Two Species within frontalis (Coleoptera: ): Evidence from Morphological, Karyological, Molecular, and Crossing Studies Author(s): Francisco Armendáriz-Toledano, Alicia Niño, Brian T. Sullivan, Jorge Macías-Sámano, Javier Víctor, Stephen R. Clarke, and Gerardo Zúñiga Source: Annals of the Entomological Society of America, 107(1):11-27. 2014. Published By: Entomological Society of America URL: http://www.bioone.org/doi/full/10.1603/AN13047

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. SYSTEMATICS Two Species Within Dendroctonus frontalis (Coleoptera: Curculionidae): Evidence From Morphological, Karyological, Molecular, and Crossing Studies

FRANCISCO ARMENDA´ RIZ-TOLEDANO,1 ALICIA NIN˜ O,2 BRIAN T. SULLIVAN,3 2,4 1 5 1,6 JORGE MACI´AS-SA´ MANO, JAVIER VI´CTOR, STEPHEN R. CLARKE, AND GERARDO ZU´ N˜ IGA

Ann. Entomol. Soc. Am. 107(1): 11Ð27 (2014); DOI: http://dx.doi.org/10.1603/AN13047 ABSTRACT Dendroctonus frontalis Zimmermann is considered one of the most important economic and ecological forest pests in the United States, Mexico, and Central America. Recently, two apparent morphological variants of this species were discovered occurring syntopically in Central America and southern Mexico. Morphotype A lack a series of Þne parallel ridges on the episternal area of the prothorax that are present on morphotype B. The goal of the present work was to clarify the taxonomic status of the morphotypes of the D. frontalis species complex. Geometric morphometric analyses of seminal rod and spermatheca shape together with the characterization of 16 attributes of external morphology revealed differences in quantitative and qualitative characters that distinguished adults of the two morphotypes from each other as well as from the closely related species Dendroctonus vitei Wood and Dendroctonus mexicanus Hopkins. Karyotype analysis of morphotype B revealed a chromosomal formula (5AA ϩ Xyp) distinct from that found in morphotype A previously reported for D. frontalis (7AA ϩ Xyp). In the laboratory, forced intermorphotype crosses produced F1 progeny but at lower frequency than intramorphotype pairings, and dissections of spermatheca revealed a lower frequency of insemination at least one type of heterotypic cross. Phylogenetic analysis of the D. frontalis species complex based on 786 bp of the cytochrome oxidase I gene indicated that morphotypes B and A are two independent groups with 98% nodal support within D. frontalis. These data provide compelling evidence that the two syntopic morphotypes represent two distinct sibling species.

KEY WORDS seminal rod, spermatheca, geometric morphometry, syntopic species, integrative

The genus Dendroctonus Erichson (Curculionidae: mis LeConte, Dendroctonus frontalis Zimmermann, Scolytinae) is widely distributed in North and Central Dendroctonus mexicanus Hopkins, and Dendroctonus America and Eurasia. They colonize and kill trees of vitei Wood (Lanier et al. 1988). The composition of the genera Larix Mill., Picea A. Dietr., Pseudotsuga the D. frontalis complex has changed virtually with Carrie`re, and Pinus L. (Wood 1982). Dendroctonus is almost every taxonomic review (Lanier et al. 1988), composed of 19 species, one of them with two sub- and species assignation has been based largely on species, which are ordered into different groups external morphological features (e.g., setal vestiture (Wood 1963, 1982; Lanier 1981; Kelley and Farrell and sculpturing of the elytral declivity, frons sculp- 1998; Zu´ n˜ iga et al. 2002a). The Dendroctonus frontalis ture, shape of epistomal process, and pronotal size and complex (sensu lato) is composed of the Pinus-infest- sculpture. Nevertheless, routine identiÞcation of the ing species Dendroctonus adjunctus Blandford, Den- more difÞcult species in the D. frontalis complex (e.g., droctonus approximatus Dietz, Dendroctonus brevico- D. frontalis, D. mexicanus, and D. vitei) cannot be performed with external morphological features 1 Instituto Polite´cnico Nacional, Escuela Nacional de Ciencias Bi- alone, because broad intraspeciÞc morphological vari- olo´gicas, Prol. de Carpio y Plan de Ayala, Col. Santo Tomas Del. Miguel Hidalgo, Me´xico D.F., CP 11340. ation exists, and key external character states can be 2 El Colegio de la Frontera Sur (ECOSUR), Carretera Antiguo distinguished only with practice. In contrast, seminal Aeropuerto km 2.5, Tapachula, Chiapas, Me´xico, CP 30700. rod shape and chromosome number have been shown 3 USDA Forest Service, Southern Research Station, 2500 Shreve- port Hwy., Pineville, LA 71360. to be reliable features for distinguishing these species 4 Current Address: Synergy Semiochemicals Corporation, 7061 (Vite´ et al. 1975; Lanier 1981; Wood 1982; Lanier et al. Meritt Ave., Burnaby, British Columbia, Canada, V5J 4R7. 1988; Zu´ n˜ iga et al. 2002a,b). With the exception of D. 5 USDA Forest Service, Forest Health Protection, 2221 N Raguet Lufkin, TX. vitei, species in the D. frontalis complex have been 6 Corresponding author, e-mail: [email protected]. characterized by karyological studies (Lanier 1981;

0013-8746/14/0011Ð0027$04.00/0 ᭧ 2014 Entomological Society of America 12 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

Fig. 1. Analyzed anatomical structures of D. frontalis in both MA, morphotype A, and MB, morphotype B. Sculpture of preepisternal area: (a) MA smooth, (b) MB striated; rows of squamiform plates on eighth tergite: (c) MA, (d) MB; acute projections on posterior edge on squamiform plates in the median proximal area of the eighth tergite: (e) MA short and uniform, (f) MB variable in size; spermatheca: (g) MA with round cornu, aggregated striae on the nodulus, and Ϸ1/4 of body covered by striations, (h) MB with oval cornu, separated striae on the nodulus, and Ϸ1/2 of body covered by striations; seminal January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 13

Lanier et al. 1988; Zu´ n˜ iga et al. 2002a,b) as well as 10% KOH. After incubation, structures were im- mtDNA sequences (Kelley and Farrell 1998). mersed in 20% acetic acid solution to neutralize the Within the frontalis complex, D. frontalis (southern KOH and subsequently rinsed with 100% ethanol. The ) is an important economic and ecological genitalia were then dissected. For females, the eighth forest pest within its extensive geographic range, tergite and spermatheca from the same individual which includes the southeastern United States, Ari- were mounted on the same slide. For males, the sem- zona, Mexico, and Central America (Clarke and inal rod was separated from the genital capsule and all Nowak 2009). During periodic outbreaks, D. frontalis parts were mounted on the same slide. All structures are capable of overcoming the defenses of healthy were mounted in HoyerÕs medium on semipermanent pine trees and deforesting patches of the landscape slides for easy manipulation throughout the study. ranging in size from small “spots” to thousands of acres Character Analysis for the Two Morphs. Based on (Billings 2011). the presence or absence of striations on the preepis- Recently, two morphologically distinct variants of ternal area of the prothorax of 389 D. frontalis D. frontalis were identiÞed in collections from Central (Midtgaard and Thunes 2002; Sullivan et al. 2012, Fig. America and southern Mexico. Morphotype A (MA) 1; this paper Fig. 1a and b), we identiÞed 268 speci- beetles were on average somewhat smaller than mor- mens striaeless (MA) and 121 with striae (MB). These photype B (MB) beetles, and MA insects of both sexes insects were exhaustively examined to identify addi- lacked of a Þne series of ridges present in the preep- tional morphological characters for the correct isternal area of the prothorax morphotype MB beetles morphs identiÞcation. From 30 reviewed attributes, 16 (Midtgaard and Thunes 2002; Sullivan et al. 2012, Fig. were taxonomic characters potentially useful for mor- 1; this paper Fig. 1a and b). Although this last character photype separation: 10 were on the external cuticle (8 has been used to separate these morphs, a formal quantitative and 2 qualitative) and the rest were ter- analysis has not been conduced to evaluate its utility minaliaÕs characters (in the eighth tergite, sper- as taxonomic character. The morphotypes have been matheca, and seminal rod). found to differ signiÞcantly in production of known Adults beetles were examined at magniÞcations up Dendroctonus pheromone components (specimens to 90ϫ and were photographed with an Environmen- from Chiapas, Mexico) and in the composition of their tal Scanning Electron Microscope ESEM (Evo 40 VP, cuticular hydrocarbons (specimens from Chiapas and Carl Zeiss, Oberkochen, Germany). All body mea- Belize; Sullivan et al. 2012). These data suggest that the surements and observations were made using a dis- D. frontalis morphotypes may represented distinct secting microscope with an ocular micrometer (20Ð species. 60ϫ). Slide-mounted features were observed with a To evaluate the taxonomic status of the two D. phase contrast microscope (400ϫ). frontalis morphotypes we analyzed: 1) variation of Character states for each of the following characters attributes of external and internal (i.e., genitalia) mor- were documented for all specimens: phology of both morphotypes; 2) the karyotypes of both morphotypes; 3) the phylogenetic relationship of 1. Outline of the frons (OF). In lateral view. 1) Trun- the two morphotypes with respect to other members cated convex, 2) completely convex. of the frontalis species complex as determined from 2. Number of rows of squamiform plates on the eighth mtDNA cytochrome oxidase I sequences; and 4) tergite (RSP). The eighth tergite displays parallel sperm transfer and reproductive success of laboratory rows of squamiform plates. The rows are better crosses between morphotypes. deÞned in the anterior area of the tergite than in the posterior, where they fade (Fig. 1c and d). 3. Acute projections on posterior edge of squamiform Materials and Methods plates in the median proximal area of the eighth Samples. In total, 389 D. frontalis, 200 D. mexicanus, tergite (AP). 1) Short and uniform (Fig. 1e), 2) and 22 D. vitei from 18, 8, and 2 geographic localities, variable in size (Fig. 1f). respectively, were analyzed (Table 1). Specimens 4. Shape of the cornu of the sphermatheca when were collected by the authors from naturally infested viewed laterally (SCS). 1) round, 2) oval (Fig. 1g Pinus spp. or borrowed from museums. Sex of the and h). The spermatheca is reniform and is divided specimens was determined by the presence of frontal into the nodulus and cornu. The nodulus is the tubercles and stridulatory apparatus in males (Lyon portion between the spermatic duct and the middle 1958, Mendoza-Correa and Zu´ n˜ iga 1991). constriction of the spermatheca, whereas the cornu Genitalia of males and females were dissected from includes the distal portion of the spermatheca be- fresh or alcohol-preserved specimens. The genitalia yond the middle constriction. The cornu presents were cleared by incubating them for 10 min at 10ЊCin variation in length relative to the width of middle

rod: (i) MA with strongly convex posterior margin of lobe, (j) MB with plane posterior margin of lobe; (k) body measures: LFT, length of frontal tubercles; EPW, epistomal process width; DE, distance between eyes; EW, eye width; LHP, length of head-pronotum; PW, width of posterior margin of pronotum; PL, pronotum length; EL, elytral length; (l) female and male genitalia: Ed, aedeagus; sp, spermatheca; sr, seminal rod; et, eighth tergite; RSP, rows of squamiform plates on eighth tergite; AP, acute projections on posterior edges on squamiform plates; n, nodulus; co, cornu; sp, spine; lb, lobe. 4A 14 Table 1. Populations of two morphotypes of D. frontalis and of D. mexicanus and D. vitei used in analyses

Number of specimens Type of analysis performedb Country Locality Coordinates Host studieda MA MB U M Sr Sp C G D. frontalis United States A Mississippi, Homochitto National Forest 31.5Њ, 91.2Њ Lindgren trap 10 X B Arizona, Coronado National Forest 31Њ 48Ј, 109Њ 24Ј Lindgren trap 6 X C Georgia, Oconee National Forest 33Њ 47Ј10Љ,83Њ 14Ј33Љ P. taeda 4X Mexico THE OF NNALS D Quere´taro Ja´lpan, La Parada 21Њ 06Ј21Љ,99Њ 05Ј99Љ P .greggii 20 X X X E Oaxaca, Santa Marõ´a Albarradas, San Pablo Villa de Mitla 16Њ 59Ј33Љ,96Њ 11Ј15Љ E P. pringlei 41 6 X X X X F Oaxaca, San Pedro Ayutla Zacatepec Dto Mixe 17Њ 07Ј12,Љ 96Њ 44Ј56Љ P. pringlei 21 8 X X X X G Oaxaca Norte 17Њ 27Ј44Љ,96Њ 41Ј38Љ P. pringlei 50 1 X X X X H Chiapas, La Trinitaria, Parque Nacional Lagos de 16Њ 08Ј42Љ,91Њ 43Ј29Љ P. oocarpa, P. maximinoi 70 70 X X X X X X Montebello I Chiapas, Teopisca 16Њ 31Ј57Љ,92Њ 28Ј12Љ Lindgren trap 3 X E NTOMOLOGICAL J Chiapas, Motozintla 15Њ 20Ј,92Њ 15Ј Lindgren trap 2 X X X Guatemala K Solola´, Finca Chuchiya 14Њ 44Ј00Љ,91Њ 07Ј50Љ P. pseudostrobus 210 X L Solola´, Colonia Maria Tecun 14Њ 48Ј50Љ,91Њ12Ј20Љ P. oocarpa 520 X M Solola´, Finca Socorro 14Њ 45Ј10Љ,91Њ 08Ј20Љ P. pseudostrobus 11 6 X N Huehuetenango, Canselaj 15Њ 16Ј10Љ,91Њ25Ј53Љ P. oocarpa 44 X Belize

O Mountain Pine Ridge Forest Reserve 16Њ 59Ј44.7Љ88Њ 46Ј23Љ Lindgren trap 8 9 X S Honduras OF OCIETY P Guaimaca 14Њ 32Ј08Љ,86Њ 49Ј00Љ P. oocarpa 17 3 X Nicaragua Q El Picacho, Cerro Tomabu 13Њ 02Ј00Љ,86Њ 18Ј00Љ P. oocarpa 10 4 X R Jalapa 13Њ 55Ј15Љ,86Њ 07Ј38Љ P. caribaea 8X Total 292 141 433 A D. mexicanus MERICA United States S Arizona, Coronado National Forest — 3X Mexico T Jalisco, Go´mez Farias 19Њ 52Ј5Љ, 103Њ 24Ј14Љ Ñ3 X U Distrito Federal,Tlalpan, Totoloapan 19Њ 18Ј00Љ,99Њ 12Ј56Љ P. montezumae 25 X V Estado de Me`xico, Nicolas Romero 19Њ 40Ј32Љ,99Њ 22Ј43Љ P. montezumae 39 X X W Puebla,Tepeyahualco, Hacienda San Roque 19Њ 29Ј52Љ,97Њ 30Ј00Љ P. cembroides 34 X X Tlaxcala, Altzayoaca, Ejido Santa Marõ´a Las Cuevas 19Њ 23Ј32Љ,97Њ 44Ј17Љ P. cembroides 30 X X Y Veracruz, Perote, Ejido Agua de los Pescados 19Њ 26Ј45Љ,97Њ 7Ј57Љ P. montezumae 37 X X Z Chiapas, Las Margaritas, Bajuco 16Њ27Ј52Љ,91Њ 53Ј18Љ P. pseudostrobus 29 X Total 200 D. vitei Guatemala Aa Sierra de las Minas Biosphere Reserve 15Њ 06Ј05Љ,89Њ 37Ј20Љ P .oocarpa 12 X 1 no. 107, Vol. Ab Chimaltenango, Patzun 14Њ 39Ј,90Њ 58Ј — 10 X Total 22

a MA, morphotype A; MB, morphotype B. b U, univariate analyses of morphological characters; M, multivariate analyses of morphological characters; Sr, geometric morphometric analysis of seminal rod; Sp, geometric morphometric analysis of spermatheca; C, karyotype analysis and crossing experiments; G, molecular phylogenetic analysis. January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 15

Fig. 2. Sr (seminal rod) and Sp (spermatheca) showing the placement of LM, landmarks, and SL, semilandmarks, for shape analysis: Sr: 1Ð8 LM; Sp: 11, 30 LM and 1Ð10, 12Ð29, and 31Ð38 SL.

constriction. In the round shape, the length is sim- were used: Principal coordinates analysis (PCoA) was ilar to width of middle constriction; in the oval used to explore multidimensional patterns of variation shape, the length is bigger than width of middle among specimens and cluster analysis to evaluate the constriction. grouping of specimens by mean of phenograms (Leg- 5. Number of striae on the spermatheca (NSS). The endre and Legendre 1998). Both methods were per- surface of the spermatheca is ornamented by trans- formed with a distance matrix computed with the verse striae that partially or completely encircle it Gower index using 15 characters deÞned in this study and are distributed predominantly in the nodulus and including the sculpture of the preepisternal area (Fig. 1g and h). This character was observed in (1] presence or 2] absence of striations on the preep- lateral view. isternal area of the prothorax; PAE) as a character 6. Density of striae in the proximal region of nodulus additional. Phenograms were built by the unweighted (DS). 1) Aggregate, 2) no-aggregate (Fig. 1g and h). pair-group method with arithmetic mean (un- 7. Proportion of the spermatheca covered by striae weighted pair-group method with arithmetic aver- (PSS). 1) Ϸ1/4 or (2) 1/2 of spermatheca (Fig. 1g age). Three different comparisons were carried out. In and h). the Þrst two, D. frontalis (both morphotypes) males 8. Shape of posterior margin of the seminal rodÕs lobe and females were analyzed separately; the third com- in lateral view (SSR). 1) Convex or 2) plane or parison was performed among D. frontalis (both mor- slightly convex (Fig. 1i and j). photypes), D. mexicanus, and D. vitei of both sexes. Continuous Characters. We measured the length of The number of characters included in each compar- the frontal tubercles (LFT), epistomal process width ison was different because of the differences in num- (EPW), distance between the eyes (DE), eye width bers of sex-speciÞc characters examined (see results). (EW), length of the head-pronotum (LHP), width of Univariate and multivariate analyses and plots were the pronotum at its posterior margin (PW), pronotum constructed using Matlab 7. 8. 0. 347 for windows (The length (PL), and elytra length (EL; Fig. 1k). MathWorks Inc.). Geometric Morphometrics. We assessed seminal rod and spermatheca shape variation in both D. Data Analyses frontalis morphotypes by means of geometric mor- Statistical Analyses. Normality of the distribution of phometrics analysis. The seminal rods of MA and each continuous character for each morphotype was MB and D. frontalis complex species D. mexicanus tested independently by the Shapiro and Wilkinson and D. vitei from multiple locations were compared (1965) test and examination of distribution histo- (Table 1). Spermatheca analysis was conducted on grams. In addition, we tested the homogeneity of vari- MA and MB specimens collected in seven localities. ances using CochranÕs test (Cochran 1941). Because Each seminal rod and spermatheca was photo- all characters had heterogeneous variances, all data graphed in lateral view using a Nikon Coolpix 5000 Ϫ were standardized (Xi Xmean/SD). camera mounted on a phase contrast microscope Univariate Analyses. Individual variation between (400ϫ). Images were identically arranged so that morphotypes was evaluated by StudentÕs t-test for the seminal rod was oriented with the spine (ventral continuous characters and the Mann-Whitney test for projection) pointing up and the lobe (or dorsal discrete characters (Zar 2010). projection) on the right-hand side of the spine; the Multivariate Analyses. Each was considered spermathecae were oriented with the termini point- an operational taxonomic unit (OTU). Two methods ing downward and the nodulus leftward (Fig. 2). 16 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

Landmark Analysis—Seminal Rod. The shape of were superimposed with GPA. Shape variation be- each seminal rod was quantiÞed from a set of two- tween morphotypes and the change in spermatheca dimensional coordinates (landmarks) with tpsDIG geometric conÞguration of each specimen were ana- 1.40 software (Rohlf 2004). Eight landmarks were lyzed as it was describe in Landmark Analysis for established: one type 1 (structures or points deÞned seminal rod. To evaluate statistical differences in sper- locally), three type 2 (points located at local cur- matheca shape between morphotypes, the relative vature maxima), and four type 3 (points distant from warps (W matrix) were analyzed through HotellingÕs the type I landmark; Zelditch et al. 2004; Fig. 2). To T test. Genitalia size differences between morpho- remove the nonshape variation (differences be- types were compared with a StudentÕs t-test using cause of location, scale, and orientation) from the spermatheca centroid size of both morphotypes. Cen- landmark conÞgurations, a Generalized Procrustes troid size was calculated in PAST 1.95 (Hammer et al. Analysis (GPA) was used (Rohlf and Slice 1990, 2001). The relation between spermatheca shape and Rohlf 1999) as implemented in the CoordGen6 pro- size was analyzed as it was describe above for seminal gram of Integrated Morphometrics Package (IMP) rod. Univariate and multivariate tests and plots (sem- (Sheets 2003) to produce a set of partial Procrustes inal rod and spermatheca) were completed using Mat- superimpositions of specimens. The coordinates ob- lab version 7.8.0.347 for windows (The MathWorks tained were transformed into relative warp scores to Inc.). produce a W matrix (Zelditch et al. 2004), and shape Karyology. The karyology of D. frontalis and its variation between morphotypes was analyzed with sister species has been studied extensively by previous a principal component analysis (ϭrelative warp anal- authors (Lanier 1981; Lanier et al. 1988; Salinas- ysis) in PCAGen 6 (Sheets 2003). The change in sem- Moreno et al. 1994; Zu´ n˜ iga et al. 1998, 2002a). Spec- inal rod geometric conÞguration of each specimen was imens of D. frontalis karyotyped by Lanier et al. visualized by thin-plate spline deformation grids in (1988), Salinas-Moreno et al. (1994), Zu´ n˜ iga et al. PAST 1.95 (Hammer et al. 2001). For evaluation of (1998), and Zu´ n˜ iga et al. (2002a) were a sample of the signiÞcant differences between species and morpho- insects set collected speciÞcally for those studies, types, the W matrix was analyzed through multivariate which were deposited at Laboratorio de Variacio´n analysis of variance (MANOVA) and post hoc pair- Biolo´gica y Evolucio´n de la Escuela Nacional de Cien- wise HotellingÕs T comparisons. For evaluation of gen- cias Bilo´gicas of Instituto Polite´cnico Nacional italia size differences, an ANOVA test was performed (ENCB-IPN) and University of Nuevo Leo´n Campus on seminal rod centroid size for each group (i.e., Linares. To classify the vouchers (20 individuals) and morphotype and species). Centroid size was calcu- specimens (302 individuals from 13 Mexican loca- lated in PAST 1.95 (Hammer et al. 2001). In addition, tions) of these collections as either MA or MB, we the relationship between seminal rod size and shape reviewed the presence or absence of prothoraxic stri- was analyzed by a PearsonÕs correlation test between ations. In addition, we performed a karyological anal- log-centroid size and relative warp one (RW1). ysis of MA and MB using live adult specimens from Semilandmark Analysis—Spermatheca. Semiland- Parque Nacional Lagunas de Montebello, Chiapas marks (arbitrary points along a curvature) were used (Table 1). to capture outline information on spermathecae be- The reproductive system of 20 males and 20 females cause of insufÞcient numbers of well-deÞned homol- each of morphotypes A and B were removed under ogous points suitable for landmarks. Semilandmarks RingerÕs solution, and then testes and ovarioles were may slide along the outline to optimally match corre- isolated and Þxed in FarmerÕs solution (3:1 ethanolÐ sponding points of other specimens. The application acetic acid). Slides were made using the aceto-car- MakeFan6 (Sheets 2003), which places alignment mine squash technique, and 100 meiotic Þgures for all “fans” at equal angular displacements along a curve, slides were analyzed from these preparations. Chro- was used to ensure consistent placement of the mosomal number was determined by counting biva- semilandmark points on spermatheca images. A fan of lent chromosomes in late prophase and metaphase, 19 radiating lines was added to all spermatheca images. and meiotic dynamics were examined for possible Radiating lines were of equal angular intervals, and the irregularities at the level of supernumerary chromo- points where they met the margin of the spermatheca somes, irregular associations in bivalents, and changes constituted semilandmarks (Bookstein 1997, Zelditch in the chromosome number. Slides were observed et al. 2004). Both landmarks and semilandmarks were under a phase-contrast microscope at 1,250ϫ, and the then digitized using tpsDIG version 1.40 software best Þgures were photographed using a Nikon Coolpix (Rohlf 2004) for, in total, 2 landmarks and 36 5000 camera. semilandmarks for each specimen (Fig. 2). When Molecular Analysis. Total genomic DNA of 10 adult semilandmarks are digitized at discrete points, a co- insects from each morphs collected in Parque Nacio- ordinates adjustment is recommended before GPA to nal Lagunas de Montebello, Chiapas, was extracted minimize the tangential variation of semilandmarks on and puriÞed (Table 1) using a DNAeasy tissue kit the spematheca contour. Therefore, the minimum (QIAGEN GmbH, Hilden, Germany). AmpliÞcation Procrustes distance criterion (Perez et al. 2006) was of an 800-bp fragment of the mitochondrial cytochrome used for an optimal superimposition with SemiLand 6 oxidase I (COI) gene was carried out using primers (Sheets 2003). To remove the nonshape variation, the TL2-N-3014 (Simon et al. 1994) and Mod-CI-J-2183 landmarkÐsemilandmark conÞgurations of specimens (CAACACTTATTTTGATTTTTTGG) designed for January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 17 this study. Polymerase chain reaction (PCR) was per- within each pit. Attacks were spaced Ͼ6 cm apart on the formed using a Biometra T Gradient thermocycler bark, with three to six attacks per log. Females were (Biometra GmbH, Hilden, Germany). AmpliÞcation conÞned for 18Ð24 h before a male was introduced into and sequencing of the mitochondrial DNA (mtDNA) the enclosure. Pairs were then excised alive from the COI fragment was carried out as described in Ruiz et gallery after 48Ð72 h. Spermathecae were dissected im- al. (2009). To assess whether pseudogenes could have mediately from females, mounted in water on slides, and been ampliÞed, a restriction fragment length poly- examined with phase contrast microscopy at 40ϫ for morphism analysis (RFLP) were carried out on frag- presence of sperm. To assess whether frequency of in- ments ampliÞed (data not shown). semination differed among the cross types, we per- The maximum-likelihood (ML) algorithm imple- formed Fisher exact tests (Zar 2010) on the 2 by 4 mented in the program aLRT-PHYML (Guindon and contingency table of the number of individuals with or Gascuel 2003) was used to infer the phylogenetic without sperm versus type of cross. All-pairwise Fisher relationships among mitochondrial DNA sequences of exact tests with a Bonferroni correction were used to both morphotypes of D. frontalis. Before the ML anal- contrast the sperm transfer of the different cross types. ysis, we determined an appropriate model of DNA In addition, we examined the spermathecae of 20 un- evolution and model parameters using both the paired females (i.e., not introduced into logs) of each Akaike Information Criterion and hierarchical likeli- morphotype to determine whether some proportion of hood ratio tests, as implemented in Modeltest v3.7 females used in the crosses might have already been (Posada and Crandall 1998). Both tests supported the inseminated before pairing. Tamura-Nei model (ϪLnL 2945.83; Tamura and Nei Gallery and Brood Production. To assess postzy- 1993) with gamma (G ϭ 0.31) distributed rate varia- gotic reproductive isolation and reproductive success tion and transitions and transvertions proportion (R ϭ in pairings between individuals of different morpho- 3.37). These G and R values, along with an optimized types, we repeated the procedures above mentioned base frequency, were used in aLRT-PHYML. To es- of the sperm transfer tests but incubated the logs at timate nodal support, the approximate likelihood ratio room temperature for 20Ð50 d before dissecting the test (aLRT; Anisimova and Gascuel 2006) with the gallery systems. For each type of cross, egg niches and ShimodairaÐHasegawa-like procedure option was larval galleries were counted per parent gallery system used. Two sequences (AF067986 and AY570903) from as approximations of the numbers of eggs laid and D. frontalis deposited in the National Center for Bio- hatched, respectively. Egg hatching percentage was technology Information (NCBI) were included in our estimated as the ratio of larval galleries to egg niches analysis (Kelley and Farrell 1998, Duan et al. 2004). in each gallery system. In addition, the length of each The sequences of D. approximatus (AF068000), D. parent gallery was measured. adjunctus (AF068001 and AF067992), D. brevicomis Variables parental gallery length and larvae per brood (AF067999 and AF068002), D. mexicanus (AF067988), (i.e., larval galleries per parental gallery that had a least and D. vitei (AFO68004; Kelley and Farrell 1998) were one larval gallery) were subjected to a square-root trans- included as outgroups. formation to meet the assumptions of parametric statis- Cross-Breeding Experiments. Logs from both in- tics, and these data along with egg niches per parent fested and uninfested Schiede ex Schitdl gallery were each evaluated separately by a two-way and H. E. Moore were obtained in the ANOVA (Zar 2010) with factors cross type (MA& ϫ Parque Nacional Lagunas de Montebello, Chiapas, Mex- MA(,MB&ϫMB(,MB&ϫMA(, and MA&ϫMB() ico, in 2009 and 2010 and transported to Colegio de la and host species (P. oocarpa and P. maximinoi). All- Forntera Sur (ϭECOSUR) Tapachula for experimenta- pairwise comparisons were carried out with TukeyÕs test. tion (Table 1). Uninfested logs for crossing tests were 30 The variable larvae per niche could not be transformed cm in length and stored under refrigeration (0Ð10ЊC) up adequately because of the large numbers of zeros, and to 10 d before use. Logs of infested P. oocarpa and P. therefore host trees were pooled and data were analyzed maximinoi were enclosed in cloth bags (50 by 70 cm), with a nonparametric one-way Kruskal-Wallis test (Zar and emerged beetles were collected daily and housed in 2010) with factor cross type: all-pairwise comparisons plastic petri dishes lined with moist paper towel and held among cross types were carried out with Mann-Whitney in refrigeration up to 5 d before use in crossing experi- tests and a Bonferroni correction. Proportions of paren- ments. Uninfested logs were allowed 4 h and beetles 1Ð2 tal galleries possessing at least one egg niche or larval h, to acclimate to room temperature after removal from gallery were separately compared among crosses by refrigeration. means of all Fisher exact tests with Bonferroni correc- Sperm Transfer. To assess possible prezygotic re- tion. productive isolation between the two morphotypes, sperm transfer among individuals was analyzed for artiÞcial crosses in the laboratory. Replicates were Results performed for all four possible crosses in both host Univariate Character Analyses of Morphotype trees. Attacks were induced on uninfested logs by conÞning females under a piece of plastic screening Outline of the Frons (OF). (&(). Thirteen per- within a Ϸ1-cm-diameter by Ϸ3-mm-depth pit cut cent of MB specimens presented a truncated convex into the outer bark with a cork borer. Within the pit, a frons and the remainder possessed fully convex Ϸ1-mm-diameter hole was drilled into the phloem frons. In MA 15% had truncated convex frons 18 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

Table 2. Statistical analysesa of continuous characters examined in two morphotypesb of D. frontalis

Characterc MA (␮m) MB (␮m) t (p) Length of frontal tubercles 32.78 Ϯ 0.9 33.6 Ϯ 1.4 1596 0.7472 Width of epistomal process 389.5 Ϯ 5.6 510.8 Ϯ 11.0 1502 Ͻ0.001 Distance between eyes 678.3 Ϯ 6.3 831.5 Ϯ 18.3 1277 Ͻ0.001 Width of left eye 201.1 Ϯ 1.2 241.6 Ϯ 3.5 1117 Ͻ0.001 Length of headÐpronotum 1383.8 Ϯ 20.8 1663.4 Ϯ 26.2 1970 Ͻ0.001 Pronotum length 812.6 Ϯ 11.0 978.1 Ϯ 14.8 1867 Ͻ0.001 Pronotum width 1186.1 Ϯ 10.0 1396.4 Ϯ 31.6 1655 Ͻ0.001 Elytra length 2004.5 Ϯ 18.5 2261 Ϯ 34.7 2275 Ͻ0.001

a t-statistic and P value of StudentÕs t-test. b MA, morphotype A; MB, morphotype B. c Illustrated in Fig. 1k. whereas the 85% had a fully convex frons. Morpho- photypes differ signiÞcantly (U ϭ 0; Z ϭϪ14.59; P Յ types did not differ signiÞcantly (U ϭ 63.9; Z ϭ 0.001). Ϫ0.14; P ϭ 0.491). Continuous Characters. (&(). For all continuous Number of Rows of Squamiform Plates on the characters, the MB possessed higher mean values than Eighth Tergite (RSP). (&). In both morphotypes, the MA (Table 2). StudentÕs t-test indicate statistically mean values for this character were similar (MB ϭ signiÞcant differences between morphotypes for all 18 Ϯ 0.4 SE; MA ϭ 16 Ϯ 0.6 SE), but differed statis- body measures analyzed (i.e., EPW, DE, WE, LHP, tically (U ϭ 190.5; Z ϭϪ2.705; P Յ 0.05). PW, PL, and E) except the LFT. Acute Projections on the Posterior Edge of Squa- miform Plates at the Median Proximal Area of The Eighth Tergite (AP). (&). All MA females possessed plates with uniformly short ornamentations and all MB Multivariate Analyses females had plates with irregularly sized ornamenta- Comparison of MA and MB Males. PCoA and clus- tions. The morphotypes differed signiÞcantly for this ter analysis were conduced with 99 MA and 65 MB ϭ ϭϪ Յ character (U 0; Z 8.055; P 0.001). using characters (EL, LHP, PL, PW, DE, WE, EPW, & Shape of the Cornu of the Spermatheca (SCS). ( ). OF, SSR, and PAE). The Þrst three principal coordi- Ninety percent of MB females possessed an ovate nates together explained 73.57% of total variation cornu whereas 10% had a round cornu. MA displayed (PCo1, 65.2%; PCo2, 4.99%; and PCo3, 3.38%). A scat- the reverse trend, with 34% of the females possessing terplot of the two Þrst components (PCo1 and PCo2) an ovate cornu and the 66% a round cornu. The mor- showed discrete and distinct phenotype separation photypes differed signiÞcantly (U ϭ 229; Z ϭϪ2.358; ϭ corresponding to each morphotype (Fig. 3a). In the P 0.019). cluster analysis dendrogram, two groups correspond- Number of Striae on Spermatheca (NSS). (&). In ing to each morphotype were obtained, with bootstrap both morphotypes, the mean values for this character values of 74% for the MA group and 73% for the MB (MB ϭ 16.7 Ϯ 0.7 SE; MA ϭ 15.6 Ϯ 0.7 SE) did not group. In the MB cluster, two subgroups were present differ statistically between morphotypes (U ϭ 348.5; with bootstrap values Ͻ50%, one with specimens ex- Z ϭϪ1.087; P ϭ 0.2769). clusively from Oaxaca state localities and the other Density of Striae at the Proximal Region of Sper- matheca’s Nodulus (DS). (&). One hundred percent with specimens from both Chiapas and Oaxaca states (Fig. 3a). The tree had a high cophenetic correlation of the all MB females had no-agglomerate striations. ϭ Յ In the MA females the 55% had aggregate striations index (r 0.917; P 0.05). and the 45% no-agglomerate striations. Mann-Whit- Comparison of MA and MB Females. PCoA and ney test showed statistically signiÞcant differences cluster analyses were carried out with 45 MA and 25 between morphotypes (U ϭ 169; Z ϭϪ3.898; P Յ MB females of morphotype using characters (EL, 0.001). LHP, PL, PW, DE, WE, EPW, OF, RSP, SCS, AP, NSS, Proportion of the Spermatheca Covered by Striae DS, PSS, and PAE). The Þrst three principal compo- (PSS). (&). Sixty-Þve percent of the MA had sper- nents together explained 62.66% of total variation matheca with a fourth part its surface covered with (PCo1, 43.87%; PCo2, 12.28%; and PCo3, 6.51%). Scat- striations, while the 35% had more than half of the ter plots of the Þrst two coordinates (PCo1 and PCo2) surface covered with striations. In the MB the 100% of showed discrete and distinct clusters corresponding to the specimens had more than half of the surface cov- each morphotype (Fig. 3b). ered by striations. Morphotypes possessed statistically In the female dendrogram, two groups were ob- signiÞcant differences by Mann-Whitney test (U ϭ tained corresponding to the two morphotypes, with a 184; Z ϭϪ4.084; P Յ 0.001). bootstrap value of 71% for the MA group and 86% for Shape of Seminal Rod Lobe (SSR). ((). All MA the MB group (Fig. 3b). Within both clusters, no specimens showed a convex shape of lobe and 100% of geographical subgroups were found, and the cophe- the MB had a plane or slightly convex lobe. The mor- netic correlation index was 0.844 (P Յ 0.05). January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 19

Fig. 3. Scatter plots from principal coordinate analysis and dendrograms from cluster analyses of morphological characters of both morphs of D. frontalis: (a) males and (b) females. Bootstrap values (1,000 pseudoreplicates) are indicated at the nodes and capital letters under the dendrograms indicate the populations analyzed from Table 1. MA, morphotype A; MB, morphotype B.

Comparison of Both Morphotypes with D. mexica- margin of the seminal rodÕs lobe from D. mexicanus and nus and D. vitei. PCoA and cluster analyses were D. vitei being concave, the charcater (SSR) for these carried out with 164 MA, 66 MB, 170 D. mexicanus, and species was coded as a different character state (3) for 10 D. vitei using characters (EL, LHP, PL, PW, DE, multivariate analyses. The Þrst three principal coor- WE, SSR, and PAE). Owing to shape of posterior dinates together explained 72.67% of total variation 20 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

Fig. 4. Scatter plot from principal coordinate analysis and dendrogram from cluster analyses of morphological characters of both morphotypes of D. frontalis, D. mexicanus, and D. vitei. Bootstrap values (1,000 pseudoreplicates) are indicated at nodes and capital letters adjacent to the dendrogram indicate the populations analyzed (Table 1). MA, morphotype A; MB, morphotype B.

(PCo1, 44.83%; PCo2, 22.6%; and PCo3, 5.24%). The comparisons: MA vs. MB (T ϭ 2503; F ϭ 138.4; P Յ scatterplot of the Þrst two coordinates showed dis- 0.001), MA vs. D. mexicanus (T ϭ 3.871 ϫ 107; F ϭ crete and distinct phenotype clusters corresponding 1.512 ϫ 105; P Յ 0.001), MA vs. D. vitei (T ϭ 3.483 ϫ to the four groups analyzed (Fig. 4). The dendrogram 1017; F ϭ 4.582 ϫ 1015; P Յ 0.001), MB vs. D. mexicanus had a cophenetic correlation index of 0.994 (P Յ 0.05) (T ϭ 1.71 ϫ 105; F ϭ 668.9; P Յ 0.05), MB vs. D. vitei and the topology showed four groups supported by a (T ϭ 5.19 ϫ 1015; F ϭ 8.11 ϫ 1013; P Յ 0.001), and D. bootstrap values Ͼ60% that correspond with the two mexicanus vs. D. vitei (T ϭ 7.098 ϫ 104; F ϭ 1109; P Յ morphotypes and two species. Both morphotypes 0.001). were clearly separated, and MB was more closely On average, MA and D. mexicanus possessed the related to the cluster formed by D. mexicanus and D. highest centroid size values (652.2 Ϯ 17.6 SE and vitei than to MA (Fig. 4). 622.7 Ϯ 30 SE, respectively) while D. vitei and MB had smaller sizes (525 Ϯ 21.7 SE and 584.4 Ϯ 39.5 SE, respectively). SigniÞcant centroid size differences Geometric Morphometrics were detected among morphs and species (ANOVA, ϭ Յ Landmark Analysis—Seminal Rod. The GPA-su- F3, 244 8.037; P 0.05). However, Tukey tests de- perimposed conÞguration of 131 MA, 114 MB, 30 D. tected signiÞcant differences only between D. vitei mexicanus, and 12 D. vitei seminal rods indicated that and the remaining groups (Q ϭ 3.83; P Յ 0.007). the positional range of variation of each landmark was Pearson correlation analysis between centroid size different. The landmarks that displayed highest vari- and the Þrst relative warp was not signiÞcant (r ϭ ation were those describing the seminal rod spine and 0.198; P Ն 0.05), suggesting that size was not an im- lobe shape. In the relative warp analysis, the Þrst three portant factor inßuencing shape differentiation. relative warps explained Ͼ73% of variation (RW1, Semilandmark Analysis—Spermatheca. The super- 46.3%; RW2, 19.3%; and RW3, 8.5%). The two-dimen- imposition conÞguration of 35 MA and 25 MB sper- sional scatterplot of the Þrst two relative warps mathecae showed that the semilandmarks with high- showed four discrete and distinct clusters correspond- est variation are those describing the curvature of the ing to each species and morphotype (Fig. 5a). The spermatheca, nodulus width, and cornu shape. The deformations in RW1 are associated to the invagina- Þrst three components of the relative warp analysis tion and evagination of the posterior margin of the explained 66.6% of shape variation (RW1, 39.5%; RW2, lobe and the relative size of the spine (Fig. 5a) and in 17.4%; and RW3, 9.7%). A two-dimensional scatterplot RW2 the changes are related to the degree of incli- of these relative warps revealed two distinct but over- nation of the spine and elongation of the lobe (Fig. lapping clusters corresponding to each morphotype 5a). (Fig. 5b). The deformations in RW1 were correlated A MANOVA on relative warp components of shape with the degree of concavity of the spermatheca and ␭ ϭ revealed statistically signiÞcant differences ( Wilks nodulus width (Fig. 5b) and in RW2, with the length ϭ Յ 0.0034; F33, 680 263.2; P 0.001) among morphs and of the nodulus and cornu (Fig. 5b). species groups. Pair-wise HotellingÕs T tests supported HotellingÕs T test on relative warp components in- statistically signiÞcant differences in shape for six dicated statistically signiÞcant differences between January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 21

Fig. 5. Scatter plots of the Þrst and second relative warps: (a) male seminal rod analysis of both morphotypes of D. frontalis, D. mexicanus, and D. vitei and (b) female spermathecae analysis of both morphotypes of D. frontalis. Changes in genitalic structures along both axes are shown by landmark conÞgurations in thin-plane-spline deformation grids. MA, morphotype A; MB, morphotype B. 22 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

frontalis corresponded to morphotype A. The chro- mosomal number of MB specimens was obtained by analyzing Þgures corresponding to meiosis I (pro- phase, metaphase, and anaphase) in testes and ovari- oles. The diploid chromosome number (2n) of MB was 12 chromosomes, with a meiotic formula of Þve auto- some pairs plus a sexual pair with an Xyp conÞguration (5AA ϩ Xyp; Fig. 6). Examination of Þgures did not reveal atypical chromosome dynamics, presence of su- pernumerary chromosomes, or chromosome alterations such as translocations or inversions. The meiotic formula for MA was 7AA ϩ Xyp; the same formula was reported for D. frontalis in previous studies (data not shown). Molecular Analysis. Ten partial sequences of each morphotypes of Х845 bp of CO1 were obtained. We did not observed double peaks, nonsynonymous mutations, indels, frameshifts, or additional stop codons, which suggest a low probability of having sequenced nuclear copies (nuclear mitochondrial DNA [NUMTs]). After the manual edition of the Fig. 6. Spermatocytes of D. frontalis morphotype B in sequences, a fragment of 786 bp was used in our prometaphase I with six chromosome pairs 1,000ϫ. Arrow analysis (GenBank KC855256Ð62). This fragment is indicates the associated sexual pair, Xyp. located between positions 2242 and 3028 in the Dro- sophila yakuba mitochondrial genome. The recovered the spermatheca shape of the two morphotypes (T ϭ phylogenetic tree showed a clear separation among 2.81 ϫ 1017; F ϭ 2.89 ϫ 1014; P Յ 0.001). MB had greater species and between both morphotypes (Fig. 7). Clus- average spermatheca centroid size than MA (1705.3 Ϯ ter I corresponded to D. adjunctus, D. brevicomis, and 26 and1579.9 Ϯ 55.4, respectively; t ϭ 2.93; P Յ 0.05). D. approximatus, with the latter two segregating as The Pearson correlation between centroid size and sibling species. Cluster II contained sister taxa D. fron- the Þrst relative warp was signiÞcant (r ϭ 0.5; P Յ talis, D. mexicanus, and D. vitei. All D. frontalis (both 0.05), suggesting that size is an important factor in- morphotypes) formed a single group with high nodal ßuencing shape differentiation. support (100%), and morphotype B was recovered as Karyology. Karyotyped vouchers and specimens a single group with 99% support. The two sequences collected for kariological studies did not show the from D. frontalis deposited in the NCBI (specimens presence of prothoraxic striations; therefore, we as- from Texas and Arizona) were integrated within of the sumed that the karyotype previously reported for D. MA group.

Fig. 7. Phylogenetic tree resulting from ML analysis of cytochrome oxidase I (COI). The model with the best Þt for nucleotide substitution was TrN-G-R. Support values at nodes were derived from 1000 pseudoreplicates. n ϭ number of individuals sequenced with similar haplotype (H). MA, morphotype A; MB, morphotype B. January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 23

Table 3. Percentage (inseminated/total) of inseminated fe- crosses. However, the reciprocal MB& ϫ MA( a males resulting from crosses of two morphotypes of D. frontalis on crosses did not differ signiÞcantly from any of the logs of two different host speciesb other cross categories in these three measurements. P. P. There was no evidence that host species inßuenced Cross type Total oocarpa maximinoi the observed differences in reproductive success MA& ϫ MA( 88% (21/24)a n/a 88% (21/24)a among the crosses, as there was not a signiÞcant in- MB& ϫ MB( 85% (11/13)ab 88% (7/8)a 86% (18/21)a teraction detected between cross type and host spe- MB& ϫ MA( 50% (7/14)ab 57% (8/14)a 54% (15/28)ab cies for the variables length of parent gallery, egg MA& ϫ MB( 36% (4/11)b 25% (1/4)a 33% (5/15)b niches per parent gallery, or larvae per brood.

a MA, morphotype A; MB, morphotype B. b Cross types associated with different letters differed signiÞcantly in the proportion of inseminated females (all-pairwise Fisher exact Discussion tests with Bonferroni correction for the number of contrasts). The present integrative approach augments and corroborates published biochemical evidence (Sulli- Cross-Breeding Experiments van et al. 2012) indicating the existence of two distinct Sperm Transfer. Spermathecae of 20 unpaired MA morphs within D. frontalis. The differentiation of mor- and MB females from the same pool collected for use photypes A and B is supported by several independent in the crossing experiments were dissected and none data sets: external morphological characters, shape of contained sperm. In total, 24 MA& ϫ MA(,21 seminal rod and spermathecae, karyological analysis of MB& ϫ MB(,28MB& ϫ MA(, and 15 MA& ϫ MB( the MB, phylogenetic analysis of mtDNA COI se- crosses were performed in which the male entered the quences, and crossing experiments between morpho- female gallery within 24 h and the female was recov- types. Our analyses strongly suggest that the MB is a ered alive afterward (Table 3). MA& ϫ MA( crosses distinct taxon that is distinguished from MA (D. fron- were obtained only on P. oocarpa, whereas the other talis sensu stricto) both genetically and morphologi- three cross categories were performed on both P. cally and from other species within the D. frontalis oocarpa and P. maximinoi. Fisher exact test showed a complex. signiÞcant association between the insemination rate Characters Analyzed. The evaluation of eight dis- and the cross type with both hosts pooled (P Ͻ 0.001) crete characters showed that Þve (PSS, DS, AP, SCS, as well as for P. oocarpa alone (P ϭ 0.004). Crosses and SSR) were also useful for separating the D. fron- within morphotype generally had a higher frequency talis morphotypes similarity to PAE. From such attri- of sperm transfer than crosses between morphotypes. butes, three could be used as diagnostic characters for With both trees included in the data, MA& ϫ MB( D. frontalis morophotype B, because they distinguish crosses had a signiÞcantly lower insemination rate it from D. frontalis morphotype A, D. mexicanus, and (33%) than either homotypic cross (88 and 86%), D. vitei by the presence of striae on preepisternal area whereas MB& ϫ MA(crosses (54%) did not differ (PAE), the irregular size of ornamentations of the signiÞcantly in insemination rate from any other cross plates on eighth tergite of females (AP), and the type (Table 3). For the three cross types attempted on shape of seminal rod (SSR). Although with some ex- both pine hosts, insemination frequencies were similar perience the morphs can be reliably distinguished by for both hosts. the shape of the seminal rod (plane or convex), the Gallery and Brood Productions. In total, 44 MA& ϫ shape variation of this character in MB should be used MA(,47MB& ϫ MB(,32MB& ϫ MA(, and 30 with caution because the lobule is sometimes slightly MA& ϫ MB( pairings were performed (Table 4), and convex. Spermathecal characters PSS, DS, and SCS in all cases males entered the female gallery entrance showed overlap of their states between morphotypes; within 24 h. SigniÞcant differences were detected however, these features can be useful in combination among cross types in the average length of the parent with nonoverlapping characters (i.e., AP and SSR) for galleries (F ϭ 4.17; df ϭ 3, 145; P ϭ 0.007), the numbers identiÞcation of D. frontalis morphotypes, as their of egg niches per parent gallery (F ϭ 3.80; df ϭ 3, 145; average states differed signiÞcantly between them. P ϭ 0.012), the number of larvae per brood (F ϭ 6.40; The statistical analysis of continuous attributes (DE, df ϭ 3, 90; P Ͻ 0.001), and the egg hatch rate (i.e., larval EL, EPW, LFT, LHP, PL, PW, and EW) also allowed galleries per egg niche; H ϭ 50.5; df ϭ 3; P Ͻ 0.001). separation of morphotypes. In general terms, MB pre- Likewise, there were signiÞcant differences among sented higher mean values than MA for these mea- cross types in the proportion of pairings that produced surements. In other cases of sibling species in the larvae (Fisher exact test, P Ͻ 0.001), but not egg niches genus and in other scolytines, quantitative measures (Fisher exact test, P ϭ 0.19). Both heterotypic crosses have sometimes been used for species separation. Ex- produced a signiÞcantly lower proportion of larvae- amples include pronotal width in Dendroctonus pon- producing parent galleries and had a signiÞcantly derosae Hopkins and Dendroctonus jeffreyi Hopkins lower egg hatch rate than either homotypic cross. The (Lanier and Wood 1968) and in species within the MA& ϫ MB( crosses produced signiÞcantly shorter frontalis complex (Lanier et al. 1988), relative length parental galleries and signiÞcantly fewer larvae per of abdominal sternites in Dendroctonus valens Le- brood than either homotypic cross and fewer egg Conte and Olivier (Pajares niches per parent gallery than the MA homotypic and Lanier 1990), and Þve different continuous char- 24 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

Table 4. Results of pairing experiments with D. frontalis morphotypesa

Total Length of parent Parent galleries with Egg niches per Parent galleries Larvae per Larvae per Cross type pairings gallery (cm)b,c,g egg nichesd,g parent galleryb,g with larvaee,g broodb,c,f,g egg niche f,g MA& ϫ MA( 44 43.1 Ϯ 3.5a 42a 66.8 Ϯ 5.9a 41a 39.0 Ϯ 3.8a 0.519 Ϯ 0.032a MB& ϫ MB( 47 38.9 Ϯ 3.4a 44a 54.1 Ϯ 6.5ab 38a 38.3 Ϯ 4.3a 0.533 Ϯ 0.040a MB& ϫ MA( 32 28.7 Ϯ 3.0ab 27a 38.3 Ϯ 6.7ab 8b 29.8 Ϯ 11.5ab 0.114 Ϯ 0.044b MA& ϫ MB( 30 26.5 Ϯ 3.9b 25a 36.1 Ϯ 5.9b 11b 12.6 Ϯ 4.5b 0.134 Ϯ 0.050b ANOVA Ͻ Ͻ P(crosses) 0.007 0.012 0.001 0.001 h P(host) 0.49 0.52 0.76 n/a P(interaction) 0.11 0.052 0.19 n/a

a MA, morphotype A; MB, morphotype B. b Means associated with the same letter did not differ signiÞcantly (TukeyÕs test; ␣ ϭ 0.05). c Both the two-way ANOVA and all-pairwise tests were performed on square-root transformed data. d Numbers of parent galleries with at least one egg niche present. Cross types associated with a different letter differed in the proportion of parent galleries with or without niches (all-pairwise Fisher exact tests with Bonferroni correction; corrected ␣ ϭ 0.0083). e Numbers of parent galleries with at least one larval gallery evident. Cross types associated with a different letter differed in the proportion of parent galleries with or without larval galleries (all-pairwise Fisher exact tests with Bonferroni correction; corrected ␣ ϭ 0.0083). f Mean numbers of larval galleries per parent gallery that produced live brood (i.e., that had at least one larval gallery). g Means associated with the same letter did not differ signiÞcantly (all-pairwise Mann-Whitney tests with Bonferrroni correction; corrected ␣ ϭ 0.0083). Results of one-way ANOVA on ranks (Kruskal-Wallis test). h Crosses were performed in similar numbers on P. oocarpa and P. maximinoi. acters in Tomicus destruens and Tomicus piniperda (L.) this structure (nodulus width, curvature of sper- (Faccoli 2006). matheca, and shape of cornu; Fig. 5b). The discrimi- Likewise, multivariate analysis of quantitative and natory power of the spermatheca shape was lower qualitative data agreed with the recovery of two dis- than that for the seminal rod, because morphometric crete groups corresponding to the morphotypes. Ro- data from the spermatheca produced merely partial bustness of these clusters was demonstrated when D. separation of the morphotypes. Although the shape vitei and D. mexicanus (species with a high morpho- and number of striae on the surface of the sper- logical similarity with D. frontalis) were incorporated matheca have been used as useful characters for sep- into the analysis and the inclusion of these additional arating D. frontalis and D. mexicanus (Rios-Reyes et al. samples did not disrupt the original clusters corre- 2008), results of the current study indicate that sper- sponding to each morphotype. Multivariate analysis matheca shape within the D. frontalis morphotypes is has been used extensively to identify and delimit mor- highly variable and consequently of limited value for phological and genetic variation within and between distinguishing them. species (Doyen 1973, Doyen and Slobodchikoff 1974, Chromosome Analysis. The chromosome number Mutanen 2005, Adeleke et al. 2008, Jeffrey et al. 2008, of the MB was different from that previously reported Padial and De la Riva 2009, Thorpe 2010). for specimens of D. frontalis (MA; Lanier et al. 1988; Shape Variation of Genitalia. Geometric morpho- Salinas-Moreno et al. 1994; Zu´ n˜ iga et al. 1998, 2002a), metric analysis of genitalic structures (seminal rod and subsequently determinated to be MA. In addition, we spermatheca) supports separation of the morphotypes conÞrmed that the chromosome number previously and also the separation of them with D. mexicanus and reported for D. frontalis was identical to that of MA D. vitei. Variation in seminal rod shape was concen- specimens collected in the same location as the karyo- trated in two speciÞc sites: the lobe and the spine (Fig. typed MB. This represents additional evidence that 5a), and was congruent with discrete seminal rod the two morphotypes are taxonomic entities. The character (i.e., SSR; posterior outline of lobe either chromosome number of MB is identical to other spe- convex and plane or slightly convex) that allowed cies in the frontalis complex including D. approxima- discrimination of the morphotypes. tus, D. brevicomis, and D. mexicanus (Lanier 1981).

Seminal rod shape is a key character used for iden- The sexual pair of MB has a Xyp conÞguration that is tiÞcation of Dendroctonus species (Vite´ et al. 1975, unlike the neo-XY conÞguration of D. approximatus Wood 1982, Lanier et al. 1988); however, a rigorous and D. brevicomis (Lanier et al. 1988, Zu´ n˜ iga et al. assessment of the taxonomic potential of this structure 2002a) but is possessed by D. mexicanus. However, had never been carried out. The current study shows differences in external and seminal rod morphology that analysis morphometric landmarks increase the (Figs. 4 and 5) as well as COI sequences (Fig. 7) discriminatory value of seminal rod morphology for between MB and D. mexicanus discount the possibility taxonomic purposes, particularly in situations where that MB and D. mexicanus are the same taxon. seminal rods cannot easily be reliably discriminated, as It is possible that the karyotype observed in the MB is the case with the D. frontalis morphotypes. specimens could represent polymorphism within D. As with the seminal rod, sites of greatest shape frontalis; however, it is unlikely that viable and fertile variation within the spermatheca indicated by geo- offspring would be produced because of the unbal- metric morphometrics were congruent with the char- anced combinations of chromosomes resulting from ϩ ϩ acter states identiÞed to categorize variation within the mixture of karyotypes 7II Xyp and 5II Xyp. The January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 25 resulting infertility would generate strong selection in parental chromosome number between morpho- against such pairings. Furthermore, chromosomal types imply that surviving F1 brood of heterotypic polymorphisms have not previously been detected in crosses would likely have been sterile. any species of Dendroctonus (Lanier 1981, Lanier et al. Forced heterospeciÞc pairings between described 1988), despite extensive chromosomal sampling of members of the frontalis complex have been reported certain species in their geographic range (Salinas- to produce parent galleries with egg niches; however, Moreno et al. 1994; Zu´ n˜ iga et al. 1998, 2002a). there are no conÞrmed reports of such crosses pro- Molecular Analysis. The COI-based tree of species ducing F1 brood (Vite´ et al. 1974, Lanier et al. 1988). in the frontalis complex is consistent with the COI- Nevertheless, selective pressures because of lower based phylogeny produced by Kelley and Farrell reproductive success of heterotypic crossings by the (1998), with the addition that the two D. frontalis D. frontalis morphotypes should support acquisition morphotypes formed distinct groups within the group and maintenance of traits that reduce incidence of of D. frontalis. The formation of two groups, one in- heterotypic pairs. Although it was possible in the lab- tegrated by the MB sequences and the other by the oratory to force or induce beetles of different mor- MA sequences derived from sympatric populations photypes to pair transfer sperm and produce larval (Chiapas), supports the presence of two different ter- brood (albeit at lower rates than homotypic crosses), minal taxa. The tree also shows a clear separation of extensive dissections of parental galleries on these taxa with respect to D. vitei and D. mexicanus. simultaneously infested by both species did not Numerous studies have debated the utility of showed incidences of heterotypic pairs (A.N. and mtDNA sequencing alone for identifying species, cor- B.T.S., unpublished data). Furthermore, differences in roborating described species, and recognizing cryptic the composition of aggregation and mate location species in groups with high morphological similarity. pheromones (Sullivan et al. 2012), spatial partitioning In this sense, the COI mtDNA differentiation level of the host (A.N. and B.T.S., unpublished data), and found between MA and MB is concordant with mor- possible differing times of arrival on the host may phological differences observed; results similar were provide important mechanisms of prezygotic repro- documented by Ruiz et al. (2009) for two subspecies ductive isolation in nature that would have been ab- from Dendroctonus pseudotsugae Hopkins. However, sent in our laboratory studies. as COI sequences from the small number of specimens Therefore, we conclude that the two morphotypes examined in this study are likely not representative of of D. frontalis are species distinct from each other as the potential variability of both morphotypes, and well as from other described species in the D. frontalis considering the wide distribution range of D. frontalis species complex. The zone of sympatry of these two (MA; Clarke and Nowak 2009) and Dendroctonus sp. species appears to include at least portions of Nica- nov. (MB; F.A.T., unpublished data), caution must be ragua, Guatemala, Honduras, Belize, and the southern taken regarding the use of mtDNA in the identiÞcation Mexican states of Michoacan, Oaxaca, and Chiapas of both morphotypes until a complete assessment of (F.AT. and G.Z., unpublished data). Morphotype B COI molecular variation is completed. insects have not been identiÞed by the authors in Crosses. Our tests to evaluate interbreeding poten- collections from elsewhere in the range of D. frontalis, tial between the morphotypes support the hypothesis which includes Sierra Gorda and Sierra Madre Ori- that they are distinct biological entities. SigniÞcantly ental (Salinas-Moreno et al. 2004), Arizona, and the higher rates of insemination in homotypic crosses southeastern United States. Specimens examined out- compared with at least one of the two heterotypic side the sympatric zone appear to be entirely MA, crosses implies the existence of at least partial prezy- which is congruent with both 1) the identical chro- gotic reproductive isolation barriers between the mor- mosome number from MA in the sympatric zone and photypes. Incompatible morphology of the genitalia D. frontalis in U.S. populations and, 2) The COI-tree (as suggested by the signiÞcant differences in seminal topology that grouped MA from Chiapas with D. fron- rod morphology between morphotypes) would have talis collected in United States. Because allopatric been one potential barrier to insemination, but be- zone of morphotype A includes the type locality of D. havioral incompatibility (e.g., unsuccessful courtship) frontalis (ÕCarolinaÕ; Wood 1982), we conclude that and other physical incompatibilities could also have MA is classiÞed appropriately as D. frontalis whereas led to a failure of pairings to copulate or transfer MB represents an undescribed species. Although sperm. Similarly, evidence of signiÞcantly lower egg these two taxa often coexist on the same individual hatch for both heterotypic crosses suggests the exis- hosts, studies suggest that they possess different col- tence of postzygotic barriers to reproduction, perhaps onization behaviors that might reduce competition. caused by the differing chromosome formulas of the Adult D. frontalis (MA) emerge from coinfested trees two morphotypes. Reduced reproductive success by earlier than Dendroctonus sp. nov. (MB), and D. fron- the heterotypic crosses was likewise implied by talis focus their attacks primarily on the upper and shorter average parental galleries, a lower frequency midbole, whereas Dendroctonus sp. nov. primarily in- of brood production, and smaller broods. Although it fest the bottom3mofthestem (Moreno 2008, un- was apparent that heterotypic crosses were capable of published data). Such differences suggest that the taxa producing viable F1, we could not ascertain whether occupy different ecological niches, which may pro- these were fertile, as brood of all cross types died mote their coexistence under conditions of syntopy before reaching adulthood. Nonetheless, differences (Rivas 1964). 26 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 107, no. 1

One important implication of our results is that the in Yunnan province (China) compared to Europe: new outbreaks currently attributed to D. fron- insights for the systematics and evolution of the genus talis in regions where Dendroctonus sp. nov. occurs Tomicus. Heredity 93: 416Ð422. may be caused instead by this latter species or by the Faccoli, M. 2006. Morphological separation of Tomicus pini- combined activity of both D. frontalis and Dendroc- perda and T. destruens (Coleoptera: Curculionidae: Sco- tonus sp. nov. Revised procedures for evaluation of lytinae): new and old characters. Eur. J. Entomol. 103: bark beetle infestations in the Central American re- 433Ð442. Guindon, S., and O. Gascuel. 2003. Simple, fast, and accu- gion will be necessary to assure that causative agents rate algorithm to estimate large phylogenies by maximum of these outbreaks are correctly identiÞed and that likelihood. Syst. Biol. 52: 696Ð704. appropriate management procedures may be devel- Hammer, O., D. Harper, and P. Ryan. 2001. Paleontological oped and implemented. statistics software package for education and data anal- ysis. Paleontol. Electron. 4: 1Ð9. Jeffrey, D. L., R. G. Foottit, G. Miller, N. J. Mills, and G. K. Acknowledgments Roderick. 2008. Molecular and morphological evalua- tion of the aphid genus Hyalopterus Koch (Insecta: We thank Juan Cruz, Efraõn Santiago, and Francisco Bo- ´ Hemiptera: Aphididae), with a description of a new spe- nilla (Comision Nacional Forestal Mexico), Juvencio ´ ´ cies. Zootaxa 1688: 1Ð19. Hernandez and Olivia Maldonado (Secretaria de Desarrollo ´ Kelley, S. T., and B. D. Farrell. 1998. Is specialization a did Agropecuario Forestal, Pesca y Acuacultura, Oaxaca, Mex- end? The phylogeny of host use in Dendroctonus bark ico), Roberto Castellanos (Parque Nacional de Lagunas de Montebello, Mexico), and Benjamin Moreno (Colegio de la beetle (Scolytidae). Evolution 52: 1731Ð1743. Frontera Sur-ECOSUR) for Þeld and lab assistance and for Lanier, G. N. 1981. Cytotaxonomy of Dendroctonus, pp. 33Ð providing insects specimens. Collection specimens were 66. In M. W. Stock (ed.), Application of Genetics and loaned by Luis Gerardo Cuellar Universidad de Nuevo Leo´n Cytology in Insects Systematics and Evolution. Forest Campus Linares. The project was funded by U.S. Depart- Wildlife and Range Experiment Station, University of ment of Agriculture Forest Service Southern Research Idaho Moscow. Station cooperative agreements 04-IC-11330129-131 and Lanier, G., and S. L. Wood. 1968. Controlled mating, kary- 07-IC-11330129-056 and partially by the Comisio´n Nacional ology, morphology and sex ratio in Dendroctonus pon- Forestal (Project 69539). F.A.T. was member of Programa derosa complex. Ann. Entomol. Soc. Am. 61: 517Ð516. Institucional de Formacio´n de Investigadores del Instituto Lanier, G., J. P. Hendrichs, and J. E Flores. 1988. Biosys- Polite´cnico Nacional (PIFI-IPN); F.A.T. and A.N. were Con- tematics of the Dendroctonus frontalis (Coleoptera: Sco- sejo Nacional de Ciencia y Tecnologõ´a fellows (267436 and lytidae) complex. Ann. Entomol. Soc. Am. 81: 403Ð418. 256629, respectively). Legendre, P., and L. Legendre. 1998. Numerical ecology: developments in environmental modeling 20, 2nd ed. Elsevier B. V. Amsterdam, The Netherlands. References Cited Lyon, R. L. 1958. A useful secondary sex character in Den- droctonus bark beetles. Can. Entomol. 90: 582Ð584. Adeleke, M. A., C. F. Mafiana, A. B. Idowu, M. F. Adekunle, Mendoza-Correa, M. G., and G. Zu´ n˜ iga. 1991. Determi- and B. M. Dansu. 2008. Morphometric studies on Culex nacio´n del sexo en Dendroctonus mexicanus Hopkins (Co- quinquefasciatus and Mansonia africana (Diptera: Culic- leoptera, Scolytidae) a trave´s del ana´lisis de cara´cteres idae) in Abeokuta, SouthÐwestern Nigeria. Tanzan. morfolo´gicos abdominales y cefa´licos. Acta Zoolo´gica J. Health Res. 10: 99Ð102. Mexicana 46: 13Ð21. Anisimova, M., and O. Gascuel. 2006. Approximate likeli- Midtgaard, F., and K. H. Thunes. 2002. Pine bark beetles in hood-ratio test for branches: a fast, 446 accurate, and the Mountain Pine Ridge Forest Reserve, Belize: descrip- powerful alternative. Systematic Biol. 55: 539Ð552. tion of the species and advice on monitoring and com- Billings, R. F. 2011. Aerial detection, ground evaluation, bating the beetle infestations. Norwegian Forestry Group and monitoring of the southern pine beetle: state per- and the Inter-American Development Bank, Isadaltø, spectives, pp. 245Ð261. In R. N. Coulson and K. D. Klepzig Norway. (eds.), Southern Pine Beetle II. U.S. Department of Ag- Moreno, C. B. 2008. Diferenciacion morfologica y compor- riculture Forest Service Southern Research Station Gen. ´ ´ tamental de dos morfos de Dendroctonus frontalis Zimmer- Tech Rep. SRS-140, Asheville, NC. Bookstein, F. L. 1997. Landmark methods for forms without mann (Coleoptera: Scolytinae) en Montebello Chiapas. Te- landmarks: morphometrics of group differences in out- sis de Maestrõ´a. El Colegio Frontera Sur ECOSUR, Chiapas, line shape. Med. Image Anal. 1: 225Ð243. Mexico. Clarke, S. R., and J. T. Nowak. 2009. Southern forest insect Mutanen, M. 2005. Delimitation difÞculties in species splits: & disease leaßet 49, pp. 1Ð8. USDA Forest Service PaciÞc a morphometrics case study on the Euxoa tritici complex Northwest Region (R6), Portland, OR. (Lepidoptera: Noctuidae). Syst. Entomol. 30: 632Ð643. Cochran, W. G. 1941. The distribution of the largest of a set Padial, J.M.N., and I. De la Riva. 2009. Integrative taxonomy of estimates variances as a fraction of their total. Ann. reveals cryptic Amazonian species of Pristimantis Hum. Genet. 11: 47Ð52. (Anura). Zool. J. Linn. Soc-Lon. 155: 97Ð122. Doyen, J. T. 1973. Systematics of the genus Coelocnemis Pajares, J. A., and G. N. Lanier. 1990. Biosystematics of the (Coleoptera: Tenebrionidae) a quantitative study of vari- turpentine beetles Dendroctonus terebrans and D. valens ation, vol. 73. University of California Publications in (Coleoptera: Scolytidae). Ann. Entomol. Soc. Am. 83: Entomology, Los Angeles, CA. 171Ð188. Doyen, J. T., and C. N. Slobodchikoff. 1974. An operational Perez, I. P., V. Bernal, and P. N. Gonzalez. 2006. Differences approach to species classiÞcation. Syst. Zool. 23: 239Ð247. between sliding semi-landmark methods in geometrics Duan, Y., C. Kerdelue´, H. Ye, and F. Lieutier. 2004. Genetic morphometrics, with an application to human craniofa- study of forest pest Tomicus piniperda (Col., Scolytinae) cial and dental variation. J. Anat. 208: 769Ð784. January 2014 ARMENDA´ RIZ-TOLEDANO ET AL.: EVALUATION OF MORPHS IN D. frontalis 27

Posada, D., and K. A. Crandall. 1998. Modeltest: testing the 2012. Biochemical evidence that Dendroctonus frontalis model of DNA evolution. Bioinformatics 14: 817Ð818. consists of two sibling species in Belize and Chiapas, Rios-Reyes, A. V., J. Valdez-Carrasco, A. Equihua-Martı´nez, Mexico. Ann. Entomol. Soc. Am. 105: 817Ð831. and G. Moya-Raygoza. 2008. IdentiÞcation of Dendroc- Tamura, K., and M. Nei. 1993. Estimation of the number of tonus frontalis (Zimmermann) and D. mexicanus (Hop- nucleotide substitutions in the control region of mito- kins) (Coleoptera: Curculionidae: Scolytinae) through chondrial DNA in humans and chimpanzees. Mol. Biol. structures of the female genitalia. Coleop. Bull. 62: 99Ð Evol. 10: 512Ð526. 103. Thorpe, R. S. 2010. Geographic variation: a synthesis of Rivas, L. R. 1964. A reinterpretation of the concepts “sym- cause, data, pattern and congruence in relation to sub- patric “and “allopatric” with proposal of additional terms species, multivariate analysis and phylogenesis. Ital. J. “syntopyc” and “allotopyc”. Syst. Zool. 13: 42Ð43. Zool. 54: 3Ð11. Rohlf, F. J. 1999. Shape statistics: procrustes superimposi- Vite´ J. P., S. Federico Islas, J.A.A. Renwich, P. R. Hughes, and tion and tangent space. J. Classif. 16: 197Ð223. R. A. Kliefoth. 1974. Biochemical and biological varia- Rohlf, F. J. 2004. TPS dig v 1.4. Department of Ecology and tion of southern pine beetle population in North and Evolution, State University of New York, Stony Brook, Central America. Z. Angew. Entomol. 75: 422Ð435. NY. (http://life.bio.sunysb.edu/morph/). Vite´, J. P., R. F. Hughes, and J.A.A. Renwich. 1975. Pine Rohlf, F. J., and D. Slice. 1990. Extension of the procrustes beetles of the genus Dendroctonus: pest population in method for the optimal superimposition of landmarks. Syst. Zool. 39: 40Ð59. Central America. FAO Bull. Plant Prot. 6: 178Ð184. Wood, S. L. 1963. A revision of the bark beetlesgenus Den- Ruiz, E., J. Victor, J. L. Hayes, and G. Zu´ n˜ iga. 2009. Molec- ular and morphological analysis of Dendroctonus pseou- droctonus Erichson (Coleoptera: Scolytidae). Great Basin dotsugae (Coleoptera: Curculionidae: Scolytinae): an as- Nat. 23: 1Ð117. sessment of the taxonomic status of subspecies. Ann. Wood, S. L. 1982. The bark and ambrosia beetles of North Entomol. Soc. Am. 6: 982Ð987. and Central America (Coleoptera: Scolytidae), a taxo- Salinas-Moreno, Y., G. Zu´ n˜ iga, and R. Cisneros. 1994. Chro- nomic monograph. Great Basin Nat. 6: 1Ð1359. mosomal variation in Mexican populations of the genus Zar, J. H. 2010. Bioestatistical analysis. Prentice Hall, NJ. Dendroctonus (Coleoptera-Scolytidae). Southwest. Nat. Zelditch L. M., D. L. Swiderski, H. D. Sheets, and W. L. Fink. 3: 283Ð286. 2004. Geometric morphometrics for biologist: a primer. Salinas-Moreno, Y., M. G. Mendoza, M. A. Barrios, R. Cis- Maple Vail, New York, NY. neros, J. E. Macı´as-Sa´mano, and G. Zu´ n˜ iga. 2004. Aerog- Zu´ n˜ iga, G., R. Cisneros, and Y. Salinas-Moreno. 1998. Chro- raphy of genus Dendroctonus (Coleoptera: Curculion- mosome number of Dendroctonus parallelocollis and D. idae: Scolytinae) in Mexico. J. Biogeogr. 31: 1163Ð1177. rhizophagus (Coleoptera: Scolytidae) and new reports of Shapiro, S. S., and M. B. Wilk. 1965. An analysis of variance Dendroctonus populations from Mexico and Guatemala. test for normality (complete samples). Biometrika 52: Ann. Entomol. Soc. Am. 91: 392Ð394. 591Ð611. Zu´ n˜ iga, G., R. Cisneros., J. L. Hayes, and J. Macı´as Sa´mano. Sheets H. D. 2003. IMP-integrated morphometrics package. 2002a. Karyology, geographic distribution and origin of Department of Physics, Canisius College, Buffalo, NY. the genus Dendroctonus Erichson (Coleoptera: Scolyt- (http://www3.canisius.edu/ϳsheets/imp7.htm). idae). Ann. Entomol. Soc. Am. 95: 267Ð275. Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. Zu´ n˜ iga, G., J. L. Hayes, J. C Gre´gorie, and R. Cisneros. 2002b. Flook. 1994. Evolution, weighting, and phylogenetics Chromosome number in Dendroctonus micans and kary- utility of mitochondrial gene sequences and a compila- ologycal divergence within the genus Dendroctonus (Co- tion of conserved polymerase chain reaction primers. leoptera: Scolytidae). Can. Entomol. 134: 503Ð510. Ann. Entomol. Soc. Am. 87: 651Ð701. Sullivan, B. T., A. Nin˜ o. B. Moreno, C. Brownie, J. Macı´as- Sa´mano, S. R. Clarke, L. R. Kirkendall, and G. Zu´ n˜ iga. Received 5 April 2013; accepted 26 September 2013.