Sex-specific effects of inbreeding in wild-caught Drosophila melanogaster under benign and stressful conditions


Laramy S. Enders, Department of Ecology, Evolution and Organismal Biology, University of California, Riverside, CA 92507, USA.
Tel.: +1 951 827 2023; fax: +1 951 827 4286;


In animal populations, sib mating is often the primary source of inbreeding depression (ID). We used recently wild-caught Drosophila melanogaster to test whether such ID is amplified by environmental stress and, in males, by sexual selection. We also investigated whether increased ID because of stress (increased larval competition) persisted beyond the stressed stage and whether the effects of stress and sexual selection interacted. Sib mating resulted in substantial cumulative fitness losses (egg to adult reproduction) of 50% (benign) and 73% (stressed). Stress increased ID during the larval period (23% vs. 63%), but not during post-stress reproductive stages (36% vs. 31%), indicating larval stress may have purged some adult genetic load (although ID was uncorrelated across stages). Sexual selection exacerbated inbreeding depression, with inbred male offspring suffering a higher reproductive cost than females, independent of stress (57% vs. 14% benign, 49% vs. 11% stress).


Inbreeding, or the mating of relatives, and the associated cost to fitness known as inbreeding depression have long been a focus in evolutionary biology, ecology and conservation biology (Charlesworth & Charlesworth, 1987; Ralls et al., 1988; Keller & Waller, 2002; Kristensen & Sorensen, 2005). Inbreeding depression is a common phenomenon in nature and has been documented in a wide range of taxa (Crnokrak & Roff, 1999; Keller & Waller, 2002). As a result, inbreeding is recognized as a potent force influencing the persistence of natural populations (Keller & Waller, 2002; Spielman et al., 2004; Frankham, 2008) as well as shaping the evolution of life history, morphology, physiology and behaviour (Charpentier et al., 2007). However, we have a poor understanding of the factors that contribute to the considerable variation in the severity of inbreeding depression that has been observed across taxa, populations and even life-history stages (Hedrick & Kalinowski, 2000; Keller & Waller, 2002; Pemberton, 2008). Currently, sexual selection and environmental stress have become the focus of much investigation owing to the potential role they play in determining the magnitude of inbreeding depression. It has been hypothesized that sexual selection can reduce inbreeding depression by increasing selection against deleterious alleles (Radwan et al., 2004; Jarzebowska & Radwan, 2009), whereas environmental stress often amplifies the negative effects of inbreeding on fitness (Armbruster & Reed, 2005). In general, both sexual selection and environmental stress are predicted to increase the effectiveness with which deleterious alleles are purged from a population (Whitlock & Bourguet, 2000; Kristensen et al., 2003; Swindell & Bouzat., 2006; Whitlock & Agrawal, 2009). However, it is unknown how these two forces interact to shape the genetic architecture and expression of inbreeding depression in natural populations.

Recent interest in the role of sexual selection in population persistence has centred around the potential for sexual selection to purge mutational load, thereby preventing mutational meltdown and reducing rates of extinction in small populations (Whitlock, 2000; Sharp & Agrawal, 2008; Jarzebowska & Radwan, 2009). Although there is some evidence that sexual selection can effectively remove deleterious mutations from populations (Whitlock & Agrawal, 2009), much less is known about the effects of sexual selection on the severity of inbreeding depression. Sexual selection theory predicts that females who choose superior males increase the chances of survival of their offspring by selection of beneficial alleles and implies that males expressing deleterious alleles are less likely to find mates (Williams, 1966; Whitlock & Bourguet, 2000). Being inbred may therefore be more costly for males than for females. Several non-Drosophila studies provide evidence that aspects of sexual selection such as male–male competition and female choice exacerbate inbreeding depression for components of male mating success and for lifetime reproductive success (Potts et al., 1994; Pray et al., 1994; Meagher et al., 2000; Slate et al., 2000; Höglund et al., 2002). One study suggests that in Drosophila, the pattern may be similar (Miller & Hedrick, 1993). Inbreeding depression was significantly greater in males when sexual selection was included as a component of male fitness (as measured by competitive male mating success). Miller & Hedrick (1993) reported substantially higher levels of inbreeding depression for competitive male mating ability (72.5%) than for female fecundity (1.5%, nonsignificant). This bias was not found when sexual selection was not incorporated as a component of male reproductive fitness (Robinson et al., 2009).

Inbreeding depression is also widely recognized as being negatively affected by environmental stress. However, the magnitude of inbreeding depression reported under stressful conditions is highly variable (see Armbruster & Reed, 2005). As a result, several authors have argued that the current body of research demonstrates a lack of consistent or predictable effects of environmental stress on the expression of inbreeding depression (Keller & Waller, 2002; Armbruster & Reed, 2005; Waller et al., 2008). Furthermore, most of this research has focused on overall survival or single fitness components, which has left another important question unanswered: what is the effect of early stress on inbreeding depression in later life-history stages? It is unclear whether stress has a long-lasting effect on the physiological functioning of an organism, exacerbating the effects of inbreeding on fitness even after the source of stress is removed. Two alternative hypotheses predict different changes in the level of inbreeding depression following exposure to stress during development. Environmental stress could amplify inbreeding depression in later life-history stages because of greater vulnerability of inbred individuals to long-lasting phenotypic effects of stress. Alternatively, stress may purge genetic load during development, thus reducing levels of inbreeding depression in later life-history stages. This can occur if stress increases selection against deleterious mutations that affect fitness at multiple life-history stages (Haldane, 1957). This hypothesis predicts that stress will amplify inbreeding depression during the stage at which individuals are exposed (e.g. larval survival) but will have either no effect or reduce inbreeding depression for later performance (e.g. reproduction), as observed by Montalvo (1994) in the blue columbine (Aquilegia caerulea). These alternatives are not mutually exclusive, but it is important to understand their relative importance.

In D. melanogaster, given that females exhibit strong sperm precedence (see review in Manier et al., 2010) and lay multiple eggs on a single fruit (Nunney, 1990), full-sib mating is expected to occur in the wild among offspring from a single fruit. However, despite the extensive use of D. melanogaster as a model for the effects of inbreeding on fitness, an accurate measure of the cost of sib mating in wild populations is lacking. The vast majority of studies have used populations maintained under laboratory conditions for well over 20 generations (Tantawy, 1957; Sharp, 1984; Mackay, 1985; Miller & Hedrick, 1993; Miller et al., 1993; Garcia et al., 1994; Hughes, 1995, 1997; Fowler & Whitlock, 2002; Hughes et al., 2002). The few exceptions are those studies that have measured the fitness effects of individuals made homozygous for various chromosomes extracted from males taken directly from wild populations. Laboratory-adapted populations may have a different genetic architecture relative to natural populations, for example because of bottleneck events occurring during the establishment or maintenance of a laboratory population. This may result in differences in the levels of inbreeding depression. Concerns have been raised regarding potential underestimation of the levels of inbreeding depression found in nature (Sheilds, 1993; Hedrick & Kalinowski, 2000; Joron & Brakefield, 2003) when laboratory populations are used. To investigate the effect of inbreeding in flies with a natural genetic architecture, the cost of sib mating needs to be measured from recently wild-caught populations.

The aim of this study was to evaluate how inbreeding depression is influenced by both sexual selection and environmental stress using two wild-caught populations of D. melanogaster. Specifically this study addresses three main questions: (i) How large is the fitness cost of sib mating when developmental and reproductive costs are included? (ii) Are reproductive costs in the inbred offspring greater in males than in females? (iii) Given a stressful environment during larval development, does inbreeding depression increase for larval survival and/or adult reproduction? The cost of one generation of full-sib mating was measured in two populations of D. melanogaster collected from Northern California after only one generation of controlled outcrossing in the laboratory (to avoid any unintentional purging of deleterious recessive alleles). We measured egg hatchability, larval-to-adult survival, female fecundity and male mating success of inbred offspring and evaluated the role of stress on inbreeding depression by comparing the effect of rearing larvae under the conditions of low and high food competition. In addition, we evaluated the effect of the age of females on inbreeding depression for female fecundity by measuring the number of offspring produced separately during the early (days 1–8) and late (days 8–16) stages in the female’s life. Previous work in Drosophila has shown that inbreeding depression increases with age (Hughes, 1995; Hughes et al., 2002).


Base population and inbreeding design

Drosophila melanogaster populations were collected from two locations in Northern California, the Galante Winery in Carmel Valley (Gala) and the Mayo Family Winery in Sonoma Valley (Mayo). To prevent modifying the genetic architecture through selection or inbreeding, 400 pairs of wild-caught flies from each location were placed in single vials and reared in the laboratory for one generation. Their progeny were outcrossed by taking a single male and female from each of the 400 pairs (per population) and mating them in a circular design, whereby each male is mated to the female from the next vial. The resulting outcrossed progeny (P generation) were then used as parents to establish families in the breeding design explained in the following sections. The initial crosses ensured that each of the original pairs contributed an equal number of progeny, that all P generation flies from a single vial were full sibs and that no inbreeding occurred since capture.

The P generation was created by collecting fifty-six virgin full sibs, twenty eight of each sex, from each of 16 Gala vials and 12 Mayo vials (each vial constitutes one outbred family). For each family, these P generation flies were divided into two equal groups (14 of each sex) to create both inbred (IB) and outbred (OB) crosses. For inbred crosses, 14 virgin females were sib-mated to 14 of their virgin brothers in a single vial, and for OB crosses, 14 virgin females were mated to 14 virgin males from another family that was chosen at random without replacement. For example, females (all sisters) from Gala family A were crossed to their brothers from family A to create the inbred cross. In addition, this same family was involved in two OB crosses: A♀ × C♂ (the maternal cross of family A) and D♀ × A♂ (the paternal cross). The results from A♀ × A♂ and A♀ × C♂ made up a maternal family lineage, and the results from A♀ × A♂ and D♀ × A♂ made up a paternal family lineage. The 16 Gala families and 12 Mayo families are a subset of the original 400 pairs that were round-robin crossed to create the P generation. As a result, the round-robin design was not complete for the families tested, and therefore, the number of maternal (28 total) and paternal (20 total) lineages used in subsequent analyses was unequal. All experiments were carried out at 25 °C.

Inbreeding depression (δ) was calculated for each family as the loss of fitness exhibited by the progeny of a sib mating, i.e. 1 – (winbred/woutbred), where (winbred/woutbred) is the fitness of the progeny from the sib-mated cross relative to the OB cross from each family. With the design described earlier, inbreeding depression can be calculated separately for the maternal and paternal lineages. For example, the value of woutbred used to calculate inbreeding depression for family A could be either from the maternal (A♀ × C♂) or from the paternal (D♀ × A♂) OB cross. Note that the maternal data and the paternal data sets are each internally independent, but that they are not independent of each other.

Egg hatchability

For each family cross (IB and OB), groups of 14 males and 14 females were set up in an empty glass bottle capped with a petri dish containing standard food medium and allowed to lay eggs for a period of 8 h. The groups were transferred to new laying dishes every 8 hours until at least 350 eggs had been laid per family cross. The number of eggs laid was counted, and the number of unhatched eggs was counted 24 h later. Inbreeding depression was calculated per family as δ = (1 – (% hatchedinbred/% hatchedoutbred)) for both maternal and paternal lineages.

Larval-adult survival

Larval-to-adult survival was measured on two concentrations of food (1× and 1/3×) in the presence of larvae from a standard competitor laboratory stock (spa). The spa competitor flies have a recessive sparkling eye phenotype that can be distinguished from the wild-type test flies. The 1× concentration of food is the standard food medium consisting of molasses, cornmeal, yeast, water and the antifungal agent Tegosept. The 1/3× concentration of food was created by adding 2/3 agar (18g/L) to 1/3 of the standard 1× food medium and was chosen based on preliminary experiments demonstrating an average 45% reduction in survival relative to the standard 1× concentration of food.

First instar larvae were collected in groups of 50 larvae per family for up to 8 h from the previously described laying dishes. This ensured that all larvae were at ± 4 h apart in development. Larvae were transferred using a paintbrush (ID #000) to vials containing 10 mL of food medium (1× or 1/3×) in the ratio of 50 test line to 150 of the standard competitor spa. Three replicate vials of each concentration of food (1× and 1/3×) were set up for each family cross (IB and OB for the 16 Gala and 12 Mayo families) and placed at 25 °C.

Following the set-up of first instar larva in vials (day 1), once the first eclosed progeny were observed (days 9–10), all emerging adults were counted and removed every 3–4 days until approximately days 20–21, by which time the number of first-generation progeny emerging per vial had diminished to zero (or nearly so) over the final 3–4 day counting interval (days approximately 19–21) and a large number of dark pupae representing the next generation were observed. There was always a clear distinction between the first and second generation.

Larval-adult survival was measured in two ways: as the proportion of test larvae surviving to eclosion (LS) and as the larval competitive index (LCI), which is the proportion that eclosed per vial of the test line relative to the proportion of spa competitors (Knight & Roberston, 1957). Inbreeding depression was calculated as (1 – (LSinbred/LSoutbred)) and (1 – (LCIinbred/LCIoutbred)) for both maternal and paternal lineages.

Fecundity of female offspring

Adult F1 females that emerged from the larval-adult survival assay were collected and used to measure female fecundity. Virgin females were collected from each replicate vial of food concentration (1× and 1/3×) and breeding treatment (OB and IB) and placed at 25 °C until they were 4–7 days old. Three females were randomly selected from each vial, and each was placed individually with two unrelated males from a standard outbred laboratory strain (MEL) for 48 h at 25 °C. Single females were transferred without males to a new vial with standard food medium and allowed to lay eggs for 8 days (vial 1), after which each female was transferred again to a new vial and allowed to lay for an additional 8 days (vial 2).

Female fecundity was measured in three ways for each individual female: (i) early fecundity from vial 1 (days 1–8) where the females were 11–14 post-eclosion at the end of the laying period; (ii) late fecundity from vial 2 (days 9–16) where females where 19–22 days post-eclosion at the end of the laying period; and (iii) total progeny production over 16 days (sum of progeny from both vials). Note that these females (IB and OB) were outcrossed so that their offspring would not exhibit inbreeding depression. As with the larval-adult survival assay, care was taken to avoid including a second generation in the progeny counts. Inbreeding depression was calculated as (1 – (avg # progenyinbred females/avg # progenyoutbred females)) for both maternal and paternal lineages. Females that were collected for testing but subsequently died or failed to produce offspring were included as having zero fecundity.

Mating success of male offspring

Adult F1 males that emerged from the larval-adult survival assay were used to measure male mating success. Virgin males were collected from each replicate vial and placed at 25 °C until they were approximately 14–18 days old. Five males (full sibs) were randomly selected from each vial and placed with 15 unrelated virgin competitor spa males and 10 unrelated virgin spa females in new vials containing 10 mL food. To control for the quality of competitor flies, spa males and females used in these mating trials were reared at low densities in 30 bottle populations of approximately 200 adults each at 25 °C and transferred every 3–4 days to avoid overcrowding. The mating trials lasted for two hours at 25 °C, after which females were removed using light anaesthesia and transferred individually to new vials. Preliminary experiments determined that this time period and ratio of test males to competitor males minimize multiple mating and maximize the overall number of successful matings, which is in agreement with previous studies (Sharp, 1984; Miller & Hedrick, 1993). After approximately 2 weeks, the progeny of each spa female were scored for eye colour to determine her mate. The spa phenotype is recessive, and therefore, if a spa female mates to a test male, 100% of the progeny will be wild type for eye colour. The short mating period (2 h) was designed to avoid females mating to multiple males, and the ratio of test males to spa males (5 : 15) provided a level of male competition that minimized the number of trials where one type of male did not mate to any of the females. A random subset of five families per population was selected for measurement.

Male mating success was measured both as a proportion of test males mating (MS) and relative to the standard competitor spa as the male competitive index (MCI). MCI was defined as the proportion of females inseminated by test males divided by the proportion of females inseminated by the standard competitor spa males. Inbreeding depression was calculated as (1 – (MSinbred male/MSoutbred male)) and (1 – (MCIinbred male/MCIoutbred male)) for both maternal and paternal lineages.

Cumulative inbreeding depression

Cumulative inbreeding depression across all life-history traits was calculated separately for both males and females and separately using either maternal or paternal lineages as the outbred reference. By multiplying % egg hatchability (EH), the LCI, and either the 16-day total female fecundity (TFF) or the MCI, cumulative average fitness was calculated for each inbred (Winbred) and outbred (Woutbred) line in both sexes. This assumes EH and LCI are equal in the two sexes, as found by Frankham & Wilcken (2006).

Cumulative inbreeding depression is expressed as:


Lethal equivalents

For each fitness trait as well as cumulative male and female fitness, we calculated β, the number of lethal equivalents per haploid genome as: β = –[ln(winbred/woutbred)]/F (Morton et al., 1956) where F is the level of inbreeding (F = 0.25). The number of lethal equivalents is defined by the rate at which the logarithm of fitness declines with inbreeding. This method is commonly used to compare the effects of inbreeding on fitness across studies, species/taxa and environments (Hedrick & Kalinowski, 2000; Armbruster & Reed, 2005).

Statistical analysis

Two separate analyses were run, one where family was assigned according to maternal lineage and one according to paternal lineage for all fitness measures (see explanation above). Larval survival, female fecundity and male mating success were analysed using an anova (SAS Version 9.1 for Windows®, SAS Institute Inc., Cary, NC, USA) with the following variables: INBREED (IB vs. OB), COMP (high vs. low larval competition), POP (Gala, Mayo), FAM (P generation family nested within population, using either the female or male lineage in separate analyses to define the OB cross), plus all two- and three-way interactions involving INBREED and COMP. INBREED and COMP are fixed effects, whereas POP and FAM are random effects. Egg hatchability was analysed using the aforementioned model but without the variable COMP. Both competitive indices (LCI and MCI) were arcsine square root transformed so that values were more normally distributed. For female fecundity, two separate analyses were run: (i) anova on total fecundity and (ii) a repeated measures anova using procedure MIXED on early and late fecundity. Unless otherwise stated, results reported are those for total female fecundity. Finally, pairwise correlations were calculated between larval survival (LCI) and each adult fitness measure (FF, MCI) across families in inbred and outbred individuals separately. In addition, the competitive index and raw percentage values were correlated for larval survival (LCI and % survival) and male mating success (MCI and % matings). Pairwise correlations were also calculated for inbreeding depression in larval survival, female fecundity and male mating success.

Cumulative male and female relative fitness, calculated as ((EH × LCI × MCI)males or (EH × LCI × FF)females), was analysed in the families that had data across all four life-history traits using the following anova model: COMP, FAM, SEX, POP, and all interactions where COMP and SEX were fixed effects and POP and FAM are random effects.

Interactions with a P value > 0.25 (Kirk, 1982) were removed from the aforementioned models, and the analysis was rerun. When a significant interaction occurred in the original model, the analysis was split (for example by POP or COMP) and a Bonferroni correction was used to adjust for multiple testing.


The data were analysed for each life-history stage separately and cumulatively across all stages. The analysis is presented first using the maternal lineage as the outcross reference (e.g. comparing A♀ × A♂ offspring to A♀ × C♂ offspring, see Methods). The robustness of these results is then examined by presenting the analysis based on the paternal lineage (e.g. comparing A♀ × A♂ offspring to C♀ × A♂ offspring).

Maternal family analysis

Inbreeding in egg hatchability caused a small but significant overall reduction in the Mayo population (P < 0.01, Table 1a), and there was significant variation in the magnitude of the inbreeding depression among families in both populations (Gala P < 0.001, Mayo P < 0.01). The overall inbreeding depression (< 2%) corresponded to < 0.1 lethal equivalents in both populations (Table 1b).

Table 1.   Analysis of inbreeding depression in egg hatchability. Outcrossing was evaluated either via the maternal lineage or via the paternal lineage. (A) Analysis of variance of the fitness loss because of sib mating versus outcrossing (INBREED). The model also included population Gala vs. Mayo (POP) and family (FAM). Analysis was split by population because of a significant interaction between INBREED and POP (F1,168 = 3.12 P < 0.001). (B) Average (± SE) egg hatchability, inbreeding depression (δ) and number of haploid lethal equivalents (β) in the two populations (Gala and Mayo). ***P < 0.001, **P < 0.01 *P < 0.05.
(A) anova Egg Hatchability
Maternal lineage
 INBREED × FAM150.01606.96***
 INBREED × FAM110.00993.82**
Paternal lineage
 INBREED × FAM(POP)170.01717.63***
 INBREED × POP10.00040.17
(B) Egg Hatchability
 Maternal97.70 ± 0.5396.62 ± 0.420.009 ± 0.04***0.04
N = 98N = 88  
 Paternal98.40 ± 0.1895.64 ± 1.060.020 ± 0.01***0.10
N = 57N = 47  
 Maternal98.94 ± 0.0997.10 ± 0.270.019 ± 0.01***0.08
N = 73N = 64  
 Paternal98.91 ± 0.1597.20 ± 0.400.020 ± 0.01***0.07
N = 62N = 55  
EffectF ratio
  1. F ratios. Unless specified above F ratios calculated as MSeffect/MSerror.


Inbreeding had a much larger effect on larval survival (P < 0.001; Table 2) of 19% under benign conditions. Taking into account the response of competitors, this estimate increased to 23%, as measured by the LCI. Under stressful conditions of limited food (high larval competition), inbreeding depression was higher, at 34% (% survival) and 63% (LCI). Based on LCI, this corresponds to a shift from roughly 1 to 4 lethal equivalents (Table 3). There were no significant differences between the populations in any of these measures, and although there was significant variation across families for larval success (P < 0.05; Table 2), there were no differences among families in inbreeding depression.

Table 2.   Analysis of variance of the fitness loss because of sib mating versus outcrossing (INBREED) for larval survival, total fecundity of surviving females and mating success of surviving males. Outcrossing was evaluated either via the female lineage or via the male lineage. The model also included population Gala vs. Mayo (POP), high vs. low competition (COMP) and family (FAM). (A) Larval survival. The analysis was split by level of competition because of a significant interaction between INBREED and COMP in the initial analysis (F1,523 = 129.83 P < 0.001). (B) Fitness measures of adult offspring. ***P < 0.001, **P < 0.01 *P < 0.05.
SourceMaternal lineagePaternal lineage
(A) Larval survival under different levels of larval competition
Low larval competition
 INBREED × FAM(POP)250.00191.04220.00240.81
 Error2900.8102 770.003 
High larval competition
 INBREED × FAM(POP)230.02801.45200.01520.91
 Error2890.0200  760.0166  
(B) Adult fitness measures
Female fecundity
 INBREED × FAM(POP)262382.42.73***172209.32.81**
 INBREED × COMP12007.82.313.20.01
 Error801871.9  464786.8 
Male mating success
 INBREED × FAM(POP)180.05901.3100.10652.09*
 INBREED × COMP10.00130.0310.02730.54
 Error1270.0454  970.0509  
EffectF ratio
  1. F ratios. Unless specified above F ratios calculated as MSeffect/MSerror.

POPMSPOP/MSFAM (Female Fecundity and Male Mating only)
Table 3.   Summary of average (± SE) larva-adult survival, total fecundity of surviving females and mating success of surviving males, as well as inbreeding depression (δ) and number of haploid lethal equivalents (β) for these traits. All traits were measured after rearing in either low or high larval competition. Larval-adult survival and male mating success were measured relative to a standard competitor (LCI or MCI) or in absolute terms as the per cent survival or the per cent of the total matings. Female fecundity is the average number of offspring produced on days 1–8 (early), days 8–16 (late) or both. N is the sample size. Traits were evaluated using the female lineage or the male lineage. ***P < 0.001, **P < 0.01 *P < 0.05.
Fitness for different levels of larval competition
Larval-adult survival
 AVG LCI1.84 ± 0.021.84 ± 0.021.45 ± 0.021.46 ± 0.030.23 ± 0.03***0.20 ± 0.02***0.950.90
 AVG % Survival89.36 ± 0.8088.97 ± 0.6472.24 ± 1.0571.98 ± 1.140.19 ± 0.01***0.18 ± 0.07***0.850.84
   N = 151   N = 128   N = 140   N = 118    
 AVG LCI5.15 ± 0.264.87 ± 0.271.84 ± 0.141.77 ± 0.130.63 ± 0.03***0.62 ± 0.03***4.123.89
 AVG % Survival43.38 ± 1.1442.14 ± 0.8127.83 ± 1.3328.53 ± 1.010.34 ± 0.04*0.32 ± 0.18***1.761.52
   N = 156   N = 126   N = 134   = 110    
Male mating success
 AVG MCI11.26 ± 0.8711.53 ± 1.145.31 ± 0.615.18 ± 0.680.57 ± 0.05***0.48 ± 0.10**3.012.58
 AVG % Matings76.98 ± 2.8074.74 ± 2.2458.30 ± 2.8055.96 ± 3.270.26 ± 0.07*0.24 ± 0.19**1.221.11
   N = 53   N = 31   N = 53   N = 33    
 AVG MCI6.89 ± 0.685.35 ± 0.573.12 ± 0.432.63 ± 0.360.51 ± 0.09***0.45 ± 0.10**3.172.39
 AVG % Matings65.04 ± 2.2560.50 ± 2.4945.92 ± 2.8742.90 ± 2.790.32 ± 0.06***0.28 ± 0.18**1.611.34
   N = 33   N = 24   N = 34   N = 25    
Female fecundity
 Early74 ± 1.2974 ± 1.6073 ± 1.2075 ± 1.56−0.04 ± 0.05−0.05 ± 0.0600
 Late83 ± 1.2483 ± 1.5258 ± 1.6255 ± 2.040.22 ± 0.05***0.25 ± 0.05***0.981.16 
 Total # offspring157 ± 1.90157 ± 2.37131 ± 2.05128 ± 2.270.14 ± 0.03***0.16 ± 0.04***0.720.71
   N = 292   N = 184   N = 245   N = 168    
 Early71 ± 1.6970 ± 2.2463 ± 1.1675 ± 2.440.01 ± 0.06-0.03 ± 0.080.020
 Late63 ± 1.1661 ± 1.4349 ± 1.2050 ± 1.320.19 ± 0.03 ***0.19 ± 0.04***0.850.80 
 Total # offspring133 ± 2.04133 ± 3.32121 ± 2.49122 ± 2.980.11 ± 0.04***0.10 ± 0.04***0.380.40
   N = 182   N = 73   N = 141   N = 77    

Both male mating success and female fecundity were reduced because of inbreeding (P < 0.01 and P < 0.001; Table 2) regardless of population. Unlike larval survival, inbreeding depression in these traits was not amplified by rearing under high larval competitive stress (Table 2) even though the absolute levels of fecundity and male mating success were substantially reduced (Table 3). Inbred females produced on average 20 (13%) fewer total offspring than outbred females, i.e. 0.56 lethal equivalents (Table 3). More inbred females (7 of 247 and 5 of 141) failed to produce offspring than outbred females (2 of 293 and 4 of 179) across both low and high larval competitive environments, respectively. Inclusion of these females with zero fecundity did not inflate the effect of inbreeding, resulting in approximately 1% greater inbreeding depression relative to when they were removed from the analysis. When fecundity was divided into early (age: 6–14 days) and late (age: 15–22 days) stages, inbreeding depression differed between the stages (P < 0.001; Table 4). Whereas early female fecundity was not different between inbred and outbred females, significant inbreeding depression was found for late female fecundity (0.98 lethal equivalents, Tables 3 and 4). Inbred males had 29% lower mating success compared to outbred males (corresponding to 1.37 lethal equivalents), a difference that rose to 54% (3.11 lethal equivalents) when their success relative to the spa males was taken into account (Table 3). Males were 14–18 days old when tested, so male inbreeding depression was greater than that found in older and younger females. Levels of inbreeding depression varied significantly across families for total female fecundity (P < 0.001), but not for male mating success (Table 2).

Table 4.   Repeated measures analysis of variance comparing the fitness loss because of sib mating for the average number of offspring produced by females on days 1–8 (early fecundity) and days 8–16 (late fecundity) under different levels of competition. TIME represents the early and late measures of fecundity.
SourceMaternal lineagePaternal lineage
  1. ***P < 0.001, **P < 0.01, *P < 0.05.

INBREED × FAM(POP)2713193.87***1610534.69***
INBREED × COMP113190.01110532.17
COMP × FAM(POP)2313195.58***1210539.34***
INBREED × COMP × FAM(POP)1313191.43710532.14*
INBREED × TIME1131969.36***11053106.36***
TIME × COMP1131928.19***1105319.75***
TIME × FAM(POP)2713192.46***1510533.17***
TIME × INBREED × COMP113194.61*1105315.57***
TIME × COMP × FAM(POP)2213195.04***1110539.11***
TIME × INBREED × POP × COMP213190.912105311.11***
TIME × INBREED × COMP × FAM(POP)3813192.75***2110533.84***

The larval and male fitness components were estimated directly as per cent success or relative to standard competitors. These estimates were highly correlated for IB and outbred offspring combined under conditions of low larval competition (larval survival: = 0.95, d.f. = 52, P < 0.001, male mating: r = 0.77, d.f. = 25, P < 0.001). These correlations were lower given high larval competition, especially for male mating (larval survival: r = 0.79, d.f. = 49, P < 0.001, male mating: r = 0.42, d.f. = 24, P < 0.05).

The three fitness components (larval survival, female fecundity and male mating) were not correlated across the outbred families: the highest correlation was between larval survival (LCI) and male mating success (MCI) under high competition (r = 0.76, d.f. = 8, P < 0.05 uncorrected for multiple testing). Furthermore, inbreeding depression was not significantly correlated across the families/lineages for larval survival, female fecundity and male mating success within or between larval competitive treatments (all P > 0.1 uncorrected for multiple testing). In particular, there was no change in the correlation between larval inbreeding depression and the inbreeding depression in male or female reproduction under benign (male r = −0.09, d.f. = 8; female r = 0.01, d.f. = 24) vs. stressed (male r = −0.06, d.f. = 8; female r = 0.13, d.f. = 13) conditions, as might be expected if there was larval-stage purging of genetic load affecting adult traits and if this purging increased under stress.

Comparison of male and female reproductive fitness of the outcrossed families showed nonsignificant negative correlations under benign and stress conditions (r = −0.58, −0.21; d.f. = 8, 8). The correlation of inbreeding depression between the sexes showed nonsignificant positive correlations under both conditions (r = 0.51, 0.30; d.f. = 8, 7).

Combining the developmental effects of hatchability and larval survival with adult reproduction (averaged across the sexes) results in a cumulative fitness loss for a sib-mated pair of 50% (2.77 lethal equivalents) under benign conditions and 73% (5.24 lethal equivalents) if their offspring experienced high larval competition (Table 5). The high inbreeding depression for male reproduction (mating success) relative to female reproduction (fecundity) resulted in cumulative inbreeding depression being significantly higher for male offspring than for female offspring under both standard conditions (low competition) and limited food (high competition) (Fig. 1, Table 5). This sex difference was not significantly altered by larval environment (Table 5).

Table 5.   Cumulative fitness loss in male and female inbred offspring from sib mating using both maternal and paternal lineage analyses. (A) Summary of average (± SE) inbreeding depression (δ) and number of haploid lethal equivalents (β), where N equals the number of families. The overall fitness loss from sib mating is the average of the fitness loss from sons and daughters. (B) Analysis of variance for cumulative offspring fitness. The model included male vs. female (SEX), high vs. low competition (COMP), population Gala vs. Mayo (POP) and family (FAM). ***P < 0.001, **P < 0.01, *P < 0.05.
(A) Summary of cumulative fitnessFemale cumulative fitness
 Low0.33 ± 0.030.36 ± 0.031.601.80
  N = 26  N = 14  
 High0.66 ± 0.040.63 ±
  N = 15  N = 7  
Male cumulative fitness
 Low0.67 ± 0.040.64 ± 0.064.434.07
  N = 11  N = 8  
 High0.81 ± 0.040.76 ± 0.076.645.71
  N = 11   N = 7  
Cumulative fitness loss from Sib mating
 Low0.50 ± 0.040.50 ± 0.052.772.77
 High0.73 ± 0.040.70 ±
 Maternal analysisPaternal analysis
(B) anova cumulative fitness
SEX × COMP10.01171.5110.04192.29
FAM(POP) × COMP80.03374.32**40.07324.00*
Error160.0078 100.0182 
EffectF ratio
  1. F ratios. Unless specified above F ratios calculated as MSeffect/MSerror.

Figure 1.

 Cumulative inbreeding depression in male and female offspring resulting from sib mating. Two levels of larval competition are shown: low and high. Cumulative relative fitness (Winbred/Woutbred) averaged across families (± SE) is plotted at each life-history stage from egg to adult. The adult fitness measure is fecundity (females) or mating success (males).

Paternal family analysis

The paternal analysis differs from the maternal approach in having the potential to create an apparent family by inbreeding interaction (FAM × INBREED), because inbred and outbred families differ in maternal environment and the sons differ in the X-chromosome that they carry. However, the effect of sib mating on fitness was qualitatively the same as the maternal analysis in almost all respects (see Tables 1–4). Differences that generated shifts in significance were observed in egg hatchability, high larval competition, male mating success and cumulative fitness. However, all inbreeding-related factors that were highly significant under the maternal analysis (< 0.01) remained so under the paternal analysis. One of these (affecting male mating success) was because of an increase in the family by inbreeding interaction (P < 0.05; Table 2), but conversely the influence of family on the cumulative inbreeding depression became nonsignificant (Table 5).


This study demonstrates a number of important features of the relationship between inbreeding and fitness under the conditions of sexual selection and environmental stress. Under standard laboratory conditions, sib mating resulted in a 50% cost to overall fitness in recently wild-caught D. melanogaster populations, corresponding to β = 2.77 lethal equivalents (Table 5). In addition, a striking difference was found in the level of inbreeding depression expressed in males and females (Fig. 1). Inbred males suffered an almost 2-fold higher cumulative loss in fitness than females, a result consistent with the study of Miller & Hedrick (1993). This difference between the sexes remained the same under high larval competition (Table 5), whereas the overall fitness cost to sib mating increased. When these cumulative effects are broken down, we found that egg hatchability was only slightly affected by inbreeding (< 2%), while even under benign conditions, relative larval survival dropped by about 20%. We also found that larval competitive stress amplified this larval inbreeding depression, but it did not increase inbreeding depression in the later life-history stages of adult reproduction. This result argues against the hypothesis that stress induces long-lasting negative phenotypic effects in inbred individuals but may be consistent with the possibility that stress, through purging some of the genetic load, can lower inbreeding depression at a later stage. All of these patterns were consistent across the two replicate populations. The only significant (but trivial) difference between them was a 2% level of inbreeding depression for egg hatchability in the Mayo population vs. 1% in Gala.

In this study, we emphasized the importance of testing wild-caught flies that retained their natural genetic architecture. However, we still tested the flies under laboratory conditions as recent advances in measuring inbreeding depression in wild populations using pedigree analysis (Pemberton, 2008) are not generally applicable to short-lived small animals such as Drosophila.

The Magnitude of inbreeding depression in recently wild-caught D. melanogaster

Table 6 summarizes literature on the effects of inbreeding on individual fitness components in D. melanogaster measured under standard benign laboratory conditions using fitness measures directly comparable to those measured in this study (see Simmons & Crow (1977) and Charlesworth & Charlesworth (1987) for reviews of estimates based on chromosomal homozygote populations). Only two studies (Miller & Hedrick, 1993; Robinson et al., 2009) measured inbreeding depression across the full spectrum of life-history stages in both sexes, and Robinson et al. (2009) excluded sexual selection in males. Our finding that mating between siblings caused a 50% reduction in overall cumulative fitness in recently caught populations of D. melanogaster under standard laboratory conditions (Fig. 1, Table 5) is consistent with results of Miller & Hedrick (1993). We found that rearing conditions that included larval competitive stress resulted in a larger 72% reduction in overall fitness.

Table 6.   Summary of the effects of inbreeding on individual fitness components in Drosophila melanogaster measured under benign laboratory conditions. The number of haploid lethal equivalents (β) was calculated as described in the text (β = −[ln(winbred/woutbred)]/F), given a single level of inbreeding (F) or as the slope of the regression of ln(fitness) on F. For chromosome homozygotes, β was calculated according to Appendix 1. For fecundity measures, the age of adult females is the age range over which females laid eggs, and for mating success, it is the age at which males were tested. Hatchability and larval survival for both inbred (IB) and outbred (OB) individuals are included when the data were available. All values for this study are from the maternal analysis and include both the competitive index and per cent success (in parenthesis) for larval survival and male mating. Cumulative fitness for each sex is a multiplicative measure that was calculated across the same life-history stages for each study (see Methods).
Life-History StageFType of InbreedingβSource
  1. *Fitness measure is relative to a standard competitor (i.e. MCI or LCI).

  2. †Male fitness used to calculate cumulative fitness was the number of progeny a male produced when mated to a stock female.

  3. ‡Male fitness used to calculate cumulative fitness was the male competitive index (MCI).

Egg hatchability  %OB%IB  
0.25Full-sib mating98.696.80.06This study
0.25Full-sib mating 96.0 90.00.26Biémont, 1978
Egg-adult survival  %OB%IB  
Egg-pupa survival0.25–0.73Full-sib mating80.060.00.43Garcia et al., 1994
Larval-adult survival0.25Full-sib mating89.472.20.95* (0.85)This study
0.25Full-sib mating85.666.71.00Ehiobu et al., 1989
0.25–0.75Full-sib mating61.949.90.37Tantawy, 1957
0.37Chrom 2 homozygotesNANA0.30Miller & Hedrick, 1993
Egg-adult survival0.25–0.88Full-sib matingNANA2.67*Latter & Roberston, 1962
0.25Full-sib mating90.070.01.01Biémont, 1978
0.25Full-sib mating80.072.00.42Robinson et al., 2009
0.37Chrom 2 homozygotesNANA0.82Bijlsma et al., 1999
0.65Chrom 3 homozygotesNANA0.89Mackay, 1985
Adult female fitness  Age   
Female fecundity0.25Full-sib mating6–14 days 0.00This study
0.25 15–22 days 0.98
0.25Full-sib mating3–6 days 1.10Ehiobu et al., 1989
0.25Full-sib mating7–8 days 0.57Robinson et al., 2009
0.50Full-sib mating2–5 days 1.12Dahlgaard & Hoffmann, 2000
0.50Full-sib mating1–5 days 0.30Miller, 1994
0.55Chrom 2 homozygotes1–4 days 0.00Miller & Hedrick, 1993
0.98Chrom 2 + 3 homozygotes1–7 days 0.34Hughes et al., 2002
  8–14 days 0.77 
  15–21 days 0.71 
0.65Chrom 3 homozygotesUnknown 1.25Mackay, 1985
Adult male fitness  Age   
Male mating success0.25Full-sib mating14–18 days 3.01* (1.22)This study
0.25–0.98Full-sib mating4–7 days 0.81*Sharp, 1984
0.25Full-sib mating5 days 5.39Pendlebury & Kidwell, 1974
0.25Full-sib mating5 days 2.25 
0.37Chrom 2 homozygotes3 days 1.02Kosuda, 1983
0.37Chrom 2 homozygotes3 days 3.49Miller & Hedrick, 1993
0.37Chrom 2 homozygotes2 days 1.86*Brittnacher, 1981
0.44Chrom 3 homozygotes3 days 0.43*Hughes, 1995
 Chrom 3 homozygotes21 days 1.14* 
0.44Chrom 3 homozygotes3 days 1.04*Patridge et al., 1985
Cumulative fitness  Sex   
 0.25Full-sib matingFemale 1.60* (1.50)This study
  Male 4.25* (2.15) 
0.25Full-sib matingFemale 0.91Robinson et al., 2009
  Male 0.92† 
0.37Chrom 2 homozygotesFemale 0.27Miller & Hedrick, 1993
  Male 3.79‡ 
0.65Chrom 3 homozygotesFemale 2.13Mackay, 1985

The contribution of egg hatchability to cumulative inbreeding depression was statistically significant but very small (2%), similar to the 6% found by Biémont (1978). The very low inbreeding depression in egg hatchability suggests that a limited fraction of the offspring’s genes are expressed at this stage and/or that mutations in the genes involved in early development are generally not fully recessive, limiting the mutation selection build-up of deleterious alleles. In contrast, reductions in fitness under low and high larval competition, respectively, in larval-to-adult survival (22%, 63%), the fecundity of female offspring (15%, 11%) and the mating success of male offspring (51%, 48%) were all substantial. Overall, our estimates of lethal equivalents for larval-adult survival (0.95) and female fecundity (0, 6–14 days; 0.98, 15–22 days) are similar to those previously reported (Table 6). The finding that inbreeding depression for female fecundity increased with female age (Table 4) is in agreement with several studies demonstrating significant age effects in both females and males (Hughes, 1995; Hughes et al., 2002).

Inbreeding depression is expected to vary among families, as they represent a random sampling of deleterious alleles in the population (Hedrick & Kalinowski, 2000; Haag et al., 2003). Interactions between inbreeding and families/lineages have been found in Drosophila (Bijlsma et al., 1999; Dahlgaard & Hoffmann, 2000; Reed et al., 2003), Daphnia (Haag et al., 2003), Peromyscus polionotus (Lacy et al., 1996) Tribolium castaneum (Pray & Goodnight, 1995), and plants (Dudash et al., 1997; Byers & Waller, 1999). In this study, highly significant interactions between inbreeding and family were observed for egg hatchability and female fecundity (Tables 1 and 2). However, this among-family variance is expected to decrease with an increase in the number of loci contributing to the trait (and hence potentially a source of deleterious alleles), a pattern found in this study for larval competition, both under low and high competition conditions, and male mating success (Table 2). This lack of a family effect for larval performance and male mating success may indicate the involvement of a large number of mildly deleterious alleles, which minimizes the sampling variance among the families.

Comparing among the traits, our results suggest that alleles having deleterious effects on the fitness of inbred larvae do not affect later life-history stages and/or that some deleterious alleles are purged at the larval stage so that they are not expressed at the later stages of female fecundity and male mating success. Under both low and high larval competition, the correlations between larval and adult traits were within the range 0–0.13. In addition, the nonsignificant positive correlation between male and female adult inbreeding depression under both benign and stressed larval conditions is consistent with the hypothesis that deleterious alleles show no consistent pattern in determining male mating success verses female fecundity.

Inbreeding depression is greater in males

Inbred male offspring showed a substantially greater loss of fitness than females, regardless of the female age (late female fecundity, β = 0.98; male mating success, β = 3.01, Table 6). Assuming no sex differences in egg-to-adult survival (Frankham & Wilcken, 2006), the cumulative relative fitness of inbred females was almost two-fold higher than inbred males (Table 5, Fig. 1), results similar to those of Miller & Hedrick (1993) (Table 6). In contrast, Robinson et al. (2009) did not incorporate sexual selection into the measure of adult male reproductive fitness and found no differences between the sexes in levels of inbreeding depression (both approximately 13%).

Marked sex differences in inbreeding depression have also been found in other animals. In house mice, Potts et al. (1994) found significant inbreeding depression for the acquisition of territories by males, whereas there was no detectable inbreeding depression for female fitness. In addition, the cost to sib mating in wild-caught mice has been shown to be almost four times greater for inbred males than for inbred females under semi-natural conditions (Meagher et al., 2000). Pray et al. (1994) also report that male red flour beetles (Tribolium castaneum) suffer greater costs of being inbred than females for proportion of offspring produced in a competitive social environment. Such results suggest that sexual selection via male–male competition and/or female choice may be responsible for amplifying the effects of inbreeding on male fitness. However, sexual dimorphism in inbreeding depression may result from competition for resources (food, water, territory) in general and therefore will not always be male biased. For example, wild female song sparrows exhibit greater inbreeding depression for lifetime reproductive success than males (Keller, 1998).

A finding that selection is greater against inbred males than inbred females can have important implications for the role of sexual selection in reducing genetic load in populations via female choice. Sexual selection is predicted to reduce the frequency of deleterious alleles in a population if males that carry a greater number of deleterious alleles are less likely to mate because of female choosiness (Whitlock & Bourguet, 2000), potentially reducing the risk of extinction in small populations (Whitlock, 2000). Recent empirical work in the bulb mite demonstrated that sexual selection can reduce both extinction rate and levels of inbreeding depression for small bottlenecked populations (Jarzebowska & Radwan, 2009), and a few other studies have demonstrated the effective removal of deleterious alleles from populations via sexual selection (Radwan, 2004; Radwan et al., 2004; Sharp & Agrawal, 2008; Hollis et al., 2009). However, our results illustrate a potential problem for the purging hypothesis, because it is assumed that deleterious alleles driven to a lower frequency by sexual selection result in an overall fitness benefit in females or in juveniles. We found no significant correlation between inbreeding depression in male reproduction and other traits. More empirical work is needed to determine whether sexual selection can in general alleviate mutational load in populations.

Inbreeding depression and environmental stress

Although we found that environmental stress (larval competition) increased inbreeding depression, we found no correlation between the benign and competitive environments in the inbreeding depression for larval survival (LCI r = −0.24) or in cumulative inbreeding depression (r = −0.09 female, r = −0.05 male). Several studies examining how purging of genetic load in different environments affects extinction rates suggest that environmental stress may increase the effectiveness of purging (Bijlsma et al., 2000; Swindell & Bouzat, 2006). However, the implications for population persistence are unclear if purging is environment specific and purging provides no fitness benefit in novel environments (Bijlsma et al., 1999; Leberg & Firmin, 2008).

A related phenomenon occurs when an environmental stress affects only one life-history stage of an organism. Does this stress amplify inbreeding depression across the entire life cycle of an organism or does it reduce inbreeding depression in later life-history stages because of genetic purging as a result of greater selection against individuals expressing deleterious recessive alleles (Armbruster & Reed, 2005; Waller et al., 2008)? We tested these hypotheses by measuring inbreeding depression during exposure to stress and after the stressor have been removed. We found that competitive stress only amplified inbreeding depression during the stage at which it was applied. Inbreeding depression affecting larval survival was substantially increased under conditions of competitive resource stress (Table 3, Fig. 1); however, female fecundity and male mating success, measured after the stress was applied, did not show increased inbreeding depression (Tables 2 and 3, Fig. 1). It is possible that the purging of individuals with the highest genetic load at the stressful stage (larval-to-adult development) could explain why inbreeding depression in the adult fitness stages was not increased (Table 3). However, the evidence for purging is weak as the among-family correlation linking inbreeding depression at the larval and reproductive stages is very close to zero under both benign and stressed conditions. This does not exclude the possibility of a purging effect, but for purging at an early stage to increase later fitness requires that some of the same deleterious alleles affected both stages, a scenario that would typically generate a positive correlation between inbreeding depression at the two stages under benign conditions.

Work in the blue columbine (Aquilegia caerulea) suggested that exposure to harsher conditions early in life (field vs. greenhouse germination) may lower inbreeding depression at later stages (Montalvo, 1994). This is presumably because of a reduction in selection against deleterious alleles under benign greenhouse conditions during the seedling period. Several other studies in plants show similar patterns of greater inbreeding depression for early life history (seed survival) than adult fitness (plant size) when seeds experienced stressful field conditions (Schoen, 1983; Kohn, 1988). Note that this role of purging cannot be detected in studies comparing lines or families that are genetically identical (e.g. when specific chromosomes are made homozygous).

Implications for inbreeding avoidance in D. melanogaster

In general, individuals that employ mechanisms to avoid mating with relatives have a selective advantage over those that do not, driving the evolution of mechanisms to avoid inbreeding (Pusey & Wolf, 1996; Panhuis & Nunney, 2007). This is because the genetic load of recessive deleterious alleles that cause inbreeding depression also creates conditions that favour genotypes that avoid inbreeding. Little is known about the levels of full-sib mating and the associated fitness costs in wild populations of Drosophila, the two factors that determine the strength of selective forces driving the evolution of avoidance mechanisms. In two cactophilic species of Drosophila and in D. melanogaster, it has been observed that females appear to reduce sperm use from related males, which may be beneficial by reducing inbreeding depression in their offspring (Markow, 1997; Panhuis & Nunney, 2007). Several other studies provide circumstantial evidence supporting the existence of post-mating, prefertilization inbreeding avoidance (PPIA) in D. melanogaster. For example, sperm competitive ability has been shown to decrease with the degree of relatedness between males and females (Clark et al., 1995, 1999; Clark & Begun, 1998; Mack et al., 2002). Our work clearly demonstrates massive fitness costs of mating with a sibling (Table 5, Fig. 1) in recently caught D. melanogaster especially if there is larval competition. Larval competition is found in nature (Nunney, 1990), and such large fitness costs provide a strong selective environment in which PPIA could evolve in this species. The level of sib mating is still unknown in wild populations of D. melanogaster; however, it appears that flies generally mate before dispersing from their natal site (unpublished data) so the frequency of such matings could be significant.

Finally, the way in which inbreeding depression is measured, using either a competitive index (LCI or MCI) versus using uncorrected per cent survival or mating success, is important in an ecological context. Estimating inbreeding depression using a competitive index is representative of situations in nature where multiple females lay eggs on a single fruit. Under these conditions, the inbred offspring of a female that has mated to a sibling are potentially competing against outbred offspring from other females. Alternatively, if only a single female lays eggs on a fruit, then raw per cent values would be an appropriate measure of fitness because inbred offspring are only competing with other inbred offspring. Inbreeding depression would be expected to be less under these conditions compared to conditions where multiple females lay eggs on a single fruit.


We thank Mariel Alvarez and Magdalene Moy for their help with this experiment and the Mayo Family and Galante wineries for providing wild populations of flies. Curt Adams, Senanu Pearson, Heather Taft, and Darren Rebar and two anonymous reviewers provided helpful comments on drafts of this manuscript. L. Enders thanks Sandy Enders, Demian Enders, Dora Kubitz and James Nitao for their continual support and Jack Werren, Allen Orr and Mr. Krull for inspiring her to pursue a career in science. This research was in part supported by a NSF Doctoral Dissertation Improvement Grant (DEB 0808416).


Appendix 1

Calculation of level of inbreeding (F) in chromosome homozygote studies in Table 6. Based on the published genome sequence of D. melanogaster (Adams et al., 2000), when chromosomes II and III are made homozygous, this corresponds to F = 1 for roughly 0.37 and 0.44 of the autosomal genome, respectively. For traits expressed in both sexes or only in males, the sex chromosomes are not expected to contribute to overall levels of inbreeding depression owing to the removal/purging of recessive deleterious alleles in males (e.g. egg-adult viability, Eanes et al. (1985)). Therefore, we defined F = 0.37 and F = 0.44 for chromosome II and III homozygous, respectively, for egg-adult viability and male fitness measure. These values are the minimum level of inbreeding (F) expected when each chromosome is made homozygous. To account for sex-limited genetic load expressed on the X-chromosome for female fertility (Tracy & Ayala, 1974; Eanes et al., 1985), we adjusted the level of expected inbreeding (F) for female fecundity (increased 18%) to include the X-chromosome being totally inbred. During the creation of lines homozygous for single chromosomes, the remaining portion of the genome is expected to become randomly inbred as a result. Mackay (1985) adjusted for this in the estimation of the level of inbreeding (F), and therefore, the calculated F = 0.65 was used in Table 6.