Department of Ecology and Systematics Division of Population Biology PO Box 65 (Viikinkaari 1) 00014 University of Helsinki

INBREEDING DEPRESSION IN THE ( cinxia)

Sari Haikola

Academic dissertation To be presented, with permission of the Faculty of Science of the University of Helsinki, for public criticism in the Auditorium 1041 of Viikki Biocentre 2 (Viikinkaari 5) on December 19th 2003 at 12 o‘clock noon.

Helsinki 2003 © Kluwer Academic Publishers (I) © Finnish Zoological and Botanical Publishing Board (II) © authors (other parts)

Cover photo by Tapio Gustafsson

Author´s address:

Department of Ecology and Systematics Division of Population Biology PO Box 65 (Viikinkaari 1) 00014 University of Helsinki Finland e-mail: [email protected]

ISBN 952-91-6679-6 (Paperback) ISBN 952-10-1501-2 (PDF, http://ethesis.helsinki.fi)

Yliopistopaino Helsinki 2003 3 INBREEDING DEPRESSION IN THE GLANVILLE FRITILLARY BUTTERFLY (MELITAEA CINXIA)

Sari Haikola

The thesis is based on the following articles or manuscripts:

I Haikola, S., Fortelius, W., O‘Hara, R.B., Kuussaari, M., Wahlberg, N., Saccheri, I.J., Singer, M.C. and Hanski, I. 2001. Inbreeding depression and the maintenance of genetic load in Melitaea cinxia metapopulations. Conservation Genetics 2: 325–335.

II Haikola, S. 2003. Effects of inbreeding in the Glanville fritillary butterfly (Melitaea cinxia). Annales Zoologici Fennici 40: 483–493.

III Haikola, S. 2003. Effects of larval group size and host plant on inbreeding depression in the Glanville fritillary butterfly. Manuscript.

IV Haikola, S. and Sarhan, A. 2003. Parental similarity, heterozygosity and offspring fitness in the Glanville fritillary butterfly (Melitaea cinxia). Manuscript.

V Haikola, S., Singer, M.C. and Pen, I. 2003. Has inbreeding depression led to avoidance of sib mating in the Glanville fritillary butterfly (Melitaea cinxia)? Submitted.

These are referred to by their Roman numerals in the text. 4 CONTRIBUTIONS

The following table shows the major contributions of authors to the original articles or manuscripts.

I II III IV V Original idea IH SH SH IH, SH SH Study design IH, IS SH SH SH SH, MS Data gathering MK, WF, MS, SH SH SH, AS SH, MS, IP SH, NW Analyses SH,RO SH SH SH SH Manuscript SH SH SH AS, SH SH, MS preparation

AS = Alia Sarhan, IH = Ilkka Hanski, IP = Ido Pen, IS = Ilik. J. Saccheri, MK = Mikko Kuussaari, MS = Michael C. Singer, NW = Niklas Wahlberg, RO = Robert B. O‘Hara, WF = Wilhelm Fortelius

Supervised by Prof. Ilkka Hanski University of Helsinki FINLAND

Reviewed by Jukka Jokela University of Oulu FINLAND

Hanna Kokko University of Helsinki FINLAND

Examined by Jouni Aspi University of Oulu FINLAND 5 CONTENTS

0 Summary Introduction 7 The study species 9 Overview of this thesis 10 Laboratory experiments 11 Results and discussion 13 Presence and severity of inbreeding depression in M. cinxia 13 Purging/maintenance of genetic load 14 The relationship between heterozygosity and parental relatedness, population connectivity and offspring fitness 14 No inbreeding avoidance 15 General discussion and conclusions 15 Acknowledgements 17 References 18

I Inbreeding depression and the maintenance of genetic load in Melitaea cinxia metapopulations Introduction Material and methods Laboratory experiments Mating success Experiment 1 Egg hatching rate Experiment 2 Experiment 3 Post- survival Experiment 4 Statistical analyses Mating success Egg hatching rate Results Mating success Egg hatching rate Experiment 2 Experiment 3 Larval survival and growth Discussion Acknowledgements References

II Effects of inbreeding in the Glanville fritillary butterfly (Melitaea cinxia) Introduction Material and methods Laboratory experiments 6

Statistical analyses The effect of the oviposition delay Comparison of the effects of inbreeding between the two generations The importance of the inbreeding history of the two sexes Results The effect of the oviposition delay Inbreeding depression in the two generations Discussion Acknowledgements References

III Effects of larval group size and host plant species on inbreeding depression in the Glanville fritillary butterfly Introduction Material and methods Material 1 Material 2 Statistical analysis Results Discussion Acknowledgements References

IV Parental similarity, heterozygosity and offspring fitness in the Glanville fritillary butterfly (Melitaea cinxia) Introduction Material and methods Study species Laboratory experiments Molecular methods Statistical analyses Results Discussion Acknowledgements References

V Has inbreeding depression led to avoidance of sib mating in the Glanville fritillary butterfly (Melitaea cinxia)? Introduction Methods Laboratory experiments Field experiments Results Discussion Acknowledgements References Introduction 7

INTRODUCTION worth 1987). Both hypotheses are based on the fact that inbreeding increases ho- Ongoing habitat loss and fragmentation mozygosity. According to the partial have caused the extinction of innumera- dominance hypothesis, inbreeding de- ble natural populations and reduced the pression is due to the expression of dele- sizes of many others (Frankham 1995). terious recessive alleles in homozygous Small and isolated populations are at risk condition in inbred individuals. The to go extinct due to several demographic overdominance hypothesis suggests that and environmental factors. For example, inbreeding depression is attributable to the extinction risk of small local popula- overdominance, the heterozygote superi- tions of the Glanville fritillary butterfly ority over homozygous genotypes (Char- (Melitaea cinxia) in the Åland Islands in lesworth and Charlesworth 1987). These southwest Finland is affected by demo- two hypotheses differ in their predictions graphic and environmental stochasticity, for the effects of prolonged inbreeding. habitat loss, parasitism, emigration from The partial dominance hypothesis pre- small habitat patches and so forth dicts that the genetic load decreases dur- (Hanski 1998). Genetic factors are also ing continuous inbreeding due to selec- important in affecting the dynamics of tion purging deleterious recessive alleles small local populations of M. cinxia from the population. On the contrary, if (Saccheri et al. 1998). Evidently, these inbreeding depression were due to over- factors are not independent: for instance, dominance, the genetic load would not patch quality may affect population size, be reduced in later rounds of inbreeding which in turn affects the genetic proper- (except in the case of asymmetrical over- ties of the population. dominance, Barrett and Charlesworth Inbreeding, mating with close rela- 1991). The partial dominance hypothesis tives, is thought to have harmful effects has gained much support and is widely on general viability of individuals. The thought to explain most of the inbreed- reduced fitness of inbred individuals rel- ing depression that occurs in natural ative to more outbred ones is called in- populations (Charlesworth and Charles- breeding depression (Hedrick and Kali- worth 1987). nowski 2000). Traits that show inbreeding The existence and severity of inbreed- depression are often closely related to fit- ing depression in different organisms ness (Frankel and Soulé 1981, Falconer have been mainly studied in laboratory/ and MacKay 1996). Thus many life-his- captivity. The importance of inbreeding tory traits of a species may be adversely in captive populations of many species affected by inbreeding, whereas metric has been firmly established (see for ex- characters may not show any inbreeding ample Ralls and Ballou 1983, Ralls et al. depression. For example, Roff (1998) 1988, Dole and Ritland 1993). Ralls et al. found that inbreeding significantly de- (1988) examined the level of inbreeding creases fecundity in the sand cricket, depression in 40 captive mammalian Gryllus firmus, but morphological traits populations. The cost of close inbreeding such as egg length did not show any or on juvenile survival was very high, al- only very weak inbreeding depression. though the severity of inbreeding de- Similarly, the effects of inbreeding can pression varied greatly between popula- differ between early and late life-stages tions (Ralls et al. 1988). The role of (Husband and Schemske 1995). inbreeding depression in natural popula- There are two main hypotheses to tions is more difficult to assess (Pusey explain the occurrence of inbreeding de- and Wolf 1996). Traditionally, the level pression (Charlesworth and Charles- of inbreeding in a given population has 8 Introduction

been determined directly using pedigree other studies such an effect was not information. With new molecular/genet- found (Dahlgaard and Loeschcke 1997, ic methods, the level of inbreeding in a Armbruster et al. 2000). In Drosophila population can also be assessed indirect- melanogaster inbreeding increases sus- ly using molecular markers (Box 1, see ceptibility to several stresses such as des- also for example Bensch et al. 1994, iccation (Dahlgaard and Hoffmann Amos et al. 2001). There are also prob- 2000). lems in assessing the level of inbreeding In theory, inbreeding depression from genetic data (Hansson and Wester- could be alleviated during repeated berg 2002), but the new techniques have rounds of inbreeding. The hypothesis of enabled researchers to study the effects continuous inbreeding purging the dele- of inbreeding in various species from terious recessive alleles from a popula- which it is not possible to obtain detailed tion has been studied in many species pedigree information. This is in fact the (Latta and Ritland 1994, Ballou 1997, case with most natural populations of Lacy and Ballou 1998), and there is some and plants. Presently studies on evidence for this (Carr and Dudash inbreeding effects in natural populations 1996). However, the strength of the ex- are rapidly accumulating. pected purging effect varies from one sit- The results of studies on inbreeding uation to another (Byers and Waller depression in natural populations have 2001), and many populations that have been variable (evidence of inbreeding experienced high level of inbreeding in depression found: Bensch et al. 1994, the past still suffer from severe inbreed- Keller et al. 1994, Brown and Brown ing depression (Charlesworth and Char- 1997, Keller 1998, Amos et al. 2001, Kel- lesworth 1987). Therefore, intentional ler et al. 2002, no inbreeding depression inbreeding of captive populations to found: Gibbs and Grant 1989, Keane et purge their genetic load may not be ad- al. 1996). However, it is safe to conclude visable (Ballou 1997, Hedrick and Ka- by now that inbreeding depression is an linowski 2000). Furthermore, even if in- important factor in many natural popu- breeding depression has been purged to lations (Crnokrak and Roff 1999, some degree in one environment, the ad- Frankham 1995), even affecting their verse effects of inbreeding may still be persistence (Saccheri et al. 1998). It has severe when conditions change (Bijlsma been suggested that inbred individuals et al. 1999). are more susceptible to harsh environ- Of particular interest in the context mental conditions than outbred ones. of inbreeding depression and purging of This means that inbreeding depression the genetic load are clusters of small ex- could be stronger under natural condi- tinction-prone populations that are con- tions than in captivity, where the condi- nected to each other via migration, i.e. tions are often near optimal (Hedrick metapopulations. Small populations in and Kalinowski 2000). Results of several metapopulations may be expected to suf- studies support the idea of stronger in- fer from inbreeding depression. On the breeding depression in natural than in other hand, migration from other nearby captive populations and/or increased populations may alleviate its impact in susceptibility of inbred individuals to en- particular populations (Thrall et al. vironmental stress (Heschel and Paige 1998, Richards 2000a,b, Ingvarsson 2001, 1995, Ramsey and Vaughton 1998, McCauley et al. 2001), which in turn Crnokrak and Roff 1999, Dahlgaard and may lower the effectiveness of purging in Hoffmann 2000, Hedrick and Kalinowski the metapopulation as a whole. 2000, Keller et al. 2002), though in some As the negative effects of inbreeding Introduction 9 are severe in many species, one might as- over its entire life-cycle. Consequently, sume that individuals in these species females might try to avoid mating with would try to avoid inbreeding. Many close relatives. To examine this issue, we species do indeed avoid inbreeding either conducted mate choice experiments, directly or indirectly (Koenig and Pitelka both in the laboratory and in the field, in 1979, Cockburn et al. 1985, Krackow and which females had an equal opportunity Matuschak 1991, Simmons 1991, Hoog- to mate with their sibs and non-sibs. land 1992, Keller and Passera 1993, Pu- sey and Wolf 1996, Dobson et al. 1997). Thus in the field cricket, Gryllus bimacu- latus, females mated preferentially with THE STUDY SPECIES unrelated males. The females also disc- minated against related males in other The Glanville fritillary butterfly, Melitaea ways, such that unrelated males had cinxia, is an endangered species in Fin- higher expected fertilization success land. Nowadays it occurs only in the (Simmons 1991). The existence of any Åland Islands, in southwestern part of mechanism to avoid inbreeding means the country (Marttila et al. 1990). In Fin- that even if a species shows strong in- land, M. cinxia has a one-year life-cycle. breeding depression in laboratory exper- The adult fly in June (Kuus- iments, inbreeding depression may not saari 1998). The females usually mate be an important factor in natural popu- only once (Kuussaari et al. 1998, Hanski lations as actual inbreeding may be very 1999: 207–232), although in high density uncommon even if the populations were mating frequency may be higher (Saasta- small. moinen, M. and Hanski, I., unpubl.). The aim of this study was to examine Mated females oviposit underneath the the existence and severity of inbreeding leaves of the larval host plants, Plantago depression in the Glanville fritillary but- lanceolata and . The eggs terfly, Melitaea cinxia, during its entire are laid in large clusters of 100–200 eggs. life cycle. The research has been con- The larvae hatch in early July and live ducted under controlled laboratory con- gregariously until early September, when ditions. In addition, I examined the sus- they start spinning a winter “nest”, a ceptibility of inbred individuals to dense silken web inside which they dia- adverse conditions, namely reduced lar- pause over winter in a group. In the val group size and rearing on a less suita- spring larvae continue feeding gregari- ble host plant species. Saccheri et al. ously until the last instar. Pupation takes (1998) found a relationship between the place in early May (Kuussaari 1998, level of homozygosity (used as a measure Hanski 1999: 207–232). of inbreeding) and extinction risk of In any one larval group, the individ- small natural populations of M. cinxia in uals are full siblings, with few exceptions the Åland Islands. The increased extinc- (two or more females have oviposited on tion risk of inbred M. cinxia populations the same host plant, Singer, M., pers. was later confirmed experimentally obs.). The size of the larval group is im- (Nieminen et al. 2001). To relate the re- portant for both early and late larval sur- sults of the laboratory experiments to vival, as bigger groups and larvae in big- these findings, we assessed the relation- ger groups have higher survival than ship between heterozygosity and related- small groups and larvae in small groups ness of the parents and several fitness-re- (Kuussaari 1998). Similarly, larval sur- lated traits of the offspring. Inbreeding vival rate may to some extent depend on affects adversely many traits of M. cinxia the host plant species used, survival be- 10 The study species

ing often though not always higher on V. hatching rate is an important fitness spicata than on P. lanceolata (van component and shows inbreeding de- Nouhuys et al. 2003). pression in many species (van Noordwijk The M. cinxia metapopulation in the and Scharloo 1981, Bensch et al. 1994, Åland Islands has been intensively stud- Saccheri et al. 1996, Su et al. 1996, Keller ied since 1991 (Hanski 1999: 207–232). 1998, Higashiura et al. 1999, Weeks et al. The larval winter webs are relatively easy 1999). We studied the effects of close in- to observe in autumn, which enables re- breeding (full-sib mating) during one cording of the numbers of larval families generation on egg hatching rate in M. in each population each year. The nests cinxia. To test the idea of inbreeding observed in autumn are checked in the purging the genetic load during continu- spring to record survival over diapause. ous inbreeding, the level of inbreeding Several features of the habitat patches are depression was compared between two also recorded. The population structure regions with different frequency of natu- is highly fragmented (Hanski 1999: 207– ral inbreeding, a highly fragmented land- 232, Nieminen et al. 2004). The genetic scape in Finland (high incidence of in- structure of the M. cinxia metapopula- breeding because of highly fragmented tion is consistent with the fragmented population structure) and more contin- habitat distribution. In a sample of 1216 uous landscape in southern France (low- larvae from 369 local populations scored er incidence of inbreeding; I). at six polymorphic allozyme loci and Heterozygosity has been used as an two polymorphic microsatellite loci, the indicator of the level of inbreeding in

global FST value among all local popula- small local populations of M. cinxia in tions was ~0.1. This demonstrates that Åland (Saccheri et al. 1998). In the first local populations are genetically differ- (I) chapter, the effects of female hetero- entiated. Furthermore, the pairwise ge- zygosity on several offspring traits were netic distance increases with increasing examined. In addition, the mating suc- geographic distance between pairs of cess of individuals originating from populations (Saccheri et al. 2004). Pres- small and isolated, presumably inbred ently ca. 4000 habitat patches are known populations, was compared with that of in Åland and monitored on a yearly ba- individuals from large and well-connect- sis. This large database and the detailed ed, presumably more outbred popula- knowledge of the biology of M. cinxia tions. provide a unique opportunity to study In the second (II) chapter I investi- genetic issues in the context of impor- gated whether the level of inbreeding de- tant demographic and ecological features pression differs between the first and the of an endangered species. second generation of full-sib mating. I also studied previously unstudied stages in the life-cycle at which the adverse ef- fects of inbreeding might be manifested. OVERVIEW OF THIS THESIS In the third (III) chapter, an experi- ment on larval survival was conducted to In the first (I) chapter, we wanted to ex- confirm previous findings of negative ef- amine the existence of inbreeding de- fects of inbreeding (II) and small larval pression in M. cinxia. Generally, traits group size (Kuussaari 1998) on early lar- that are closely related to fitness are ex- val survival. In addition, I wanted to find pected to be most susceptible to inbreed- out whether inbred and outbred individ- ing depression (Frankel and Soulé 1981, uals would differ in their responses to Falconer and MacKay 1996). The egg adverse conditions. Specifically, I exam- Overview of this thesis 11 ined whether inbred individuals are cannot be much shortened from the nat- more susceptible than outbred ones to ural length of one year. The larval groups the adverse effect of reduced larval group in the laboratory are susceptible to viral size. I also examined whether the larval outbreaks and the larvae often suffer survival differed between the two host high mortality during diapause. The but- plant species, P. lanceolata and V. spicata, terflies do not readily mate under artifi- and whether there was any interaction cial light and there are often difficulties between host plant and the inbreeding to get mated females to lay eggs without status of larvae. bright natural light. This means that The fourth (IV) chapter is focused conducting large-scale breeding experi- more on the influence of population ments with M. cinxia in the laboratory is structure and dynamics on inbreeding not possible. However, given the interest- depression in the M. cinxia metapopula- ing ecological setting and the detailed tion in the Åland Islands. The aim of this knowledge of the biology and popula- chapter is to connect the results of the tion dynamics of the species, even limit- laboratory studies on inbreeding depres- ed laboratory studies on the effects of in- sion to the finding of increased extinc- breeding are valuable. This thesis tion risk of less heterozygous, small nat- consists of small-scale laboratory studies ural populations (Saccheri et al. 1998). on several aspects of inbreeding and in- The effects of the heterozygosity and de- breeding depression in M. cinxia. gree of genetic similarity between par- Material for chapters I, II and IV was ents on their fertility/fecundity and off- originally collected as eggs or post-dia- spring fitness were examined. We also pause larvae from natural populations in assessed the effects of these genetic fac- the Åland Islands. In addition, for the tors in a more indirect way by studying comparison between two areas, Åland the importance of population connectiv- and southern France (II), mated butter- ity on inbreeding depression. flies from different areas in southern If a species possesses any inbreeding France were collected. The material in avoidance mechanism, natural popula- chapter IV is a subset of the material in tions would not greatly suffer from in- the chapter II. In experiment 4 in chap- breeding even if inbreeding depression ter I, mothers of the larvae used in the would be severe when tested under labo- experiment were collected from natural ratory conditions. In the last (V) chapter populations in the Åland Islands. In we studied whether M. cinxia avoids chapter III the material originated partly mating with close kin to escape the nega- from previous laboratory populations tive effects of inbreeding that are evident and partly from natural populations. from previous experiments (I, II). The rearing procedure was generally as follows, with minor modifications in each experiment. The post-diapause lar- vae were reared on potted host plants or LABORATORY in small plastic boxes (in which case lar- EXPERIMENTS vae were fed with cut host plant leaves) in the laboratory under artificial light. There are several difficulties in studying One day after pupation, the pupae were the effects of inbreeding in M. cinxia in gently removed to plastic boxes covered the laboratory. The larvae have to go with a nylon mesh. Newly-emerged but- through a diapause that lasts for several terflies were removed to net cages and months in order to develop into pupae. fed with mild honeywater. Matings were This means that the generation length mostly conducted indoors and mated fe- 12 Laboratory experiments

males were placed into small net cages to lay eggs. The cages contained either a liv- ing host plant or a cut leave of the host plant. Egg batches were collected daily or every second day and placed onto petri dishes, one group per dish. After a cou- ple of days the group was gently spread out with a brush. An enlarged photocopy of the dish was taken, from which the clutch size (number of eggs) was counted (Saccheri et al. 1998, Fig. 1). In most cas- es the number of hatched larvae was counted in a similar fashion. The pre-di- apause larvae were reared on potted host plants. When ready for diapause, the lar- val groups were placed into small film cans, which were kept either indoors at 2–5°C or in an outdoor shed over the winter. In experiment 4 in chapter I, the av- erage heterozygosity of the mothers at seven enzyme loci and one microsatellite locus was scored and their offspring were placed in groups on a large meadow on an isolated island in soutwestern Fin- land, to have them in a common natural environment. In the spring the larvae were collected and reared to butterflies in the laboratory (I). In chapter III new- Figure 1. Photocopies of petri dishes contain- ly-hatched larvae were reared in groups ing eggs (upper picture) and newly-hatched of 10 or 50 on a potted host plant in the larvae (lower picture). Each counted egg/larva laboratory. Survival rate and the quality has been marked with a pen. of the silken web of each group were as- sessed. In chapter V each female was given 120

an equal opportunity to mate with her 100 brothers or with unrelated males. This was done in large mating cages in the 80 laboratory with Finnish butterflies and 60 in the field with French butterflies. 40 Frequency

20

0 1234567891011121314 Number of larval nests

Figure 2. Number of larval groups in newly- founded populations of M. cinxia in the first au- tumn after colonization (data for 2002, n=205). Results and discussion 13

RESULTS AND DISCUSSION grants in a mark-recapture study was 330 m, and the maximum distance was Presence and severity about 3 km (Hanski et al. 1994, 1995, of inbreeding depression Hanski 1999: 207–232). in M. cinxia The new populations are likely to be established by a single female from a The newly-established populations of M. nearby and most probably a relatively cinxia in the Åland Islands mainly con- large population. In this case the off- sist of one larval (Fig. 2). In gen- spring are presumably relatively outbred eral, the size of local populations is and the level of inbreeding is not expect- small, with the mean of 4.3 larval groups ed to differ from the overall level of in- in years 1993–2002. These figures sug- breeding in the large populations. On gest that inbreeding is frequent in the the other hand, small populations that metapopulation. The population turno- have consisted of only a few larval ver rate is high: each year about 40% of groups for two or more generations are the populations go extinct and about expected to be more or less inbred. This 10% of currently empty meadows be- means that the average heterozygosity of come colonised. Migration as well as col- the individuals in these two types of onisation distances are mostly short populations should differ: the newly- (Hanski et al. 1994, 1995, Hanski founded and large old populations are 1999: 207–232, Nieminen et al. 2004, Fig. expected to be less homozygous than the 3): the median distance moved by immi- old small populations. Saccheri et al.

Figure 3. The upper panels show the minimum colonisation distances (distance to the nearest population) of M. cinxia in years 1999-2001. The black dots give the ratio of colonised patches over the total number of empty patches in each distance class. The lower panels give the number of patches that remained empty (gray bars) and the number of colonised patches (black bars) in each connectivity class (connectivity measures the effective distance to all possible source populations and decreases to the right in this figure). From Nieminen et al. (2004). 14 Results and discussion

(1998) found that the less heterozygous Purging/maintenance populations have increased extinction of genetic load risk compared to more heterozygous populations. The increased extinction To examine whether inbreeding leads to risk of more homozygous M. cinxia pop- purging of deleterious recessive alleles ulations has been experimentally con- from populations, we compared the se- firmed by Nieminen et al. (2001). verity of inbreeding depression between Inbreeding has adverse effects on two areas that differ in their overall pop- several fitness-related traits in M. cinxia ulation structure and presumed inbreed- (I). One generation of full-sib mating ing history (I). In the metapopulation in decreases egg hatching rate (I, II) and Åland, a large fraction of individuals oc- larval survival (II, III). After another curs at any one time in small popula- generation of full-sib mating there is a tions and are exposed to inbreeding. In marked increase in the proportion of contrast, in southern France the popula- completely infertile pairings (II). Simi- tion structure is much more continuous larly to M. cinxia, the butterfly Bicyclus (I). Therefore, one could assume that anynana shows severe inbreeding de- purging should have been more efficient pression in egg hatching rate (Saccheri et in Åland, which indeed appears to be the al. 1996). In B. anynana, there was a case. Full-sib mating lowered egg hatch- marked increase in the proportion of en- ing rate significantly more in France tirely infertile clutches when the male (~40% decrease) than in Åland (~10% parent was inbred (Saccheri, I.J., Lloyd, decrease, I). However, part of the genetic H.D., Helyar, S.J. and Brakefield, P.M., load is still maintained in the Åland Is- unpubl.). Although my sample size was lands, as the inbreeding depression in egg very small, there was a similar tendency hatching rate is still substantial (I, II). for infertility of the clutches of inbred This could be explained by part of the males in M. cinxia (II). The average genetic load being due to overdominance. clutch size decreases due to inbreeding in It is more likely, however, that migration M. cinxia, which effect is seen only when between populations that carry different the parents are inbred (II). Furthermore, deleterious recessives could impair the butterflies originating from small and effectiveness of purging and thus main- presumably inbred populations suffer tains the genetic load in the system (I). from lowered mating success compared to more outbred butterflies from large populations (I). This indicates that the mating success may be negatively affect- The relationship between ed by inbreeding. Furthermore, there heterozygosity and parental was a tendency for the inbred groups to relatedness, population suffer more from adverse conditions, i.e. connectivity and offspring from reduced larval group size and rear- fitness ing on a less suitable host plant species, P. lanceolata. This implies that in natural Considering the population structure of populations, which suffer from stressful M. cinxia in the Åland Islands, one can environmental conditions, the effects of assume that average heterozygosity is a inbreeding could be even more severe relatively reliable measure of inbreeding than in the studies reported here. in the system (Saccheri et al. 1998). Sac- cheri et al. (1998) studied the relation- ship between population heterozygosity and extinction risk. After accounting for Results and discussion 15 relevant ecological factors, there was a larval survival and overall fitness in this significant negative relationship between population class (IV). Population con- heterozygosity and the extinction risk of nectivity index (Si, Moilanen and Niem- local populations (Saccheri et al. 1998). inen 2002) was not related to the genetic In our laboratory study, female heterozy- measures or to any of the fitness traits gosity was positively associated with lar- (IV). This may be due to the small size of val weight following diapause and nega- the populations studied, most popula- tively associated with the length of the tions consisting of only one or two larval pupal period (I). The latter can be an groups. Among such small populations important factor in natural populations frequent random extinction and recolo- as long pupal period may increase the nisation events may obscure the impact risk of parasitism (Lei et al. 1997). of population connectivity. In order to further clarify the rela- tionship between the laboratory results on inbreeding depression (I) and the as- No inbreeding avoidance sociation between heterozygosity and ex- tiction risk in the field (Saccheri et al. Given the strength of inbreeding depres- 1998, Nieminen et al 2001), we exam- sion in M. cinxia that was discovered in ined the relationship between genetic laboratory experiments (I, II, III, IV), parameters – heterozygosity and parental one might expect that some mechanism relatedness – and offspring fitness (IV). to escape the negative effects of inbreed- In this case no significant association be- ing would have evolved in this species. tween heterozygosity and fitness was However, based on the results of chapter found. In contrast, increasing parental V it is apparent that inbreeding avoid- relatedness was significantly associated ance does not occur in M. cinxia. This with lowered egg hatching rate and larval can perhaps be understood considering survival. Considering this result and the the small size of local populations in findings of other laboratory studies (I, II, Åland. Kuussaari et al. (1998) found that III), the interpretation by Saccheri et al. the fraction of mated females was related (1998) of inbreeding causing increased to population size, such that in small extinction risk of small populations is populations the fraction of mated fe- supported. One should also take into ac- males was lower than in large popula- count that inbreeding depression is likely tions. If close relatives were avoided as to be more severe in the field than in the mates, the females in small inbred popu- laboratory. Furthermore, as inbreeding lations would be at risk of finding no decreases egg hatching rate, the inbred mate at all, which would be even worse larval groups are expected to be smaller than to suffer from the consequences of than outbred ones. As small larval groups inbreeding. experience elevated mortality, there is fur- ther amplification of the negative effects of inbreeding on larval survival in natural GENERAL DISCUSSION populations (III, Kuussaari 1998). AND CONCLUSIONS In accordance with our expectation, old small populations in Åland were on Inbreeding depression has recently at- average more homozygous than small tracted much attention in the context of newly-founded populations (IV). Simi- metapopulations. Severe inbreeding de- larly, parental relatedness was highest in pression has been reported for example the old small populations, which was as- in studies on Daphnia (Ebert et al. 2002, sociated with lowered egg hatching rate, Haag et al. 2002) and Silene alba metap- 16 General discussion and conclusions

opulations (Thrall et al. 1998, Richards Inbreeding is expected to be frequent 2000a,b, McCauley et al. 2001). in small local populations of M. cinxia in The rockpool metapopulation of D. the Åland Islands. Therefore, selection magna, a freshwater crustacean, in the during continuous inbreeding could in archipelago in southern Finland, shares theory have purged the genetic load due several similar features with M. cinxia to deleterious recessive alleles from the metapopulation in Åland. Like the latter metapopulation. However, this study species, D. magna has a highly fragment- shows that even though purging has to ed metapopulation structure and high some extent occurred, part of the genetic turnover rate: each year about 20% of load still remains in the system (I). In the the populations go extinct and 5% of the laboratory, several different traits at vari- empty patches are colonised (Ebert et al. ous life stages show strong inbreeding 2002, Haag et al. 2002). The colonisa- depression. These traits include impor- tions may be accomplished by only one tant fitness components such as egg or a few individuals as is also the case in hatching rate and larval survival (I, II, M. cinxia in Åland. Consequently, in- III, IV, Table 1). Inbreeding decreases breeding is common in the D. magna both early larval survival and survival metapopulation. Under competitive con- over diapause. In chapter I, there was an ditions, inbreeding depression is severe, indication of lowered larval weight in the at least partly due to reduced fertility of spring in the progeny of less hetero- inbred individuals (Haag et al. 2002). zygous compared to more heterozygous The plant S. alba often has a patchy females, though Nieminen et al. (2001) distribution. In a long-term study of its found no effect of inbreeding on post- metapopulation in Virginia, the popula- diapause larval weight. Results on in- tion turnover rate was found to be high breeding effects on adult longevity are (Thrall et al. 1998, Richards 2000a,b). also unclear. In the laboratory, no in- Newly-founded populations usually con- breeding depression in longevity was sists of less than six individuals. In spite found, whereas the field data suggested of frequent occurrence of inbreeding, that more heterozygous females live the metapopulation shows substantial in- longer than less heterozygous females. In breeding depression (Richards 2000b), brief, inbreeding depression is manifest- just as is the case with M. cinxia. Inbreed- ed over the entire life-cycle of M. cinxia ing depression in S. alba is manifested as (Table 1). As there is no discrimination lowered germination success of inbred in- against close relatives as mates, inbreeding dividuals. Gene flow between populations depression is expected to be an influential alleviates inbreeding depression, but is factor also in natural populations (V). also presumed to maintain genetic load in Saccheri et al. (1998) found that in- the metapopulation (Richards 2000b). creased homozygosity was associated with The three metapopulations described elevated extinction risk of small popula- above, those of M. cinxia, D. magna and tions of M. cinxia. This study has clarified S. alba, share several common features. the relationship between homozygosity All have high yearly extinction risk and and extinction risk of small natural popu- highly fragmented population structure. lations. The overall conclusion is that in- Newly-founded populations are generally breeding is likely to have a substantial im- small and inbreeding is expected to be pact on the dynamics of the M. cinxia frequent in all three systems. In addition, metapopulation in the Åland Islands. all of these metapopulations show strong inbreeding depression. General discussion and conclusions 17

Table 1. Effects of inbreeding on fitness-related traits of M. cinxia over its entire life-cycle. A ques- tion mark refers to controversial or unclear results.

Trait Inbreeding depression References Egg hatching rate YES (I, II, IV, Nieminen et al. 2001) Pre-diapause survival YES (II, III, IV) Survival over diapause YES (Nieminen et al. 2001) Post-diapause weight ? (I, Nieminen et al. 2001) Larval development rate NO (I) Survival from diapause to pupa NO (I) Pupal weight NO (I) Pupal development rate YES (I) Survival from pupa to adult NO (I) Adult longevity ? (I, Saccheri et al. 1998, Nieminen et al. 2001) Mating success YES (I) Clutch size YES (II)

ACKNOWLEDGEMENTS from Åland. Also several other people have been involved in collecting larval material, First I want to thank my supervisor Ilkka thanks to everyone. I want also to thank the Hanski for his guidance over the four years girls in our office – Leena, Anna-Liisa, Jenni that I have been struggling with this study. and Tarja – for friendship and support. Everything has not gone smoothly, there have Thanks to others in MRG as well! been enormous difficulties in conducting the Riitta Rantala worked with me in the experiments: occasionally a viral outbreak Tvärminne Zoological Station and was a destroyed a whole butterfly generation (and good companion – thanks to Riitta. The staff ruined a months‘ work), often the experi- in Tvärminne were very helpful and deserve ments failed due to difficulties in getting the warm thanks. Several other people – Wilhelm butterflies to reproduce etc. I have to admit Fortelius, Mikko Kuussaari, Niklas Wahlberg, that I have been desperate, and more than Mike Singer and many others – have also giv- once the thought has crossed my mind, that I en their contribution to the experiments and should give up… Then Ilkka‘s support and I want to thank them all. Special thanks to encouragement has helped me to come to Mike! second thoughts. I just can not express how In the darkest moments I wouldn‘t have grateful to him I am now! I have been really managed without my parents and my sister lucky to have had such an excellent supervi- Päivi. Their support and love have been cru- sor! cial – it means everything to me! I also want I am grateful to the whole Metapopula- to thank my closest friend Jonna who I have tion Research Group: it has been a privilege known since we were small kids attending the to belong to such a group of wonderful peo- same kindergarten. She has been a wonderful ple. Whenever I have needed help there have friend and has given her support whenever I always been people that are ready to share have needed it the most! their knowledge and skills. Especially I want The work was financially supported by to thank Bob O‘Hara for assistance with the the University of Helsinki, the Academy of statistics part of my work and Marko Nie- Finland and the EU TMR network Fragland. minen who has collected material for me 18 Acknowledgements

REFERENCES Review of Ecology and Systematics 18: 237–268. Amos, W., Wilmer, J.W., Fullard, K., Burg, Cockburn, A., Scott, M.P. and Scotts, D.J. T.M., Croxall, J.P., Bloch, D. and Coul- 1985. Inbreeding avoidance and male- son, T. 2001. The influence of parental biased natal dispersal in Antechinus relatedness on reproductive success. spp. (Marsupialia: Dasyuridae). Ani- Proceedings of the Royal Society of mal Behaviour 33: 908–915. London, Series B: Biological Sciences Coltman, D.W., Pilkington, J.G., Smith, J.A. 268: 2021–2027. and Pemberton, J.M. 1999. Parasite- Armbruster, P., Hutchinson, R.A. and Lin- mediated selection against inbred soay vell, T. 2000. Equivalent inbreeding de- sheep in a free-living, island popula- pression under laboratory and field tion. Evolution 53: 1259–1267. conditions in a tree-hole-breeding Coulson, T.N., Pemberton, J.M., Albon, mosquito. Proceedings of the Royal S.D., Beaumont, M., Marshall, T.C., Society of London, Series B: Biological Slate, J., Guinness, F.E. and Clutton- Sciences 267: 1939–1945. Brock, T.H. 1998. Microsatellites reveal Ballou, J.D. 1997. Ancestral inbreeding heterosis in red deer. Proceedings of only minimally affects inbreeding de- the Royal Society of London, Series B: pression in mammalian populations. Biological Sciences 265: 489–495. Journal of Heredity 88: 169–178. Crnokrak, P. and Roff, D.A. 1999. Inbreed- Barrett, S.C.H. and Charlesworth, D. 1991. ing depression in the wild. Heredity Effects of a change in the level of in- 83: 260–270. breeding on the genetic load. Nature Dahlgaard, J. and Hoffmann, A.A. 2000. 352: 522–524. Stress resistance and environmental Bensch, S., Hasselquist, D. and von dependency of inbreeding depression Schantz, T. 1994. Genetic similarity be- in Drosophila melanogaster. Conserva- tween parents predicts hatching fail- tion Biology 14: 1187–1192. ure: nonincestuous inbreeding in the Dahlgaard, J. and Loeschcke, V. 1997. Ef- great reed warbler. Evolution 48: 317– fects of inbreeding in three life stages 326. of Drosophila buzzatii after embryos Bijlsma, R., Bundgaard, J. and van Putten, were exposed to high temperature W. 1999. Environmental dependence stress. Heredity 78: 410–416. of inbreeding depression and purging Dobson, F.S., Chesser, R.K., Hoogland, J.L., in Drosophila melanogaster. Journal of Sugg, D.W. and Foltz, D.W. 1997. Do Evolutionary Biology 12: 1125–1137. black-tailed prairie dogs minimize in- Brown, J.L. and Brown, E.R. 1997. Are in- breeding? Evolution 51: 970–978. bred offspring less fit? Survival in a Dole, J. and Ritland, K. 1993. Inbreeding natural population of Mexican jays. depression in two Mimulus taxa meas- Behavioral Ecology 9: 60–63. ured by multigenerational changes in Byers, D.L. and Waller, D.M. 2001. Do the inbreeding coefficient. Evolution plant populations purge their genetic 47: 361–373. load? Effects of population size and Ebert, D., Haag, C., Kirkpatrick, M., Riek, mating history on inbreeding depres- M., Hottinger, J.W. and Pajunen, V.I. sion. Annual Review of Ecology and 2002. A selective advantage to immi- Systematics 30: 479–513. grant genes in a Daphnia metapopula- Carr, D.E. and Dudash M.R. 1996. In- tion. Science 295: 485–488. breeding depression in two species of Falconer, D.S. and Mackay, T.F.C. 1996. In- Mimulus (Scrophulariaceae) with con- troduction to quantitative genetics, trasting mating systems. American 4th edition. Longman, Harlow, Eng- Journal of Botany 83: 586–593. land. Charlesworth, D. and Charlesworth, B. Frankel, O.H. and Soulé, M.E. 1981. Con- 1987. Inbreeding depression and its servation and evolution. Cambridge evolutionary consequences. Annual University Press, New York. References 19

Frankham, R. 1995. Conservation genetics. hypothesis. Trends in Ecology and Ev- Annual Review of Genetics 29: 305– olution 16: 62–63. 327. Keane, B., Creel, S.R. and Waser, P.M. 1996. Gibbs, H.L. and Grant, P.R. 1989. Inbreed- No evidence of inbreeding avoidance ing in Darwin‘s medium ground or inbreeding depression in a social finches (Geospiza fortis). Evolution 43: carnivore. Behavioral Ecology 7: 480– 1273–1284. 489. Haag, C.R., Hottinger, J.W., Riek, M. and Keller, L.F. 1998. Inbreeding and its fitness Ebert, D. 2002. Strong inbreeding de- effects in an insular population of pression in a Daphnia metapopulation. song sparrows (Melospiza melodia). Evolution 56: 518–526. Evolution 52: 240–250. Hanski, I. 1998. Metapopulation dynamics. Keller, L. and Passera, L. 1993. Incest avoid- Nature 396: 41–49. ance, fluctuating asymmetry, and the Hanski, I. 1999. Metapopulation ecology. consequences of inbreeding in Irid- Oxford University Press, Oxford. omyrmex humilis , an ant with multi- Hanski, I., Kuussaari, M. and Nieminen, ple queen colonies. Behavioral Ecology M. 1994. Metapopulation structure and Sociobiology 33: 191–199. and migration in the butterfly Melitaea Keller, L.F., Grant, P.R., Grant, B.R. and cinxia. Ecology 75: 747–762. Petren, K. 2002. Environmental condi- Hanski, I., Pakkala, T., Kuussaari, M. and tions affect the magnitude of inbreed- Lei, G. 1995. Metapopulation persist- ing depression in survival of Darwin‘s ence of an endangered butterfly in a finches. Evolution 56: 1229–1239. fragmented landscape. Oikos 72: 21– Keller, L.F., Arcese, P., Smith, J.N.M., 28. Hochachka, W.M. and Stearns, S.C. Hansson, B. and Westerberg, L. 2002. On 1994. Selection against inbred song the correlation between heterozygosity sparrows during a natural population and fitness in natural populations. bottleneck. Nature 372: 356–357. Molecular Ecology 11: 2467–2474. Koenig, W.D. and Pitelka, F.A. 1979. Relat- Hedrick, P.W. and Kalinowski, S.T. 2000. edness and inbreeding avoidance: Inbreeding depression in conservation counterploys in the communally nest- biology. Annual Review of Ecology ing acorn woodpecker. Science 206: and Systematics 31: 139–162. 1103–1105. Heschel, M.S. and Paige, K.N. 1995. In- Krackow, S. and Matuschak, B. 1991. Mate breeding depression, environmental choice for non-siblings in wild house stress, and population size variation in mice: evidence from a choice test and a scarlet gilia (Ipomopsis aggregata). reproductive test. Ethology 88: 99–108. Conservation Biology 9: 126–133. Kuussaari, M. 1998. Biology of the Glan- Higashiura, Y., Ishihara, M. and Schaefer, ville fritillary butterfly (Melitaea cinx- P.W. 1999. Sex ratio distortion and se- ia). PhD thesis, University of Helsinki. vere inbreeding depression in the gyp- Kuussaari, M., Saccheri, I., Camara, M. and sy moth Lymantria dispar L. in Hanski, I. 1998. Allee effect and popu- Hokkaido, Japan. Heredity 83: 290– lation dynamics in the Glanville fritil- 297. lary butterfly. Oikos 82: 384–392. Hoogland, J.L. 1992. Levels of inbreeding Lacy, R.C. and Ballou, J.D. 1998. Effective- among prairie dogs. American natural- ness of selection in reducing the genet- ist 139: 591–602. ic load in populations of Peromyscus Husband, B.C. and Schemske, D.W. 1995. polionotus during generations of in- Magnitude and timing of inbreeding breeding. Evolution 52: 900–909. depression in a diploid population of Latta, R. and Ritland, K. 1994. The rela- Epilobium angustifolium (Onagraceae). tionship between inbreeding depres- Heredity 75: 206–215. sion and prior inbreeding among pop- Ingvarsson, P.K. 2001. Restoration of ge- ulations of four Mimulus taxa. Evolu- netic variation lost-the genetic rescue tion 48: 806–817. 20 References

Lei, G.C., Vikberg, V., Nieminen, M. and of environment on the magnitude of Kuussaari, M. 1997. The parasitoid inbreeding depression in seed germi- complex attacking Finnish populations nation in a partially self-fertile peren- of the Glanville fritillary Melitaea cinx- nial herb (Blandfordia grandiflora, Lil- ia (Lep: ), an endangered iaceae). International Journal of Plant butterfly. Journal of Natural History Sciences 159: 98–104. 31: 635–648. Richards, C.M. 2000a. Genetic and demo- Marttila, O., Haahtela, T., Aarnio, H. and graphic influences on population per- Ojalainen, P. 1990. Suomen päiväper- sistence: gene flow and genetic rescue hoset. Kirjayhtymä Oy, Helsinki. in Silene alba. In Genetics, demogra- McCauley, D.E., Richards, C.M., Emery, phy and viability of fragmented popu- S.N., Smith, R.A. and McGlothlin, J.W. lations (Young, A.G. and Clarke, G.M., 2001. The interaction of genetic and eds.), pp. 271–290. Cambridge Univer- demographic processes in plant meta- sity Press, Cambridge. populations: a case study of Silene Richards, C.M. 2000b. Inbreeding depres- alba. In Integrating ecology and evolu- sion and genetic rescue in a plant tion in a spatial context (Silvertown, J. metapopulation. American naturalist and Antonovics, J., eds.), pp. 177–196. 155: 383–394. Blackwell Science. Roff, D.A. 1998. Effects of inbreeding on Moilanen, A. and Nieminen, M. 2002. Sim- morphological and life history traits of ple connectivity measures in spatial the sand cricket, Gryllus firmus. Hered- ecology. Ecology 83: 1131–1145. ity 81: 28–37. Nieminen, M., Siljander, M and Hanski, I. Saccheri, I.J., Brakefield, P.M. and Nichols, 2004. Structure and dynamics of Meli- R.A. 1996. Severe inbreeding depres- taea cinxia metapopulations. Chapter sion and rapid fitness rebound in the 4 in On the wings of checkerspots: a butterfly Bicyclus anynana (Satyridae). model system for population biology Evolution 50: 2000–2013. (Ehrlich, P.R. and Hanski, I., eds.). Ox- Saccheri, I.J., Boggs, C.L., Hanski, I. and ford University Press, in press. Ehrlich, P.R. 2004. Genetics of check- Nieminen, M., Singer, M.C., Fortelius, W., erspot populations. Chapter 10 in On Schöps, K. and Hanski, I. 2001. Experi- the wings of checkerspots: a model mental confirmation that inbreeding system for population biology (Ehr- depression increases extinction risk in lich, P.R. and Hanski, I., eds.). Oxford butterfly populations. American Natu- University Press, in press. ralist 157: 237–244. Saccheri, I., Kuussaari, M., Kankare, M., Pusey, A. and Wolf, M. 1996. Inbreeding Vikman, P., Fortelius, W. and Hanski, I. avoidance in animals. Trends in Ecolo- 1998. Inbreeding and extinction in a gy and Evolution 11: 201–206. butterfly metapopulation. Nature 392: Ralls, K. and Ballou, J. 1983. Extinction: 491–494. lessons from zoos. In Genetics and Simmons, L.W. 1991. Female choice and conservation: A reference for manag- the relatedness of mates in the field ing wild and plant populations cricket, Gryllus bimaculatus. Animal (Schonewald-Cox, C.M., Chambers, Behaviour 41: 493–501. S.M., MacBryde, B. and Thomas, W.L., Su, G.S., Liljedahl, L.E. and Gall, G.A.E. eds.), pp. 164–184. The Benjamin/ 1996. Effects of inbreeding on growth Cummings Publishing Company, and reproductive traits in rainbow Menlo Park, California. trout (Oncorhynchus mykiss). Aquacul- Ralls, K., Ballou, J.D. and Templeton, A. ture 142: 139–148. 1988. Estimates of lethal equivalents Thrall, P.H., Richards, C.M., McCauley, and the cost of inbreeding in mam- D.E. and Antonovics, J. 1998. Metap- mals. Conservation Biology 2: 185– opulation collapse: the consequences 193. of limited gene-flow in spatially struc- Ramsey, M. and Vaughton, G. 1998. Effect tured populations. In Modeling spati- References 21

otemporal dynamics in ecology (Bas- the suitability of two host plant spe- compte, J. and Solé, R.V., eds.), pp.83– cies. Ecological Entomology 28: 193– 104. Springer-Verlag, Berlin. 202. van Noordwijk, A.J. and Scharloo, W. 1981. Weeks, S.C., Marcus, V. and Crosser, B.R. Inbreeding in an island population of 1999. Inbreeding depression in a self- the great tit. Evolution 35: 674–688. compatible, androdioecious crusta- van Nouhuys, S., Singer, M.C. and Niemin- cean, Eulimnadia texana. Evolution 53: en, M. 2003. Spatial and temporal pat- 472–483. terns of caterpillar performance and