Morphological and Genetic Reappraisal of the Orius Fauna of the Western United States (Hemiptera: Heteroptera: Anthocoridae)

Annals of the Entomological Society of America, 109(2), 2016, 319–334 doi: 10.1093/aesa/sav155 Advance Access Publication Date: 7 January 2016 Systematics Research article

David R. Horton, Tamera M. Lewis, Stephen F. Garczynski, Kelly Thomsen-Archer, and Thomas R. Unruh Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

USDA-ARS, 5230 Konnowac Pass Rd., Wapato, WA 98951 ([email protected]; [email protected]; steve. [email protected]; [email protected]; [email protected]) and 1Corresponding author, e-mail: [email protected]

Received 27 August 2015; Accepted 8 December 2015

Abstract Examination of minute pirate bugs, Orius spp. (Hemiptera: Heteroptera: Anthocoridae), from a broad geographic range in the western United States prompted a reappraisal of the taxonomic composition of the fauna native to the western United States and Canada. Current syntheses and catalogs list three species of Orius native to this region. In a previous study, we showed how geographic variation in external traits of one of these species, Orius diespeter Herring, 1966, had led to mistakes in identification of species within this complex. More extensive collecting efforts have now led to the discovery of specimens having traits not fully consistent with descriptions of any described species. We provisionally categorized these unresolved specimens into one of eight phenotypic groups based upon combinations of body size, visual appearance of genitalia, association with specific plant taxa, and geographic source. Genitalia from 382 specimens were then measured to determine whether phenotypic groupings were confirmed by statistical analysis of genitalic morphology. Principal components analysis showed that size and shape of the male’s paramere differed among phenotypes. The paramere of unresolved specimens often diverged from the paramere of described species. Length of the female’s copulatory tube differed between several of the unresolved phenotypes and described species. Analysis of DNA sequences showed that five of the eight phenotypes diverged genetically from other phenotypes and from described species. DNA sequence data did not separate two described species (Orius tristicolor (White, 1879) and Orius harpocrates Herring, 1966), suggesting that these two species are a single species. The combined morphological and genetic evidence indicates that the Orius fauna of the western United States is composed of a mix of two described species and possibly five undescribed cryptic species. We summarize the known distributions of described and cryptic undescribed species, and discuss the implications of our work for the biological control community.

Key words: Orius, genitalia, mtDNA, cryptic species

Insects in the Family Anthocoridae (Hemiptera: Heteroptera) are Hawaii (Davis and Krauss 1963) as components of classical biologi- predators of small soft-bodied arthropods in crop and natural sys- cal control attempts. tems worldwide (Lattin 1999). In the continental United States and In many areas worldwide, the Orius fauna is composed of com- Canada, the Anthocoridae includes about 90 species (Henry 1988), plexes of often highly similar species. These complexes may occupy of which members of the genus Orius Wolff, 1811, are the most the same agricultural fields, stands of vegetation, or even the same familiar. There are 70 known species of Orius worldwide individual plant (e.g., Knowlton 1949, Lykouressis and Perdikis (Yasunaga 1997). Species of Orius are important sources of biologi- 1997, Ohno and Takemoto 1997, Lewis et al. 2005, Shapiro et al. cal control on row crops, in greenhouses, and on flowering orna- 2009, Lewis and Horton 2010). Because life histories and impact on mentals (Lattin 1999). Several species of Orius are reared prey almost certainly vary species-to-species within complexes (Ito commercially for release in North America and Europe (Cranshaw and Nakata 1998, Tommasini et al. 2004), it is important that spe- et al. 1996, van Lenteren et al. 1997). North American species of cies within complexes are correctly identified. Similarity in appear- Orius have been released in Europe (Tommasini et al. 2004) and ance among species in some Orius complexes can make it very

Published by Oxford University Press on behalf of Entomological Society of America 2016. This work is written by US Government employees and is in the public domain in the US. 319 320 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 difficult to separate them, and mistakes in identification are not size, visual appearance of genitalia, plant source, and geographic uncommon (e.g., Shapiro and Ferkovich 2009, Lewis and Horton source. We then measured genitalia of a subset of identified and 2010, Lewis and Lattin 2010). unresolved specimens, and examined whether size and shape of geni- The two senior authors of this paper recently conducted an talia differed among phenotypes or between unresolved phenotypes extensive reevaluation of the morphology and geographic distribu- and described species. Lastly, we determined whether phenotypic tion of a North American species of minute pirate bug, Orius dies- groupings were consistent with differences among groups in DNA peter Herring, 1966 (Lewis and Horton 2010). Several hundred sequences. Variation in morphology of genitalia accompanied by specimens of Orius from throughout North America were examined clear differences in DNA sequences would be evidence for the exis- in that study, and it became clear from those specimens that the tence of a cryptic species complex within the North American fauna Orius complex in western North America includes insects whose of Orius. external traits and genitalic morphology are not fully consistent with descriptions of O. diespeter or other described species. The dis- crepancies between literature descriptions and specimens were often Genitalia of Orius Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 quite subtle, and a number of specimens whose external appearance An important structure for identifying species of Orius is the left identified them as either O. diespeter or a common western species, paramere of the male, which has been used in taxonomic treatments Orius tristicolor (White, 1879), were found upon examination of of Orius beginning with the work of Ribaut (1923). As in other male or female genitalia to depart from descriptions of genitalia Anthocoridae, the right paramere has been lost, while the left func- available in earlier publications. tions as a copulatory organ (Carayon 1972, Pe´ricart 1972). The scle- These observations prompted us to conduct a quantitative rotized organ is situated in a small recessed area on the dorsal part examination of genitalic morphology for the Orius fauna native to of the male’s ninth abdominal segment (Fig. 1A). During mating, the western North America. We examined Orius collected from male extends his abdomen beneath the female, and rotates the para- throughout the western half of the United States, with more limited mere until the flagellum and associated, membranous endosoma can collecting of specimens from the southcentral and southeastern part be inserted into the female’s genitalia. The paramere is a complex, of the country. Specimens that could not readily be identified as three-dimensional structure, accompanied on its dorsal surface by described species by descriptions and illustrations in the literature an apically tapered cone and flagellum (Fig. 1A). The cone and flag- (White 1879; Kelton 1963, 1978; Herring 1966; Lewis and Horton ellum often differ in appearance among species, and the two struc- 2010) were assigned provisionally to one of eight phenotypic tures are used extensively in taxonomic treatments of Orius. In groups. These groups were delineated by a combination of body some especially difficult complexes, species cannot be identified

Fig. 1. (A) Abdomen and paramere (in dorsal view) of two Old World species of Orius (after Yasunaga 1997); (B) copulatory tube of O. tristicolor (in dorsal view); and (C) photograph of the copulatory tube of O. tristicolor showing thickened basal region and membranous region terminating in sperm pouch (in dorsal view). Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 321 without examination of parameres (Ghauri 1972, Pe´ricart 1972, misidentifications of specimens (Fig. 3A; see the discussion of Yasunaga 1997, Lewis and Horton 2010). O. diespeter below). Less often, morphology of the female’s genitalia is used to identify The taxonomic status of O. tristicolor was uncertain throughout species of Orius (Pe´ricart 1972, Yasunaga 1997, Postle et al. 2001, the late 1800s and first half of the 1900s, due to questions about Lewis and Horton 2010). A specialized structure (copulatory tube; whether White had described a valid species or whether the species Carayon 1953) of female Orius acts to receive the genitalia of the was merely a color variant of the common eastern minute pirate male and to channel the male’s intromittent organ to the sperm stor- bug, Orius insidiosus (Say, 1832). Indeed, many early 1900s cita- age organ located at the terminal end of her copulatory tube (Fig. 1B tions actually refer to O. tristicolor as a color variant of O. insidio- and C). The structure opens on the ventral surface of the female sus (Parshley 1920, 1922, 1923; Blatchley 1926; Torre-Bueno within the membrane between abdominal segments VII and VIII, typi- 1930). Kelton (1963) examined the paramere of both species, pro- cally offset from the midline of the insect. The copulatory tube varies viding the first illustration of this organ for both taxa, and con- extensively among species in length, width, and curvature, and can be cluded that O. tristicolor and O. insidiosus were indeed valid used to identify difficult species in the absence of male specimens species (Fig. 4A). The paramere of O. tristicolor has a long, sweep- (Pe´ricart 1972, Yasunaga 1997, Hernandez and Stonedahl 1999, Bu ing, and slender flagellum, in contrast to the much stouter and Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 and Zheng 2001, Lewis and Horton 2010, Shapiro et al. 2010). shorter flagellum of the paramere of O. insidiosus (Fig. 4A). Orius diespeter and O. harpocrates were both described by J. L. Herring in 1966. The description of O. diespeter was based upon Taxonomic History of the Western North one male and one female specimen collected in the 1930s from western British Columbia, Canada. Orius harpocrates was described American Fauna from a single male specimen collected near San Francisco in 1935. The Orius fauna of the western United States and Canada includes According to Herring (1966), O. harpocrates is larger than both three species native to North America (Henry 1988, Maw et al. O. tristicolor and O. diespeter, and (unlike O. tristicolor) is a deep 2000)1: Orius tristicolor (White, 1879), Orius diespeter Herring, chestnut brown color (Fig. 2B). We have seen but one literature 1966, and Orius harpocrates Herring, 1966. Orius tristicolor is the reference to O. harpocrates since Herring’s description in 1966 (a widespread and familiar bicolored species (Fig. 2A) seen throughout listing of the species from an island off of the coast of California; the western United States and northwards into western Canada Miller and Davis 1985); however, we readily collect O. harpocrates (Fig. 3A). Its range southwards extends through Central and South from flowering herbaceous plants at the type location (T.M.L. and America as far south as Argentina (Carpintero 2002). The species D.R.H., unpublished data). Herring (1966) separated O. diespeter was described by F. Buchanan White in 1879 (as a species of from O. harpocrates by size and pubescence, although we find that Triphleps Fieber, 1860) from two specimens collected in California. body size is variable enough to overlap between the two species The source location in California is unknown. Location of the type (T.M.L., unpublished data). Orius diespeter was distinguished from specimens is also unknown (Herring 1966). White’s description of O. tristicolor by its dark brown color (Fig. 2C). Herring (1966) Triphleps tristicolor was based entirely on external traits, including included illustrations of the paramere of O. tristicolor and O. har- size, color, pronotal shape, and punctation. Triphleps Fieber, 1860 pocrates (Fig. 4B). Herring’s illustration of the O. tristicolor para- was later synonymized with Orius Wolff, 1811 (Schumacher 1922). mere is similar to that in Kelton (1963). The paramere of Catalog listings and taxonomic treatments have suggested that O. O. harpocrates is similar in shape to that of O. tristicolor (Fig. 4B). tristicolor occurs throughout western North America, northwards Herring’s illustration of the O. diespeter paramere was incorrect into Alaska and the northwestern Canadian provinces, and east- (see next paragraph), so it is not reproduced here. wards into the New England states and eastern Canada (Kelton Kelton (1978) reexamined the type specimen of O. diespeter, 1963, 1978; Herring 1966; Henry 1988; Maw et al. 2000). and found that the paramere used by Herring (1966) as the source However, our examination of specimens during the last decade sug- of his illustration had a damaged flagellum. Kelton provided a new gests that the geographic range of O. tristicolor in North America is illustration of the O. diespeter paramere from specimens collected at substantially more limited than this, due to widespread the type locality in British Columbia, and showed that the paramere of O. diespeter has a slender, long flagellum similar to that of the O. 1 A very common species in eastern North America, Orius insidiosus tristicolor paramere (Fig. 4C). Kelton (1978) updated the geographic (Say, 1832), has been collected in some geographic locations west of distribution of O. diespeter to include Alberta, Canada, and Scudder the Great Plains (Knowlton 1949, Herring 1966, Gillespie and Quiring (1997) recorded O. diespeter from Yukon, Canada. 2006, T.M.L., unpublished data). Some records in western North Lewis and Horton (2010) showed that O. diespeter is consider- America may be due to introductions, as this species is sold commer- ably more variable in color than indicated by Herring (1966) or cially for use in biological control programs. Genitalic morphology of Kelton (1978), and that specimens of O. diespeter often have a this species is so different from other western species (Herring 1966, bicolored appearance (Fig. 2D) rather than the dark coloration Kelton 1978) and from our unresolved phenotypes, that we have not described by Herring (1966). Bicolored specimens are similar in included O. insidiosus as part of this synthesis. A Holarctic species external appearance to O. tristicolor, and examination of genitalia not included in North American faunal lists, Orius sibiricus Wagner, is often required to separate the two species (Lewis and Horton 2010). In both species, the flagellum is arc-shaped, but the flagellum 1952, is also excluded from the present study. This species was previ- of the O. diespeter paramere is shorter than that of the O. tristicolor ously known only from the Palearctic Region but was recently found paramere, and completes a lesser portion of a semicircle than shown to occur in Yukon, Canada (Lewis et al. 2015). While O. sibiricus is by the O. tristicolor flagellum (Fig. 4D, panels 1-2). Length of the quite similar in color to the melanic form of O. diespeter, and has flagellum can easily be compared between species by locating the tip been misidentiﬁed as O. diespeter (Lewis et al. 2015), it differs from of the structure in relationship to an imaginary horizontal line O. diespeter in sculpture of the dorsum, shape of the pronotum, and extending from the base of the flagellum through the apex of the shape of the paramere. paramere cone (this imaginary line shown as a dotted line in 322 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 2. External appearance of (A) O. tristicolor (Bear Creek, CO); (B) O. harpocrates (San Francisco, CA); (C) melanic O. diespeter (Peshastin, WA); and (D) nonmelanic O. diespeter (Lovell, WY). Photographs by T.M.L.

Fig. 4D, panel 1). The flagellum of the O. tristicolor paramere that there is no conclusive evidence that O. tristicolor occurs in east- reaches that imaginary line, whereas the flagellum of the O. dies- ern North America (Fig. 3A). As far as we have been able to deter- peter paramere terminates well short of the line (Fig. 4D, panel 2). mine, records of O. tristicolor from Alaska and northwestern Lewis and Horton (2010) also provided photographs of the cop- Canada are also in error, and again are misidentifications of O. dies- ulatory tube of O. diespeter and O. tristicolor (Fig. 4D, panel 3). peter (T.M.L., unpublished data). Our studies suggest that the The copulatory tube of both species is composed of three parts northern range of O. tristicolor in western North America may not (Figs. 1B and 4D): a thickened, relatively rigid section that begins at extend much beyond central British Columbia (Fig. 3A). the tube’s opening and extends most of the way to the sperm pouch; a small, relatively indistinct collar at the distal end of the sclerotized basal section; and a short membranous section that enters the sperm Characteristics of Unidentified Specimens pouch. The copulatory tube of O. diespeter is shorter and is posi- Specimens of Orius were collected from multiple sites primarily tioned nearer the insect’s midline than the copulatory tube of west of the Great Plains (Fig. 5). A few locations from outside of O. tristicolor (see the vertical lines in Fig. 4D, panel 3). Finally, this targeted region were also sampled, prompted by our discovery Lewis and Horton (2010) updated the distribution of O. diespeter in Texas and Florida of specimens having genitalia superficially sim- well beyond the limited range indicated by Herring (1966), Kelton ilar in appearance to genitalia of O. diespeter, but occurring well (1978), and Scudder (1997) to include Alaska, most of the northern outside of the geographic range of O. diespeter (Fig. 3B). The major- states of the continental United States, and western, southcentral, ity of specimens were collected as adults from the field over the pre- and southeastern Canada (Fig. 3B). They also suggested that the vious 15 yrs by T.M.L. and D.R.H. A few specimens were records of O. tristicolor from eastern North America made in the collected as nymphs and reared to adulthood in the laboratory on a early- to mid-1900s (Parshley 1920, Torre-Bueno 1930, Kelton diet of Ephestia (Lepidoptera) eggs. Insects were collected using 1963, Herring 1966) actually were of nonmelanic O. diespeter, and sweep nets or by aspirating specimens from beating sheets. Plant Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 323 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 3. Current best estimates of geographic distribution of O. tristicolor (A: gray shading), O. harpocrates (A: red circle), and O. diespeter (B) in the continental United States and Canada. The ranges shown are based upon our collecting efforts, examination of museum specimens, and records in Lewis and Horton (2010).

Fig. 4. Line drawings and photographs of the male and female genitalia of O. tristicolor, O. diespeter, and O. harpocrates taken directly from the listed publications. Parameres in panels (A–C) are not to scale. Vertical lines in panel (D3) show location of midline of insect. Line drawings in panels A and C are used with per- mission of The Canadian Entomologist, and Agriculture and Agri-Food Canada. 324 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 5. Geographic sources of specimens (full sample) from which the subsample of measured specimens was selected. source of specimens was recorded. None of the specimens had traits the sources of these atypically small specimens, and were only rarely that would identify them as species known only from Mexico, sources of O. diespeter. We assigned these small-bodied and plant- Central America, or South America (Champion 1900, Herring specific specimens to two phenotypes, designated as Nettle and Sage 1966). Our full sample numbers over 3,000 specimens, now (Fig. 6B). Both phenotypes occur within the geographic range of mounted on points and housed at the USDA-ARS, Yakima O. diespeter. Finally, a number of unresolved specimens had genita- Agricultural Research Laboratory, Wapato, WA. lia that were noticeably different in appearance from the genitalia of A subset of 194 male and 188 female specimens were selected O. diespeter. Three phenotypes were identified, designated here as from the full sample for measurement of genitalia. Specimens were HookS, HookL, and CO (Fig. 6C). All three of these genitalic phe- selected to include the three described species collected from multi- notypes overlap geographically with O. diespeter. ple geographic regions, and specimens whose external traits, visual appearance of genitalia, geographic source, or plant source raised questions of identification (see following paragraph). From the full Statistical Comparison of Parameres sample, we first identified specimens of O. tristicolor and O. harpo- Methods for preparing a paramere for measurement are described crates using descriptions and illustrations in White (1879), Kelton elsewhere (Lewis and Horton 2010). As in that earlier study, the (1963, 1978), and Herring (1966). Specimens of O. harpocrates male genitalia were not placed in potassium hydroxide (KOH) dur- (N ¼ 10) were identified using external color and geography, while ing preparation of the slide mount, as we found that KOH alters the specimens of O. tristicolor (N ¼ 158) were identified by external orientation of the flagellum and leads to errors in measurement. traits in combination with size and shape of the paramere. All 168 Each paramere was photographed at 200 with a digital camera specimens were subsequently used in the statistical analyses of geni- (Spot Insight, Diagnostic Instruments, Sterling Heights, MI) talia (described in following sections). The remaining specimens had attached to a compound microscope (Leica DMLS, Bannockburn, genitalia noticeably different from the genitalia of O. tristicolor and IL). Five measurements were obtained from each photograph by use O. harpocrates. From this pool of specimens, O. diespeter (N ¼ 47) of digitizing software (Fig. 7): 1) flagellum length; 2) width of clas- was identified by external traits, genitalia, and geography (Lewis per body; 3) distance between cone tip and flagellum tip (a function and Horton 2010), leaving a mix of specimens (N ¼ 167) whose of “flare” at tip of flagellum combined with flagellum length and identifications could not readily be resolved. shallowness of the flagellum’s arch); 4) area of the space between The 167 unresolved specimens were provisionally assigned to the flagellum and the paramere body (a function of flagellum length, one of eight groups (hereafter, “phenotypes”) based upon geogra- flagellum arch, and size of paramere body); and 5) area of the semi- phy, body size, plant source, and visual appearance of genitalia (Fig. circle defined by the flagellum and an imaginary line connecting 6). Specimens that were collected from geographic regions well out- base and tip of the flagellum (a function of flagellum length and side of the geographic range of O. diespeter (Fig. 3B) fell into one of arch of flagellum). three mutually separate regions, designated here as TX, FL, or S.W. Principal components analysis (PCA) was used to examine (Fig. 6A). Second, a number of unresolved specimens were of whether the five paramere measurements led to separation of speci- obvious smaller body size than O. diespeter. Stinging nettle (Urtica mens in agreement with species’ identification or with phenotypic dioica; Urticaceae) and sagebrush/rabbitbrush (Artemisia, grouping (Fig. 6). The analyses were done using PROC PRINCOMP Ericamaria,orChrysothamnus: Asteraceae) were almost invariably (SAS Institute 2012). We first analyzed all 194 specimens to look for Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 325 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 6. Eight categories (“phenotypes”) of unresolved specimens deﬁned by (A) geography, (B) body size and plant source, or (C) visual appearance of genitalia. Colored symbols are used in subsequent ﬁgures to depict the different phenotypes.

Fig. 7. Five measurements taken from each paramere for use in principal components analysis. separation of O. tristicolor and O. harpocrates from O. diespeter among specimens in PCA scores, to the extent that some specimens and unresolved specimens. Orius tristicolor and O. harpocrates clus- of O. tristicolor overlapped along the size axis with O. harpocrates tered together on the first principal component separately from (Fig. 8). O. diespeter and unresolved specimens (Fig. 8). The first two axes Our next analysis was limited to parameres of O. diespeter and explained 97.3% of the variation in measurements. All loadings unresolved specimens (N ¼ 108). Visual examination showed that along the first principal component were positive (Table 1), suggest- the paramere of each of the unresolved phenotypes had the general ing that axis 1 is a measure of overall paramere size. The paramere appearance of the O. diespeter paramere, including presence of a of O. diespeter was smallest, of intermediate size for O. tristicolor, thin flagellum that terminated sooner in its arc than what is found and was largest for O. harpocrates, which agrees with visual assess- for O. tristicolor and O. harpocrates. However, there were visual ment (Fig. 8: photographs). Axis 2 scores were positively correlated differences between O. diespeter and several unresolved phenotypes with size of gap between flagellum tip and paramere cone (Table 1). in shape and size of the paramere (Fig. 9). The two boxes enclosing Thus, for a given paramere size (axis 1), a specimen with a large neg- each paramere in Fig. 9 visually show differences between the ative score on axis 2 had a paramere with a small distance between O. diespeter paramere and the same organ for each of the eight flagellum tip and cone. The second principal component is useful for unresolved phenotypes. The larger of the two boxes encompasses separating phenotypes of unresolved specimens, as shown in the fol- the entire paramere of O. diespeter (flagellum þ body), while the lowing paragraphs. Orius tristicolor showed substantial variation smaller box within the larger box defines width of the paramere 326 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 8. Scatter plot of principal component scores (PC2 vs. PC1) from the analysis of all 194 specimens and photographs of parameres of O. harpocrates, O. tristicolor, and O. diespeter showing visually differences among species in size of paramere. Sources of specimens for photographs: O. harpocrates (Moss Beach, CA); O. tristicolor (Hanford National Research Park, Benton County, WA); O. diespeter (Sapinero, CO).

among phenotypic groups (PC1: F ¼ 71.3, P < 0.0001; PC2: Table 1. Eigenvectors for the first two principal component axes 8,99 F8,99 ¼ 12.5, P < 0.0001). A Dunnett’s test was used to show statisti- Measure All Orius diespeter cal separation of the unresolved phenotypes from O. diespeter specimens and unresolved (shown as asterisks along top axis or right axis in each of the four specimens small panels in Fig. 10). Five of the phenotypes had a statistically smaller (Nettle) or larger (TX, S.W. CO, HookL) paramere than the PRIN1 PRIN2 PRIN1 PRIN2 paramere of O. diespeter (Fig. 10, panels B–E; see asterisks along Flagellum length (1) 0.50 0.17 0.53 0.17 top axis). We also observed separation of phenotypes along axis 2 Width of paramere body (2) 0.48 0.06 0.44 0.21 (Fig. 10). Axis 2 scores depict variation in distance between the flag- Flagellum tip to cone tip (3) 0.15 0.98 0.05 0.89 ellum tip and cone tip for a given paramere size. Three phenotypes Area beneath flagellum (4) 0.50 0.05 0.50 0.29 (FL, Sage, and HookS) having mean size (PC1) scores similar to the Area 2 beneath flagellum (5) 0.50 0.13 0.52 0.21 O. diespeter mean had statistically smaller gaps between flagellum

Numbers in parentheses refer to the trait numbers shown in Fig. 7. and cone tip compared to the mean O. diespeter gap (Fig. 10B, C, and E; note location of symbols vertically along right axes of panels and accompanying asterisks). The HookL paramere had a small gap body. With use of this visual tool, it is apparent that the paramere of in relation to its overall large size (Fig. 10E). several of the unresolved phenotypes departed from the O. diespeter paramere in size or shape (Fig. 9). Principal components analysis was used to statistically examine Statistical Comparison of Copulatory Tubes variation in size and shape of parameres for O. diespeter and unre- The copulatory tube and associated abdominal segment were posi- solved specimens (Fig. 10). The first two axes explained 89% of the tioned with the internal abdominal body wall facing up, flattened variation in measurements. The first axis appears again to be a meas- with a cover slip, and photographed at 200 (Lewis and Horton ure of paramere size, while the second axis again was most highly cor- 2010). In all specimens, the copulatory tube was of the same general related with gap size (Table 1). Convex hulls in panels B–E enclose appearance as the structure of O. diespeter and O. tristicolor our aprioridefined phenotypes. The four smaller panels show the (Fig. 4D, panel 3), and was composed of a somewhat thickened and three geographic phenotypes (FL, TX, and S.W.) contrasted with O. relatively rigid basal portion, an indistinct collar at the termination diespeter (Fig. 10, panels B and D); the two body size or host plant of the rigid basal section, and a short membranous section entering phenotypes (Nettle and Sage) contrasted with O. diespeter (Fig. 10, the sperm pouch. Length of the tube from its opening at the body panel C); and the three genitalic phenotypes (HookL, HookS, and wall to its membranous distal end was measured using digitizing CO) contrasted with O. diespeter (Fig. 10,panelE). software. Sample size was 188 specimens. The copulatory tube Mean PC1 scores for each phenotype are shown in panels B–E as showed a threefold difference between its maximum length in speci- colored symbols horizontally along the top axis of each panel, while mens of O. tristicolor and O. harpocrates, and its minimum length mean PC2 scores are shown vertically along the right axis. Analysis in specimens of the Nettle phenotype (Fig. 11). Mean length of the of variance showed that mean scores along either axis differed copulatory tube differed among the eight unresolved phenotypes Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 327 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 9. Photographs of paramere of O. diespeter and unresolved phenotypes. The two boxes defined by the dashed lines are used to visually contrast the O. diespeter paramere with parameres of the unresolved phenotypes. The larger box defines overall size of the O. diespeter paramere; the smaller box defines width of the paramere body. Sources of specimens: O. diespeter (Sapinero, CO); FL (Naples, FL); TX (Weslaco, TX); S.W. (Calimesa, CA); HookL (Banning, CA); HookS (Delta, CO); CO (Saguache, CO); Nettle (Ontario, OR); Sage (Biggs, OR).

and O. diespeter (Fig. 11; F8,97 ¼ 65.6, P < 0.0001; analysis excludes Clara, CA). Products were sequenced in both directions by MCLAB O. tristicolor and O. harpocrates). A Dunnett’s test showed that five (South San Francisco, CA) using an ABI 3730XL sequencer. of the phenotypes had a statistically shorter (Nettle) or longer (TX, Sequences were aligned and edited in versions 6.2 and 7.12 of S.W. CO, HookL) copulatory tube than the mean length of the Geneious (Kearse et al. 2012). Consensus sequences were trimmed O. diespeter copulatory tube (Fig. 11; asterisks). to 394 bp (CO1), 433 bp (CytB), and 477 bp (EF1-a) for use in the phylogenetic analyses. All sequences have been deposited in GenBank (Supp. Appendix 2 [online only] shows specimen data and Molecular Phylogenetic Analysis GenBank Accession numbers). A sample of 66 specimens containing examples of each phenotype A phylogenetic tree incorporating all three genes was created (Fig. 6) was used for molecular analysis. DNA segments from two using a Bayesian-Monte Carlo approach implemented in mitochondrial genes (cytochrome oxidase subunit one [COI; 63 MrBayes3.2 (Ronquist et al. 2012). Orius insidiosus was assigned as specimens] and cytochrome B [CytB; 66 specimens]) and one nuclear the outgroup. Two program runs each produced 211 trees; 159 trees gene (elongation factor one alpha [EF1-a; 59 specimens]) were used were sampled from the second run and used to produce a 50% con- for phylogenetic analysis. Both mitochondrial genes have been used sensus tree (Fig. 12). Phylogenetic trees were also created using a to separate difficult species of Old World Orius (Muraji et al. maximum parsimony approach rather than a Bayesian approach for 2000a, b). DNA was extracted from whole bodies of specimens comparison to the consensus tree in Fig. 12 (Supp. Fig. 1 [online using the DNeasy Blood & Tissue Kit (QIAGEN, Valencia CA); for only]). A clear pattern in clustering of phenotypes is evident, with dried specimens, a Chelex 100 resin protocol was first used to stabi- several of our unresolved phenotypes from Fig. 6 departing from lize DNA (Barcenas et al. 2005). PCR was conducted using described species and from other phenotypes (Fig. 12). Notably, the Titanium Taq (Clontech Laboratories, Inc., Mountain View, CA). S.W., Sage, TX (with FL), and HookL (with HookS) phenotypes are PCR protocols consisted of 1 min at 95C followed by cycling of well supported. There is additionally clear separation of O. diespeter 95C for 15 s, annealing at 50–58C for 45 s, extension at 72C for from O. tristicolor and O. harpocrates. The Nettle phenotype did 30 s, with the cycling followed by1 min of 72C. Annealing was not separate from O. diespeter, and together formed a sister group 55C for CytB, 58C for EF1-a, and 50C for COI with forward pri- with the Sage phenotype. The CO phenotype formed a sister group mer 1718, otherwise 58C was used for COI (primer sets listed in with the S.W. phenotype (Fig. 12). There were only modest sequence Supp. Appendix 1 [online only]). PCR product was directly differences between the TX and FL phenotypes, and between the sequenced following a clean-up using ExoSAP-IT (Affymetrix, Santa HookL and HookS phenotypes. These modest differences reflect in 328 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 10. (A) Scatter plot of principal component scores (PC2 vs. PC1) in analysis which excludes the O. tristicolor and O. harpocrates specimens; N ¼ 108 specimens. (B–E) Convex hulls show unresolved specimens grouped by phenotype; the convex hull for the Sage phenotype (panel C) excludes one specimen from Idaho having atypically small genitalia. Panels B and D: contrasts the three geographic phenotypes with O. diespeter; panel C: contrasts the two body size/plant source phenotypes with O. diespeter; panel E: contrasts the three genitalic phenotypes with O. diespeter. Colored symbols distributed horizontally along the top axis of each panel or distributed vertically along the right axis of each panel show mean PC1 or PC2 scores, respectively, for each phenotype. part geographic effects in sequence differences within clades: Texas separation (90 to 111 base pairs) between the outgroup taxon vs. Florida specimens within the TX/FL clade; and California vs. (O. insidiosus) and the remaining taxa (Table 2). The CO and S.W. Washington and Colorado specimens within the HookL/HookS clades were relatively close to one another, with an average of 43 clade (Fig. 12). No genetic separation was observed between O. har- base pair differences (Table 2). The largest separation of clades pocrates and O. tristicolor (Fig. 12). The average distance between excluding comparisons involving O. insidiosus was between the TX/ pairs of phenotypic groups varied from a low of 21 base pair differ- FL clade and the CO clade (Table 2). Each distinct clade in Fig. 12 is ences (O. diespeter/Nettle and Sage groups) to substantially larger also seen in the four maximum parsimony trees, with minor Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 329 Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 11. Scatter plot showing total length of the copulatory tube for 188 specimens of Orius. Colored symbols as in Figs. 6 and 10. Presence of an asterisk indicates that mean length of the copulatory tube for that phenotype is signiﬁcantly different than the O. diespeter mean (by Dunnett’s test). variation evident among the three trees created from single-gene seg- harpocrates is readily collected from herbaceous and shrubby plants ments (Supp. Fig. 1 [online only]). at multiple locations surrounding the Bay Area, and these dark- colored specimens co-occur at these locations with bicolored specimens identified as O. tristicolor. Orius harpocrates is reproductively Reappraisal of the Western U.S. Fauna compatible with O. tristicolor collected from regions of sympatry or Described Species allopatry (T.M.L. and D.R.H., unpublished data). We believe that O. harpocrates is merely a color morphotype of O. tristicolor, albeit Morphometric analysis of the paramere and copulatory tube led to generally with larger genitalia (Figs. 8 and 11) than O. tristicolor. complete separation of O. diespeter from O. tristicolor and O. harpocrates (Figs. 8 and 11). The paramere of O. diespeter is noticeably smaller than the paramere of O. tristicolor and O. harpocrates, Unresolved Specimens while the copulatory tube of female O. diespeter is shorter than the Statistical analyses of genitalic morphology led to sorting of speci- same structure for the other two species. The flagellum of the mens into defined clusters, albeit often with overlap along one or O. diespeter paramere terminates earlier in its arc than the flagellum both PCA axes (Fig. 10) or with overlap in length of the copulatory of either the O. tristicolor or O. harpocrates paramere (Fig. 8). tube (Fig. 11). Gene trees often supported our groupings (Fig. 12), Morphological separation between O. diespeter and O. tristicolor/ although there were exceptions. O. harpocrates was accompanied by genetic separation (Fig. 12). We also showed separation of genitalic morphology between O. tris- Nettle phenotype ticolor and O. harpocrates, albeit with overlap between O. harpo- Genitalia of the Nettle phenotype were much smaller than genitalia crates specimens and specimens of O. tristicolor having a large of O. diespeter (Figs. 10 and 11). Nettle specimens did not diverge paramere (Fig. 8) or a long copulatory tube (Fig. 11). The genitalia genetically from O. diespeter (Fig. 12). We have collected the Nettle of O. tristicolor appeared to be substantially more variable than the phenotype from Washington, Oregon, and Idaho (Fig. 13), always genitalia of O. diespeter or O. harpocrates (Figs. 8 and 11). The in association with stinging nettle, Urtica dioica (Urticaceae). copulatory tube of O. tristicolor showed a 1.5-fold difference Specimens have a smaller overall body size than specimens of between its minimum length in specimens from Washington State O. diespeter. The smaller size of genitalia appears not to be caused and its maximum length in a specimen from southern California by feeding on nettle or on nettle-associated arthropods, as rearing of (Fig. 11). Whether variation in genitalia of O. tristicolor is somehow the Nettle phenotype in the laboratory on a diet of Ephestia associated with geography is not yet known. We have collected (Lepidoptera) eggs did not lead to the production of specimens hav- specimens from southern California in which the copulatory tube ing genitalia that were larger than genitalia of field-collected insects was almost as short as the structure in Washington State specimens. (T.M.L. and D.R.H., unpublished data). The Nettle phenotype is DNA sequence data did not provide any evidence that O. tristicolor partially compatible reproductively with O. diespeter (T.M.L. and from northern latitudes differed genetically from southern specimens D.R.H., unpublished data). Geographic distribution of the Nettle (Fig. 12). phenotype overlaps distribution of O. diespeter (Figs. 3B and 13). Of interest is our observation that the differences between O. tristicolor and O. harpocrates in size or shape of the paramere and copulatory tube were not accompanied by genetic divergence Sage phenotype between these two species. Specimens of Orius having typical Specimens of the Sage phenotype are smaller in body size than O. harpocrates traits (dark coloration [Fig. 2B] and large paramere O. diespeter. Genitalia are very similar in appearance to the genitalia or long copulatory tube), collected from the type location near San of O. diespeter (Figs. 9-–11). However, the Sage phenotype showed Francisco, did not diverge genetically from O. tristicolor collected modest (2%) but consistent genetic differences from O. diespeter from any of several geographic locations (Fig. 12). Orius (Fig. 12). We have collected specimens of the Sage phenotype 330 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2

FL and TX phenotypes Genitalia showed either complete (male) or partial (female) morphometric separation between these two phenotypes (Figs. 10 and 11). DNA sequences grouped the two phenotypes together in the same clade (Fig. 12). Both phenotypes diverged from O. diespeter in morphology of genitalia and in DNA sequences (Figs. 10-–12). Both phenotypes also separated genetically from all other unresolved specimens (Fig. 12). The FL and TX phenotypes have hemelytral markings very similar to markings of O. tristicolor, and both phenotypes can be confused with this species. Indeed, we suspect that old records for O. tristicolor in Florida (Barber 1914) are actually for specimens of the FL phenotype. We have collected the TX phenotype from southern Texas, northwards into Oklahoma and Nebraska (Fig. 13); of these sites, only Nebraska may be within the Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 geographic range of O. diespeter. Our FL specimens were collected from two locations in Florida (Fig. 13), both well outside of the known range of O. diespeter. Plant records for FL and TX specimens included weedy Asteraceae, weedy Brassicaceae, cotton, and potatoes. Insects of the TX phenotype are reproductively compatible with insects of the FL phenotype (T.M.L., unpublished data).

HookS and HookL phenotypes Specimens of these two phenotypes have quite distinctive genitalia not observed in any of the described species or in any of the other unresolved phenotypes. The two phenotypes also diverged genetically from described species and from all other unresolved phenotypes (Fig. 12). The flagellum of the male’s paramere exhibits a noticeable flare at its tip (Fig. 9), while the copulatory tube of females is enlarged near its base, which is not found in described species or in any of the unresolved phenotypes. While the genitalia of HookS specimens were noticeably smaller than genitalia of the HookL specimens, the two phenotypes were genetically similar (Fig. 12). We collected HookL specimens from Asteraceae, Polygonaceae, and Brassicaceae in arid rangelands of Washington, southern California, and eastern Nevada (Fig. 13), sometimes together with O. tristicolor. The genetically similar HookS was collected from sites in western Colorado (Fig. 13), generally from shrubby Chenopodiaceae (Atriplex and Sarcobatus). We have not collected HookL specimens from Colorado, and it is unclear whether the HookL and HookS phenotypes overlap geographically. We have yet to determine whether the two phenotypes are reproductively compatible.

S.W. phenotype Specimens of this phenotype have a larger paramere (especially in length of the flagellum) than that of O. diespeter, and a longer copulatory tube than the copulatory tube of O. diespeter (Figs. 10 and 11). The S.W. phenotype separates genetically from described species and other phenotypes (Fig. 12). We collected this phenotype Fig. 12. Phylogenetic tree developed in MrBayes using segments of two mito- from a variety of shrubby and herbaceous plant taxa in southern chondrial genes (cytochrome oxidase subunit one and cytochrome B) and a California eastwards into Nevada and Arizona (Fig. 13), well out- segment of one nuclear gene (elongation factor 1 alpha). Numbers above hor- side of the known range of O. diespeter. The phenotype was col- izontal branches represent the probability of base pair substitutions along the branches. Numbers below branches in italics represent clade credibility esti- lected together with O. tristicolor at multiple locations in southern mated in MrBayes. Colored symbols as in Fig. 6. California, and may actually be more common in some of these locations than O. tristicolor. on sagebrush, Artemisia spp., and rabbitbrush, Ericameria/ Chrysothamnus spp. (all Asteraceae), from sites in the northwestern CO phenotype United States southwards into Colorado (Fig. 13). Orius diespeter is We collected this phenotype only in Colorado (Fig. 13). Specimens rarely collected from these plant taxa. Geographic range of the were collected from shrubby and herbaceous plant taxa Sage phenotype overlaps extensively with geographic range of (Cercocarpus [Rosaceae]; Cirsium [Asteraceae]; Pericome caudata O. diespeter. [Asteraceae]) over a range of altitudes between 1,500 and 3,500 m. Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 331

Table 2. Average number of DNA base pair differences (% difference) among phenotypic groups

Phenotype O. diespeter Sage CO S.W. HookL & O. tristicolor & TX & FL & Nettle HookS O. harpocrates

Sage 21 (2%) – CO 54 (4%) 66 (5%) – S.W. 59 (5%) 68 (5%) 43 (3%) – HookL & HookS 66 (5%) 74 (6%) 76 (6%) 79 (6%) – O. tristicolor & O. harpocrates 69 (5%) 76 (6%) 71 (5%) 71 (5%) 62 (5%) – TX & FL 71 (5%) 79 (6%) 89 (7%) 78 (6%) 71 (5%) 74 (6%) – O. insidiosus 92 (7%) 106 (8%) 95 (7%) 100 (7%) 93 (7%) 90 (7%) 111 (8%) Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021

Fig. 13. Locations from which phenotypes deﬁned in Fig. 6 have been collected.

The CO phenotype has a larger paramere and longer copulatory composed of one widespread species, O. tristicolor, plus two less tube than O. diespeter (Figs. 10 and 11), and with the S.W. pheno- common species, O. harpocrates and O. diespeter, which are limited type diverges genetically from described species and from other phe- in range to the region around the Bay Area, California, and a few notypes (Fig. 12). localized areas in western and northwestern Canada (Herring 1966, Kelton 1978, Henry 1988, Scudder 1997, Maw et al. 2000). This conclusion appears often to lead to an automatic assumption in the A Cryptic Species Complex biological control literature that composition of the Orius fauna in Historical summaries in catalogs or other syntheses lead to the con- natural or managed systems of western North America comprises clusion that the Orius fauna native to the western United States is but a single species, O. tristicolor. Only recently has information 332 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 begun to accumulate to contradict that impression (Lewis and entirely of O. diespeter (T.M.L. and D.R.H., unpublished data). Horton 2010). Historical failure to appreciate that O. diespeter is Whether errors in identifications have occurred in crop systems for highly variable in color led to substantial underestimation of the any of the specimen groupings in Fig. 6 is unknown. geographic and ecological range of this species, and to overestima- Second, we are not yet sure of the geographic range of most of tion of the range of O. tristicolor (Lewis and Horton 2010). The these phenotypes. For example, we assume that our southern U.S. morphological and molecular results of the current study now sug- phenotypes (TX, S.W., and HookL) extend southwards from our gest that the native Orius fauna of the western and southcentral collecting regions into crop and noncrop habitats in Mexico. The United States is actually a complex of two described species (we are extent of that range southwards has yet to be determined. We are assuming that O. harpocrates is merely a color variant of O. tristi- equally uncertain whether the geographic distribution of central color) and perhaps five undescribed cryptic species (Sage, CO, S.W., U.S. phenotypes such as TX or CO extends eastwards from our col- TX/FL, and HookL/HookS). lection sites to include portions of the crop growing regions within Previous work from our laboratory showed that a second genus or east of the Great Plains. Examination of specimens from regions of Anthocoridae in North American includes hidden species com- east of our area of focus is needed to better define the eastern limits plexes. Anthocoris antevolens White, 1879 is a common source of of described species and unresolved phenotypes. Finally, it is not Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 biological control in fruit orchards of western North America clear whether the TX and FL phenotypes overlap geographically (Anderson 1962, Kelton 1978). We now know that this species is across the Gulf Coast States (Fig. 13), as we have not yet examined actually part of a cryptic species complex whose members are exter- specimens of Orius from these states. nally similar to one another (Horton and Lewis 2005; Horton et al. A third issue concerns the use of Orius in classical biological 2007, 2008). Insects identified as A. antevolens include specimens control programs. Specimens of Orius from the western United which depart from other specimens of A. antevolens in morphology States have been used in classical biological control efforts, involving of genitalia, DNA sequences, and host plant specificity, much as we releases of specimens identified as O. tristicolor into Florida (Frank have shown here for the Orius complex. Crosses between nonlike and McCoy 1994) and into Hawaii on multiple occasions (Davis phenotypes of A. antevolens often fail to produce offspring. The and Krauss 1963, 1965; Culliney and Nagamine 2000). Whether lack of successful reproduction occurs even between geographically these releases included species other than O. tristicolor is unknown. sympatric pairs of nonlike phenotypes (Horton and Lewis 2005). Releases made in Hawaii included insects collected from multiple We concluded that A. antevolens is actually a cryptic species com- geographic regions (California and multiple locations in Arizona); plex (Horton et al. 2007, 2008). thus, the releases could have included one or more of the phenotypes In the present study, we show that specimens of North American or species that inhabit the southwestern United States. Orius tristi- Orius often differed in morphological, genetic, and behavioral color is established in Hawaii (Lattin 2007; confirmed by examina- (plant specificity) traits. While we have yet to conduct all of the mat- tion of genitalia [T.M.L., unpublished data]), but because we do not ing crosses needed to confirm reproductive isolation among unre- know what species actually were released in the release programs, it solved phenotypes, it seems likely that the North American fauna of is unclear whether establishment of O. tristicolor in Hawaii was Orius (like certain Anthocoris) is indeed a complex of reproduc- from these planned releases or through accidental introduction. tively isolated cryptic species. Support for this statement includes Lastly, our results complicate understanding the role of Orius as the obvious departure of some groupings of specimens from other sources of biological control in western U.S. crops, and impede groups in genitalic morphology (e.g., the existence of a “flare” at efforts at managing or conserving these insects in crops. For exam- the tip of the flagellum for the HookL and HookS specimens; Fig. 9) ple, the presence of Orius in a crop field and in an adjacent noncrop and the significant divergence in DNA sequences among several of habitat might be interpreted as evidence that there is movement by our a priori defined groups (Fig. 12). We anticipate that mating tri- these bugs between habitats, especially if we assume erroneously als will prove that several of the phenotypes listed in Fig. 6 are that the Orius fauna in most agricultural regions of the western reproductively isolated from other phenotypes as well as from United States is composed of a single species, O. tristicolor. described species. However, confirmation that cryptic species externally similar to The discovery that the western fauna of Orius likely is a cryptic O. tristicolor occur in many regions of the western United States species complex has several implications for the biological control would force us to reevaluate this idea. That is, certain of the phe- community. First, our unresolved specimens had the bicolored exter- notypes in Fig. 6 appear to be relatively specialized on a few nal appearance that could (without thorough examination) easily noncrop plant taxa, including especially the Nettle phenotype lead to identification as O. tristicolor or as the nonmelanic form of (associated with Urtica dioica), the Sage phenotype (Artemisia, O. diespeter. This observation suggests that mistakes of identifica- Ericameria,andChrysothamnus), and the HookL/HookS pheno- tion in the biological control literature might be relatively common, type (weedy Asteraceae and Chenopodiaceae). These insects may although the frequency of these mistakes would presumably depend be unlikely to move from noncrop habitats into crop habitats. upon how regularly these undescribed species actually occur in crop Thus, even large densities of Orius in habitats adjacent to crops systems. While our samples included only a limited number of speci- may provide little in the way of practical benefit to agriculture if mens collected from agricultural fields, those samples did include these insects happen to be as specialized as our collecting efforts two phenotypes listed in Fig. 6: TX (potato and cotton) and S.W. may indicate. In sum, until we have more complete geographic (cotton). We believe that misidentification of O. diespeter as O. tris- and biological data for these cryptic species, it will be difficult to ticolor has led to a historical lack of appreciation of the importance optimally incorporate these insects into biological control of O. diespeter in agriculture (Lewis and Horton 2010). Our collect- programs. ing records for O. diespeter now include a number of agricultural crops, including corn, sugarbeets, eggplant, potatoes, pears, and alfalfa (Lewis and Horton 2010; T.M.L., unpublished data). Indeed, Supplementary Data a 2013 collection of several hundred specimens of Orius from multi- Supplementary data are available at Annals of the Entomological Society of ple alfalfa fields in eastern Montana was found to be composed America online. Annals of the Entomological Society of America, 2016, Vol. 109, No. 2 333

Acknowledgments Henry, T. J. 1988. Family Anthocoridae Fieber, 1837. The minute pirate bugs, pp. 12–28. In T. J. Henry and R. C. Froeschner (eds), Catalog of the We thank several colleagues for collecting and forwarding specimens of Heteroptera, or true bugs, of Canada and the continental United States. E.J. Orius: Mary Sorensen, Alexis Park, and Joshua Oliva (University of Brill, Leiden, The Netherlands. California, Riverside, CA); Steve Naranjo (USDA-ARS, Maricopa, AZ); Bob Hernandez, L. M., and G. M. Stonedahl. 1999. A review of the economically Pfannenstiel (USDA-ARS, Weslaco, TX, and Manhattan, KS); Susan Halbert important species of the genus Orius (Heteroptera: Anthocoridae) in East (Florida Department of Agriculture, Gainesville, FL); Brad Higbee Africa. J. Nat. Hist. 33: 543–568. (Paramount Farms, Shafter, CA); Andy Jensen (Northwest Potato Research Herring, J. L. 1966. The genus Orius of the western Hemisphere (Hemiptera: Consortium, Eagle, ID); and Tatyana Rand (USDA-ARS, Sidney, MT). We Anthocoridae). Ann. Entomol. Soc. Am. 59: 1093–1109. are especially grateful to our coworkers Deb Broers and Merilee Bayer for Horton, D. R., and T. M. Lewis. 2005. Size and shape differences in genitalia their energetic and enthusiastic collecting efforts over the past decade, efforts of males from sympatric and reproductively isolated populations of that provided a wealth of specimens and plant records from locations Anthocoris antevolens White (Heteroptera: Anthocoridae) in the Yakima throughout the United States. Additional collecting assistance was provided Valley, Washington. Ann. Entomol. Soc. Am. 98: 527–535. by other past and present members of the Horton lab, especially Richard Horton, D. R., T. R. Unruh, T. M. Lewis, and K. Thomsen-Archer. 2007. Lewis, Gene Miliczky, and Lila Scaife. We thank Rodney Cooper (USDA- Morphological and genetic divergence in three populations of Anthocoris Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 ARS, Yakima Agricultural Research Laboratory, Wapato, WA), Thomas antevolens (Hemiptera: Heteroptera: Anthocoridae). Ann. Entomol. Soc. Henry (Systematic Entomology Laboratory, USDA-ARS, National Museum Am. 100: 403–412. of Natural History, Washington, D.C.), and Richard Zack (Department of Horton, D. R., T. M. Lewis, K. Thomsen-Archer, and T. R. Unruh. 2008. Entomology, Washington State University, Pullman, WA) for review of an Morphology, genetics, and male mating success compared between earlier version of this manuscript. Deb Broers provided the line illustration in Anthocoris musculus and A. antevolens (Hemiptera: Heteroptera: Fig. 1B. Anthocoridae). Proc. Entomol. Soc. Wash. 110: 960–977. Ito, K., and T. Nakata. 1998. Diapause and survival in winter in two species References Cited of predatory bugs, Orius sauteri and O. minutus. Entomol. Exp. Appl. 89: 271–276. Anderson, N. H. 1962. Anthocoridae of the Pacific Northwest with notes on Kearse, M., R. Moir, A. Wilson, S. Stones-Havas, M. Cheung, S. Sturrock, S. distributions, life histories, and habits (Heteroptera). Can. Entomol. 94: Buxton, A. Cooper, S. Markowitz, C. Duran, et al. 2012. Geneious basic: 1325–1334. An integrated and extendable desktop software platform for the organiza- Barber, H. G. 1914. Insects of Florida. II. Hemiptera. Bull. Am. Mus. Nat. tion and analysis of sequence data. Bioinformatics 28: 1647–1649. Hist. 33: 495–535. Kelton, L. A. 1963. Synopsis of the genus Orius Wolff in America north of Barcenas, N. M., T. R. Unruh, and L. G. Neven. 2005. DNA diagnostics to Mexico (Heteroptera: Anthocoridae). Can. Entomol. 95: 631–636. identify internal feeders (Lepidoptera: Tortricidae) of pome fruits of quaran- Kelton, L. A. 1978. The anthocoridae of Canada and Alaska. Heteroptera tine importance. J. Econ. Entomol. 98: 299–306. Anthocoridae. The Insects and Arachnids of Canada. Part 4. Agriculture Blatchley, W. S. 1926. Heteroptera or true bugs of eastern North America. Canada Research Publication 1639. Ottawa, Ontario, Canada. The Nature Publishing Company, Indianapolis, IN. Knowlton, G. F. 1949. Orius feeding notes. Bull. Brooklyn Entomol. Soc. 44: Bu, W.-J., and L.-Y. Zheng. 2001. Fauna Sinica. Insecta Vol. 24. Hemiptera. 53–55. Lasiochilidae Lyctocoride Anthocoridae. Chinese Academy of Sciences, Lattin, J. D. 1999. Bionomics of the Anthocoridae. Annu. Rev. Entomol. 44: Science Press, Beijing, China. 207–231. Carayon, J. 1953. Existence d’un double orifice geńital et d’un tissu conduc- Lattin, J. D. 2007. The Lasiochilidae, Lyctocoridae, and Anthocoridae teur des spermatozoı¨des chez les Anthocorinae (Hemipt. Anthocoridae). (Hemiptera: Heteroptera) of the Hawaiian Islands: Native or introduced? C.R. Acad. Sci. Paris 236: 1206–1208. Proc. Entomol. Soc. Wash. 109: 75–80. Carayon, J. 1972. Caracte`res syste´matiques et classification des Anthocoridae Lewis, T. M., and D. R. Horton. 2010. Orius diespeter Herring in North (Hemipt.). Ann. Soc. Entomol. France (N.S.) 8: 309–349. America: Color variation and updated distribution (Hemiptera: Carpintero, D. L. 2002. Catalogue of the Neotropical Anthocoridae Heteroptera: Anthocoridae). Proc. Entomol. Soc. Wash. 112: 541–554. (Heteroptera). Rev. Soc. Entomol. Argent. 61: 25–44. Lewis, T. M., and J. D. Lattin. 2010. Orius (Heterorius) vicinus (Ribaut) Champion, G. C. 1900. Anthocoridae: Insecta: Rhynchota: Hemiptera- (Hemiptera: Heteroptera: Anthocoridae) in Western North America, a cor- Heteroptera, pp. 306–335. In F. D. Godman and O. Salvin (eds), Biologia rection of the past. Proc. Entomol. Soc. Wash. 112: 69–80. Centrali-Americana. Taylor & Francis, London. Lewis, T. M., D. R. Horton, and D. A. Broers. 2005. New state and United Cranshaw, W., D. C. Sclar, and D. Cooper. 1996. A review of 1994 pricing States records for Anthocoridae (Hemiptera: Heteroptera). Pan Pac. and marketing by suppliers of organisms for biological control of arthro- Entomol. 81: 59–67. pods in the United States. Biol. Control 6: 291–296. Lewis, T. M., D. R. Horton, and J. D. Lattin. 2015. First Nearctic records for Culliney, T. W., and W. T. Nagamine. 2000. Introductions for biological con- Orius (Dimorphella) sibiricus Wagner (Hemiptera: Heteroptera: Anthocori- trol in Hawaii, 1987–1996. Proc. Hawaii. Entomol. Soc. 34: 121–133. dae), a Eurasian Steppe inhabitant. Proc. Entomol. Soc. Wash. 117: 389– Davis, C. J., and N.L.H. Krauss. 1963. Recent introductions for biological 399. control in Hawaii – VIII. Proc. Hawai. Entomol. Soc. 18: 245–249. Lykouressis, D. P., and D. C. Perdikis. 1997. The phenology and abundance of Davis, C. J., and N.L.H. Krauss. 1965. Recent introductions for biological certain species of Orius (Hemiptera: Anthocoridae) that occur in Greece. control in Hawaii – X. Proc Hawai. Entomol. Soc. 19: 87–90. Israel J. Entomol. 31: 47–54. Fieber, F. X. 1860. Exegesen in Hemipteren. Wien. Entomol. Zeit. 4: 257–272. Maw, H.E.L., R. G. Foottit, K.G.A. Hamilton, and G.G.E. Scudder. 2000. Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA pri- Checklist of the Hemiptera of Canada and Alaska. National Research mers for amplification of mitochondrial cytochrome c oxidase subunit I Council of Canada, NRC Research Press, Ottawa, ON, Canada. from diverse metazoan invertebrates. Mol. Mar. Biol. Biotech. 3: 294–299. Miller, S. E., and W. S. Davis. 1985. Insects associated with the flowers of two Frank, J. H., and E. D. McCoy. 1994. Commercial importation into Florida of species of Malacothrix (Asteraceae) on San Miguel Island, California. invertebrate animals as biological control agents. Fla. Entomol. 77: 1–20. Psyche 92: 547–556. Ghauri, M.S.K. 1972. The identity of Orius tantillus (Motschulsky) and notes Muraji, M., K. Kawasaki, and T. Shimizu. 2000a. Nucleotide sequence on other Oriental Anthocoridae (Hemiptera, Heteroptera). J. Nat. Hist. 6: variation and phylogenetic utility of the mitochondrial COI fragment in 409–421. anthocorid bugs (Hemiptera: Anthocoridae). Appl. Entomol. Zool. 35: Gillespie, D. R., and D.M.J. Quiring. 2006. Orius insidiosus (Hemiptera: 301–307. Anthocoridae) in the lower Fraser Valley, British Columbia, Canada. Can. Muraji, M., K. Kawasaki, and T. Shimizu. 2000b. Phylogenetic utility of nu- Entomol. 138: 590–593. cleotide sequences of mitochondrial 16S ribosomal RNA and cytochrome b 334 Annals of the Entomological Society of America, 2016, Vol. 109, No. 2

genes in anthocorid bugs (Heteroptera: Anthocoridae). Appl. Entomol. Scudder, G.G.E. 1997. True bugs (Heteroptera) of the Yukon, pp. 241–336. Zool. 35: 293–300. In H. V. Danks and J. A. Downes (eds), Insects of the Yukon. Biological sur- Ohno, K., and H. Takemoto. 1997. Species composition and seasonal occur- vey of Canada (Terrestrial Arthropods). Ottawa, Canada1034 pp. rence of Orius spp. (Heteroptera: Anthocoridae), predacious natural ene- Shapiro, J. P., and S. M. Ferkovich. 2009. Corrected species identiﬁcation of mies of Thrips palmi (Thysanoptera: Thripidae), in eggplant ﬁelds and the predator Orius pumilio (Heteroptera: Anthocoridae) in a research col- surrounding habitats. Appl. Entomol. Zool. 32: 27–35. ony. Fla. Entomol. 92: 399. Palumbi, S. R. 1996. Nucleic acids II: The polymerase chain reaction, pp. 205– Shapiro, J. P., P. D. Shirk, S. R. Reitz, and R. Koenig. 2009. Sympatry of Orius 247. In D. M. Hillis, C. Moritz, and B. K. Mable (eds), Molecular systemat- insidiosus and O. pumilio (Hemiptera: Anthocoridae) in North Central ics, 2nd edn. Sinauer Press, Sunderland, MA. Florida. Fla. Entomol. 92: 362–366. Parshley, H. M. 1920. Hemiptera from Peaks Island, Maine, collected by Mr. Shapiro, J. P., P. D. Shirk, K. Kelley, T. M. Lewis, and D. R. Horton. 2010. G.A. Moore. Can. Entomol. 52: 80–87. Identity of two sympatric species of Orius (Hemiptera: Heteroptera: Parshley, H. M. 1922. Records of Nova Scotian Hemiptera-Heteroptera. Proc. Anthocoridae). J. Insect Sci. 10: 189. (http://www.insectscience.org/ Acadian Entomol. Soc. 1922: 102–108. 10.189). Parshley, H. M. 1923. Family anthocoridae, pp. 665–668. In W. E. Britton Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. Flook. 1994.

(ed.), The hemiptera or sucking insects of Connecticut. State Geological and Evolution, weighting, and phylogenetic utility of mitochondrial gene se- Downloaded from https://academic.oup.com/aesa/article/109/2/319/2195305 by guest on 23 September 2021 Natural History Survey, Hartford, CT. quences and a compilation of conserved polymerase chain reaction primers. Pe´ricart, J. 1972. He´mipte`res. Anthocoridae, Cimicidae et Mircophysidae de Ann. Entomol. Soc. Am. 87: 651–701. L’Ouest-Paleárctique, pp. 1–402. In Faune de L’Europe et du Bassin Tommasini, M. G., J. C. van Lenteren, and G. Burgio. 2004. Biological traits Me´diterraneén, Vol. 7. Masson et Cie, Paris. and predation capacity of four Orius species on two prey species. Bull. Postle, A. C., M. Y. Steiner, and S. Goodwin. 2001. Oriini (Hemiptera: Insect. 57: 79–93. Anthocoridae) new to Australia. Aust. J. Entomol.40: 231–244. Torre-Bueno, J. R. 1930. Records of Anthocoridae, particularly from New Ribaut, H. 1923. E´ tude sur le genre Triphleps (Heteroptera-Anthocoridae). York. Bull. Brooklyn Entomol. Soc. 25: 11–20. Bull. Soc. Hist. Nat. Toulouse 51: 522–538. van Lenteren, J. C., M. M. Roskam, and R. Timmer. 1997. Commercial mass Ronquist, F., M. Teslenko, P. van der Mark, D. L. Ayres, A. Darling, S. production and pricing of organisms for biological control of pests in Ho¨ hna, B. Larget, L. Liu, M. A. Suchard, and J. P. Huelsenbeck. 2012. Europe. Biol. Control 10: 143–149. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice Wagner, E. 1952. Die europaïschen Arten der Gattung Orius Wff. (Hem. Het. across a large model space. Syst. Biol. 61: 539–542. Anthocoridae). Notulae Entomol. 32: 22–59. SAS Institute. 2012. SAS 9.3 for Windows. SAS Institute Inc., Cary, NC. White, F. B. 1879. Descriptions of new Anthocoridae. Entomol. Mon. Mag. Say, T. 1832. Descriptions of new species of heteropterous Hemiptera of 16: 142–148. North America. New Harmony, IN. 39 pp. Yasunaga, T. 1997. The flower bug genus Orius Wolff (Heteroptera: Schumacher, F. 1922. U¨ bersehene Hemipteren-Gattungen. Deutsche Ent. Anthocoridae) from Japan and Taiwan. Appl. Entomol. Zool. 32: 355–364, Zeits. 1922: 337–338. 379–386, 387–394.