from one of the ballfields with an active Cerceris colony.This bee makes nests with a tumulus similar to those of C. californica, although the species within can be determined by gently agitating the nest entrance with a stick and eliciting a defensive showing from the occupant.Of the species commonly encountered with C. californica, Philanthus gibbosus Fabricius, 1775, was the most similar in appearance and behavior.It can be distinguished by its irregular tumulus (Fig. 1) and its bee prey.A colony of Eucerceris flavocincta Cresson, 1865, was found at a survey site west of the Cascade Range, where C. californica is not known to occur.However, the known range of E. flavocincta includes much of C. californica's range (Scullen 1968, Bohart andGrissell 1975).While similar to C. californica in color and behavior, E. flavocinta preys upon Curculionidae (Bohart andGrissell 1975, Scullen 1968).Cerceris nigrescens Smith, 1856 was also found at one site; it too preys upon weevils (Bohart and Grissell 1975).
Following the novel suggestion that beetle-hunting Cerceris wasps could be a viable tool in buprestid surveys (Marshall et al. 2005), significant effort has been made to develop and prescribe useful methods of exploiting this phenomenon.Careless et al. (2014) identified three general considerations for evaluating the potential effectiveness of Cerceris as a buprestid survey tool: accessibility, productivity, and sustainability."Accessibility" refers to the temporal, geographic, and behavioral occurrence of the wasp species of interest.Cerceris californica nests located in this study were rare, and did not seem to occur in enough sampling locations to be a major source of buprestid survey records.Large colonies were even rarer, limiting effectiveness of this monitoring tool.Colonies of C. californica were fairly active during the day, although their seasonal activity may be abbreviated when compared with C. fumipennis.
While far fewer beetles were recovered in this survey than most utilizing Cerceris fumipennis, C. californica was still relatively "productive", i.e. collected beetles at a rate that equals or exceeds other methods (Careless et al. 2014).We cannot address how effective hand sampling or beating might have been for buprestids, but in 2013, more beetles were recovered from C. californica colonies than nearby purple sticky traps.The sticky traps were sited to maximize opportunity to catch EAB -in the upper canopy of ash trees -so this comparison doesn't directly address their targeted effectiveness for Agrilus planipennis.Even so, the discrepancy in 2013 between beetles captured on traps (0) and those collected in only 19 hours of buprestid monitoring (18) suggests that C. californica could play at least a small role in Washington as a supplemental survey technique.Linsley and MacSwain (1956) report hundreds of buprestids recovered from 25 C. californica wasps and excavated burrows, indicating that with greater monitoring effort in the right habitat this wasp is productive indeed.No doubt more species and individuals would have been recovered from even the few sites monitored here had there been more collecting time.The last of the effectiveness criteria, "sustainability", concerns the ability of a Cerceris species to tolerate repeated interference by humans.This aspect of C. californica biology was not directly examined in this study, although Linsley and MacSwain (1956) report that females would continue to embark upon foraging flights, even after having five beetles in a row removed.We can only add the general observation that in this study individual wasps also embarked upon a foraging flight after having a beetle removed.Ultimately, developing a Cerceris monitoring program for exotic buprestids in the western states may be more productive in Oregon or California, where host plant (e.g.Fraxinus, Quercus) habitat and buprestid-hunting wasp habitat overlap.However, as was the case for Washington before this study, no significant efforts have been made to locate and map colonies of C. californica or buprestid-feeding Cerceris.This would be a critical first step to monitoring with this species in the western states, particularly since colony frequency and size seem to be the major limiting factors in Washington.

Conclusions
Washington Cerceris californica colonies are less common and smaller than C. fumipennis colonies in the eastern states.Colonies of C. californica appear to be restricted to east of the Cascade Range in Washington state, and were seldom located near EAB trapping sites.Cerceris californica captures beetles within the size and weight range of target Agrilus spp.Despite low total numbers, C. californica wasps were more effective tools for general buprestid sampling in the Yakima area when compared with nearby EAB purple traps, and citizen monitoring may be a useful supplemental monitoring program in select cities in eastern Washington given such effective foraging.The monitoring potential of buprestid-hunting Cerceris spp.may be greater in other western states and should be studied.

Figure 3 .
Figure 3. Monitored Cerceris californica colony locations compared with 2013 EAB grids and purple prism trap locations.

Figure 4 .
Figure 4. Hymenoptera species commonly found at Cerceris californica colonies.The non-Cerceris species were observed at the majority of surveyed sites, irrespective of C. californica's presence.Philanthus gibbosus was the most notable and similar species likely to be found with C. californica.