Ann. Zool. Fennici 42: 141–160 ISSN 0003-455X Helsinki 26 April 2005 © Finnish Zoological and Botanical Publishing Board 2005

Phylogeny and evolution of parthenogenesis in Finnish bagworm (: Psychidae: Naryciinae) based on mtDNA-markers

Alessandro Grapputo*, Tomi Kumpulainen & Johanna Mappes

Department of Biological and Environmental Science, P.O. Box 35, FI-40014 University of Jyväskylä, Finland (*e-mail: [email protected].fi)

Received 25 Oct. 2004, revised version received 8 Dec. 2004, accepted 6 Dec. 2004

Grapputo, A., Kumpulainen, T. & Mappes, J. 2005: Phylogeny and the evolution of parthenogene- sis in Finnish species (Lepidoptera: Psychidae: Naryciinae) based on mtDNA- markers. — Ann. Zool. Fennici 42: 141–160.

We investigated species diversity and evolution of parthenogenesis among bagworm moth species of Dahlica and using mitochondrial DNA sequencing. Par- thenogenesis is rare among Lepidoptera other than Psychidae. Genera Dahlica and Siederia form a confusing group with controversial species boundaries and widely overlapping morphological features that make species determination difficult. We evaluated the reliability of species determination based on wing scale morphology by comparing it with a phylogenetic tree obtained using mtDNA. Species determination based on morphological characteristics did not correspond to species determination based on mtDNA markers. On the basis of the molecular phylogeny, the status of these two genera is questionable. Our results indicate that parthenogenetic D. fennicella, D. triquetrella and D. lichenella evolved independently from different sexual ancestors suggesting that asexual reproduction is favoured in this group.

Introduction organelles, or essential genes. Among Lepidop- tera, asexual reproduction is extremely rare and, Sexual reproduction is the dominant strategy in in most known examples, parthenogenesis is a higher (and plants) making the para- secondary reproductive strategy (Suomalainen dox of sex one of the most fundamental ques- 1962, Bell 1982). Among Naryciinae (Lepidop- tions in evolutionary research. One hypothesis tera: Psychidae) however, several parthenoge- is that females of sexually selected lineages netic species are known. These species can be are in a male-dependent trap and that the trap classified as strictly asexual lineages since no alone should often be enough to maintain sex males are known among them. Interestingly, both since females need males for normal reproduc- sexual and asexual species often co-exist and tion (West-Eberhard 2003). Depending on the there are no reliable morphological or ecological species, females may rely upon males to stimu- cues to separate parthenogenetically and sexu- late ovulation, to initiate embryonic develop- ally reproducing females (Kumpulainen et al. ment, or to provide essential genetic transcripts, 2004, Kumpulainen 2004). Thus, psychid 142 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42 offer an attractive opportunity to investigate the scale morphology of female pupa and abdominal evolution of parthenogenesis. If parthenogenesis spines of adult females. Also, certain differences has evolved several times in this group, we can in the structures of the front legs of adult males assume that some ecological, physiological or have been given as grounds for separating the environmental factors may favor parthenogen- species (Hättenschwiler 1997). However, there is esis in this group. too much variation among the species in the head Species determination, status and phyloge- scale morphology of the pupa for clear species netic relationships of many species in this sub- determination and all other feature differences are are, however, still controversial (Suoma- also insufficient to reliably determine all species lainen 1970, Lokki et al. 1975, Suomalainen (Hättenschwiler 1997). Genital indices have been 1980, Hermann 1994, Hättenschwiler 1997, described for most of the species but these also Kullberg et al. 2002). The group is characterised overlap considerably (Hättenschwiler 1997). So by high endemism in mountain ranges in south- far, only the shape of wing scales from the central ern Europe and central Asia. There is a great area of the tip of male fore wings (Suomalainen number of morphologically similar species with 1980) has been shown to contain sufficient vari- several problematic species pairs and still unre- ation between adult specimens of most species, solved relationships between parthenogenetic even though the use of wing scales for species and sexual forms, which have caused confu- determination has led to controversial interpreta- sion in the nomenclature and ecological studies tion of species boundaries (Sauter 1956, Suoma- involving well known taxa as well. lainen 1980, Hättenschwiler 1997). Moreover, In Europe, the tribus Dahlicini (Naryciinae) this feature is inadequate for species determina- consist of 49 species in six genera (Sauter & tion of different parthenogenetic species. Hättenschwiler 1991). To date, 22 species of Current phylogenetic methods allow for test- bagworm moths (Dahlicini) have been recorded ing the validity of morphological and ecological in Finland. Of these, seven species are included features in constructing phylogenetic trees for a in two genera — Dahlica Enderlein and Sied- given taxon. The well-known problems of distin- eria Meier — which formerly were grouped in guishing between homologous and homoplasi- the Solenobia Duponchel) (Suomalainen ous features have made it difficult to interpret 1980). All Dahlica and Siederia moths are rela- the true phylogenetic relations of many systems tively small-sized (female body length 3–6 mm) (Brooks & McLennan 1991, Avise 1994). The and their larval cases are particularly similar in use of genetic markers together with ecologi- size. Most species have a wide range of similari- cal and morphological features, however, has ties in both morphological and ecological fea- revealed new information on the phylogeny of tures. Dahlica and Siederia are most abundant many controversial groups of animals, such as near forest edges and other semi-open parts of pandas (O’Brien 1987), flightless birds (Had- their habitats, where the microclimate is favour- drath & Baker 2001) and many groups able and larval food plants (several moss genera) (Brower 1997, Kruse & Sperling 2002). Among are abundant (Kumpulainen 2004). All species , and particularly among Microlepidop- have a very short adult phase, only 3–6 days. tera, many controversial groups of species still The longest stage of their life cycle, from one to remain (Hättenschwiler 1997, Kruse & Sperling two years, consists of five phases of larval devel- 2002). The phylogeny of such groups as well as opment (Hättenschwiler 1985, 1997). Larvae of the status of many species is subject to change Dahlica and Siederia moths usually live and feed when new studies applying more precise and freely on mosses on the forest floor, on stones or diverse methods are published. other similar semi-open habitat patches, often Due to the apparent lack of reliable mor- close to tree trunks. phological features for species determination in Species determination among Dahlica and the genera Dahlica and Siederia we conducted Siederia is very difficult due to the lack of distinct a phylogenetic analysis based on partial DNA morphological or ecological features. Current sequences of mitochondrial Cytochrome Oxidase attempts to classify the species are based on head genes and small ribosomal subunit gene (12S) to ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 143 examine species diversity of solenobid bagworm throughout most temperate areas of Europe, from moths. We were particularly interested in com- Spain to Great Britain and Ireland in the west and paring the classification based on wing scale to Finland in the north. In southern and central morphology to the phylogeny based on mtDNA Finland, D. triquetrella is not rare, but can nor- markers. We were also interested in studying mally be found only in small numbers (Suoma- the evolution of parthenogenesis in this group. lainen 1980). This species has both sexually and MtDNA analysis provided a test of species deter- parthenogenetically reproducing forms of which mination independent of wing scale morphology, only the parthenogenetic, tetraploid form occurs reproductive strategy and ecological data. in Finland (Suomalainen 1980). Lokki et al. (1975) found two genotypically different forms (an eastern and a western form) in Finland using Materials and methods electrophoresis. D. triquetrella occurs mainly on the edges of warmer forest habitats, possibly For the phylogenetic study we included all preferring sandy and often south-facing habitats, seven species previously described in Finland. sometimes with a human influence (Arnscheid To obtain a more accurate phylogenetic analy- 1985, Hättenschwiler 1997). D. triquetrella is sis of Naryciinae, we also included samples slightly larger than other Dahlica or Siederia of Siederia rupicolella, Dahlica triquetrella species and as larvae it should be relatively easy and D. lichenella from the Alps. These species to distinguish by its larger size and also by the also occur in Finland. Moreover we added sev- triangulate larval case (Suomalainen 1980, Hät- eral endemic species (D. sauteri, D. wockei, D. tenschwiler 1997). vaudella and D. generosensis) from the Alps. From Sweden we included samples of D. triquet- rella, S. rupicolella and S. listerella and finally Dahlica lichenella (Linnaeus, 1761) samples of D. triquetrella and D. lichenella from western Canada, which were introduced D. lichenella is sometimes considered to contain to Canada in the 1940s (Arnscheid 1985, Hät- a sexual form D. lichenella f. fumosella and a tenschwiler 1985, 1997, Hermann 1994). Sample parthenogenetic form D. lichenella f. lichenella locations and the number of samples per loca- (Sauter 1956, Arnscheid 1985, Hermann 1994, tion are given in Table 1. We collected bagworm Hättenschwiler 1997). Some authors (Suoma- moth larvae by placing tape traps on the trunks lainen 1980, Kullberg et al. 2002), however, of trees, mostly silver birch (Betula pendula consider D. lichenella to be purely partheno- Roth) and Norway spruce (Picea abies), and in genetic and sexual D. lichenella f. fumosella to one population (Rauma) by searching for larvae belong to a separate species Dahlica fumosella in the shadow and underside of larger stones. (Heinemann, 1870) or Dahlica lazuri (Clerck, In early spring, larvae of Siederia and Dahlica 1759). Both “forms” of this species are widely species climb up tree trunks and other similar distributed in Europe, from Spain to Finland. The places for pupation, as snow usually still covers sexual form is relatively common in the Nordic the ground at this time of a year. Climbing larvae countries (Suomalainen 1980, Hättenschwiler stick to tape traps and are easy to collect. After 1997). D. lichenella is often found in rocky habi- collection, we reared the larvae until adulthood tats with mostly scarce vegetation (Suomalainen in the laboratory. A brief description of the spe- 1980). In Finland, D. lichenella is a strictly par- cies used in this study is given below. thenogenetic and relatively rare species occur- ring only in the coastal areas of southern and south-western parts of the country (Suomalainen Study species 1980). D. lichenella is quite difficult to dis- tinguish from all other species of Dahlica and Dahlica triquetrella (Hübner, 1813) Siederia (except D. triquetrella) (Suomalainen 1980, Hättenschwiler 1997). The larval case is D. triquetrella is widely distributed and occurs not triangulate (see D. triquetrella) and it is often 144 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

fumosella (1) (1) f. (1) (2) S. listerella D. charlottae / D. lichenella D. to mtDNA S. rupicolella S. listerella S. rupicolella D. fennicella S. listerella S. listerella D. vaudella D. sauteri D. wockei D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. lichenella D. lichenella D. lichenella D. lichenella D. fennicella D. fennicella D. charlottae D. lazuri D. charlottae D. generosensis /

2003 Male Sexual 2001 Female Parthenogenetic 2004 2000 Male 2004 Female 2003 Sexual 2003 Male 2003 Sexual Male Sexual Male 2003 Male Sexual 2003 Sexual Female 2001 Sexual 2004 Female 2004 Parthenogenetic Female 2004 Female Parthenogenetic Female Parthenogenetic Female Parthenogenetic 2002 Parthenogenetic Parthenogenetic 2004 Female 2000 Female Parthenogenetic 2004 Parthenogenetic Male Male Sexual Sexual 2003 Male Sexual 2000/1 Female Parthenogenetic 2000/1 Female/Male Sexual 2000/1 Male Sexual Year of Sex Reproduction Species according collection

Kroisbach, Austria Martigny, Switzerland Jyväskylä, Finland Stockholm, Sweden 1997 Jyväskylä, Finland Stockholm, Sweden Female St. Cergue, Switzerland Steyr, Austria Sexual Dürnstain, Austria 1994 Linz, Austria Stanserhorn, Switzerland Female Kroisbach, Austria 1994 Sexual Steyr, Austria Jyväskylä, Finland Female Åminsby, Finland Turku, Finland Parthenogenetic Öland, Sweden Sorgan Banhof, Switzerland Agassis BC, Canada 2003 Yvonand, Switzerland Rauma, Finland Female 1991 Åminsby, Finland 1983 Parthenogenetic Jyväskylä, Finland Female Female Stockholm, Sweden Parthenogenetic Parthenogenetic Locality

DTp4, DTp5, DTp7, DTp7a, DTp7b, DTp8, DTp9

4 SRp1, SRp2; SRp2a, SRp2b Jyväskylä, Finland 2 DFu1, DFu2 1 SR1 3 SR2, SR2a, SR2b 2 SR3, SR4 1 SL3 2 SL1, SL2 1 DV 2 DS1, DS2 3 DW1, DW2, DW3 3 DTs1, DTs2, DTs3 1 DTp1 3 DTp2, DTp2a, DTp3 7 1 DTp6 2 Uster, Switzerland DTp6a, DTp10 2 DTp6c, DTp12 1 DTp6b 1 DTp11 2 1994–2003 DLi4, DLi4a 2 Female DLi1, DLi2 1 DLi3 2 DLi4b, DLi5 4 Parthenogenetic DF1, DF1a, DF2, DF3 2 DF4, DF4a 4 DC1, DC1a, DC1b, DC2 3 DL1, DL2, DL3 Jyväskylä, Finland 2 Jyväskylä, Finland DL4, DL4a 4 DG1, DG2, DG3, DG4 Altanssee, Austria N Specimen code

fumosella asex f. Samples, number of individuals, specimen codes, location, sex, mode of reproduction, and mitochondrial classification of bagworm moths used in this study.

able 1. Specimen codes were assigned on the basis of the morphological classification. Individuals originally classified as the same species and sharing the same haplotype are indicated by the same code followed by a different letter (e.g. SRp2, SRp2a and b represent three different specimens of the same species sharing the same haplotype

T SRp2). Species according to morphology S. rupicolella S. rupicolella S. rupicolella S. rupicolella S. listerella S. listerella D. vaudella D. sauteri D. wockei D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. triquetrella D. lichenella D. lichenella D. lichenella D. lichenella D. fennicella D. fennicella D. charlottae D. lazuri D. lazuri D. lichenella D. generosensis

ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 145 more dark-coloured than that of D. fennicella or other Dahlica species (Suomalainen 1980).

Dahlica fennicella (Suomalainen, 1980)

This species was previously considered to be a form of D. lichenella, but was described as a separate species by Suomalainen (1980). How- Fig. 1. The male wing scale morphology from the ever, it is still possible that tetraploid, partheno- central area of the tip of the front wing (redrawn from genetic D. fennicella could be an asexual form of Sauter (1956) and Suomalainen (1980)). a known sexual Dahlica species (Suomalainen 1980). In Finland, D. fennicella is reported to be rare with only a few known populations in south- Dahlica charlottae (Meier, 1957) ern Finland and in the southern parts of central Finland (Suomalainen 1980). It is not known to D. charlottae is distributed over the eastern parts occur in other parts of Finland or in other coun- of northern and central Europe (Hermann 1994, tries. D. fennicella is a small species, which is Hättenschwiler 1997). In Finland, this sexual difficult to separate from other Dahlica species species is fairly common all over the country, and, as larvae, also from Siederia species. The but is usually not very abundant (Suomalainen shape and size of the larval case can be very 1980). According to Suomalainen (1980) and similar to those of D. lazuri and S. rupicolella Hättenschwiler (1997), this species can be sepa- (Sauter 1954) (Suomalainen 1980). rated from other Dahlica and Siederia species by the narrow shape of wing scales and by the number of spines per wing scale (Fig. 1), and Dahlica lazuri (Clerck, 1759) sometimes by the relatively triangulate shape of the larval case also. D. charlottae occurs in The status of this species is confusing (Arnscheid different forest habitats but can most commonly 1985) as D. lazuri is possibly a synonym for D. be found in habitats with pine (Pinus sylvestris), fumosella (Heinemann 1870), which also occurs swamps and sandy ridges (Suomalainen 1980, in central Europe where some authors consider it Hermann 1994). to be a sexual form (f. fumosella) of D. lichenella (Hermann 1994, Hättenschwiler 1997). However, other authors treat it as an independent sexual Siederia listerella (Linnaeus, 1758) species (Suomalainen 1980, Kullberg et al. 2002). This species is widespread in central Europe, but This sexual species has also been known as S. seems to be more common in the northern than pineti (Zeller, 1852) and some authors have also in the southern Europe (Hermann 1994, Hät- used the name S. cembrella (Linnaeus, 1761). tenschwiler 1997). In Finland, D. lazuri is a very S. listerella is widely distributed in Europe common sexual species occurring throughout the from England to north-western Russia (Her- whole country (Suomalainen 1980). This species mann 1994), especially in the northern parts of is morphologically very similar to D. charlot- the continent (Hättenschwiler 1997). In Finland, tae but, according to Hättenschwiler (1997) and S. listerella is quite rare, occurring mainly in Suomalainen (1980), it can be separated by the the southern parts of the country (Suomalainen shape of male wing scales (see Fig. 1). The larval 1980). This species is morphologically very case of D. lazuri is often very similar to D. fen- similar to S. rupicolella but, according to Hät- nicella and S. rupicolella. This species can be tenschwiler (1997) and Suomalainen (1980), can found in many different forest habitats, but may be distinguished by the wider shape of the male prefer sites with more than average air humidity wing scales (Fig. 1). This species occurs in vari- (Hättenschwiler 1997). ous forest habitats, often with Norway spruce 146 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

(Picea abies), and it may prefer habitats with a very wide and comparable to those of Siederia warmer microclimate (Hermann 1994). rupicolella (Hättenschiler 1997) but the front wings are narrow and have a slightly round wing tip. Wings are also dark coloured with irregular Siederia rupicolella (Sauter, 1954) cream-coloured spots. The species occurs only on rocks (Hättenschwiler 1997). This species is considered to be sexual (Suoma- lainen 1980, Hättenschwiler 1997) although many collectors and taxonomists have assumed Dahlica sauteri (Hättenschwiler, 1977) it also has a parthenogenetic form in northern Europe. S. rupicolella is known to occur in D. sauteri is a relatively widely distributed spe- mountainous regions (in the Alps in Switzer- cies occurring in northern Switzerland, Ger- land, Austria, and Germany) and in Scandina- many and most of western and central Europe via (Hermann 1994, Hättenschwiler 1997). It up to the North Sea (Arnscheid 1985, Hät- is considered rare in central Europe (Hermann tenschwiler 1997, http://www.lepidoptera.bai. 1994, Hättenschwiler 1997), but in Finland, pl/start.php?lang=GB). The form of the male S. rupicolella is reported to be quite common wings is variable but generally narrower at the and has spread almost throughout the country base than in other species (Hättenschwiler 1997, (Suomalainen 1980). It occurs in many different Arnscheid 1985). Wings are grey with very small forest habitats, often with spruce and is often white spots. Wing scales are also variable but most abundant at forest edges, although it is not mostly of medium width with two or more known to prefer any certain forest type (Suoma- spines (Arncheid 1985, Hättenschwiler 1997). D. lainen 1980, Hermann 1994), except in northern sauteri inhabits sunny forest edges or sparse for- Finland, where it can be found in pine swamps ests (Herrmann 1994) and is often found in the (J. Itämies pers. obs.). Males of this species have same habitats as those of D. triquetrella (Hät- narrower wing scales than S. listerella (see Fig. tenschwiler 1997). 1) (Suomalainen 1980, Hättenschwiler 1997). The larval case is often difficult to distinguish from that of S. listerella, D. lazuri or D. fen- Dahlica wockei (Heinemann, 1870) nicella. This species occurs in central Europe and has been recorded in Germany, Poland and possibly Dahlica generosensis (Sauter, 1954) Austria (Arnscheid 1985, Herrmann 1994, http:// www.lepidoptera.bai.pl/start.php?lang=GB). This sexual species occurs in Austria, Italy and Male front wings have ground colouration, in southernmost Switzerland (Arnscheid 1985, slightly yellowish and wide light coloured spots Hättenschwiler 1997). Male wing scales resem- (see Arnscheid 1985). The species lives in rela- ble those of D. lazuri (Hättenschwiler 1997) tively dry deciduous forests and feeds on lichens but they can be distinguished because wings are (Arnscheid 1985). It is the first Dahlica species wider and grey with slightly yellowish (Hätten- to occur in early spring (Herrmann 1994). scwhiler 1997) or brownish (Arnscheid 1985) shading. The species is mostly found in open grassy mountain habitats with rocks, stones and Species identification stone walls (Hättenschwiler 1997). We primarily used males for identification of sexual species. However, for S. rupicolella, we Dahlica vaudella (Hättenschwiler, 1990) also used sexual females. We classified males into different species using the wing scales (Fig. D. vaudella is known to occur only on the Swiss 1). For sexual females, we used mate choice and part of the Jura mountains. Wing scales are acceptance to reliably determine the species. ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 147

S. rupicolella females were paired with males methods described above. The outgroup species, (one at a time) starting with those males whose Psyche norvegica (Heylaerts, 1882), Diplodoma larval case was morphologically most similar laichartingella (Goeze, 1783) and to their own. We observed the reactions of the borealis (Wocke, 1862), were determined by males towards offered females. When presented the shape, material and size of their larval cases. to females of their own species, males reacted All outgroup species are very easy to determine, immediately to the female pheromone and tried particularly when compared with Dahlica and to initiate copulation. Reactions towards females Siederia species (Hättenschwiler 1997). of other species vary from avoidance to no We assigned a specimen code to each sample reaction at all. We were able to confirm success- (Table 1) based on the morphological classifica- ful fertilisation when fertilised eggs hatched. In tion. Following the sequence analysis, speci- our analysis we used only sexual females that mens originally classified as the same species had successfully copulated with males that had and sharing the same haplotypes were assigned already been identified according to their wing to the same code followed by a different letter scale morphology. We determined parthenoge- (a, b, c etc.). netic females on the basis of their behaviour. After hatching, all females climb on top of their own larval case. Parthenogenetic females soon Laboratory procedures start to lay eggs inside their larval case whereas sexual females stay on their larval case and begin Samples were frozen at –20 °C until DNA to release pheromones to attract males. Once extraction. Additional samples from Switzerland parthenogenesis was confirmed, we also used and Canada were preserved in 70% ethanol. We the form of the larval case and the distribution isolated total genomic DNA from the entire indi- area as identification methods. D. triquetrella vidual with a solution of 5% Chelex chelating has a large larval case with an easily recognis- resin (Pearce et al. 1997). A fragment of about able triangular shape. It was more problematic 600 bp spanning from the end of the COI to the to identify the small sized, non-triangular larval COII was amplified using the primers S2792 case of parthenogenetic Dahlica. We assigned (Brower 1994) and C2-N-3389 (Simon et al. other Dahlica parthenogenetic forms to the dif- 1994). Moreover, a fragment of about 400 bp ferent species according to the form and material of the 12S rRNA gene was amplified using the forming the sack and their geographic location. primers SR-J-14233 and SR-N-14588 (Simon et All specimens collected from the south-western al. 1994). We carried out the amplifications in a coast of Finland were assigned to D. lichenella. total volume of 25 µl, with the use of 10 mM of

This species is not known to occur in central Tris-HCl, 1.5 mM of MgCl2, 5 pmoles of each Finland. Parthenogenetic specimens collected in primer, 200 µM of each dNTPs, 1 unit of Taq central Finland were considered asexual S. rupi- polymerase (Boehringer Mannheim) and 20–50 colella according to collectors’ common belief. ng of DNA. The forward and reverse primers Moreover, D. fennicella, the other possible par- labelled with a fluorescent dye (IDR-800 and thenogenetic species, is considered very rare and IDR-700 respectively; Li-Cor Inc.) were used to occur only in southern Finland (Suomalainen with the Thermo Sequenase DYEnamic Direct 1980). Finally, specimens collected in Åminsby, cycle sequencing kit (Amersham) as described the location where Suomalainen (1980) first in the technical bulletin #71 (Li-Cor Inc.). We described D. fennicella, were classified as D. then loaded the sequencing reactions in a Li-Cor fennicella. Endemic species from the Alps (Swit- DNA bidirectional sequencer 4200 and ran them zerland and Austria) were obtained and classified overnight. COI and COII sequences translated by P. Hättenschwiler and E. Hauser. Species correctly. We, therefore, feel confident that the from Sweden, Canada and the Alps that are also sequences were from the mitochondrial DNA present in Finland were classified by P. Hättern- genome. The sequences reported here have been schwiler, E. Hauser and G. Palmqvist and were deposited in GenBank (data accession numbers: confirmed by one of the authors (TK) with the AY449388–AY449457). 148 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

Sequence analyses available at: http://www.agapow.net/software/ mac5/) which allow gaps to be treated as 5th We aligned the sequences with ClustalX (Thomp- state and MrBayes 3.0 (Huelsenbeck & Ronquist son et al. 1997) using the “full multi-alignment” 2001) where gaps are treated as missing data. option and dynamic programming (slow-accu- The effects of different models of nucleotide rate pairwise alignment method). Gap opening substitutions were explored. In MAC5 both the penalty was 15 and gap extension penalty 6.66 Jukes and Cantor (Jukes & Cantor 1969) (JC) in both pairwise and multiple alignment param- and the F84 model of substitution (Kishino & eters. In the latter, delay divergent sequence was Hasegawa 1989), which is equivalent to the 30% and DNA transition weight 0.50. The align- HKY model, were applied, while in MrBayes the ment was checked visually. A gap of 16 bases JC, HKY and the general time reversible (GTR) followed by a T and another gap of 6 bases were (Lanave et al. 1984) models with g were applied. introduced by the alignment program between An analysis setting of one million generations, the COI and tRNAleu genes of the TB1, TB2 and sampling every 100 trees and with a burn-in DLa1 and DLa2 sequences. However, the T was of 200 000 was applied for both programs. In clearly the first T of the tRNAleu, thus we manu- MAC5 the weighting option was set either to off ally modified the alignment creating a unique (all gaps treated equally) or on (for down weight- gap of 22 bases in these four sequences. ing gap positions in the alignment. The weight is Test of stationary across sequence was per- the reciprocal of the average gap length at that formed with the program Treefinder (Jobb et column of the alignment; if there are no gaps, the al. 2004) on all sites. The sequence of Psyche weight is one). Three independent runs were used norvegica did not pass the stationary test and so to check that the chains had converged properly. it was excluded from the analyses. Phylogenetic reconstructions were obtained with Bayesian inference. The method was chosen Results because Bayesian inference allows the imple- mentation of specific models of DNA substitu- We used a concatenated sequence of COI (188 tion and it also performs exceptionally well in bp), intergenic sequence (27 bp), tRNAleu (68 supporting correct grouping as compared with bp), COII (295 bp) and 12S (364 bp) to deter- traditional ML and MP methods (Alfaro et al. mine the diversity of bagworm moth species in 2003). However, we need to note that the full Finland. A total of 271 (28.8%) of the sites were meaning of Bayesian support values is still under variable, of which 235 (24.9%) were parsimony- debate (Simmons et al. 2004). informative. The variable sites in the different To select the model of DNA substitution that genes were: 83 (44%) in the COI fragment, 9 best fitted the data, we employed the hierarchi- (13%) in the tRNAleu, 79 (27%) in the COII seg- cal likelihood ratio test (Huelsenbeck & Cran- ment and 91 (25%) in the 12S fragment. The TS: dall 1997) implemented in MODELTEST 3.06 TV ratio was estimated to be 2.7 for the entire (Posada & Crandall 1998). For the combined sequence. Corrected sequence divergence values data set of COI, tRNAleu, COII and 12S, the calculated with the HKY + G model are shown in Hasegawa, Kishino and Yano model of nucle- Table 2. The gamma shape parameter was 0.21. otide substitution, with the gamma distribution All D. triquetrella sequences were characterised parameter to correct for the rate variation among by an insertion of three bases (CTA or CCA in sites (HKY + g) (Hasegawa et al. 1985), was the DTp1) in the intergenic sequence between COI most appropriate based on the hierarchical likeli- and tRNAleu. hood ratio test. Under the Akaike information The trees produced by the Bayesian meth- criterion (Akaike 1974), the K81uf + g model ods of estimating the phylogenetic relationships of nucleotide substitution (Tamura & Nei 1993) among bagworm moths with JC and F84 models was the best model. of nucleotide substitutions and with a gap treated Bayesian analyses were performed with as 5th state (either gap no-weighted or weighted MAC5 (written by McGuire and Agapow and in MAC5) are shown in Figs. 2, 3, 4 and 5. The ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 149 trees obtained by treating gaps as missing data gap and with no gap weight, D. triquetrella was (MrBayes) are shown in Figs. 6 (JC + g model) the sister group of the D. wockei/fennicella/ and 7 (HKY + g model). The GTR + g model lichenella cluster. With the JC model and no gap produced an identical tree to the HKY + g tree. weight, D. triquetrella haplotypes were the sister In all trees the haplotypes of the parthenoge- group of the D. charlottae/lazuri/wockei group netic D. lichenella form a monophyletic group (Fig. 2). When the gap weight was applied in which was the sister group of the haplotypes of the JC model, the tree obtained was similar to the sexual D. fumosella. D. lichenella showed those obtained with the gap treated as missing very low intraspecific variation although it data: D. triquetrella haplotypes were the sister included samples from Canada (Table 2). D. gen- group of the S. rupicolella/listerlla/vaudella/sau- erosensis haplotypes also formed a monophyletic teri group (Fig. 4). Basal to this group was the group which was basal to the D. lichenella/ D.wockei/charlottae/lazuri group. In MrBayes fumosella groups. D. fennicella and all samples trees this relationship was resolved only with classified as the parthenogenetic form of S. rupi- the JC + g model (Fig. 6), while with the more colella clustered in the same group. Intraspecific complex model of nucleotide substitution the variation was very low (0% to 0.3% divergence) relationships among major clusters were unre- and these species shared the most common hap- solved (Fig. 7). The position of Siederia haplo- lotypes (DF1). D. fennicella group was the sister types in the trees varied along with the model group of two haplotypes (DL1 and DL2) of the of nucleotide substitution implemented and how sexual species D. lazuri. The cluster D. fenni- gaps were treated. Nevertheless, Siederia always cella/rupicolella/lazuri was the sister cluster of formed two well-supported distinct groups (Figs. D. fumosella/lichenella/generosensis group. All 2, 3, 4, 5, 6 and 7). A Swiss and a Swedish S. the relationships listed above were highly sup- rupicolella haplotypes formed one cluster (SR1 ported in all trees. D. lazuri haplotypes did not and SR3 respectively) and all other specimens form a monophyletic group. Three more speci- from Finland and Sweden classified as S. rupi- mens (DL3, DL4 and DL4a) clustered with D. colella and S. listerella formed another cluster. charlottae haplotypes. The D. charlottae/lazuri We found very little genetic variation among the cluster was the sister group of D. wockei in all different sexual Siederia in Finland and Sweden, trees but in the tree obtained using the model whereas we observed high divergence (10%) F84 + gap and no gap weight in MAC5 (Fig. 3). between the two Siederia groups. In all analyses D. wockei haplotypes always formed a mono- with gaps treated as missing data and with the phyletic group. The D. wockei/charlottae/lazuri JC model + gaps (Figs. 2, 4, 6 and 7), the two S. group was the sister group of the D. lichenella/ rupicolella haplotypes (SR1 and 3) were basal to fennicella cluster when the F84 model with gap all the Dahlica and Siederia species. In contrast, weight was used in Mac5 (Fig. 5). With the same using the F84 model + gaps and no gaps weight model but no gap weight D. charlottae/lazuri the other S. rupicolella and S. listerella haplo- was the basal group of the cluster formed by D. types were basal to the tree (Fig. 5). The position wockei/fennicella/lichenella (Fig. 3). Intraspe- of D. vaudella and D. sauteri also varied accord- cific variation among D. triquetrella was higher ing to the model used in the phylogenetic recon- than that observed for the other species, with struction (see e.g., Figs. 3 and 4). a divergence of up to 4.2% between Austrian and Swiss haplotypes (DTp1 and DTp4 respec- tively) (see Table 2). D. triquetrella formed a Discussion monophyletic group when gaps were treated as 5th state (Figs. 2, 3, 4 and 5). When gaps were Finnish samples of bagworm moths clustered in treated as missing data, the relationships among six separate groups with all models of substitu- D. triquetrella haplotypes were unresolved form- tion used to infer their phylogenetic relation- ing a polytomy with those of the S. rupicolella/ ships. These relationships allowed us to compare listerella/D. sauteri/vaudella group (Figs. 6 and the species’ identity with the identifications based 7). In MAC5, using the F84 model both with on their wing morphology, the most used charac- 150 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42 . The genetic distances genetic The 12S . g model of nucleotide substitution. Specimens not included in the matrix were: SL3 because it was identical to SR2, SRp1 and 2 because DFu1 DFu2 0.686 0.687 0.319 0.309 0.087 0.084 0.115 0.111 0.112 0.108 0.106 0.102 0.104 0.100 0.087 0.083 0.093 0.090 0.074 0.073 0.074 0.073 0.057 0.054 0.062 0.059 0.064 0.061 0.079 0.078 0.075 0.075 0.070 0.069 0.072 0.072 DFu1 DFu2 0.007 0.004 0.008 0.006 0.007 0.004 0.006 0.003 0.007 0.005 0.027 0.024 0.028 0.026 0.026 0.024 0.075 0.071 0.076 0.073 0.072 0.069 0.027 0.025 0.029 0.026 0.002 DLa1 SR1 0.486 PN SR2 0.641 DLa1 DV 0.682 0.287 DS1 SR1 0.334 0.640 DS2 0.617 0.113 DW1 0.354 SR2 0.625 0.316 DW2 0.101 0.644 0.320 0.103 DTs1 0.647 DV 0.322 0.106 0.105 DTs2 0.086 0.644 0.323 0.110 DS1 DTp1 0.084 0.679 0.029 0.274 0.110 0.098 DTp2 0.701 0.030 0.287 0.083 DS2 0.091 DTp3 0.107 0.712 0.295 0.090 0.001 0.083 DTp4 0.103 0.701 DW1 0.275 0.093 0.108 0.089 0.084 DTp5 0.703 0.289 0.073 0.101 0.106 0.084 0.090 DW2 DTp6 0.686 0.073 0.300 0.078 0.099 0.061 0.085 DTp7 0.654 0.079 0.282 0.095 0.071 DTs1 0.065 0.076 0.005 DLi1 0.666 0.071 0.269 0.091 0.077 0.095 0.078 0.081 DLi2 0.051 0.271 0.085 0.069 DTs2 0.723 0.091 0.078 0.096 DLi3 0.055 0.085 0.078 0.049 0.726 0.331 0.078 0.060 0.092 DTp1 DLi4 0.084 0.084 0.054 0.716 0.326 0.089 0.080 0.057 0.079 DLi5 0.080 0.063 0.082 DTp2 0.701 0.328 0.086 0.059 0.002 0.079 0.114 DF1 0.074 0.057 0.078 0.738 0.322 0.089 0.083 0.037 DTp3 0.111 DF2 0.074 0.110 0.059 0.072 0.622 0.339 0.086 0.079 0.032 0.114 DF3 0.106 0.089 0.072 DTp4 0.108 0.037 0.640 0.091 0.300 0.073 0.036 0.110 DC1 0.114 0.086 0.105 0.032 0.645 0.106 0.307 0.087 0.075 0.008 DTp5 0.117 DC2 0.110 0.073 0.654 0.108 0.036 0.103 0.311 0.091 0.006 DC3 0.035 0.089 0.098 0.116 0.076 0.680 0.104 0.008 DTp6 0.318 0.106 0.092 0.007 DL1 0.037 0.086 0.101 0.109 0.673 0.110 0.006 0.096 0.304 0.102 0.078 DTp7 0.007 DL2 0.041 0.088 0.103 0.113 0.003 0.009 0.620 0.093 0.103 0.311 0.108 0.073 0.038 0.085 0.087 0.078 0.114 0.036 0.010 0.632 0.095 0.106 0.075 0.291 0.101 0.037 0.090 0.096 0.076 0.089 0.033 0.092 0.108 0.297 0.104 0.080 0.078 0.040 DG1 0.038 0.091 0.078 0.088 0.087 0.032 0.074 0.098 0.106 0.082 0.076 0.037 DG2 0.075 0.091 0.086 0.085 0.055 0.667 0.033 0.073 0.072 0.088 0.078 0.037 0.079 0.086 0.088 0.053 0.104 0.662 0.004 0.071 0.316 0.071 0.091 0.064 0.075 0.037 0.065 0.071 0.058 0.106 0.013 0.312 0.080 0.094 0.069 0.093 0.061 0.079 0.063 0.073 0.065 0.055 0.013 0.078 0.065 0.097 0.063 0.092 0.011 0.076 0.106 0.062 0.075 0.063 0.059 0.063 0.083 0.061 0.095 0.011 0.079 0.107 0.060 0.103 0.083 0.065 0.060 0.063 0.081 0.065 0.075 0.080 0.058 0.101 0.085 0.063 0.062 0.091 0.000 0.083 0.083 0.066 0.055 0.077 0.058 0.067 0.059 0.091 0.089 0.080 0.085 0.074 0.063 0.057 0.080 0.062 0.063 0.056 0.089 0.085 0.072 0.077 0.063 0.054 0.085 0.077 0.063 0.060 0.059 0.074 0.044 0.074 0.081 0.067 0.085 0.081 0.060 0.056 0.086 0.071 0.042 0.076 0.084 0.069 0.078 0.048 0.053 0.086 0.075 0.042 0.074 0.080 0.070 0.081 0.046 0.055 0.070 0.077 0.067 0.049 0.068 0.076 0.046 0.068 0.075 0.070 0.051 0.069 0.067 0.051 0.068 0.075 0.071 0.072 0.065 0.055 0.052 0.051 0.072 0.073 0.064 0.054 0.056 0.049 0.052 0.074 0.068 0.056 0.049 0.049 0.070 0.058 0.050 0.058 0.056 0.080 0.060 0.060 0.080 0.061 0.077 0.077 0.066 0.066 0.068 0.067 DLi1 DLi2 DLi3 DLi4 DLi5 DF1 DF2 DF3 DC1 DC2 DLi2 DC3 DLi3 0.003 DLi4 0.002 DL1 DLi5 0.001 0.003 DF1 0.002 DL2 0.002 DF2 0.027 0.003 0.001 DF3 DFu1 0.029 0.002 0.029 DC1 DFu3 0.027 0.031 0.024 0.001 DC2 0.076 0.029 0.026 DC3 0.026 DG1 0.078 0.074 0.024 DL1 0.027 0.025 0.073 0.075 0.076 DL2 0.025 DG2 0.026 0.022 0.071 0.078 0.073 0.024 0.023 0.001 0.073 0.023 0.075 0.077 0.002 0.025 0.019 DG1 0.070 0.078 0.069 0.003 0.020 DG2 0.020 0.074 0.011 0.070 0.072 0.022 0.022 0.009 0.066 0.012 0.073 0.073 0.023 0.011 0.011 0.009 0.069 0.075 0.009 0.010 0.010 0.009 0.071 0.002 0.012 0.009 0.011 0.011 0.001 0.010 0.013 0.026 0.062 0.001 0.024 0.027 0.064 0.063 0.027 0.028 0.065 0.060 0.028 0.068 0.061 0.066 0.069 0.001 0.068 0.065 0.064 0.024 0.023 0.026 0.024 0.010 0.010 0.008 0.008 0.002 PN PN DLa1 SR1 SR2 DV DS1 DS2 DW1 DW2 DTs1 DTs2 DTp1 DTp2 DTp3 DTp4 DTp5 DTp6 DTp7 DLi1 fumosella fumosella fumosella fumosella f. f. f. f. were calculated with the HKY + Pairwise sequence divergence among haplotypes used in this study based on the concatenated sequence of COI, tRNAleu, COII and COII tRNAleu, COI, of sequence concatenated the on based study this in used haplotypes among divergence sequence Pairwise 2. Table they were identical to DF1 and DL4, DTs2 because it differed by one indel from DC1 DTs1 respectively. Psyche norvegica Diplodoma laichartingella Siederia rupicolella Siederia rupicolella Dahlica vaudella Dahlica sauteri Dahlica sauteri Dahlica wockei Dahlica wockei Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica fennicella Dahlica fennicella Dahlica fennicella Dahlica charlottae Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lichenella Dahlica lichenella Dahlica generosensis Dahlica generosensis Psyche norvegica Diplodoma laichartingella Siederia rupicolella Siederia rupicolella Dahlica vaudella Dahlica sauteri Dahlica sauteri Dahlica wockei Dahlica wockei Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica fennicella Dahlica fennicella Dahlica fennicella Dahlica charlottae Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lichenella Dahlica lichenella Dahlica generosensis Dahlica generosensis

ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 151 . The genetic distances genetic The 12S . g model of nucleotide substitution. Specimens not included in the matrix were: SL3 because it was identical to SR2, SRp1 and 2 because DFu1 DFu2 0.686 0.687 0.319 0.309 0.087 0.084 0.115 0.111 0.112 0.108 0.106 0.102 0.104 0.100 0.087 0.083 0.093 0.090 0.074 0.073 0.074 0.073 0.057 0.054 0.062 0.059 0.064 0.061 0.079 0.078 0.075 0.075 0.070 0.069 0.072 0.072 DFu1 DFu2 0.007 0.004 0.008 0.006 0.007 0.004 0.006 0.003 0.007 0.005 0.027 0.024 0.028 0.026 0.026 0.024 0.075 0.071 0.076 0.073 0.072 0.069 0.027 0.025 0.029 0.026 0.002 DLa1 SR1 0.486 PN SR2 0.641 DLa1 DV 0.682 0.287 DS1 SR1 0.334 0.640 DS2 0.617 0.113 DW1 0.354 SR2 0.625 0.316 DW2 0.101 0.644 0.320 0.103 DTs1 0.647 DV 0.322 0.106 0.105 DTs2 0.086 0.644 0.323 0.110 DS1 DTp1 0.084 0.679 0.029 0.274 0.110 0.098 DTp2 0.701 0.030 0.287 0.083 DS2 0.091 DTp3 0.107 0.712 0.295 0.090 0.001 0.083 DTp4 0.103 0.701 DW1 0.275 0.093 0.108 0.089 0.084 DTp5 0.703 0.289 0.073 0.101 0.106 0.084 0.090 DW2 DTp6 0.686 0.073 0.300 0.078 0.099 0.061 0.085 DTp7 0.654 0.079 0.282 0.095 0.071 DTs1 0.065 0.076 0.005 DLi1 0.666 0.071 0.269 0.091 0.077 0.095 0.078 0.081 DLi2 0.051 0.271 0.085 0.069 DTs2 0.723 0.091 0.078 0.096 DLi3 0.055 0.085 0.078 0.049 0.726 0.331 0.078 0.060 0.092 DTp1 DLi4 0.084 0.084 0.054 0.716 0.326 0.089 0.080 0.057 0.079 DLi5 0.080 0.063 0.082 DTp2 0.701 0.328 0.086 0.059 0.002 0.079 0.114 DF1 0.074 0.057 0.078 0.738 0.322 0.089 0.083 0.037 DTp3 0.111 DF2 0.074 0.110 0.059 0.072 0.622 0.339 0.086 0.079 0.032 0.114 DF3 0.106 0.089 0.072 DTp4 0.108 0.037 0.640 0.091 0.300 0.073 0.036 0.110 DC1 0.114 0.086 0.105 0.032 0.645 0.106 0.307 0.087 0.075 0.008 DTp5 0.117 DC2 0.110 0.073 0.654 0.108 0.036 0.103 0.311 0.091 0.006 DC3 0.035 0.089 0.098 0.116 0.076 0.680 0.104 0.008 DTp6 0.318 0.106 0.092 0.007 DL1 0.037 0.086 0.101 0.109 0.673 0.110 0.006 0.096 0.304 0.102 0.078 DTp7 0.007 DL2 0.041 0.088 0.103 0.113 0.003 0.009 0.620 0.093 0.103 0.311 0.108 0.073 0.038 0.085 0.087 0.078 0.114 0.036 0.010 0.632 0.095 0.106 0.075 0.291 0.101 0.037 0.090 0.096 0.076 0.089 0.033 0.092 0.108 0.297 0.104 0.080 0.078 0.040 DG1 0.038 0.091 0.078 0.088 0.087 0.032 0.074 0.098 0.106 0.082 0.076 0.037 DG2 0.075 0.091 0.086 0.085 0.055 0.667 0.033 0.073 0.072 0.088 0.078 0.037 0.079 0.086 0.088 0.053 0.104 0.662 0.004 0.071 0.316 0.071 0.091 0.064 0.075 0.037 0.065 0.071 0.058 0.106 0.013 0.312 0.080 0.094 0.069 0.093 0.061 0.079 0.063 0.073 0.065 0.055 0.013 0.078 0.065 0.097 0.063 0.092 0.011 0.076 0.106 0.062 0.075 0.063 0.059 0.063 0.083 0.061 0.095 0.011 0.079 0.107 0.060 0.103 0.083 0.065 0.060 0.063 0.081 0.065 0.075 0.080 0.058 0.101 0.085 0.063 0.062 0.091 0.000 0.083 0.083 0.066 0.055 0.077 0.058 0.067 0.059 0.091 0.089 0.080 0.085 0.074 0.063 0.057 0.080 0.062 0.063 0.056 0.089 0.085 0.072 0.077 0.063 0.054 0.085 0.077 0.063 0.060 0.059 0.074 0.044 0.074 0.081 0.067 0.085 0.081 0.060 0.056 0.086 0.071 0.042 0.076 0.084 0.069 0.078 0.048 0.053 0.086 0.075 0.042 0.074 0.080 0.070 0.081 0.046 0.055 0.070 0.077 0.067 0.049 0.068 0.076 0.046 0.068 0.075 0.070 0.051 0.069 0.067 0.051 0.068 0.075 0.071 0.072 0.065 0.055 0.052 0.051 0.072 0.073 0.064 0.054 0.056 0.049 0.052 0.074 0.068 0.056 0.049 0.049 0.070 0.058 0.050 0.058 0.056 0.080 0.060 0.060 0.080 0.061 0.077 0.077 0.066 0.066 0.068 0.067 DLi1 DLi2 DLi3 DLi4 DLi5 DF1 DF2 DF3 DC1 DC2 DLi2 DC3 DLi3 0.003 DLi4 0.002 DL1 DLi5 0.001 0.003 DF1 0.002 DL2 0.002 DF2 0.027 0.003 0.001 DF3 DFu1 0.029 0.002 0.029 DC1 DFu3 0.027 0.031 0.024 0.001 DC2 0.076 0.029 0.026 DC3 0.026 DG1 0.078 0.074 0.024 DL1 0.027 0.025 0.073 0.075 0.076 DL2 0.025 DG2 0.026 0.022 0.071 0.078 0.073 0.024 0.023 0.001 0.073 0.023 0.075 0.077 0.002 0.025 0.019 DG1 0.070 0.078 0.069 0.003 0.020 DG2 0.020 0.074 0.011 0.070 0.072 0.022 0.022 0.009 0.066 0.012 0.073 0.073 0.023 0.011 0.011 0.009 0.069 0.075 0.009 0.010 0.010 0.009 0.071 0.002 0.012 0.009 0.011 0.011 0.001 0.010 0.013 0.026 0.062 0.001 0.024 0.027 0.064 0.063 0.027 0.028 0.065 0.060 0.028 0.068 0.061 0.066 0.069 0.001 0.068 0.065 0.064 0.024 0.023 0.026 0.024 0.010 0.010 0.008 0.008 0.002 PN PN DLa1 SR1 SR2 DV DS1 DS2 DW1 DW2 DTs1 DTs2 DTp1 DTp2 DTp3 DTp4 DTp5 DTp6 DTp7 DLi1 fumosella fumosella fumosella fumosella f. f. f. f. were calculated with the HKY + Pairwise sequence divergence among haplotypes used in this study based on the concatenated sequence of COI, tRNAleu, COII and COII tRNAleu, COI, of sequence concatenated the on based study this in used haplotypes among divergence sequence Pairwise 2. Table they were identical to DF1 and DL4, DTs2 because it differed by one indel from DC1 DTs1 respectively. Psyche norvegica Diplodoma laichartingella Siederia rupicolella Siederia rupicolella Dahlica vaudella Dahlica sauteri Dahlica sauteri Dahlica wockei Dahlica wockei Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica fennicella Dahlica fennicella Dahlica fennicella Dahlica charlottae Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lichenella Dahlica lichenella Dahlica generosensis Dahlica generosensis Psyche norvegica Diplodoma laichartingella Siederia rupicolella Siederia rupicolella Dahlica vaudella Dahlica sauteri Dahlica sauteri Dahlica wockei Dahlica wockei Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica triquetrella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica lichenella Dahlica fennicella Dahlica fennicella Dahlica fennicella Dahlica charlottae Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lazuri Dahlica lichenella Dahlica lichenella Dahlica generosensis Dahlica generosensis

152 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

TB1 TB2 100 DLa2 DLa1 100 SR1 SR3 DTp1 DTp12 DTp8 100 DTp5 DTp7a 100 99 DTp7b 100 99 DTp7 99 DTp9 DTp4 100 99 DTs3 DTs2 99 99 DTs1 89 99 DTp6 DTp6c 100 DTp6a 72 DTp10 DTp11 DTp6b 100 DTp3 100 DTp2 DTp2a SL2 100 SL1 SR2a 92 SR2b 100 SR2 51 SR4 SL3 100 DV DS2 100 DS1 100 DW1 DW2 99 97 DW3 DL3 100 DC2 DC1a DL4a 99 DC1 100 99 DC1c DL4 DC1b 99 DFu2 Fig. 2. Bayesian tree DFu1 obtained from the com- 100 DLi2 DLi4a bined sequence of COI, DLi5 DLi4b tRNAleu, COII and 12S, 100 100 DLi3 using the JC + gaps model DLi4 DLi1 of nucleotide substitu- DG3 100 100 DG2 tion with no gap weight DG1 in MAC5. Numbers at DG4 100 DL1 the nodes represent the DL2 support values. Nodes 100 DF4a DF4 with 50% or less support DF3 SRp2b were collapsed. Black and 100 DF1 dashed branches indicate SRp2 100 SRp2a sexual and parthenoge- DF2 netic lineages, respec- DF1a SRp1 tively. ter for identification of these species. Although two distinct groups which were never mono- the sequence data were unable to completely phyletic, questioning the validity of classifying resolve the relationships because of incongru- these bagworm moths in two different genera ence among different trees obtained with differ- (Dahlica and Siederia). ent models of substitution, our results still raise Although the Finnish Siederia samples could significant questions about the current be separated into two clearly different mor- and relationships of Naryciinae species. phological groups on the basis of their wing The position of Siederia haplotypes in the scales (S. rupicolella and S. listerella), in our trees varied along with the model of nucleotide mtDNA analysis, all Finnish specimens of sexual substitution implemented and how gaps were Siederia formed a homogenous, well-supported treated. Nevertheless, Siederia always formed group with the inclusion of two S. listerella and ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 153

TB1 TB2 100 DLa2 DLa1 100 DV DS2 100 DS1 SL1 100 100 63 SL2 SL3 SR2a SR2b 100 57 SR4 SR2 100 SR3 52 SR1 100 DTp3 DTp2a 100 DTp2 DTp1 100 DTp12 DTp6 100 DTp10 100 DTP6a 100 DTp6c DTp6b DTp11 92 97 DTs3 Dts1 100 98 Dts2 DTp8 98 100 DTp9 100 DTp4 DTp7b DTp5 96 DTp7a DTp7 DL3 100 DC2 DC1 DC1b 98 DC1a 96 DC1c DL4a 99 DL4 100 DW1 DW3 DW2 Fig. 3. Bayesian tree 100 DL1 DL2 obtained from the com- 100 98 58 DF4 bined sequence of COI, DF4a DF2 tRNAleu, COII and 12S SRp1 using the F84 + gaps 100 SRp2b SRp2a model of nucleotide sub- DF1a 100 100 SRp2 stitution with no gap DF3 weight in MAC5. Numbers DF1 DG1 at the nodes represent the 100 DG2 support values. Nodes DG3 100 DG4 with 50% or less support 96 DFu1 DFu2 were collapsed. Black and 99 DLi4b dashed branches indicate DLi5 DLi1 sexual and parthenoge- DLi2 100 DLi3 netic lineages, respec- 86 DLi4 tively. DLi4a one S. rupicolella from Sweden. Very different these two species. S. rupicolella is a rare species (10% divergence) were two haplotypes of S. in the Alps and in Scandinavia but clearly has rupicolella collected in Switzerland and Sweden. wide distribution as previously reported (Hät- From these results it seems that all Finnish sam- tenschwiler 1997). ples can be classified as S. listerella according to D. fennicella was considered by Suomalainen the mtDNA. Thus, contrary to the general under- (1980) a separate species, locally restricted and standing, S. listerella seems very common in very rare. Many collectors instead considered Finland. Its wing scale morphology is more vari- parthenogenetic moths in other parts of Finland able than previously thought and widely over- as the parthenogenetic form of S. rupicolella. In laps with that of S. rupicolella, making this mor- our analysis, all specimens identified as “parthe- phological character not a diagnostic to separate nogenetic S. rupicolella” form a monophyletic 154 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

TB1 TB2 100 DLa2 DLa1 100 SR1 SR3 100 DL1 DL2 100 99 DF4a DF4 100 DF3 SRp1 100 DF1a SRp2 100 DF2 100 SRp2a DF1 SRp2b 100 DG1 100 DG3 DG2 100 DG4 99 DFu2 DFu1 100 DLi2 DLi1 DLi4 DLi3 100 DLi4b 100 DLi5 DLi4a 100 DW1 DW2 99 97 DW3 DL3 100 DC2 DL4a DC1a 99 DC1c DC1b 99 DL4 92 DC1 100 DV DS2 100 100 DS1 SL2 SL1 SR2b 100 SR2a 72 SL3 Fig. 4. Bayesian tree 51 SR4 SR2 obtained from the com- DTp1 bined sequence of COI, 99 100 DTp2 DTp3 tRNA , COII and 12S, DTp2a leu DTp12 using the JC + gaps model DTp6 of nucleotide substitution 89 DTp6b 100 100 DTp11 with gap weight in MAC5. DTp10 DTp6a Numbers at the nodes DTp6c represent the support 99 99 DTs3 DTs2 values. Nodes with 50% 100 99 DTs1 DTp8 or less support were col- 100 DTp7a lapsed. Black and dashed DTp5 DTp7 branches indicate sexual 99 DTp7b and parthenogenetic line- 99 DTp9 DTp4 ages, respectively. group with D. fennicella sharing the most found these two asexual moths (from now on common haplotype. The question thus arises: are called D. fennicella) in six of eight randomly these two asexual bagworm moths the same spe- chosen potentially suitable habitat patches. In cies? Most likely they are. In consequence, mor- most of these, D. fennicella was very numerous phology, cover material and size of larvae sack (being the most abundant species of Dahlica or in D. fennicella are more variable than earlier Siederia). This species is often found in warm thought. Habitat requirements also vary consid- and half-open habitats such as parks, scattered erably since this species has earlier been found plantations of forest or wooded garden areas, only on rocks and stones (Suomalainen 1980). where the ground layers of vegetation are not We found most of our samples of D. fennicella too dense. Most of the habitats are close to lake and “asexual S. rupicolella” on tree trunks. We shores (Kumpulainen 2004). However, there is ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 155

TB1 TB2 100 DLa2 DLa1 SL1 SL2 100 SR2 SR2a SL3 SR2b 100 SR4 100 SR1 SR3 100 DV DS1 100 100 DS2 DTp1 61 100 DTp2 DTp3 DTp2a DTp12 DTp6c 53 55 DTp10 100 100 DTp6b DTp6a DTp11 DTp6 92 96 DTs3 DTs1 DTs2 DTp8 100 DTp7b 100 DTp7 DTp5 98 DTp7a 99 DTp4 DTp9 100 DW1 DW2 72 DW3 DL3 100 DC2 DC1c DL4a DC1b DC1 DC1a DL4 Fig. 5. Bayesian tree 62 100 DL2 obtained from the com- DL1 100 96 DF4a bined sequence of COI, DF4 DF1 tRNAleu, COII and 12S SRp1 using the F84 + gaps 100 SRp2a DF3 model of nucleotide sub- 100 DF1a 100 DF2 stitution with no gap SRp2b weight in MAC5. Numbers SRp2 DG1 at the nodes represent the 100 DG2 support values. Nodes DG4 100 DG3 with 50% or less support 97 DFu2 DFu1 were collapsed. Black and 96 DLi3 dashed branches indicate DLi4b DLi5 sexual and parthenoge- DLi4 netic lineages, respec- 100 DLi2 DLi4a tively. DLi1 considerable overlap in habitat preferences of to hybridisation involving the same maternal many Dahlica and Siederia species and, thus, species. In many animals (including insects), habitat alone cannot be used for species deter- asexual reproduction has been associated with mination, except between D. fennicella and D. interspecific hybridisation. Hybridisation implies lichenella, which are not known to co-occur. reticulation of lineages and the possibility of It is, however, possible that “asexual S. rupi- introgression of genetic markers resulting in colella” is a morphologically differentiated clone incongruence between gene trees and species of D. fennicella that has separated from ancestral trees (Normark & Lanteri 1998). Interspecific D. fennicella so recently that the genetic differ- hybridization, however, is unlikely to be at the ences are too slight to be detected in our data. point of origin of parthenogenetic species in Finally, asexuality in both species could be due the Dahlica/Siederia group. Asexual Dahlica 156 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

TB1 TB2 1.00 DLa1 DLa2 1.00 SR1 SR3 1.00 DTp3 DTp2a DTp2 SL1 1.00 SL2 SR2b 1.00 SL3 0.98 1.00 SR2 0.79 SR2a SR4 1.00 DV 1.00 DS1 DS2 DTp1 DTp12 1.00 DTp6a 0.94 DTp6 1.00 1.00 DTp10 DTp11 DTp6b 1.00 DTp6c 0.86 DTp8 0.98 DTp7a DTp7 DTp7b 0.99 DTp5 DTp9 1.00 DTp4 0.97 DTs3 0.96 DTs2 0.98 DTs1 DL3 1.00 DC2 DC1a DC1 0.95 DC1b 1.00 DC1c 1.00 0.95 DL4 DL4a 1.00 DW1 DW2 0.69 DW3 DG1 Fig. 6. Bayesian tree 1.00 DG2 DG3 obtained from the com- DG4 bined sequence of COI, DLi3 1.00 DLi1 tRNAleu, COII and 12S, 1.00 DLi2 using the JC + g model DLi4a DLi4 of nucleotide substitution 0.1 DLi4b 1.00 0.99 DLi5 and gap treated as miss- DFu1 ing data in MrBayes ver. 0.97 DFu2 1.00 DL1 3.0. Numbers at the nodes 0.99 DL2 represent the support DF4 DF4a values. Nodes with 50% 1.00 SRp1 SRp2 or less support were col- 0.96 SRp2a lapsed. Black and dashed SRp2b 1.00 DF2 branches indicate sexual DF3 and parthenogenetic line- DF1 DF1a ages, respectively. triquetrella (and most likely all other species) develop and the zygoid phase is restored with the reproduce by automictic parthenogenesis (Narbel fusion of the sperm nucleus (Seiler 1967). 1950). In this type of parthenogenesis, the early From our results D. lichenella is a separate, stages of meiosis in the egg are similar to those in obligatorily parthenogenetic species. There is very the fertilized eggs of sexual species. The zygoid little genetic variation among the Finnish, Swiss phase is restored by the fusion of the two central and Canadian samples. In Finland, according to azygoid nuclei. This results in the formation Suomalainen (1980), larvae of this species can be of the so called “Richtung-Kopulations-Kerns” found on stones but we observed numerous larvae (RKK). The fusion always leads to heterogamy also on walls of buildings and on tree trunks (T. and production of females. RRK also forms in Kumpulainen, K. Kulmala & J. Mappes pers. the sexual form of D. triquetrella, but does not obs.). In coastal Finland, D. lichenella seems to ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 157

TB1 TB2 1.00 DLa1 DLa2 1.00 SR1 SR3 1.00 DW1 DW2 0.71 0.67 DW3 DL3 1.00 1.00 DC2 DC1a 0.94 DC1 DC1b 0.95 DC1c DL4 1.00 DL4a DTp1 1.00 DTp3 DTp2a DTp2 DTp12 1.00 DTp6a DTp6 1.00 DTp10 DTp11 DTp6b 1.00 DTp6c 0.52 DTp8 DTp7a 0.92 DTp7 DTp7b DTp5 0.99 0.97 1.00 DTp9 DTp4 0.98 DTs3 0.89 DTs2 0.96 DTs1 0.55 SL1 SL2 1.00 SR2b SL3 0.94 SR2 0.52 SR2a SR4 1.00 DV DS1 1.00 DS2 Fig. 7. Bayesian tree DG1 obtained from the com- 1.00 DG2 DG3 bined sequence of COI, DG4 DLi3 1.00 tRNAleu, COII and 12S, DLi1 using the HKY85 + g model 1.00 DLi2 DLi4a of nucleotide substitution 0.1 DLi4 and gap treated as miss- DLi4b 1.00 0.93 DLi5 ing data in MrBayes ver. DFu1 0.90 DFu2 3.0. Numbers at the nodes 1.00 DL1 represent the support 0.91 DL2 DF4 values. Nodes with 50% DF4a 0.99 SRp1 or less support were col- SRp2 lapsed. Black and dashed 0.95 SRp2a SRp2b branches indicate sexual 0.99 DF2 and parthenogenetic line- DF3 DF1 ages, respectively. DF1a prefer habitats with humid and temperate micro- according to the wing scale morphology, clus- climates such as sea- and lakeshore areas (T. tered with D. charlottae. Two other “D. lazuri” Kumpulainen & A. Grapputo pers. obs.). Even specimens, in contrast, clustered together as in the absence of any reliable morphological or the sister group of the asexual D. fennicella. ecological characteristics that would separate D. These specimens could not be separated by male lichenella from D. fennicella, since the differ- wing scale morphology from those that grouped ences in the species distribution are surely helpful together with D. charlottae. Thus, some other for species identification (see above). character is needed for determining the species Moths classified morphologically as D. lazuri based on their morphology. D. lazuri and D. did not form a monophyletic group. In our lichenella (f. fumosella) did not cluster together analysis four specimens considered “D. lazuri” in any trees, indicating that D. lazuri is a separate 158 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42 species from the sexual morph of D. lichenella as compared with that of sexually produced off- (Arnscheid 1985, Hermann 1994, Hättenschwiler spring. In psychid moths instead, parthenogenetic 1997). Sexual D. lichenella f. fumosella samples reproduction seems to have evolved several times from Switzerland formed the sister species of the as suggested by the alternation of parthenoge- asexual form D. lichenella confirming the previ- netic and sexual species along the phylogenetic ously established relationships. trees. According to the mtDNA data, the parthe- Intraspecific variation among D. triquetrella nogenetic species D. fennicella, D. triquetrella was higher than that observed among the other and D. lichenella evolved independently from species and similar to the interspecific varia- different sexual ancestors. Parthenogenetic Dahl- tion observed among Dahlica species (Table 2). ica species have a reproductive output equal Despite the high genetic distances among hap- to that of their sexual relatives, and Wolbachia lotypes, D. triquetrella formed a monophyletic is not known to affect the reproductive strat- group when gaps were treated as 5th state. How- egy (Kumpulainen et al. 2004). Wolbachia is a ever the phylogenetic relationships among D. feminizing bacterium which can cause parthe- triquetrella and other Dahlica species were unre- nogenesis in many insect species (Stouthamer solved when gaps were treated as missing data. et al. 1999). Multiple evolutionary events of All D. triquetrella haplotypes have an insertion parthenogensis suggest that asexual reproduction of three bases (see Materials and methods) that is favoured in this group. Recently, it has been no other haplotypes from other species have. hypothesized that loss of sex, including faculta- Gaps contain historical information suitable for tive or cyclic parthenogenesis, should be associ- phylogenetic analysis and such information is not ated with weak sexual selection when not accom- recovered by those methods that omit gaps from panied by male parental contributions (West- their calculations (Giribet & Wheeler 1999). This Eberhard 2005). Psychidae moths fit this picture is clearly shown in D. triquetrella. From the phy- well since there is little opportunity for sexual logenetic analysis using gaps, results show that selection by direct female choice, since the wing- asexual D. triquetrella in Finland are more simi- less females mate only once, with the first male lar to sexual and asexual D. triquetrella found to attempt copulation (personal observation). in Austria than to asexual samples from Swit- Overlapping sexual characteristics, like complex zerland. High variation in the species has also genitalia in Siederia and Dahlica also support the been observed in the head morphology of female hypothesis of weak sexual selection. Moreover, pupae and in the morphology of males (Hätten- sexual populations of this group are very small, schwiler 1997). It is, however, not known if there genetically isolated and highly inbred, and males is a direct association between genetic variation are severely sperm-limited. Although about 50% and such morphological variation. Seiler (1961, of males mate more than once, females that mate 1963) reported that different populations of par- with already mated males lose 30%–100% of thenogenetic D. triquetrella show different pat- viable offspring (Kumpulainen 2004). terns of egg development. Differences in life-his- The closest sexual species to asexual D. tory traits between D. triquetrella populations of fennicella and D. lichenella are quite rare as distant geographical areas led Seiler to conclude compared with the asexual relatives. Many theo- that parthenogenesis may have evolved inde- retical models assume that the advantage of sex pendently several times in different populations is primarily due to its ability, through recombina- (Lokki et al. 1975). Our phylogenetic analyses tion, to generate greater genetic diversity than (with gaps) and the position of sexual D. triquet- asexuality and to spread good genes rapidly rella with respect to the other parthenogenetic through a population by unlinking them from samples in the trees support Seiler’s view. Other- bad genotypes, and thereby enhancing adapta- wise, we have to assume that sexual reproduction tion in a changing environment (Hamilton 1980, has reverted from asexuality. Kondrashov 1993). Recently, we found lower In general, parthenogenesis is very rare among genotype diversity in asexual D. fennicella as Lepidoptera and often regarded as a secondary compared with that of the sexual S. rupicolella reproductive strategy with low offspring fitness and D. charlottae. Nevertheless, clonal diversity ANN. ZOOL. FENNICI Vol. 42 • Phylogeny and parthenogenesis in Psychidae 159 was very high in all asexual populations which Austria, Sweden and Canada, and to A.-M. Koskela and T. may allow them to successfully compete with Koskela for their help in collecting material in the archi- pelago of Rauma. We also thank K.E. Knott, S. Mills, J. sexual species (A. Grapputo, T. Kumpulainen & Itämies, J. Kullberg, M. Björklund, P. Mutikainen, an anony- J. Mappes unpubl. data). It is reasonable then to mous referee and all the people at “the round table meeting” ask how sexual reproduction can be maintained of the Evolutionary Ecology research unit, for their valuable in this group given the obvious disadvantages comments on a previous version of the manuscript. We thank of it (cost of males, cost of mating and severe Kari Kulmala for assistance in field and laboratory work. inbreeding depression). The most tempting cur- This study was financed by the Academy of Finland (to JM, project numbers 768945, 779874) and the Centre of Excel- rent hypothesis for the maintenance of sex is that lence in Evolutionary Ecology. parasitoids and/or diseases can favour sexual reproduction due to their higher genetic diver- sity. Indeed, we found very strong evidence that References sexuals cope better in environments with high parasitoid prevalence while asexual reproduc- Akaike, H. 1974: A new look at the statistical model identifi- tion dominates in areas with very low parasitoid cation. — IEEE Trans. Automat. Contr. 19: 716–723. prevalence (Kumpulainen et al. 2004). Although Alfaro, M. E., Zoller, S. & Lutzoni, F. 2003: Bayes or clonal diversity of D. fennicella is extremely bootstrap? A simulation study comparing the perform- high and populations of sexual S. rupicolella ance of Bayesian Markov Chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. and D. charlottae are inbred, sexual species still — Mol. Biol. Evol. 20: 255–266. show higher genotype diversity than asexual D. Arnscheid, W. R. 1985: Ein Beitrag zur Systematik der euro- fennicella which may explain why sexual repro- päischen Arten der Gattung Solenobia Duponchel, 1842 duction still exists in this group. (Lepid., Psychidae, Teleporiinae). — Nachrichten des In conclusion, wing scale morphology misled entomologischen Vereins Apollo 4: 1–56. Avise, J. C. 1994: Molecular markers, natural history and the identification of many specimens of Dahlica evolution. — Chapman & Hall, New York. lazuri and sexual Siederia. Our study indicates Bell, G. 1982: The masterpiece of nature: the evolution and that intraspecific variation for this character is genetics of sexuality. — University of California Press, higher than previously suggested making it a Berkeley. poor diagnostic. In cases of parthenogenetic spe- Brooks, D. R. & McLennan, D. A. 1991: Phylogeny, ecology, cies, ecological and distributional data are helpful and behavior: a research program in comparative biol- ogy. — The University of Chicago Press, Chicago. for species determination in Finland. However, Brower, A. V. Z. 1994: Rapid morphological radiation and because of the high number of different species in convergence among races of the butterfly Heliconius central Europe, it is questionable if ecological or erato inferred from patterns of mitochondrial DNA evo- distributional data is generally useful. This study lution. — PNAS 91: 6491–6495. also raised many new questions and revealed Brower, V. Z. 1997: Cladistic analysis of Heliconius butter- novel problems. Firstly, our results from mtDNA flies and relatives (Nymphalidae: Heliconiiti): a revised phylogenetic position for Eueides based on sequences markers question the existence of two separate from mtDNA and a nuclear gene. — Proceedings: Bio- genera. Secondly, the high genetic differentiation logical Sciences 264: 969–977. for certain species and the incongruence between Giribet, G. & Wheeler, W. C. 1999: On gaps. — Mol. Phylo- morphological characters and mtDNA sequences genet. Evol. 13: 132–143. suggest that several samples from different popu- Haddrath, O. & Baker, A. J. 2001: Complete mitochondrial DNA genome sequences of extinct birds: ratite phy- lations should be analysed in any attempt to logenetics and the vicariance biogeography hypothesis. elucidate the relationships among psychid moths. — Proc. R. Soc. Lond. B 268: 939–945. Multiple evolutionary events of parthenogenesis Hamilton, W. D. 1980: Sex versus non-sex versus parasite. support the current understanding that asexuality — Oikos 35: 282–290. is favourable strategy in Psychidae moths. Hasegawa, M., Kishino, H. & Yano, T. A. 1985: Dating of the human ape splitting by a molecular clock of mitochon- drial-DNA. — J. Mol. Evol. 22: 160–174. Hättenschwiler, P. 1985: Psychidae. — In: Heath, J. & Emmet, Acknowledgements A. M. (eds.), The moths and butterflies of Great Britain and Ireland 2: 128–151. Harley Books, Colchester, UK. We are grateful to P. Hättenschwiler, E. Hauser and G. Hättenschwiler, P. 1997: Die Sackträger der Schweiz (Lepi- Palmqvist for providing us with samples from Switzerland, doptera, Psychidae). — Schmetterlinge und ihre Lebens- 160 Grapputo et al. • ANN. ZOOL. FENNICI Vol. 42

räume Arten–Gefährdung–Schutz, Band 2: 165–308. Pro Sauter, W. 1956: Morphologie und Systematik der schwei- Natura, Basel, Switzerland. zerischen Solenobia-Arten (Lep. Psychidae). — Rev. Hermann, R. 1994: Psychidae: Naryciinae. — In: Ebert, G., Suisse Zool. 63: 451–550. Esche, T. & Hermann, R. (eds.), Die Schmetterlinge Sauter, W. & Hättenschwiler, P. 1991: Zum System der palae- Baden-Württembergs, Band 3, Nachtfalter: 356–402. arktischen Pyschiden (Lep. Psychidae) 1. Teil: Liste der Ulmer Verlag, Stuttgart. palaearktischen Arten. — Nota Lepid. 14: 69–89. Huelsenbeck, J. P. & Ronquist, F. 2001: MRBAYES: Baye- Seiler, J. 1961: Untersuchungen über die Entstehung der sian inference of phylogenetic trees. — Bioinformatics Parthenogenese bei Solenobia triquetrella F.R. (Lepido- 17: 754–755. ptera, Psychidae) III. Die geographische Verbreitung der Huelsenbeck, J. P. & Crandall, K. A. 1997: Phylogeny esti- drei Rassen von Solenobia triquetrella (bisexuell, dipo- mation and hypothesis testing using maximum likeli- lid und tetraploid parthenogenetisch) in der Schweiz und hood. — Ann. Rev. Ecol. Syst. 28: 437–466. in den angrenzenden Ländern und die Beziehung zur Jobb, G., von Haeseler, A. & Strimmer, K. 2004: TREEF- Eiszeit. Bemerkungen über die Entstehung der Patheno- INDER: a powerful graphical analysis environment for genese. — Zeitschrift für Vererbungslehre 92: 261–316. molecular phylogenetics. — BMC Evol. Biol. 4: 18. Seiler, J. 1963: Untersuchungen über die Entstehung der Par- Jukes, T. H. & Cantor, C. R. 1969: Evolution of protein mol- thenogenese bei Solenobia triquetrella F.R. (Lepidoptera, ecules. — In: Munro, H. N. (ed.), Mammaliam protein Psychidae) IV. Wie besamen begattete dipolid und tetra- metabolism: 21–132. Academic Press, New York. ploid parthenogenetische Weibchen von S. triquetrella Kishino, H. & Hasegawa, M. 1989: Evaluation of the maxi- ihre Eier? Schicksal der Richtunggskörper im unbesam- mum likelihood estimate of the evolutionary tree topolo- ten und besamten Ei. Vergleich der Ergebnisse mit F1 gies from DNA sequence data, and the branching order Aufzuchten und beziehngen zur Genese der Parthenoge- in Hominoidea. — J. Mol. Evol. 29: 170–179. nese. — Zeitschrift für Vererbungslehre 94: 29–66. Kondrashov, A. S. 1993: Classification of hypotheses on the Seiler, J. 1967: Untersuchungen über die Entstehung der advantage of amphimixis. — J. Hered. 84: 372–387. Parthenogenese bei Solenobia triquetrella F.R. (Lepi- Kruse, J. J. & Sperling, F. A. H. 2002: Phylogeny of Nearc- doptera, Psychidae) VII. Versuch einer experimentellen tic species of the Xylosteana group of Archips Hübner Analyse der Genetick der Pathenogenese. — Mol. Gen. (Lepidoptera: Tortricidae) based on combined analysis Genet. 99: 274–310. of morphological and mitochondrial DNA data sets. Simmons, M. P., Pickett, K. M. & Miya, M. 2004: How — Ann. Entomol. Soc. Am. 95: 288–301. meaningful are Bayesian support values? — Mol. Biol. Kullberg, J., Albrecht, A., Kaila, A. & Varis, V. 2002: Check- Evol. 21: 188–199. list of Finnish Lepidoptera — Suomen perhosten luet- Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H. & telo. — Sahlbergia 6: 45–190. Flook, P. 1994: Evolution, weighting, and phylogenetic Kumpulainen, T. 2004: The evolution and maintenance of utility of mitochondrial gene-sequences and a compila- reproductive strategies in bag worm moths (Lepidoptera: tion of conserved polymerase chain reaction primers. Psychidae). — Jyväskylä Studies in Biological and Envi- — Ann. Entomol. Soc. Am. 87: 651–701. ronmental Science 132: 1–42. Stouthamer, R., Breeuwer, J. A. & Hurst, G. D. 1999: Wol- Kumpulainen, T., Grapputo, A. & Mappes, J. 2004: Parasites bachia pipientis: microbial manipulator of and sexual reproduction in psychid moths. — Evolution reproduction. — Annu. Rev. Microbiol. 53: 71–102. 58: 1511–1520. Suomalainen, E. 1962: Significance of parthenogenesis in Lanave, C., Preparata, G., Saccone, C. & Serio, G. 1984: A the evolution of insects. — Annu. Rev. Entomol. 7: new method for calculating evolutionary substitution 349–365. rates. — J. Mol. Evol. 20: 86–93. Suomalainen, E. 1970: Über die Solenobia-Arten Finnlands Lokki, J., Suomalainen, E., Saura, A. & Lankinen, P. 1975: (Lepidoptera: Psychidae). — Ann. Entomol. Fenn. 36: Genetic polymorphism and evolution in parthenogenetic 139–142. animals. II. Diploid and polyploid Solenobia triquetrella Suomalainen, E. 1980: The Solenobiinae species of Finland (Lepidoptera: Psychidae). — Genetics 79: 513–525. (Lepidoptera: Psychidae), with a description of a new Narbel, M. 1950: La cytologie de la parthénogénèse chez species. — Ent. Scand. 11: 458–466. Solenobia sp. (Lepidopteres Psychides). — Chromo- Tamura, K. & Nei, M. 1993: Estimation of the number of soma 4: 56–90. nucleotide substitutions in the control region of mito- Normark, B. & Lanteri, A. A. 1998: Incongruence between chondrial-DNA in humans and chimpanzees. — Mol. morphological and mitochondrial-DNA characters sug- Biol. Evol. 10: 512–526. gests hybrid origins of parthenogenetic weevil lineages Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, (Genus Aramigus). — Syst. Biol. 47: 475–494. F. & Higgins, D. G. 1997: The ClustalX windows inter- O’Brien, S. J. 1987: The ancestry of the giant panda. — Sci. face: flexible strategies for multiple sequence alignment Am. 257: 102–107. aided by quality analysis tools. — Nucl. Acids. Res. 24: Pearce, J. M., Fields, R. L. & Scribner, K. T. 1997: Nest 4876–4882. materials as a source of genetic data for avian ecological West-Eberhard, M. J. 2003: Developmental plasticity and studies. — J. Field Ornithol. 68: 471–481. evolution. — Oxford University Press, New York. Posada, D. & Crandall, K. A. 1998: Modeltest: testing West-Eberhard, M. J. 2005: The maintenance of sex as a the model of DNA substitution. — Bioinformatics 14: trap due to sexual selection. — Q. Rev. Biol. 80. [In 817–818. press].

This article is also available in pdf format at http://www.sekj.org/AnnZool.html