We based this study on more than 12000 specimens. Most specimens are preserved in collections, but many (over 3000 specimens) were part of surveys conducted in eastern Canada and south of the Great Lakes in the United States following the establishment of Sirex noctilio Fabricius. Most of these specimens were not retained. The following is a list of collections with their respective curators.

Collection of samples: Woodwasps for the DNA analysis portion of this study were collected by numerous collaborators or the authors using 3 different methods. They were netted or hand-collected, especially at forest fires; reared from host material; or collected in Lindgren funnel or panel traps baited with terpenes and/or ethanol. The trapped specimens were mostly collected as by-products of bark beetle trapping programs. Specimens were frozen, preserved directly in 70%-95% ethanol or collected into diluted ethylene glycol or similar preservative and then transferred to 70%-95% ethanol. Specimens were accumulated at the USFS–MS, CNC, and PFRC for DNA analysis.

Most specimens were studied and images taken with a MZ16 Leica binocular microscope and an attached Leica DFC420 digital camera. Some specimens were photographed using a DSLR Canon Rebel Xti camera with a 100 mm macro lens. Multiple images through the focal plane were taken of a structure and these combined using Combine ZM or ZP designed by Alan Hadley to produce a single, focused image. Specimens were illuminated with a 13 watt daylight fluorescent lamp.

DNA Isolation. DNA was isolated, amplified and sequenced both in Guelph and Stoneville, MS. DNA from specimens from Ottawa and Victoria were sequenced in the Biodiversity Institute of Ontario, Guelph, ON, according to standard protocols (as detailed in Fernandez-Triana et al. 2011). Protocols used in Stoneville were as follows. Tissue for extraction was collected from the thorax either by pulling off a hind leg and collecting the muscle tissue still attached to the coxa or by digging tissue directly from the thorax with a pair of forceps. Genomic DNA was isolated from the tissue using either a slightly modified Quiagen DNeasy spin-column protocol for animal tissues or the Masterpure™ Yeast DNA Purification kit by Epicentre (Madison, WI). We modified the DNeasy spin–column protocol by changing the conditions of the proteinase K incubation from 1–3 hrs at 56° C to 1 hr at 70° C and by changing the final elution solution from 200μl Buffer AE to 50μl Buffer AE plus 200μl Ambion nuclease free water. In all extractions, care was taken to avoid digestive tract tissue and eggs which might contain microbial contaminants such as Wohlbachia sp. Early in the study, a Wohlbachia species was sequenced from a woodwasp but not from a species used in this study. We have sequenced more than 1000 woodwasps (leg or thorax tissue) since then with no further discovery of Wohlbachia.

Amplification and clean up. Over the course of the study several PCR reaction amplification protocols were used successfully. The most evolved and preferred protocol is very similar to that used by Roe et al. (2006). PCR reactions containing 10μl of DNA template, 9μl of Ambion nuclease free water, 2.5 μl Advantage 2 10X buffer (Clontech, Mountain View, CA), 2 μl of each oligo (each at 10mM), 1.5 μl of dNTP mix (each at 10mM) and 0.4 μl of Advantage 2 Taq, were amplified in a PTC-100 Programmable Thermal Controller (M. J. Research Inc.) as follows; an initial denaturation step at 94°C for 2 minutes followed by 35 cycles of 94°C for 30 seconds, 45°C for 30 seconds and 68°C for 2 minutes, followed by a final extension at 68°C for 10 minutes. The extension steps were at 68°C rather than 72°C because Advantage 2 Taq is more efficient at the lower temperature (Manufacturer’s instructions). The oligos used were LCO1490: 5’-ggtcaacaaatcataaagatattgg-3’and HCO2198: 5’-taaacttcagggtgaccaaaaaatca-3’of Folmer et al. (1994) where the numbers refer to the position of the Drosophila yakuba 5’ nucleotide. PCR Products were visualized on 30% acrylamide/bis gels (mini Protean II electrophoresis cell by BioRad) stained with either ethidium bromide or preferably EZ-Vision 2 (N650-Kit by Amresco Inc.). PCR products were cleaned using an Exo-SAP protocol. Up to 20 μl of PCR product was mixed with 8μl of Exo-SAP (2μl Exonuclease I at 10U/μl, USB product no. 70073Z, Cleveland, OH; 20 μl Shrimp Alkaline Phosphatase at 1U/μl USB product no. 70092Z, Cleveland, OH; 78 μl ddH₂O) and heated to 37°C for one hour followed by 15 minutes at 80°C.

Sequencing. Double stranded PCR products (at least 20ng/μl) were sequenced on an ABI 3730xl sequencer (Applied Biosystems, Foster City, CA) using BigDye 3.1 in 10μl reactions (1.75μl 5X sequencing buffer, 0.5 μl BigDye 3.1, 0.8 μl 10 μM primer, at least 20 ng DNA template and water up to 10 μl). DNA template was quantified by comparison to Low DNA Mass Ladder (Invitrogen cat. No. 10068-013, Carlsbad, CA), at least 1 μl of template was used even if the concentration of DNA appeared to be significantly greater than 20 ng/μl. The cycle sequencing reaction was 2 minutes at 96°C followed by 25 cycles of 96°C for 30 seconds, 50 °C for one minute and 60°C for 4 minutes. The sequencing reaction (10μl) was stopped by addition of 2.5 μl 0.125 M EDTA (pH 8.0) followed by centrifugation at 4000 rpm for one minute. The products were precipitated for 30 minutes in the dark by addition of 30μl of 100% ethanol followed by centrifugation at 4000 rpm for 30 min at 4°C. The samples were washed with 100μl of 70% ethanol spun for 15 minutes at 1650 rpm for 15 minutes and then air-dried in the dark for 15 minutes. Dried products were stored at -20°C until injection. Products were re-upped in 100μl of deionized water, centrifuged at 4000 rpm for 2 minutes and injected immediately into the sequencer using the ABI default injection module appropriate for the installed capillary array, but decreasing the injection time to 2 sec.

Data Manipulation. Sequences were captured using Data Collection Software v3.0 with Dye set Z_BigDyeV3 from Applied Biosystems which gave us ab1. sequence trace files and seq. sequence text files. Templates were sequenced in both directions and the corresponding sequences were paired into individual specimen contigs using Lasergene Seqman by DNAStar. To obtain full length sequences it was sometimes necessary to sequence individual specimens several times and combine the partial sequences to form the final sequence used for analysis. Individual specimen contigs were aligned using Clustal V, and built into trees (Neighbor Joining) (Saitou and Nei 1987) using Megalign also by DNAStar.

Exclusion of Numts and Heteroplasmy. Two of the potential pitfalls of using mitochondrial sequences for identification include mistakenly sequencing nuclear pseudogenes of mitochondrial origin (NUMTs), or obtaining multiple sequences from heteroplasmic individuals. To reduce the risk of NUMTs we were careful to select only muscle (mitochondrial rich) tissue from specimens and all sequences were translated and inspected for stop codons and insertions and deletions (characteristics of pseudogenes). To date, all siricid sequences have been free of stop codons, insertions and deletions. Heteroplasmy is when an individual has more than one mitochondrial haplotype (sequence). To reduce possible variation due to heteroplasmy we sequenced double stranded PCR products directly rather than sequencing clones. If there were rare alternate haplotypes they would be masked by the most common haplotype. We further sequenced many individuals multiple times with no variation (data not shown).

Although siricids are large and colorful insects, they are not commonly encountered in general collecting in forests and more specialized techniques are often used to obtain them. These methods fall into three general categories: collecting in specific habitats based on knowledge of siricid behavior, trapping using a variety of different traps, and rearing from infested wood. With the recent discovery of Sirex noctilio in North America (Hoebeke et al. 2005, deGroot et al. 2006) there has been increased interest in surveys for S. noctilio and other siricids and the techniques below are evaluated in light of their utility for survey work.

Active collecting. Like many wood-boring insects, S. noctilio and presumably other siricids are attracted to the volatiles produced by wounded, stressed or dying trees (Madden 1971, Newmann et al. 1982). In some circumstances a single, cut tree can be attractive. NMS and Paul Lago collected more than 100 specimens of S. nigricornis and many other wood borers and parasitoids over a 3-day period in October, 2001, on a single loblolly pine (DBH approximately 30 cm) that was cut into approximately 50 cm bolts at a semi rural-setting in Oxford, Mississippi. Unfortunately, this was a rare occurrence; NMS has attended many freshly cut trees that were not visited by siricids. Presumably, in Oxford, there was a local population of recently emerged S. nigricornis and the cut loblolly pine was the only local source of volatiles.

Most often, siricids are attracted to areas where there are many wounded trees. In Western North America, siricids are commonly found at forest fires. Males form mating aggregations high up on unburned trees at the edge of forest fires and females can be found ovipositing into freshly burned stumps or trees (Middlekauf 1960, Middlekauf 1962, Westcott 1971, Schiff unpublished data). Larvae can develop in the fire-killed trees and adults sometimes emerge from houses built with salvaged lumber (Middlekauf 1962, Lynn Kimsey personal communication). Siricids are also found at logging decks and at mills where the cut trees presumably release attractive volatiles (Wickman 1964, Wood Johnson personal communication). Siricids can be surveyed at fires and mills but these are not always located in the study area of interest.

Siricids are also known to “hill-top”. Males and females of many widely dispersed insect species find mates at prominent landscape features like the tops of hills. Typically, there are more males than females and the host plants do need to be present as the females can fly to the host after mating. “Hilltopping” is probably much more common than has been reported because it is unusual to find a hill top with short vegetation where it can be observed (for general information, see Skevington (2008)). Similar behaviour has also been noted on fire towers. Specimens of Urocerus sah and Xeris melancholichus were collected over several years at the top of Mount Rigaud in eastern Canada (Fig. A2.1). At the same site, males of many species of Diptera, Lepidoptera, other Hymenoptera and Coleoptera were observed in similar aggregations. Among Hymenoptera, males of Xiphydria spp., Trichiosoma triangulum Kirby and Cimbex americana Leach were commonly collected with only occasional females being collected. This phenomenon is widespread. J. O’Hara, a dipterist, collected many males of Sirex obesus Bradley on hill tops in Arizona and New Mexico, Chapman (1954) recorded numerous males of Urocerus flavicornis on a mountain top in western Montana, and Jennings and Austin collected or recorded nine males of Austrocyrta fasciculata Jennings and Austin (Xiphydriidae) aggregating on top of Mount Moffatt and Mount Rugged in Queensland, Australia (Jennings et al. 2009).

Trapping. Siricids are most commonly collected by three trapping methods: 1) flight intercept trapping, 2) using artificial tree-mimicking traps baited with a chemical lure and 3) using trap or lure trees.

1. The most commonly used flight intercept trap is the Townes style Malaise trap (Townes 1972). Although Malaise style traps were designed to catch Hymenoptera, including Symphyta, they only occasionally catch siricids (Smith and Schiff 2002) and are generally considered to be too expensive to use for siricid surveys.
2. The use of artificial tree-mimicking traps with lures for siricids is largely a byproduct of bark beetle trapping programs. In fact, the discovery of S. noctilio in the United States resulted from the identification of a siricid caught in an exotic bark beetle survey funnel trap (Hoebeke et al. 2005). Almost all the survey work since the discovery of S. noctilio in North America has used artificial traps. The traps most commonly used are the Lindgren multiple-funnel trap and the cheaper cross-vane trap (Figs. A2.2 and A2.3). In silhouette, the traps mimic tree trunks and both use liquid filled collecting vessels. Typically the traps are baited with lures that mimic host volatiles of a wounded tree, namely a combination of monoterpenes and/or ethanol. These traps are relatively cheap and easy to assemble and service but like the Malaise trap they are not particularly efficient. In a 1999 study of five types of traps, 1661 siricids were collected over 5300 trap days for a trapping rate of approximately one siricid every three days. Presumably these are optimal results because the traps were located around a mill considered to be a wood-borer rich environment (McIntosh et al. 2001). The relatively low efficiency of these traps may be a function of the type of lure. These baited traps likely compete with all the stressed or damaged trees in the area, which reduces their effectiveness. Presumably trapping would be more efficient if the traps were baited with specific sex pheromone lures but none have been identified for Siricidae to date although components of contact sex pheromones for S. noctilio have recently been reported (Böröczky et al. 2009). An anomaly of artificial traps is that they seldom catch male siricids. We believe this is because traps are normally positioned with the top approximately two meters from the ground to facilitate collecting samples and male siricids spend most of their time in tree tops.
3. Originally, “trap” trees were used as a means to detect the presence of S. noctilio in Southern Hemisphere Pinus radiata plantations. Selected trees that were mechanically wounded were found to be attractive to S. noctilio, depending on the season and degree of wounding. Felled trees were attractive immediately but only susceptible to attack for about 2 weeks whereas girdled trees were not attractive for 9–12 days but remained attractive for a season or more (Madden 1971, Madden and Irvine 1971). The method was later refined by switching to use of a chemical herbicide instead of mechanical wounding (Morgan and Stewart 1972, Minko 1981, Newmann et al. 1982) and the trap trees evolved into a delivery system for parasitic nematodes as well as a means of detecting S. noctilio. Once the wounded trap tree was infested with S. noctilio, it would be felled and inoculated with nematodes. The nematodes would attack the larvae and be distributed when the adult woodwasps emerged. In the northern United States, the suitability and attractiveness of trap trees for S. noctilio is dependent on timing of herbicide injection and host tree species (Zylstra et al. 2010). Although this is the preferred method for detecting S. noctilio and delivering the parasitic nematode to control infestations in the Southern Hemisphere, it is labor intensive for survey work and requires landowner consent to wound trees. As far as we know trap tree methods have not been developed for any native species.

Rearing. Perhaps the best way to collect siricids is by rearing them from infested logs. The advantages of this method are that males are often reared along with females, the host tree can often be positively identified, and living specimens can be obtained for biological studies. This method can also be proactive. Specimens of Urocerus taxodii for this study were reared by wounding three bald cypress trees in the Delta National Forest, Sharkey Co., Mississippi, waiting for them to be attacked and later caging 1.5 meter bolts from the trees at the USFS–MS. Many other specimens in this study were also reared from wounded trees as part of a decade long Canadian Forest Service wood borer survey (as in Figs. A2.4, A2.5 and A2.6). Disadvantages include difficulty finding suitably infested trees and the space and time required for rearing. NMS has found siricid-infested trees by following siricid specific parasitoids like the giant ichneumonid wasps Megarhyssa spp., and looking for siricid damage such as perfectly round emergence holes. In some cases, after multiple drillings, female siricids and/or Megarhyssa can no longer withdraw their ovipositors and they become stuck and die. Ants or birds dispose of the bodies but the ovipositors sometimes remain protruding from the wood, indicating siricid infested trees (Spradberry and Kirk 1978, Schiff, unpublished data). Another clue is to look for the characteristic brown staining in cut timber resulting from the symbiotic fungus, Amylostereum sp. (Spradberry and Kirk 1978, Tabata and Abe 1997).