JournalofResearchon theLepidoptera 31(l-2):16-23, 1992 Direction Of Spring Migration Of Vanessa cardui (Nymphalidae) In Colorado James A. Scott 60 Estes St., Lakewood, Colorado 80226 Abstract. Springmigrationof3016Vanessacardui(L.)migrantswas studiedinColoradoin 1992 and 1983,usingstandardvectormethods. Migrationwastotheeast-northeast/northeast(averaging31°in 1992, 51° in 1983, east being 0° and north 90°), and this direction did not changesignificantlyduringtheday,disprovingatheorythatmigrants maintainaconstantangletothesun. Theefficiency(unidirectionality) ofmigrationisabout80%duringpeakmigration,butdropstonearzero afterward. Some spring adults in central Colo, overwintered there. Migrants mate-locate all day long, versus late in the day for non- migrants. Key Words: Vanessa cardui, migration, sun-compass mechanism. Introduction The study of migration of butterflies is filled v^ith numerous brief reports ofvast s’svarms heading in a certain direction. There are few^er studies in v^hichthe direction ofeach individual v^as charted, and even fewer in which the time ofday was recorded. The latter is ofinterest becauseofatheorythatmigrantsflyataconstantangletothesunBaker (1968a, b, 1969) and thus change direction from morning to afternoon. Recent migrations ofVanessa cardui allowed the testing ofthis theory. Methods Time ofdayanddirectionwererecordedforeachof3016 adults: 2725in 1992 and 291 in 1983. I walked about open areas (mostly large grassy areas) in metropolitanDenver,Colorado,duringmosthoursonpeakflightdaysbutduring onlypartofthedayonsomedays,andestimateddirectionofeachindividualseen as one of16 compass directions (N, NNE, NE, ENE, E, ESE, etc.). Standardvectormethodswereused. Eachobservationrepresentsanarrowof length 1, and the individual arrows can be drawn connected end-to-start (the start ofthe next vector connected to the pointed end ofthe previous vector) on graph paper. Then alinecanbedrawnconnectingstartoffirstvectortoendof lastvector,whichformsthetotalvectorwhoselengthandanglecanbemeasured from the graph paper (which represents the overall length and direction of migration). Thefollowingequationsareasimplerwaytocalculatethelengthand direction ofthe total vector, where E is the numberofadults fl5dngtoward the east, SSWthe numberflyingtoward the south-southwest, etc. (The derivation ofthe equations is this: for ENE for instance the angle is 22.5° and a vectorof length 1 has a Y-component [height] ofsin 22.5° or .3827 and anX-component [widthlofcos22.5°or.9239;and.70711isthesinandcosof45°fortheNEvector, etc.) 31(l-2):16-23, 1992 17 W X component oftotal vector= E - + .70711(NE + SE) - .70711(SW + NW) + .3827(NNE + SSE)- .3827(SSW+NNW)+ .9239(ENE + ESE)- .9239(WSW WNW) + Y component oftotal vector= N - S + .70711(NE + NW) - .70711(SE + SW) + .9239(NNE +NNW)- .9239(SSE +SSW)+ .3827(ENE +WNW)- .3827(ESE + WSW) V(length ofvector) = squareroot(X^ +Y^) Bycustominmathematics, thedirections and axes are arranged as inFig. 1, and angles are counterclockwise from due east (east being 0°, north 90°, west 180°,south270°). Theoveralldirectionofmigrationcanbefoundbyplottingthe individualvectors, orfromelementarytrigonometry: AngleA= tanHY/X) Theefficiencyofmigrationisgivenbythelengthofthetotalvectordividedby the numberofadults observed (N). Itrepresents the proportion ofindividuals thatflyintheoveralldirectionofmigration(thedirectionality),andrangesfrom 1.0 if all individuals fly the same direction, to zero ifflights are random in direction. I included in sample size only flying adults, and ignored the ones feedingonflowers orresting. Thus: Directionality(Efficiency) ofmigration E = V/N N = sample size N, 90° 1992 Fig. 1. Migration April 26-30 1992 during one-hour periods starting at the times indicated. A dashed line pointstothe 8:00vector, and “7:00” is left ofthe 7:00vectorwhichishiddenamongthebasesoftheothervectors. Numbers on Y and X axes represent numberof migrating individuals. 18 J. Res. Lepid. Results In 1992, veryfewmigrants (n=13)werenotedfromApril 10-25,which A flew predominantly eastward. vast migration occurred April 26-30, fewwere seenMay 1-2, a small migration occurred May3, few May4-5, few(N=18)flewnortheastwardMay6,andveryfew(N=8)flewpredomi- nantlynortheastwardMay8-18. Table1detailsmigrationfromApril26- May 5. The April 26-30 peak was massive (especiallyApril 28-30), the overalldirectionforthefivedayswasalittlenorthofENE(31°,V=1945., N=2395), and efficiency was great (81%). May 1 the migration was ceasing (flying to NNE with only 50% efficiency), and May 2 the few adults seenwere essentiallynonmigratorywithavery smallvector and almostzeroefficiency. May3asmallmigration(N=167)againoccurred, butoddlyitsdirectionwas southeastwardathighefficiency(79%). May 4-5fewadultswereseenandmostwerenonmigratoryjudgingbythelow efficiency (36% and 24%); they flew predominantly eastward May 4 indicating perhaps a mixture of southeastern migrants like the day before and northeastward migrants like the next day May 5 when they flew northeastward as before. Table 1. Daily migration 1992, N=2725. Day N Direction Vector Unidirectionality April 26 57 25.02° 40.27 .71 April 27 124 16.52° 104.66 .84 April 28 1065 29.25° 882.09 .83 April 29 474 39.51° 382.63 .81 April 30 677 30.26° 543.17 .80 May 1 53 63.89° 26.65 .50 May 2 20 19.75° 1.06 .05 May 3 167 320.34° 132.08 .79 May 4 23 357.87° 8.24 .36 May 5 26 41.81° 6.28 .24 Table 2. Migration each hourApril 26-May 5, 1992, N=2683. Times are military standard time of start of one hour periods (7:00 means 7:00-7:59). Hour N Direction Vector Unidirectionality 7 00+ 20 45.84° 5.18 .26 : 8:00-h 75 33.94° 53.73 .72 9:00+ 286 28.28° 241.11 .84 10:00+ 361 22.35° 295.22 .82 11:00+ 373 30.10° 321.60 .86 12:00+ 223 40.40° 185.51 .83 13:00+ 282 31.38° 218.47 .77 14:00+ 422 23.12° 307.54 .73 15:00+ 157 328.94° 113.10 .72 16:00+ 335 35.29° 250.15 .75 17:00+ 109 49.12° 68.19 .63 18:00+ 40 39.54° 38.98 .97 31(l-2):16-23, 1992 19 Table 3. Migration each hourApril 26-30, 1992, N=2395. Hour N Direction Vector Unidirectionality 7:004- 19 45.71° 6.18 .33 8:004- 75 33.94° 53.73 .72 9:004- 274 26.71° 238.13 .87 10:004- 317 19.42° 277.88 .88 11:004- 373 30.10° 321.60 .86 12:00-h 223 40.40° 185.51 .83 13:004- 268 31.52° 214.07 .80 14:00-^ 361 26.27° 289.26 .80 15:00+ 27 10.24° 20.24 .75 16:00+ 309 39.85° 249.05 .81 17:00+ 109 49.12° 68.19 .63 18:00+ 40 39.54° 38.98 .97 Total 2395 30.77° 1944.63 .81 Table 4. Migration 1983, in two-hour periods. Hour N Direction Vector Unidirectionality 8:00+ 45 51° 11 .25 10:00+ 61 45° 25 .41 12:00+ 55 59° 34 .61 14:00+ 73 54° 49 .67 16:00+ 57 42° 42 .74 Total 291 51° 161 .55 In1983,afewadultswereseenApril22-24,withamoderatemigration April25-27,fewmigrantsApril28-30,andasmallmigrationMay4. The overalldirection ofmigrationwastothenortheast(51°, Table 4, Fig. 2). Evidently the southwestern deserts such as the Mojave and Sonoran Deserts are the major source for springV. cardui in central Colorado, if the butterflies maintained the same ENE-NE flight in Arizona-Utah- California thattheyflew in Colorado. Migration in insects is generally considered to be “post-teneral, pre- reproductive” (Johnson, 1969), althoughthere are afew observations of oviposition then continuing migration, and oviposition during north- wardmigrationistypical ofDanausplexippus (L.)(Cockrell et al. 1993); thus migration obviously starts after the cuticle is hardened for flight, andstopsastheeggsbecomereadytobelaid. Thisgeneralizationseems true for V. cardui also, because the relative lack ofmigration on some days in 1992 (May 1-2, 4-5) suggests that adults had stopped migrating then,presumablytoreproduce(severalovipositionswereobservedApril 28 by a nonmigrating female). In 1983 theefficiencyofmigrationgraduallyincreasedduringtheday (Table 4), evidentlybecause mornings were cool and adults oftenfed on flowersinthemorning. Butin1992efficiencywasaboutthesameallday (Tables2-3)exceptveryearlyinthemorning,probablybecausemornings were warmer in 1992. 20 J. Res. Lepid. Fig. 2. Migration 1983 during two-hour periods starting atthe times indicated. The butterfly sun-compass mechanism Baker(1968a,b, 1969)claimedthatbutterfliesmigratebymaintaining a constant angle to the sun, so that their direction changes clockwise duringthedayinnorthernlatitudeswherethesun’spositionmoveseast to south to west during the day. Vanessa cardui, however, does not change its direction ofmigration duringthe day (Tables 2-4, Figs. 1-2). AdultsflewalittlenorthofENEin 1992,NEin 1983,withnosignificant change of direction during the day. Obviously V. cardui has some neurological mechanism that produces a constant direction of flight during the day despite the change of direction ofthe sun as it moves acrossthesky. Probablythismechanismisthesamesun-compassclock that has been demonstrated in honeybees. The maindifferencebetweentheobservations ofBakerandmyownis that Baker recorded flight directions in ordinarypopulations thatwere notin migratoryflight, andthen assumed(wrongly)thattheirbehavior was like that ofmigratory butterflies, whereas my observations are of My obvious, definitemigrations. opinionisthatBaker’s observations of nonmigratory butterflies showed a change in direction during the day because glare from lookingtoward the sun affected the abilityto notice 31(1-2):16=23, 1992 21 butterflies fl5dng in certain directions, causing a systematic bias that followedthechangeofdirectionofthesunduringtheday. Possiblysome non-migratorybutterfliesdochangedirectionsomewhatduringtheday, but this seems doubtful. Since no other authors have proved any significant change ofdirection ofmigration during the day, and other studies provedthat direction does notchange duringthe day(Arbogast 1966onAgraulisvanillae, Balciunas&Knopf1977on Urhanusproteus, Walker 1978 onPreciscoenia^Phoehis sennae^ K. Adamsin Baker 1978 onBelenois aurota, D. Lawrie 1984 Lepid. News #1 p. 7 and M. Myres Can. Field Nat. 99:147-155 on Vanessa cardui), clearly Baker’s theory must be discarded. Tomore-objectivelydeterminethesun’sinfluenceonobserverbias,and to obtain more accurate data, more sophisticated apparatus will be necessary: the observerwould siton aplatformthatrotatedfrequently, and would note for each migrant the compass angle printed on a surrounding deck and enter it into a computer, while the computer recordedthe angle ofthe observer’s seattothe sun; the computerwould usethetimeofdaytocalculatetheanglesofbutterflyandobservertodue north,andthenumberofadultsseenateachdirectioncouldbecompared to the angle ofobserver to the sun to determine any observer bias. Adult overwintering in Colorado Most authorshaveperhaps wronglyassumedthat V. cardui overwin- SW tersonlyinsouthernareaslike ArizonaandMexico(Williams 1970). But in central Colorado, my 30 years’ observations suggest that adults almost never migrate southward; the very few southward migrations that have been observed were high in the mountains in midsummer (Emmel&Wobus 1966in Colo., andI saw asouthwardmigrationinthe alpine zone ofWind River Mts. Wyo. early Aug. 1983 and another in alpine central Colo.). Southward migrationhas neverbeen seen on the Colorado plains/foothills, where every year during September large numbers nectar on Chrysothamnus nauseosus and other flowers until frost, withoutmigrating. Whenwild-caughtSept, adults areplacedin a homefreezer,theydienoearlierthantheotherVanessa,Polygonia, and Nymphalis which are known to overwinter as adults (all die within 30 minutes in the freezer, suggesting that adults in nature must require manyhours days orweeks ofgraduallycoldertemperatures to increase theirinternalconcentrationofglyceroltosurvivefreezing). In addition, adults are generallypresent in spring even in years without noticeable in-migration, and Cockerell (1934) once found an adult in January in nature in Boulder Co. Colo. Evidence that V. cardui overwinters in Britain, where it was once thoughtto bejust a temporary migrant, has been found (Baker 1978). But analysis ofwing length proves that many spring adults must be migrants from southward. In Europe, spring adults average smaller than summer adults (Baker 1978), because smaller adults bred in the 22 J. Res. Lepid. south migrate north in spring. This is also true in central Colo., where forewinglengthinthemonthsfromApril-Oct. averages 28.9, 28.0, 29.6, 32.6, 31.2, 32.1, 31.4 mm. The main change in size is about mid June, mm thusApril-June10adultsaverage28.1 (S.D.2.9,range21-33,N=59), mm whereas June 17-Oct. adults average 31.8 (S.D. 2.0, range 25-37, N=182). This highly significant difference (P<.01) proves that spring Colo,adultsmustbesupplementedbymigrationfromthesouth,because if spring adults were solely overwinterers they would have the same forewinglengthastheadultsthepreviousfall. Butsomeadultsarelarge mm in spring (32-33 in both April and May) and variation in size is greater in spring (larger S.D.), and these large adults still could be overwinterers. ProbablyadultshibernateincentralColoradoeveryyear, asinEurope whereBaker(1978)concludedthat“eachautumnsomeindividualsenter hibernation throughout the breeding range.” The strength of spring versusfallmigrationsisreversedbetweenV. carduiandD.plexippus: V. cardui migrates stronglyin springandweaklyornot at allinfall,while D.plexippusmigratesslowlynorthwardinspringoveraperiodofseveral generations, stronglysouthward duringonegenerationinfall (Cockrell etal. 1993). ThusmigrationinV.carduishouldnotbeviewedasastrictly seasonalmovementlikethatofDanausplexippus]itmustbeviewedalso as apopulationoutbreakinwhichduringafewoutbreakyearsadultsfly to regions (mostly northward in spring) where prospects of rearing offspring are presumed to be better. Mate-locatingbehavior When not migrating, V. cardui males mateJocate (chase others and court)onlylateinthe day, preferablylate afternoon-earlyevening. But during migrations they mate-locate all day: the number ofchases seen each hour ofmigration in 1983 & 1992 was 8:00 (8:00-8:59)-4, 9:00-16, 10:00-18, 11:00-15, 12:00-12, 13:00-9, 14:00-18, 15:00-12,16:00-25,17:00- 14, 18:00-14. No courtships were seen. Many migrations would be towardregionswithalowpopulationofadults,somatingenrouteseems preferable to waiting until arrival and gambling on the presence of a suitable mate there. Flower-feedingbehavior — Adults presumably both migratory and nonmigratory—often fed during the migrations. Yellow and white flowers may be preferred (although this may be an artifact since cultivated purple Buddleja davidii is enormouslypopular). Numberofvisits andflowercolorwere: Taraxacum officinale (yellow) 190 visits, Syringa vulgaris (blue) 45 (white) 1, Prunus cerasus (white) 24, Prunus virginiana melanocarpa (white) 14,Penstemonsecundiflorus(blue)6,Malus sp. crabapple(pink) 7, Astragalus drummondii (white) 4, “pot ofgold” mustard (yellow) 4, Erigeroncompositus(white)3,Thlaspiarvense(yellow) Prunuspissardi 31(l-2):16-23, 1992 23 rosea(white)2,Cryptanthaminima(white)2,Malussylvestris2(white), Erysimum (yellow) 2, Oxytropis lamberti (purple) 1. Literature Cited Arbogast,R., 1966. MigrationofAgraulisvanillaeinFlorida. Fla.Ent.49:141- 145. Baker, R. R., 1968a. Apossible method ofevolution ofthe migratoryhabit in butterflies. Phil. Trans. Roy. Ent. Soc. London (B) 253:309-341. 1968b. SunorientationduringmigrationinsomeBritishbutterflies. Proc. , Royal Ent. Soc. London (A), 43:89-95. , 1969. Theevolutionofthe migratoryhabitinbutterflies. J. Anim. Ecol. —38:703-746. ”, 1978. The evolutionary ecology ofanimal migration. Holmes & Meier Publ. Inc., N. Y., 1012 p. Balciunas,j. &K. Knopf, 1977. Orientation,flightspeeds, andtracksofthree species ofmigratingbutterflies. Fla. Ent. 60:37-39. Cockerell,T.D., 1934. PyrameiscarduionNewYear’sDay. Ent.News45:174. Cockrell,B.,S.Malcolm,&L.Brower,1993. Time,temperature,andlatitudinal constraints on the annual recolonization ofeastern North America by the monarchbutterfly. Pp.233-251In:Biologyandconservationofthemonarch butterfly, S. MalcolmandM. Zaluckieds..No. 38Science Series, Nat. Hist. Mus. Los Angeles County. Emmel,T.C. &R.A.WoBUS, 1966. AsouthwardmigrationofVanessacarduiin late summerand fall, 1965. J. Lepid. Soc. 20:123-124. Johnson,C.G., 1969. Migration&dispersalofinsectsbyflight. Methuen&Co. Ltd., London, 763 p. Walker, T., 1978. Migration and re-migration of butterflies through north peninsularFlorida:quantificationwithmalaisetraps. J.Lepid.Soc.32:178- 190. Williams, C. B., 1970. The migrations ofthe painted lady butterfly, Vanessa cardui (Nymphalidae), with special reference to North America. J. Lepid. Soc. 24: 157-175.