© Entomologica Fennica. 3.XII. 1990

Polymorphism in Noctua pronuba (, ) in southern Norway

Geir E. E. S0li & Trond Andersen

S¢li , G. E. E. & Andersen, T. 1990: Polymorphism in Noctua pronuba (Lepido­ ptera, Noctuidae) in southern Norway. - Entomologica Fenni ca l: 155- 161. Forty-five light trap samples of Noctua pronuba (Linnaeus, 1758), consisting of nearl y 11 000 individuals, from 38 localities in western and southeastern Norway, were scored for three morphs, rufous, ochre and silver, according to the forewing co1ouration. Rufous constituted 37% to 62% of the samples from western Norway, and 50% to 73% of the samples from southeastern Norway. Males constituted 93% of the specimens. In 12 samples, with a sufficient number of females (n>25), no sex difference in the freq uency of rufous was established. The seasonal variation in the frequency of rufous was considered negligible and no annual variation was apparent. No geographic variation in the freq uency of rufous was demonstrated in western Norway. In southeastern Norway one sample differed significantl y from all the others. When samples from adjacent localities were lumped, significant differences in the frequency of rufous were found between three regions in western Norway, and between southeastern Norway and all the regions in western Norway. In southeastern Norway, regression analyses showed a significant increase in th e frequency of rufous towards the east. Geir E. E. S¢1i & Trond Andersen, Museum of Zoology, University of Bergen, Museplass 3, N- 5007 Bergen, Norway

1. Introduction the country (Cook & Sarsam 198 1), whil e geo­ graphic vari ation in morph frequencies was dem­ The , Noctua pronuba onstrated in France (Poitout & Bues 1976). A (Linnaeus, 1758) is polymorphic for fo rewing single sample anal ysed from Finland (Mikkola & colouration. Several phenotypes have been de­ Jalas 1977) showed 67 % rufous, a somewhat scribed (see e. g. Heath & Emmet 1979). The hi gher value than those obtained from the British polymorphism is considered to have a genotypic Isles and France. basis, the morphs being controll ed by several The reason for the lack of geographic variati on all eles at one or more loci (Cook & Sarsam 198 1). in morph frequencies has been sought in the Geographic variation in the distribution of th e mobility of the adults. In the British Isles, Cook & three most common phenotypes, rufous, ochre Sarsam ( 198 1) reported that N. pronuba is hi ghl y and silver, has been studied on the British Isles and mobile, but in France Poi tout & Bues ( 1976) in France. On the British Isles the frequencies considered N. pronuba to be a moderately mobile have been found to be fairly constant throughout species. 156 S¢li & Andersen: Polymorphism in Noctua pronuba • ENTOMOL. FENNICA Vol. 1

• Rufous 0 Ochre 0 Silver

0 100km

Fig. 1. The frequency of the three morphs of Noctua pronuba at 38 localities in southern Norway. The numbers refer to the localities in Table 1.

N. pronuba is common and widely distributed 2. Study area in the coastal areas of southern Norway (Nordstrom et a!. 1969), and large samples are readily ob­ The material compri ses samples from 38 locali­ tained with light traps. In Grenland, southeastern ties, of which 29 are situated in western Norway, Norway, the duration of the flight period was and 9 in southeastern Norway (Fig. 1; Table 1). estimated at 74 days, the median day for males The samples from western Norway have been being 13 August (S01i 1987), and in Sotra, western grouped in five regions: Norway, at about 65 days, with the median day for Sotra: Eight samples (localities 1-6) collected in males on 12 August (Andersen 1982). the northern part of the island of Sotra, where The present study of N. pronuba outlines the the landscape is rather flat and the sampling sexual differences, and temporal, annual and sites are not separated by more than 7 km. geographical variation in the morph frequencies Bergen: Eleven samples (localities 7- 15 ) col­ in southern Norway. As adult mobility reduces lected in Bergen and its sun·oundings, from geographic variation, a higher degree of variation Saudalskleivane in the north to Lundatne in may be expected in western Norway, where geo­ the south. graphic barriers, such as hi gh mountains and fjords, Central Hordaland: Four samples (localities 16-19) are likely to reduce dispersion. In southeastern collected from different localities in the cen­ Norway, where the topography is less interrupted, tral, mountainous regions of northern Horda­ the morphs ought to be more evenl y distributed. land. ENTOMOL. FENNICA Vol. 1 • S¢/i & Andersen: Polymorphism in Noctua pronuba 157

Table 1. Samples of Noctua pronuba from 381ocalities in southern Norway, scored for the three morphs rufous, ochre and silver.

Males Females

Locality Year rut och sit %rut rut och sit % rut

Landro 1978 17 13 0 57 2 2 Austre Loftmyra 1978 34 35 8 44 4 4 1 3 Eidesvag 1975 43 31 8 52 3 1 1976 56 30 5 62 4 6 0 4 Ongeltveit 1976 361 368 80 45 28 35 9 39 5 Austervagen 1977 228 186 38 50 23 24 10 40 1978 38 34 8 48 6 5 0 6 Geitneset 1978 14 10 2 54 0 0 7 Saudalskleivane 1976 99 102 21 45 11 18 5 32 8 Straume 1975 271 239 50 48 15 16 6 41 1976 129 126 20 47 6 8 0 1980 14 18 3 40 0 0 0 9 Ervik 1976 212 192 49 47 22 10 8 55 10 Paradis 1977 26 16 6 54 4 4 11 Ole lrgensvei 1976 176 148 37 49 12 4 12 Kalandseid 1976 101 104 24 44 11 19 2 34 13 Lundatr

Outer Hard anger: Eight samples (localities 20- 25) this whiteness/silverness may be due to air-filled collected along the southern side of the Har­ cavities within the scales. dangerfjord and on the island of Stord, from The variation in morph frequencies was tested Rommetveit in the southwest to Rosendal in for independence using G statistics on RxC tables, the northeast. and applying Williams' correction to the estimated Inner Hardanger: Three samples (localities 26-28) G values. As rufous was by far the most abundant collected along the inner part of the Har­ morph, the tests were focused on the distribution dangerfjord; in addition one sample from of rufous vs. non-rufous. Regression analysis was Haugesund (Joe. 29). used fo r testing the clinal ity of the geographic variation, and t statistics for testing the signifi­ In southeastern Norway the 9 localities range cance of the regression coefficient (see Sokal & from K vernhusvannet (Joe. 30) in the south west to Rohlf 1981). Uller!(ly (loc. 38) in the northeast (Fig. I). One locality, Mostranda (Joe. 37), was situated on the island of Tj0me. 4. Results

4.1. Sex differences 3. Material and methods In none of the 12 samples with a sufficient num­ The material was coll ected with modified Robin­ ber of females did the frequency of rufous differ son traps, fitted with mercury vapour bulbs, dur­ significantly between males and females. ing the period from 197 5 to 1986. The material consisted of 45 samples and 10 623 individuals; only samples with at least 25 males were included. 4.2. Seasonal and annual variation Males constituted 93 % of the specimens. Sex differences were tested only in samples with at As most samples were coll ected during a com­ least 25 females, i. e. in 12 samples; geographical paratively short period, only 11 samples yielded vari ation was studied in males only. data suitable for analysis of seasonal variation. The specimens were sexed, and scored fo r Only one of these, Ole lrgensvei (Joe. 11 ), showed three morphs according to their forewing colour a near-significant difference (?<0.05) in the dis­ (see e.g. Heath & Emmet 1979). The three morphs tribution of rufous throughout the season (Table are characteri zed by 2). The seasonal variation is therefore considered negligible. I) a fairly even browni sh/ reddish colour on the Samples from more than one year are avail ­ forewings; able from five localiti es (Table 3). No signifi cant 2) a more mottled ochreous colour, or annual differences were found. 3) a grayish/silveri sh colour. Each morph has a different expression in the two sexes, the females always being paler, and for the 4.3. Geographic variation ochreous and grayish morphs also less mottled than the males. The nomenclature used by Cook & Samples within each of the fiv e regions in western Sarsam ( 198 1) is adopted, the darkest morph Norway and within southeastern Norway were being named rufous, the paler morph ochre, and tested for independence in the frequency of rufous. the third morph silver. Silver was identified by the In western Norway significant or near-significant presence of scales, whi ch give the its silver­ differences were found in two instances only: ish appearance. These scales seem to be poorly between Eidesvag (Joe. 3) and Ongeltveit (Joe. 4) pigmented, and are most di stinct along the costal on Sotra (P

between homogeneous samples, due to chance higher than in all five regions in western Norway. alone. However, as th e test on material from Sotra The frequency of rufous in Outer Hardanger was yielded a highly significant difference, it must be significa ntly hi gher than in Bergen and on assumed that differences do reall y ex ist in the Sotra, on the northern side of the Hardanger­ frequency of rufous on this island. In spite of this, fjord . Likewise, the frequency of rufous in Inner we have lumped the samples from Sotra and Hardanger was fo un d to differ significantly from li kewise all the samples within the other fo ur that on the island of Sotra. These two regions are western Norwegian regions. separated by more than 80 km. Differences in In southeastem Norway the frequency of ru­ morph frequencies were not demonstrated be­ fous in the sample from Mostranda (Joe. 37) was tween Bergen and the two regions in the central found to be significantly to highly significantly parts of western Norway, Inner Hardanger and different from all other samples taken in the re­ Central Hordaland, or between the latter two gion. Hence, this sample has been excluded from regions. further tests. None of the other samples differ To test clinal variation, the frequencies were signi ficantly from each other, and they have been arranged according to the positi on of th e locali ty lumped. along a west-east gradient, separately for samples Differences in the freq uencies of rufous were from western Norway and southeastern Norway. tested between the five regions in western Nor­ In western Norway the frequency of rufous ranged way, and between these regions and southeastern from 37 to 62%. A sli ght decline in the frequency Norway. The frequency of rufous in southeastern of rufous towards the east, i.e. inland, was ob­ Norway was significantly to hi ghl y significantly served; 12% of this colour variation can be ex-

Table 2. Seasonal va ri ation in th e frequency of the rufous morph in males of Noctua pronuba in 11 localities in southern Norway.

-30 July 1- 15 Aug . 16-31 Aug . 1-30 Sept Sampl e % rut Sample % rut Sample % rut Sample % rut

Ongeltveit 47 55 343 45 250 40 169 48 Saudalskleivane 6 40 68 142 45 14 Straume (1975) 91 47 168 49 276 49 25 40 Straume (1976) 64 45 118 45 58 52 35 49 Ervik 66 50 196 45 165 47 26 50 Ole lrgensvei 51 53 65 52 226 45 19 Kalandseid 6 27 37 176 46 20 Lii 32 41 84 56 52 48 31 36 Rosendal (1977) 7 329 41 100 50 2 Hasselasen (1980) 136 54 173 47 123 59 Gravastranda 63 62 573 56 259 54

Table 3. Annual variation in the frequency of the rufous morph in males of Noctua pronuba in five localities in southern Norway.

1975 1976 1977 1978 1980 1981 Sample %rut Sample %rut Sample %rut Sample %rut Sample %rut Sample % ruf

Eidesvag 82 52 91 62 Austervagen 452 50 80 48 Straume 560 49 275 47 35 40 Rosendal 153 41 529 44 438 43 Hasselasen 432 53 134 51 160 S¢li & Andersen: Polymmphism in Noctua pronuba • ENTOMOL. FENNICA Vol. 1

plained by the gradient (r2 = 0.12) although the cause rapid spreading of genes in populations and, coefficient is near-significantly different from hence, to reduce regional variation (e.g. Hartl

zero (t33=-2. 131; P< 0.05). 1980). The importance of gene flow in the main­ In southeastern Norway the frequency of ru­ tenance of this cline is emphasized by the high fous ranged from 50 to 73% and showed an in­ frequency of rufous obtained in the more isolated crease towards the east, 44% ofthe variation being locality on the island of Tj0me. explained by the gradient (r2=0.44, t =2.513; Significant differences were obtained between 8 P< 0.05). If we leave out Mostranda (Joe. 37), whose three regions in western Norway, all separated by frequency of 73% differs significantly from the mountains and wide fjords. Likewise, significant others, the analysis yields an r2 value of 0.72 differences were found between southeastern

(t7=4.225; P< 0.01). Norway and all the regions in western Norway. These results are in accordance with observations from France (Poitout & Bues 1976), indicating 5. Discussion reduced mobility, most probably caused by dis­ tance and geographical barriers. According to As the seasonal and annual variation in the morph Poi tout & Bues ( 1976), N. pronuba is intermedi­ frequencies was negligible, the frequencies could ate in behaviour between migrating and non­ be compared between samples obtained in differ­ migrating species. ent years and at different times of the season. In general, polymorphism is maintained by The frequency of rufous was found to vary differences in the selection pressures on the significantly between different regions in South morphs, in many instances differences in preda­ Norway. This is not in accordance with results tion pressures. The selective forces acting on N. obtained from the British Isles. On the British pronuba are still obscure. However, Cook & Isles, 80 samples (9147 individuals) of N. pro­ Sarsam (1981) suggested that the polymorphism nuba, collected from Scotland to Wales, were in N. pronuba is the result of opposing selection scored, and the frequencies of the three morphs pressures in the two sexes, as they obtained dif­ were found to be fairly constant throughout the ferent fitness values for rufous males and females. country (Cook & Sarsam 1981). In France, how­ In the present study no difference was found ever, where more than 57 000 individuals from between the sexes. The reason for this may partly four geographical provinces in the southwestern be sought in the restricted number of samples and southeastern parts of the country were scored, available for comparison. the distribution of the three morphs varied signifi­ Breeding experiments on the British Isles and cantly between the geographical provinces (Poitout in France have yielded contradictory results con­ & Bues 1976). The frequencies of rufous were cerning the inheritance of the phenotypes in N. somewhat higher in southern Norway than those pronuba. After performing breeding experiments reported from the British Isles and France, while with 62 couples, Poitout & Bues (1976) reported the frequencies of silver were similar. In Finland, however, the frequency of rufous was relatively high (67% ), but this result is based on one sample Table 4. The range in frequencies(%) of the rufous and only, consisting of 238 specimens (Mikkola & silver morphs of Noctua pronuba in western Norway, Jalas 1977) (Table 4). southeastern Norway, England, France and Finland. The results from Norway suggest a relatively Reg ion rufous silver high degree of mobility inN. pronuba. In south­ eastern Norway, where the sampling sites were West Norway 37-62 0-17 not separated by major geographical barriers, the Southeast Norway 50-73 8-13 samples are most probably taken from a contin­ England • 21 - 41 0 -8 uous population. The observed cline, with an in­ France b 39 - 49 9 - 14 Finland c 67 13 crease in the frequency of rufous toward the east, is probably smoothed by gene flow caused by the •Cook & Sarsam 1981 ; bPoitout & Bues 1976; cMikkola mobility of the ad ults. Migration is known to & Jalas 1977. ENTOMOL. FENNICA Vol. 1 • S¢ /i & Andersen: Polym01phism in Noctua pronuba 161 that rufous is dominant over silver, which again is Hartl, D. L. 1980: Principles of populati on genetics. - dominant over ochre. On the other hand, Cook & Sinauer Assoc. Sunderland. 488 pp. Sarsam (1981) suggested that rufous is dominant Heath , J. & Emmet, A. M. 1979: The and butterflies of Great Britain and Ireland. 9. - Curven Books, over ochre, and that silver is recessive to non­ Oxford. 288 pp. silver, situated at another locus. These conclu­ Mikkola, K. & Jalas, I. 1977: Suomen perhoset. Yi:ikki:iset sions are based upon the progeny of a seri es of 14 I. - Otava, Keuruu. 256 pp. crosses, of which the parentage was unknown in Nordsti:i m, F. , Kaaber, S. , Opheim, M. & Sotavalta, 0. 1969: eight instances. Further breeding experiments are De fennoskandiska och danska nattflynas utbredning necessary to clarify the inheritance of the different (Noctuidae). - CWK. Gleerup, Lund. 160 pp + 403 maps. morphs in N. pronuba. Poitout, S & Bues, R. 1976: L' incidence des mi grations d 'ad ultes sur le degre et sur Ia variabilite structurale de Acknowledgemems. For contributing materi al to thi s study I ' heterogeneite genetiquedans les populations nature lies we are indebted to Per Andersen, Arne A. Grimenes, Sigmund de lepidopteres Noctuidae.- Ann. Zooi. -Ecol. Anim. Hansen, Tom Kleppaker, Wigard Nil sen, Gustav Pedersen, 8:69-8 1. Torstein So lh s<~y and Karin Swane. Robinson, R. 1973 : Lepidoptera genetics. - Pergamon Press, Oxford. 687 pp. Sokal, R. R. & Rohlf, F. J. 198 1: Biometrics. - W. H. Freeman, New York. 860 pp. References Ss<~li , G. E. E. 1987: Li ght trap catches of Bombycoidea, Sphingoidea, Notodontoidea and (Lepi­ Andersen, T. 1982: Some studies on Macrolepidoptera in doptera) in two forest communities in Southern Nor­ coastal heath land habitats in western Norway. - Faun a way. - Cand. scient. thesis, Uni versity of Bergen. 39 Norv. Ser. B. 29:85-104. pp. Cook, L. M. & Sarsam, V. 198 1: Polymorphi sm in the moth Noctua pronuba (L.). - Heredit y 46:443-447. Received 3. 111.1 989