MORPHOLOGY, HISTOLOGY, AND FINE STRUCTURE Differenes in Wing Morphometris of Lymntri dispr (Lepidopter: Ereide) Between Popultions Tht Vry in Femle Flight Cpility JUAN SHI, 1 FANG CHEN, 1 AND MELODY A. KEENA 2,3 Ann. Entomol. So. Am. 108(4): 528 535 (2015); DOI: 10.1093/es/sv045 ABSTRACT All mle gypsy moths, Lymntri dispr L., re ple of strong direted flight, ut flight in femles vries, inresing from west to est geogrphilly ross Eursi. To etter understnd how the wings differ etween femle flight ple nd flightless strins, wing morphometri nlysis of 821 gypsy moths from eight geogrphi strins (three Lymntri dispr dispr L., four Lymntri dispr siti Vnukovskij, one Lymntri dispr jponi Motshulsky) ws performed. Body mss; length nd width of oth fore- nd hindwing; nd wing re, spet, nd lods were mesured on oth sexes from eh strin. Gypsy moths were sexully dimorphi; femles hd higher wing lod, lrger spet rtios, igger wing re, nd hevier ody mss thn mles. Wing lods of femles, ut not mles, differed signifintly mong geogrphi strins nd were lower in flight ple strins. Wing spet ws less vrile within eh sex mong the strins. Femle fore- nd hindwing re were oth lrger in strins with strong direted flight pilities ompred with flightless strins, suggesting oth fore- nd hindwing res ply signifint roles in flight. A logisti regression model using femle forewing length nd wing lod orretly predited the femle flight pility of the soure strins >97% of the time nd my e useful tool to use in onjuntion with moleulr methods for deteting introdutions of Asin gypsy moth. None of the mle morphometri wing hrters were found to relily predit the femle flight pility of the soure popultion. KEY WORDS gypsy moth, wing morphology, flight pility, Ereide Gypsy moth is one of the most serious defoliting forest pests, ple of widespred outreks in temperte Holrti regions (Dvidson et l. 2001, Orozumekov et l. 2009). The gypsy moth, Lymntri dispr L., hs three reognized suspeies, Lymntri dispr dispr L. lled the Europen gypsy moth, Lymntri dispr siti Vnukovskij lled the Asin gypsy moth, nd Lymntri dispr jponi Motshulsky lled the Jpnese gypsy moth (Pogue nd Shefer 2007). Asin gypsy moth is distriuted in Asi est of the Url Mountins, widely in Russi, Chin, Jpn, nd Kore. Jpnese gypsy moth is distriuted on ll min islnds in Jpn, ut only in limited re on Hokkido (Keen et l. 2008). Europen gypsy moth ws originlly found in Europe, ws identlly introdued into North Ameri in 1869 (Forush nd Fernld 1896), nd hs sine spred with the leding edge of the infesttion rehing Mine, Minnesot, Wisonsin, Illinois, The use of trde, firm, or orportion nmes in this pulition is for the informtion nd onveniene of the reder. Suh use does not onstitute n offiil endorsement or pprovl y the U.S. Deprtment of Agriulture or the Forest Servie of ny produt or servie to the exlusion of others tht my e suitle. 1 Beijing Key Lortory for Forest Pest Control, Beijing Forestry University 35 Qinghu Est Rod, Beijing, Chin, 100083. 2 Northern Reserh Sttion, Forest Servie, United Sttes Deprtment of Agriulture 51 Mill Pond Rod, Hmden, CT 06514. 3 Corresponding uthor, e-mil: mkeen@fs.fed.us. Indin, Ohio, West Virgini, Virgini, North Crolin, Ontrio, Quée, New Brunswik, nd Nov Soti. The infesttion of Europen gypsy moth in the estern United Sttes is too well estlished to erdite, ut mesures to slow the spred nd ontrol lol outreks re eing tken (Toin 2008). Asin gypsy moth is onsidered to pose more signifint thret glolly, owing to its preferene for roder rnge of host speies (Brnhikov 1989), shortened egg hill requirements (Keen 1996, Wei et l. 2014), nd the flight pility of femles (Keen et l. 2008). However, femle gypsy moths ple of strong direted flight were lso found outside of Asi nd Sieri in the northestern prt of Europe, nd Jpnese gypsy moth femles lso re ple of flight (Keen et l. 2008). Flying femle gypsy moths re ttrted to lights in port res nd their egg msses hve een interepted in North Ameri on ships nd their rgo (U.S. Deprtment of Agriulture Animl nd Plnt Helth Inspetion Servie Plnt Protetion nd Qurntine [USDA-APHIS-PPQ] 2014). To prevent introdutions of flight ple femle gypsy moths from Asi, interntionl ollortions hve een set up to monitor port res nd ertify ships leving these ports during moth flights re free of egg msses (USDA-APHIS-PPQ 2014). Pheromone trps re used in North Ameri to monitor round ports for introdutions nd to delimit estlished popultions so they n e erdited. For regultory purposes, the USDA Pulished y Oxford University Press on ehlf of Entomologil Soiety of Ameri 2015. This work is written y US Government employees nd is in the puli domin in the US.
July 2015 SHI ET AL.: ASIAN GYPSY MOTH WINGS 529 refers to ny iotype of L. dispr possessing femle flight pility s the Asin gypsy moth, ut in this rtile, only the L. dispr siti suspeies will e lled Asin gypsy moth. Moleulr methods re used to determine the origin of mle gypsy moths ught in pheromone trps, s relile morphologil hrters re lking, ut no dequte methods urrently exist tht n predit the femle flight pility of the soure popultion (Keen et l. 2008). Intrspeifi polymorphisms in wing length, flight musles, flight ehvior, or ll hve een shown to exist etween dispersing nd nondispersing inset individuls (Hrrison 1980). In femle gypsy moths the two iotypes (flight ple nd flightless) hve een shown to differ in wing size, size of flight musles, nd preflight ehviors (Shields et l 1997; Keen et l. 2001, 2008). Previous studies of gypsy moths hve suessfully used disriminte nlysis to lssify femles sed on wing size (forewing length nd hindwing width) nd segregte for ody mss (mximum dominl width) into three flight pility groups (flightless, flight ple, or F 1 ) tht exhiit gliding-type flight (Keen et l. 2008). No nlysis of wing shpe (spet rtio) or wing loding (ody mss divided y totl wing re) hs een done for gypsy moth femles. Wing shpe hs een shown to hve lrge impt on the flight performne in other Lepidopter; longer slender wings hve een ssoited with long-durtion flight nd short rod wings with slow or more gile flight (Betts nd Wootton 1988, DeVries et l. 2010). As wings re omplex strutures, wing shpe nd size do not neessrily sle proportionlly with ody size, so oth need to e evluted independently (Outomuro et l. 2013). In ddition, there hs een little done to evlute the mle gypsy moths from the three suspeies to determine if there re ny mle trits tht re orrelted with the femle flight pility of the soure popultion. For exmple, it might e possile tht mles from popultions with femles ple of flight hve longer wings thn those from popultions with flightless femles if the inheritne of the trit is purely utosoml. In this study, morphometri pproh ws utilized to ssess vrition in wing shpe nd size for oth mle nd femle gypsy moths mong eight geogrphi popultions (two Europen gypsy moths, one Jpnese gypsy moths, nd five Asin gypsy moths) tht vry in femle flight pility. This morphometri informtion ws then used to determine if there re femle, mle, or oth hrters tht ould e used to relily distinguish etween popultions with nd without femles ple of sustined flight. Mterils nd Methods Gypsy Moth Strins nd Rering. We nlyzed 821 speimens (37 73 individuls of eh sex from eh strin) of L. dispr soured from eight sites: two sites in Chin, one site in Jpn, two sites in Russi, two sites in Europe, nd one site in the United Sttes. Detils of eh of the eight strins re given in Tle 1. Bsed on the reent review of Lymntri (Pogue nd Shefer 2007), the JN strin is the jponi suspeies, ll Russin nd Chinese strins re the siti suspeies, nd ll remining strins should e the dispr suspeies. All gypsy moths were trnsported under permit to the Forest Servie qurntine fility in Ansoni, CT. Vouher speimens for eh strin were deposited t the Entomology Division, Yle Peody Museum of Nturl History, New Hven, CT. Lrve from 10 to 14 different egg msses were rered to produe the dults used in this study. Lrve were held in wlk-in environmentl hmers mintined t 25 6 1 C, 60 6 5% reltive humidity, nd photoperiod of 16:8 (L:D) h. Lrve were rered in groups of eight (one to two groups per egg mss) in 177-ml ler plsti ups with unwxed pper lids for 35 47 d. Eh up ontined 90 ml of high whet germ diet (Bell et l. 1981) mde with Wesson slt mix without iron, nd dding the pproprite mount of morphous FePO 4 per liter of diet. The following mounts of morphous FePO 4 per liter of diet were used: 0.21 g for the CJ, CR, JN, nd RM strins; 0.17 g for the LJ nd RS strins; nd 0.13 g for the UC nd KG strins. Pupe were hrvested, sexed, nd stored y sex, egg mss (fmily), nd strin in 473-ml unwxed squt pper ups with ler plsti lids until dult elosion. Adults were removed dily, weighed, nd mesured. Wing Mesurement nd Morphometris Clultions. Moths were llowed to fully hrden their expnded wings efore mesurements were tken. Tle 1. Approximte lotion (ltitude nd longitude) of soure popultions nd designtions for strins of gypsy moth evluted in this study, rrnged y longitude from est to west Strin Country Colletion lotion Ltitude Longitude Smple size Femle Mle JN Jpn Ngoy, Honshu 35.15N 137.08 E 39 73 RM Russi Minerlni, Primorski 44.10 N 133.15 E 57 46 CR Chin Hrin, Heilongjing 45.78 N 126.61 E 42 44 CJ Chin Ynzikou, Beijng 40.32 N 116.15 E 55 60 RS Russi Shir, Khkssi 54.41 N 90.00 E 55 54 LJ Lithuni Juodkrnte, KuzsinNezijos 55.31 N 21.06 E 37 63 KG Germny Knigserg in Byer Bvri 50.10 N 10.34 E 38 45 UC United Sttes Bethny, New Hven County, CT 41.25 N 73.00 W 67 46 These strins re the sme ones tht were used in Keen et l. 2008 nd more detils on the olletion dte nd numer of egg msses used to strt the strin n e found there. The CR strin ws strted from 15 egg msses olleted in 2012 nd the CJ strin ws strted from 15 egg msses olleted in 2011.
530 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 108, no. 4 Body mss ws determined y weighing 1 24-h-old moths in pretred glssine envelopes on n eletroni lne urte to 0.0001 g. Moths were then frozen for 24 h efore the right fore- nd hindwings were refully removed from eh individul t the wing se using fine foreps nd sissors. The two wings from n individul were then seured to fom ord using inset pins prior to tking mesurements. Using digitl liper, the following mesurements were tken: forewing length, forewing width, hindwing length, nd hindwing width. The wing lndmrks used for eh mesurement were s follows: forewing length ws from se of the ostl vein to the tip of the pil ngle (pproximtely the pil end of the fourth rdil rnh); forewing width ws from the pil end of the third rdil rnh to the nl mrgin ( line pproximtely prllel to the ody); hindwing length ws from the se of the uitl vein to the pil end of the third medil rnh; nd hindwing width ws from the pil end of the suostl vein to the nl end of the third nl rnh (Fig. 1). The dethed wings were then photogrphed with Cnon 450D (EF-S18-55 mm, Tokyo, Jpn) digitl mer while they were held flt y glss plte. A lk kground ws used for femle wings nd white one for mle wings to improve the imge qulity nd ontrst. A ruler ws inluded in eh photo on the sme plne s the wings to provide tool for 1 5 8 4 2 3 Fig. 1. The lndmrks used for eh wing mesurement were s follows: forewing length ws from se of the ostl vein (1) to the tip of pil ngle, pproximtely the pil end of the fourth rdil rnh (3); forewing width ws from the pil end of the third rdil rnh (2) to the nl mrgin (4); hindwing length ws from the se of the uitl vein (5) to the pil end of the third medil rnh (7); nd hindwing width ws from the pil end of the suostl vein (6) to the end of the third nl rnh (8). 6 7 lirting the reltionship etween pixels nd distne. To ensure the sme zoom nd sme height ove glss, the mer ws set s follows: mer height: 34 m ove glss; zoom: 35 mm; mro-setting:? 1/ 100 exposure, 5.6 f-stop; /1/60 exposure, 4.5 f-stop. The digitl imges were leled with individul numer, sex, nd soure popultion. The wing re ws determined using the softwre IMAGEJ 1.47 (Ferreir nd Rsnd 2012). The imges were first onverted to inry nd the kground olor ws deleted to leve white spe round the lk wing imge (Fig. 1). The outline of the imge ws defined to eliminte minor dmge to the mrgin of the wings. With the ruler of known sle, the pixels per mm ws lirted using the sle on the ruler in the imge nd then forewing nd hindwing re ws otined. The totl wing re ws lulted s twie the mesured right forewing nd hindwing res omined. Wing spet rtios were lulted s the squre of forewing length, divided y the forewing re. Wing loding ws lulted y dividing ody mss y wing re. Sttistil Anlysis. The mle nd femle wing dt were nlyzed seprtely. The fit of eh dt set to vrious distriutions ws evluted using PROC UNIVARIATE (SAS Institute 1999, Cry, NC). The Shpiro Wilk nd the Anderson Drling test were used to ssess normlity. However, in ses where no distriution met the normlity ssumption, the distriution tht gve the est fit sed on the Kolmogorov Smirnov test ws used. The following dependent ontinuous vriles were nlyzed in PROC GLIM- MIX (SAS Institute 1999): wing lod, forewing spet, hindwing spet, totl wing re, forewing re, hindwing re, ody mss, forewing length, forewing width, hindwing length, nd hindwing width. A ompletely rndomized design ws used with geogrphil strin s the fixed effet nd the mternl fmily (egg mss) s the rndom effet. The norml distriution with n identity link ws used for the hindwing spet, forewing re, hindwing re, forewing length, forewing width, hindwing length, nd hindwing width. The gmm distriution with log link funtion ws used for the wing lod, wing spet, totl wing re, nd ody mss. The gmm distriution ws hosen euse it gve the est fit to these dt, whih hd long right tils euse of overdispersion. The ody mss mesurements were multiplied y 1,000 nd the lod ws multiplied y 100 efore nlysis to void tking the logs of deiml numers. For eh model, residuls were evluted for normlity nd the homogeneity of vrine ws ssessed using Levene s test. The group option ws used in the rndom sttement to ount for unequl vrines mong geogrphil strins (ll the femle prmeters exept hindwing re nd spet, nd ll the mle prmeters exept ody mss, totl wing re nd forewing spet) if they existed. Differenes mong mens were determined y the lest-squres mens test with ¼ 0.05 nd onservtive Tukey Krmer grouping (SAS Institute 1999). A Person orreltion nlysis (PROC CORR, SAS Institute 1999) ws done independently for eh sex on
July 2015 SHI ET AL.: ASIAN GYPSY MOTH WINGS 531 the wing morphometri hrters (wing lod, forewing spet, hindwing spet, totl wing re, forewing length, forewing width, hindwing length, nd hindwing width) nd ody mss to determine whih trits were sttistilly independent enough (hd solute r vlues of <0.4) to e inluded in model. The wing size hrters were ll strongly orrelted nd they were lso orrelted with ody mss so forewing length ws seleted to represent this group. Fore- nd hindwing spet rtios were orrelted with eh other ut not with wing lod or the size hrters so oth wing lod nd forewing spet were hosen to e inluded. The geogrphil strins were grouped y femle flight pility: flightless (KG nd UC), mixed flight (30% not strong fliers, LJ), nd strong flight (CJ, CR, RM, RS, nd JN). The flight pility lssifitions were either sed on wht ws reported y Keen et l. Wing Lod (mg/mm 2 ) 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Femles Mles CR LJ RM CJ RS JN UC KG Geogrphil Strin Fig. 2. Comprison of wing lod mong eight gypsy moth strins (geometri men 6 95% CI error rs), whih re ordered y inresing wing lod. Different letters indite signifint differenes in wing lod mong gypsy moth strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. (2008) or on MAK unpulished dt. Hlf of the individuls from eh strin within the flightless nd the strong flight groups were seleted for use in reting model to distinguish etween the two groups nd the other hlf of the dt were held to evlute the model. The mixed flight group ws lso retined for use in model evlution. Seprte logistil regression nlyses (PROC LOGISTIC, SAS Institute 1999) were run for eh sex with flight group s the response vrile nd ll possile omintions of forewing length, wing lod, nd forewing spet rtio s the ontinuous vriles to ssess their preditive power. The results of the est model nd the one for forewing length lone re presented euse forewing length is esy to mesure, is often mentioned s distinguishing trit for femles, nd wing loding nnot e lulted for the ded dry mle moths trpped in pheromone trps, whih my e the only mteril ville for use in identifying the suspeies of n introdution. Akike s informtion riteri (AIC; Akike 1973) were used to evlute the fit of the models: the smller the AIC vlue the etter the fit. The mximum likelihood prmeter estimtes in eh model were used to lulte the proility of femle flight (positive vlues inditing flight) for eh mle nd femle moth from the reserved set nd the perentges of orret nd inorret lssifitions were determined. Results The mle nd femle L. dispr were sexully dimorphi oth in wing morphometris nd olors (femles white nd mles rown; Tles 2 nd 3). Mles hd smller wings, weighed less, nd hd smller wing lods thn femles. The wing lod of femles ws signifintly different mong geogrphil strins (F ¼ 31.02; df ¼ 7, 82; P < 0.0001; Fig. 2). CT nd KG femles wing lod ws signifintly greter thn tht of femles from the other strins. JN femles wing lod ws signifintly greter thn CR nd RM femles wing lods. There were no signifint differenes etween strins in wing lods of mles (F ¼ 2.08; df ¼ 7, 82; Tle 2. Differenes in femle gypsy moth wing morphometris mong geogrphi strins Strin Sex BM (g) FWL (mm) FWW (mm) HWL (mm) HWW (mm) FWA (m 2 ) HWA (m 2 ) CR Femle 0.88 6 0.12 34.78 6 1.00 20.06 6 0.53 25.36 6 0.72 20.05 6 0.82 4.46 6 0.22 3.32 6 0.12 LJ Femle 0.55 6 0.08d 27.65 6 0.80de 14.39 6 0.31 19.57 6 0.53d 14.04 6 0.38 2.68 6 0.13d 1.82 6 0.12d RM Femle 1.09 6 0.05 35.89 6 0.62 20.44 6 0.44 25.42 6 0.31 20.99 6 0.32 4.74 6 0.16 3.55 6 0.12 CJ Femle 1.31 6 0.05 38.48 6 0.54 21.60 6 0.31 28.14 6 0.30 21.91 6 0.42 5.75 6 0.16 3.97 6 0.11 RS Femle 1.04 6 0.05 33.72 6 0.52 18.48 6 0.25 24.65 6 0.30 19.81 6 0.31 4.40 6 0.10 3.26 6 0.12 JN Femle 1.47 6 0.06 38.57 6 0.92 21.21 6 0.35 28.07 6 0.64 22.05 6 0.72 5.90 6 0.26 4.08 6 0.13 UC Femle 1.04 6 0.05 27.87 6 0.35d 14.44 6 0.23 19.75 6 0.27d 14.97 6 0.28 2.97 6 0.08 2.09 6 0.11d KG Femle 0.84 6 0.08 25.49 6 0.56e 14.26 6 0.44 18.47 6 0.48d 14.37 6 0.44 2.42 6 0.11d 1.75 6 0.12d Sttistis Femle F ¼ 19.19 F ¼ 78.87 F ¼ 100.66 F ¼ 100.94 F ¼ 76.65 F ¼ 87.86 F ¼ 64.46 df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7,300) P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 BM, ody mss; FWL, forewing length; FWW, forewing width; HWL, hindwing length; HWW, hindwing width; FWA, forewing re; HWA, hindwing re. For eh trit the men 6 SE is given, different letters in eh olumn indite signifint differenes etween strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. The differenes in the denomintor degrees of freedom re due to the use of the group y strin tht hd to e used when vrines were not equl etween strins.
532 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 108, no. 4 Tle 3. Differenes in mle gypsy moth wing morphometris mong geogrphi strins Strin Sex BM (g) FWL (mm) FWW (mm) HWL (mm) HWW (mm) FWA (m 2 ) HWA (m 2 ) CR Mle 0.17 6 0.05 25.00 6 0.43 15.60 6 0.27 19.35 6 0.28 15.20 6 0.24 2.44 6 0.08 2.03 6 0.06 LJ Mle 0.13 6 0.05 21.69 6 0.22d 13.18 6 0.14 16.32 6 0.18d 13.09 6 0.16e 1.71 6 0.04e 1.45 6 0.03e RM Mle 0.17 6 0.05 24.38 6 0.26 15.37 6 0.29 18.29 6 0.22 15.54 6 0.17 2.42 6 0.06 2.05 6 0.05 CJ Mle 0.25 6 0.05 28.63 6 0.36 17.69 6 0.33 21.69 6 0.33 17.75 6 0.38 3.41 6 0.07 2.88 6 0.09 RS Mle 0.18 6 0.05 23.14 6 0.24 13.79 6 0.17 17.27 6 0.16 14.53 6 0.17d 2.19 6 0.04 1.83 6 0.04d JN Mle 0.25 6 0.05 29.08 6 0.36 17.89 6 0.25 21.78 6 0.22 18.27 6 0.24 3.52 6 0.07 2.99 6 0.07 UC Mle 0.16 6 0.05 21.69 6 0.26d 13.15 6 0.20 16.54 6 0.25d 13.91 6 0.24de 1.99 6 0.05d 1.77 6 0.05d KG Mle 0.15 6 0.05 21.16 6 0.21d 13.15 6 0.16 16.12 6 0.17d 13.71 6 0.17e 1.80 6 0.04de 1.50 6 0.04e Sttistis Mle F ¼ 26.54 F ¼ 101.97 F ¼ 73.01 F ¼ 98.80 F ¼ 67.41 F ¼ 159.05 F ¼ 90.22 df ¼ (7,341) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 82) df ¼ (7, 341) df ¼ (7, 82) P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 BM, ody mss; FWL, forewing length; FWW, forewing width; HWL, hindwing length; HWW, hindwing width; FWA, forewing re; HWA, hindwing re. For eh trit the men 6 stndrd error is given, different letters in eh olumn indite signifint differenes etween strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. The differenes in the denomintor degrees of freedom re due to the use of the group y strin tht hd to e used when vrines were not equl etween strins. Forewing Aspet Rtio 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Femles Mles CR LJ RM CJ RS JN UC KG Geogrphil Strin Fig. 3. Comprison of forewing spet rtio mong eight gypsy moth strins (geometri men 6 95% CI error rs). Different letters indite signifint differenes in wing spet rtio mong gypsy moth strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. Hindwing Aspet Rtio 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Femles d Mles d d d d CR LJ RM CJ RS JN UC KG Geogrphil Strin Fig. 4. Comprison of hindwing spet rtio mong eight gypsy moth strins (geometri men 6 95% CI error rs). Different letters indite signifint differenes in wing spet rtio mong gypsy moth strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. P ¼ 0.0553), ut iologilly signifint differenes my exist. Within eh geogrphil strin the femle moths tended to hve lrger wing spet rtios thn mles oth for fore- nd hindwings. Aspet rtios differed signifintly y strin for femles (forewing: F ¼ 8.79; df ¼ 7, 300; P < 0.0001; hindwing: F ¼ 7.36; df ¼ 7, 300; P < 0.0001). The LJ strin hd lrger forewing spet rtio thn ll ut the CR strin (Fig. 3) nd lrger hindwing spet rtio thn ll ut the KG strin (Fig 4). The LJ strin mles lso hd the lrgest forewing spet rtios nd strin hd signifint impt on the rtio (F ¼ 24.43; df ¼ 7, 341; P < 0.0001). The LJ nd CR strin mles hd the lrgest hindwing spet rtios nd there ws signifint strin effet on the rtio (F ¼ 28.19; df ¼ 7, 82; P < 0.0001). The totl wing re of mle gypsy moths ws smller thn tht of femle moths ross ll eight geogrphil strins (Fig. 5). Totl wing re of oth femles (F ¼ 76.57; df ¼ 7, 300; P < 0.0001) nd mles (F ¼ 159.05; df ¼ 7, 341; P < 0.0001) vried signifintly mong strins. Individuls from the CJ nd JN strins hd signifintly lrger totl wing re thn ll other strins in mles nd lrger thn ll ut the RM strin in the femles. Individuls from the KG nd LJ strins hd signifintly smller totl wing re thn ll other strins in mles nd smller thn ll ut the UC strin in the femles. Geogrphil strin hd signifint effet on ody mss in oth the femle (F ¼ 19.19; df ¼ 7, 82; P < 0.0001) nd mle (F ¼ 26.54; df ¼ 7, 341; P < 0.0001).TheodymssofJNfemleswssignifintly hevier thn tht of ll ut the CJ strin (Tle 2).
July 2015 SHI ET AL.: ASIAN GYPSY MOTH WINGS 533 LJ femles weighed less thn the femles of ll the other strins. The ody mss of CJ nd JN mles ws signifintly hevier thn tht of mles of the other strins (Tle 3). Tle 4 provides the prmeter estimtes, odds rtios, nd model fit sttistis for the two est models for prediting the femle flight pility of strin sed on mle or femle wing morphometri hrters. Both femle forewing length nd wing loding were signifint preditors of femle flight pility nd model ontining oth hd the lowest AIC fit sttisti. This model orretly predited femle flight for 96.7% of the reserved femles from flight ple strins nd lk of flight for 98.1% of the femles from the flightless strins. When the model ws evluted using the LJ femles, it predited tht 97.3% would e ple of flight. When the other model tht only used femle forewing length ws Totl Wing Are (m 2 ) 25.0 20.0 15.0 10.0 5.0 0.0 d d Femles Mles CR LJ RM CJ RS JN UC KG Geogrphil Strin d d Fig. 5. Comprison of totl wing re mong eight gypsy moth strins (geometri men 6 95% CI error rs). Different letters indite signifint differenes in totl wing re mong gypsy moth strins within eh sex sed on Tukey Krmer s multiple omprison tests t the 5% signifine level. evluted using the reserved femles, 90.2% of flight ple nd 94.3% of flightless femles were orretly predited, while only 40.5% of the LJ femles were predited to e flight ple. Of the hrters evluted, only mle forewing length ws found to e good preditor of the femle flight pility of the soure strin (Tle 4). When the model tht only used mle forewing length ws evluted, it orretly predited femle flight for 90.6% of the reserved mles from strins with flight ple femles. The model only predited no flight for 65.3% of the reserved mles from strins with flightless femles. When the model ws evluted using the LJ mles, it only predited tht 33.3% of the mles were from strin tht the femles hd flight pility. Disussion Femles from geogrphil strins tht were known to e ple of strong direted flight hd signifintly smller wing lods thn the strins with flightless femles. There were no signifint differenes etween the geogrphil strins in the wing lods of mles, lthough the differenes my e lrge enough to e iologilly signifint. Mles with lower wing lods my e ple of stronger or longer distne flight thn those with higher wing lods. Within eh geogrphil strin, the femle moths tended to hve lrger wing spet rtio thn mles. Although wing spet rtios did differ etween strins for oth mles nd femles, the differenes did not orrespond with the known femle flight pility of the strins. Wing lod nd spet rtio re predited to e linked to mneuverility; other Lepidopter with low wing lod nd rounder wings (lower spet rtio) hve een shown to hve greter mneuverility (Betts nd Wootton 1988, DeVries et l. 2010). As gypsy moth mles ll fly in zigzg pttern to sty within the pheromone plume to find femles to mte with (Crdé nd Hgmn 1979), low wing lods nd spet rtios would e expeted. Also, wing size (totl wing re) vries onsiderly more thn shpe (wing spet rtios) etween strins for femles, whih would e expeted if the heritility (proportion of the vrition tht is ttriutle to genetis) of these two trits Tle 4. Logisti regression prmeter estimtes (6SE), odds rtios (95% Wld CL), nd model fit sttistis for prediting the femle flight pility of the strin using mle or femle wing forewing length (FWL) nd wing loding (WL) Logisti models Femle Mle Prmeter Estimtes Interept 25.57 6 4.34 18.54 6 6.57 26.88 6 4.72 27.01 6 4.84 FWL 0.85 6 0.14 1.00 6 0.28 1.20 6 0.21 1.21 6 0.21 Wld v 2 34.86 Wld v 2 12.60 Wld v 2 33.61 Wld v 2 33.53 df 1, P < 0.0001 df 1, P ¼ 0.0004 df 1, P < 0.0001 df 1, P < 0.0001 WL 0.12 6 0.04 0.004 6 0.03 Wld v 2 11.80 Wld v 2 0.01 df 1, P ¼ 0.0006 df 1, P ¼ 0.9044 Odds Rtios FWL 2.34 (1.76 3.09) 2.71 (1.56 4.69) 3.34 (2.22 5.01) 3.34 (2.22 5.02) WL 0.88 (0.83 0.95) 1.00 (0.94 1.07) AIC sttisti 61.19 26.95 92.63 94.62 AIC, Akike s informtion riteri (Akike 1973) used to evlute the fit of the models, lower vlues re etter fits.
534 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 108, no. 4 differs, whih ould led to different ptterns of evolution, s ws found for Drosophil mediopuntt (Bitner-Mthé nd Klzko 1999,). Further work would e needed to estimte heritility for these trits. Averge forewing length within strin ws longer in femlesthninmles,ndlonger in femles from L. d. siti (34 39 mm) strins thn L. d. dispr strins (25 28 mm). This is onsistent with the dignosti femle forewing lengths for the two suspeies, L. d. siti 28 41 mm nd L. d. dispr 20 30 mm (Pogue nd Shefer 2007). Additionlly, the femles from the strins within the siti suspeies hve longer hindwings thn the femles from the dispr suspeies, whih suggests tht oth wings ply signifint roles in flight pility. This is different from the finding tht only forewings re relted to dispersl in Melite inxi (L.), nother lepidoptern (Breuker et l. 2007). The logisti regression model using femle forewing length nd wing loding orretly predited the femle flight pility of the soure strins >97% of the time, whih ws etter thn just using femle forewing length lone. In ft, some femles from some of the Asin strins should not e predited to e ple of flight euse flight pility is not fixed (some individuls with only gliding-type flight) in ny of the popultions tht were evluted (Keen et l. 2008). This model ould e useful tool (if live femles re ville so fresh weights n e otined) to use in onjuntion with the moleulr methods to predit the flight pility of femles in the soure popultion. This would e espeilly useful if the moleulr methods suggest the soure popultion is from Europe euse mny Europen popultions hve only flightless femles, while others, like the LJ strin, n hve >70% of the femles tht n fly. Even though ll mles fly, mles from the Fr Est Asin popultions (femles fly) tended to hve lrger wings thn those from Europe (some femles n fly) or North Ameri (femles don t fly) when rered in ommon environment. However, the logisti regression model sed on mle forewing length ( signifint preditor of femle flight) inorretly predited the femle flight pility of the soure strin frequently (up to 67% in the LJ strin) nd so n not e relily used. Previous studies demonstrted tht flight pility of femle gypsy moth hs heritility of 0.60 nd forewing length hs heritility of 0.70, oth with no evidene of sex linkge or mternl effets (Keen et l. 2007). This would suggest tht the genes for ritil flight hrteristis suh s wing length re most likely loted on the utosomes rther thn mitohondril DNA or the sex hromosomes. Mle moths, ll of whih n fly, did not seprte relily into groups sed on the femle flight pility of the strins, suggesting tht there hs to e one or more genes on the W hromosome (femles re the heterogeneti sex in Lepidopter) tht diretly regulte flight genes in femles or indiretly regulte it y oding for sexully dimorphi ody msses. Further reserh to identify the genes involved in flight is needed to determine extly how flight is regulted in the two sexes. Aknowledgments We thnk P. Gri, P. Moore, nd A. Vndel for rering ssistne t the Forest Servie qurntine fility in Ansoni, CT. We thnk J. Rihrds, R. Simmons, V. Snhez, nd the two nonymous reviewers for offering suggestions tht improved this mnusript. This work ws funded y grnt from the Ntionl Nturl Siene Foundtion of Chin (31170613) nd Beijing Higher Edution Young Elite Teher Projet (YETP0740). Referenes Cited Akike,H.1973.Informtion theory nd n extension of the mximum likelihood priniple, pp. 267 281. In B. N. Petrov nd F. Cski (eds.), Seond interntionl symposium on informtion theory. Ademii Kido, Budpest. Brnhikov, Y. N. 1989. Eologil sis of the evolution of host reltionships in Eursin gypsy moth popultions, pp. 319 338. In W. E. Wllner nd K. A. MMnus [Teh. Coord.], Proeedings, Lymntriide: A Comprison of Fetures of New nd Old World Tussok Moths, USDA Gen. Teh. Rep. NE-123. Broomll, PA. Bell,R.A.,D.C.Owens,M.Shpiro,ndJ.R.Trdif.1981. Development of mss rering tehnology, pp. 599 633. In C. C. Done nd M. L. MMnus (eds.), The Gypsy Moth: Reserh Towrd Integrted Pest Mngement. USDA Teh. Bull. 1584, Wshington, DC. Betts,C.R.,ndR.J.Wootton.1988.Wing shpe nd flight ehviour in utterflies (Lepidopter: Ppilionoide nd Hesperioide): A preliminry nlysis. J. Exp. Biol. 138: 271 288. DeVries, P. J., C. M. Penz, nd R. I. Hill. 2010. Vertil distriution, flight ehviour nd evolution of wing morphology in Morpho utterflies. J. Anim. Eol. 79: 1077 1085. Bitner-Mthé, B. C., nd L. B. Klzko. 1999. Heritility, phenotypi nd geneti orreltions of size nd shpe of Drosophil mediopuntt wings. Heredity 83: 688 696. Bitner-Mthé, B. C., nd L. B. Klzko. 1999. Size nd shpe heritility in nturl popultions of Drosophil mediopuntt: temporl nd mirogeogrphil vrition. Geneti 105: 35 42. Breuker,C.J.,P.M.Brkefield,ndM.Gis.2007.The ssoition etween wing morphology nd dispersl is sexspeifi in the glnville fritillry utterfly Melite inxi (Lepidopter: Nymphlide). Eur. J. Entomol. 104: 445 452. Crdé, R., nd T. Hgmn. 1979. Behviorl responses of the gypsy moth in wind tunnel to ir-orne enntiomers of disprlure. Environ. Entomol. 8: 475 484. Dvidson, C. B., K. W. Gottshlk, nd J. E. Johnson. 2001. Europen gypsy moth (Lymntri dispr L.) outreks: review of the literture, p. 15. USDA Gen. Teh. Rep. NE-278. U.S. Deprtment of Agriulture, Forest Servie, Northestern Reserh Sttion, Newtown Squre, PA Ferreir, T., nd W. S. Rsnd. 2012. ImgeJ User Guide IJ/1.46. (imgej.nih.gov/ij/dos/guide/), 2010 2012 Forush, E. H., nd C. H. Fernld. 1896. The gypsy moth. Porthetri dispr (Linn.). Wright & Potter, Boston, p. 495. Hrrison, R. G. 1980. Dispersl polymorphisms in insets. Annu. Rev. Eol. Syst. 11: 95 118. Keen, M. A. 1996. Comprison of the hth of Lymntri dispr (Lepidopter: Lymntriide) eggs from Russi nd the United Sttes fter exposure to different tempertures nd durtions of low temperture. Ann. Entomol. So. Am. 89: 564 572. Keen,M.A.,W.E.Wllner,P.S.Grinerg,ndR.T. Crdé. 2001. Femle flight propensity nd pility in Lymntri dispr (Lepidopter: Lymntriide) from Russi, North Ameri, nd their reiprol F1 hyrids. Environ. Entomol. 30: 380 387.
July 2015 SHI ET AL.: ASIAN GYPSY MOTH WINGS 535 Keen, M., P. Grinerg, nd W. Wllner. 2007. Inheritne of femle flight in Lymntri dispr (Lepidopter: Lymntriide). Environ. Entomol. 36: 484 494. Keen,M.A.,M.J.Côté, P. S. Grinerg, nd W. E. Wllner. 2008. World distriution of femle flight nd geneti vrition in Lymntri dispr (Lepidopter: Lymntriide). Environ. Entomol. 37: 636 649. Orozumekov, A. A., A. M. Liehold, V. I. Ponomrev, nd P. C. Toin. 2009. Gypsy moth (Lepidopter: Lymntriide) in Centrl Asi. Am. Entomol. 55: 258 264. Outomuro, D., D. C. Adms, nd F. Johnsson. 2013. Wing shpe llometry nd erodynmis in lopterygid dmselflies: omprtive pproh. BMC Evol. Biol. 13: 118. Pogue, M. G., nd P. W. Shefer. 2007. Areviewofseleted speies of Lymntri Hüner [1819] inluding three new speies (Lepidopter: Notuide: Lymntriine). USDA Forest Helth Tehnology Enterprise Tem. FHTET-2006 2007 SAS Institute. 1999. SAS/STAT user s guide, version 8, 6th ed. SAS Institute, Cry, NC. Shields, K. S., R. T. Hirth, R. F. Young, nd M. A. Keen. 1997. Comprison of thori musulture of AsinndNorthAmeringypsymoths,p.78.In S.L.C. Fosroke nd K. W. Gottshlk (eds.), Proeedings, U. S. Deprtment of Agriulture Intergeny Gypsy Moth Reserh Forum 1997; Annpolis, MD. USDA Gen. Teh.Rep.NE-240 Toin, P. C. 2008. Cost nlysis nd iologil rmifitions for implementing the gypsy moth Slow the Spred Progrm, p. 21. Gen. Teh. Rep. NRS-37. U.S. Deprtment of Agriulture, Forest Servie, Northern Reserh Sttion, Newtown Squre, PA (USDA-APHIS-PPQ 2014). U.S. Deprtment of Agriulture, Animl nd Plnt Helth Inspetion Servie, Plnt Protetion nd Qurntine. 2014. Asin gypsy moth response guidelines. (http://www.phis.usd.gov/plnt_helth/plnt_pest_info/ gypsy_moth/downlods/agmsurveyresponseguidelines. pdf) (essed 15 April 2015). Wei, J., Y.-Q. Luo, J. Shi, D.-P. Wng, nd S.-W. Shen. 2014. Impt of temperture on postdipuse nd dipuse of the Asin gypsy moth, Lymntri dispr siti. J. Inset Si.14: 5. (www.insetsiene.org/14.5) (essed 14 My 2015). Reeived 2 Deemer 2014; epted 29 April 2015.