Ancestral populations perform better in a novel environment: domestication of medfly populations from five global regions

Abstract

Geographically isolated populations of a species may differ in several aspects of life-history, morphology, behavior, and genetic structure as a result of adaptation in ecologically diverse habitats. We used a global invasive species, the Mediterranean fruit fly to investigate, whether adaptation to a novel environment differs among geographically isolated populations that vary in major life history components such as life span and reproduction. We used wild populations from five global regions (Kenya, Hawaii, Guatemala, Portugal, and Greece). Adult demographic traits were monitored in F₂, F₅, F₇ and F₉ generations in captivity. Although domestication in constant laboratory conditions had a different effect on the mortality and reproductive rates of the different populations, a general trend of decreasing life span and age of first reproduction was observed for most medfly populations tested. However, taking into account longevity of both sexes, age-specific reproductive schedules, and average reproductive rates we found that the ancestral Kenyan population kept the above life history traits stable during domestication compared to the other populations tested. These findings provide important insights in the life-history evolution of this model species, and suggest that ancestral medfly populations perform better than the derived – invasive ones in a novel environment.

Introduction

Domestication defined as “…the evolutionary genetic change arising from the transition of a population from nature to deliberate human cultivation” (Simoes et al., 2007) provides both a theoretical and experimental framework for testing biological hypotheses on the effects of important evolutionary processes such as selection, adaptation, genetic drift, and adaptation on the life history and behaviour of model organisms (Prasad & Joshi, 2003). Apart from its significance in many fields of basic research, domestication itself is also of great practical importance in both captive breeding of endangered species in conservation programs (Frankham, 2008), and mass rearing of insect strains for potential use in field releases as biological control agents (Miyatake, 1998; Miyatake & Yamagishi, 1999). However, long-term adaptation to captivity is often related to problems that render difficult the interpretation of findings derived from such studies. For example, the loss of genetic variability often observed in long-established laboratory insect populations, may alter not only the expression of genes related to important life history traits, but also the genetic correlations between them (Hoffmann et al., 2001; Sgro & Partridge, 2000). Therefore, the documentation and understanding of patterns of change during adaptation to captivity in key life history traits is a prerequisite for the assessment of the true biological significance of results obtained by studies utilizing domesticated populations (Sgro & Partridge, 2000; 2001).

Evolutionary theory predicts that a population introduced to a new environment (e.g. laboratory) is expected to undergo adaptation that is an increase in mean fitness due to genetic change (Matos et al., 2000). Different populations evolving in the same environment may ultimately converge in many behavioural and life history traits (Futuyma, 2005; Matos et al., 2004) However, it is possible populations of a species originating from different geographic regions with a wide range of climatic conditions, to follow different evolutionary pathways during adaptation to the laboratory, due to differences in their genetic background (Griffiths, Schiffer & Hoffmann, 2005).

The Mediterranean fruit fly (medfly), Ceratitis capitata (Diptera: Tephritidae) is a highly polyphagous insect pest (Liquido, Cunningham & Nakagawa, 1990) that today exhibits an almost cosmopolitan geographical distribution due to its high invasive potential (Bonizzoni et al., 2001). It also constitutes a significant model species in demographic and ageing research (Carey, 2003). Genetic studies utilizing extensive sampling throughout the medfly range of distribution suggest that the region of sub-Sahara East Africa (Kenya) probably represents the source area of the species, since Kenyan populations carry the highest levels of genetic variability (Gasperi et al., 2002). The gradual loss of genetic variability from eastern and southern Africa to the Mediterranean combined with the historical aspects of medfly global invasion point out that Spain was most likely the species entry pathway along the Mediterranean Basin to the East (Fimiani, 1989; Gomulski et al., 1998). Independent and repeated colonization events from both the Mediterranean and Africa due to increased human mobility and trading activities probably account for the more recent invasion of medfly to Latin America and the Pacific (Malacrida et al., 2007). Medfly presence has also been confirmed in North America with the species believed to be established in California (Carey, 1991).

Medfly adaptation to captivity may require several generations (Economopoulos, 1992; Rossler, 1975; Souza, Matioli & Souza, 1988). The successful establishment of a wild medfly population to the laboratory and long term rearing on an artificial diet, is usually accompanied with an increase in adult fitness attributed to both a reduction in preoviposition period and an increase of female fecundity (Vargas & Carey, 1989). This is supported by a wealth of studies focusing on the demographic components of both laboratory adapted (Economopoulos, 1992; Krainacker, Carey & Vargas, 1987; Vargas & Carey, 1989; Vargas et al., 1997), and wild medfly populations (Carey, 1984; Diamantidis et al., 2009; Harris & Carey, 1989; Papadopoulos, Katsoyannos & Carey, 2002). However, a comparison among medfly populations from different geographic localities in key life history traits during domestication is not yet available.

In a series of previous articles we demonstrated that medfly populations from geographically distant locales exhibit differences in major demographic and behavioural traits (Diamantidis, Carey & Papadopoulos, 2008; Diamantidis, Papadopoulos & Carey, 2008; Diamantidis et al., 2009). Here we test the hypothesis that geographically isolated medfly populations with a different genetic background share a similar demographic response during domestication. We predict that in the novel, constant laboratory environment (a) adult longevity and age of first reproduction decreases, (b) reproductive rates increase, and (c) changes in the previous traits would be proportionally similar in all populations. Finally, (d) age-specific reproductive schedules are expected to follow a similar pattern during domestication among all populations.

Materials and Methods

Experimental conditions and flies used

The experiments were conducted in the laboratory of Entomology and Agricultural Zoology at the University of Thessaly during autumn 2005-spring 2008, at 25 ± 1°C, 65 ± 5% R.H., and a photoperiod of L14:D10 with photophase starting at 0700 h. Light was provided by daylight fluorescent tubes and by natural light from four windows with the intensity inside the test cages ranging from 1500 to 2000 Lux.

We tested five medfly populations originating from the following global regions: (a) Extra–Mediterranean (Portugal, Madeira, lat: 32.74, lon: −16.98, host: Prunus persica), (b) Africa (Kenya, Nairobi, lat: −1.27, lon: 36.8, host: Coffea arabica), (c) Mediterranean (Greece, Chios, lat: 38.47, lon: 25.99, host: Citrus aurantium), (d) Central America (Guatemala, Antigua, lat: 14.56, lon: −90.74, host: C. arabica), and (e) Pacific (Hawaii, Kauai, lat: 22.03, lon: −159.32, host: C. arabica). Collections of wild medflies from the above fruit hosts were made by harvesting infested fruits and allowing the larvae to pupate under laboratory conditions. Pupae retrieved were transported by a courier agency to our laboratory in Volos, Greece. A total of 500–1000 pupae were introduced for each medfly population. On a previous study, we demonstrated that the above populations have evolved different survival and reproductive schedules (Diamantidis et al., 2009). Female cohorts were classified into either short-lived (Guatemala, Hawaii, and Kenya) or long-lived (Portugal, Greece). Males on the other hand were grouped differently with only the Guatemalan population being short-lived and the remaining four forming a group of long-lived populations. Although lifetime fecundity rates were similar among populations large differences were observed in their age-specific reproductive patterns.

Rearing methods

We reared all five medfly populations for 10 generations under identical laboratory conditions. Rearing of adult flies was accomplished by keeping them, after emergence, in groups of about 100 individuals (approximately 50 males and 50 females) in wooden, (30 by 30 by 30 cm), wire-screened cages provided with water and a standard adult diet consisting of a mixture of yeast hydrolysate, sugar, and water (YS) at 4:1:5 ratio respectively. Females were allowed to oviposit on 5 cm diameter hollow, plastic hemispheres of red colour (domes) that were artificially punctured with 40–50 evenly distributed holes on their surface. Eggs were deposited on the inner surface of the dome. Each dome was fitted in a 5 cm diameter hole made on the cover of a 5.5 cm diameter plastic petri dish. Water was placed in the base of the petri dish in order to maintain humidity levels beneath the dome adequate enough for female oviposition (Boller, 1985). A plastic cup containing 0.5 ml of orange juice was also placed in the base of the petri dish to stimulate oviposition. Immatures were reared (same density of 50–100 eggs per food amount for all populations) on an artificial diet consisting of 200 g sugar, 200 g brewer’s yeast, 100 g soybean flour, 4 g salt mixture, 16 g ascorbic acid, 16 g citric acid, 3 g sodium propionate, and 1 l water (Boller, 1985).

Collecting eggs from female medflies of specific age-classes (e.g. young or old flies) after long-term laboratory adaptation, may lead to over-representation of their offspring in each new generation (Meats, Holmes & Kelly, 2004). This fact may affect important adult life history traits, such as survival and age-specific reproductive schedules. In order to minimize the effects of such an “inadvertent” selection on adult fitness of the following generations, we collected eggs for all five populations during their rearing in the laboratory for 10 generations, from females that belonged to three different age-classes: 15–20, 25–30, and finally 35–40 days old flies. Adults obtained after rearing immatures from egg collections of the different age-classes mentioned above were randomly mixed to produce the adult population of the next generation.

Adult demographic traits

Demographic data for adults were collected in F₂, F₅, F₇, and F₉ generation for all five populations used. Immediately after adult emergence pairs consisting of a male and a female from each medfly population and generation mentioned above, were placed into cages containing adult diet, water, and oviposition substrates (domes). Mortality and female fecundity were recorded as a daily egg count until death. We ran 50 replicates for each generation and medfly population tested, with the exception of F₂ flies from Guatemala where 100 replicates were used. Therefore, a total of 2100 individuals (1050 individuals for each sex) were tested.

Statistical analyses

Survival analysis methodology was employed in an effort to quantify the effect of possible explanatory covariates on survival. Effects of sex, population (categorical variables), and adaptation to laboratory (quantitive variable) on life span were assessed using the Cox proportional hazards model, λ(t) = λ₀ (t) e^{ηi (Xi)} (1) (Collett, 2003). The Cox proportional hazards model relates survival to one or more explanatory covariates that may be associated with it. The model consists of the baseline hazard function λ₀ (t), which describes how the hazard (risk of death) changes over time at baseline levels of covariates, and the effect parameters, describing how the hazard varies in response to explanatory covariates. The estimated Cox model with sex, population, and adaptation to laboratory as covariates is given by (1) where η_i(X_i) =β₁population(1)_i + β₂population(2)_i + β₃population(3)_i + β₄population(4)_i + β₅adaptation_i + β₆adaptation_ipopulation(1)_i + β₇adaptation_ipopulation (2)_I + β₈adaptation_ipopulation(3)_i + β₉adaptation_ipopulation(4)_I + β₁₀sex_i + β₁₁population(1)_isex_i + β₁₂population(2)_isex_i + β₁₃population(3)_isex_i + β₁₄population(4)_isex_i + β₁₅adaptation_isex_i + β₁₆adaptation_ipopulation(1)_isex_i + β₁₇adaptation_ipopulation(2)_isex_i + β₁₈adaptation_ipopulation(3)_isex_i + β₁₉adaptation_ipopulation(4)_isex_i (1a) is the linear component of the model, λ(t) and λ₀ (t) denote the hazard rate conditional on the observed covariates and baseline hazard respectively. All covariates are indicator functions with sex_i = 1 when subject is male (0 else), population1_i, population2_i, population3_i, and population4_i = 1 when subject originated from Greece, Portugal, Guatemala, and Hawaii respectively (0 else), so that females from Kenya form the baseline. The covariate effects on life span are measured by the beta coefficients in the linear component of the model. Kaplan–Meier estimators of pre-oviposition, oviposition and post-oviposition periods were calculated in all generations (F₂, F₅, F₇, and F₉) for each population. Pair-wise comparisons of the respective periods between different generations for each population were conducted using the logrank (Mantel–Cox) test. The effects of population and adaptation to the laboratory on fecundity rates were determined by two-way analyses of variance (ANOVA) with interaction (Sokal & Rohlf, 1995). Receiver Operating Characteristics (ROC) curve analysis was used for the comparisons of oviposition distributions within each medfly population along the lines described in Alonzo et al. (2009). The significance of the separation of oviposition distributions was assessed based on the index of the area under the ROC curve (AUC). A significant result for the AUC implies that two generations of the same population have significantly different oviposition distributions. Data analysis was performed using R 2.10.1 (http://www.r-project.org) and SPSS 17.0 (SPSS Inc., Chicago, IL).

Results

Survival

Female and male average and maximum life span in F₂, F₅, F₇, and F₉ generations are given in Tables 1 and 2 respectively. With the exception of females from Kenya, average life span of females was reduced from F₂ to F₉ generation for all populations tested (Table 1). However, this reduction was significant only for females from Portugal. Male life span was also significantly reduced during laboratory adaptation from F₂ to F₉ generation for all five populations, except for those from Guatemala and Kenya (Table 2). The reduction in both female and male life span was especially noteworthy (≈ 22 and 45 days respectively) for the Portuguese population. Males outlived females in all generations and populations tested (Tables 1 and 2).

Table 1.

Average female life span of five medfly populations in F₂, F₅, F₇, and F₉ generation. Fifty pairs of adults (replicates) were used for each population and generation, apart from F₂ generation from Guatemala where 100 replicates were used.

Population   origin	Average life span (days ± SE)
	F2		F5		F7		F9
	Average	Maximum	Average	Maximum	Average	Maximum	Average	Maximum
Kenya	39.2 ± 2.2 a	81	36.9 ± 2.2 a	82	42.6 ± 1.9 a	71	41.4 ± 2.0 a	74
Portugal	60.2 ± 3.4 a	130	52.3 ± 4.7 a	150	46.6 ± 4.2 ab	140	38.1 ± 2.6 b	94
Greece	59.8 ± 3.3 a	108	51.4 ± 2.7 a	99	54.5 ± 1.9 a	78	52.0 ± 3.0 a	87
Hawaii	31.9 ± 1.8 a	69	28.5 ± 1.3 a	56	31.5 ± 1.5 a	72	27.5 ± 1.5 a	61
Guatemala	33.9 ± 1.7 a	73	33.9 ± 2.7 a	89	37.0 ± 2.4 a	88	33.1 ± 2.4 a	85

Averages within rows followed by the same letter are not significantly differed (log-rank test; P > 0.05)

Table 2.

Average male life span of five medfly populations in F₂, F₅, F₇, and F₉ generation. Fifty pairs of adults (replicates) were used for each population and generation, apart from F₂ generation from Guatemala where 100 replicates were used.

Population   Origin	Average life span (days ± SE)
	F2		F5		F7		F9
	Average	Maximum	Average	Maximum	Average	Maximum	Average	Maximum
Kenya	77.4 ± 6.7 a	245	71.2 ± 4.0 a	140	73.1 ± 6.7 a	213	70.6 ± 6.1 a	172
Portugal	105.6 ± 7.6 a	192	64.9 ± 7.9 b	192	50.7 ± 5.4 b	169	60.9 ± 5.1 b	194
Greece	83.1 ± 6.0 a	169	64.6 ± 4.0 b	148	69.1 ± 4.2 b	120	68.0 ± 4.9 b	155
Hawaii	48.6 ± 4.7 a	163	51.5 ± 4.0 a	109	54.2 ± 4.4 a	133	35.6 ± 3.1 b	103
Guatemala	52.1 ± 3.7 a	167	46.2 ± 5.5 a	154	52.4 ± 6.6 a	176	52.4 ± 6.8 a	161

Averages within rows followed by the same letter are not significantly differed (log-rank test; P > 0.05)

The estimation of beta coefficients (Appendix 1) of the model (1a) shows a significant effect of population (x² = 48.1; df = 4; P < 0.001), sex (x² = 31.3; df = 1; P < 0.001), and laboratory adaptation (x² = 13.8; df = 1; P < 0.001) on mortality rates. In the same analysis, the interaction between population and laboratory adaptation was also significant (x² = 12.3; df = 4; P < 0.05), pointing towards a differential effect of laboratory adaptation on mortality rates of the different populations.

Reproduction

Domestication reduced preoviposition period from F₂ to F₉ generation for all five populations tested but Kenyan (Table 3). However, the other two segments of reproductive life span did not follow a consistent trend while adapting to captivity among all populations used. Nonetheless, both oviposition and post-oviposition periods did not differ between the F₂ and the F₉ in Kenyan flies (Table 3).

Table 3.

Reproductive periods of females for five medfly populations in F₂, F₅, F₇ and F₉ laboratory generations. Females were ovipositing in hollow, plastic hemispheres.

Reproductive   periods-   Origin	Average duration (days ± SE)
	F2	F5	F7	F9
Preoviposition
Kenya	9.6 ± 0.6 ab	7.6 ± 0.3 c	10.8 ± 0.7 a	8.7 ± 0.4 bc
Portugal	19.3 ± 2.3 a	20.4 ± 2.8 a	17.3 ± 1.9 a	11.1 ± 0.6 b
Greece	18.7 ± 1.2 a	17.8 ± 1.1 ab	17.2 ± 1.4 ab	15.7 ± 0.7 b
Hawaii	13.0 ± 0.8 a	11.6 ± 0.8 ab	9.9 ± 0.5 b	9.8 ± 1.1 b
Guatemala	17.8 ± 0.9 a	15.0 ± 1.3 ab	13.6 ± 1.1 bc	10.9 ± 0.7 c
Oviposition
Kenya	21.7 ± 1.5 a	24.6 ± 1.6 a	25.9 ± 1.6 a	26.8 ± 1.4 a
Portugal	36.7 ± 2.7 a	31.3 ± 3.5 ab	28.8 ± 3.6 ab	25.3 ± 2.3 b
Greece	35.2 ± 2.3 a	31.8 ± 2.4 a	32.2 ± 1.7 a	36.4 ± 2.6 a
Hawaii	15.9 ± 1.6 ab	13.8 ± 0.9 b	18.5 ± 1.4 a	12.8 ± 1.3 b
Guatemala	19.8 ± 1.6 a	16.8 ± 2.1 a	21.8 ± 1.8 a	20.2 ± 2.2 a
Post oviposition
Kenya	7.8 ± 1.0 a	4.6 ± 0.6 b	6.4 ± 1.2 ab	6.0 ± 0.9 ab
Portugal	4.9 ± 1.1 a	4.8 ± 0.8 a	6.3 ± 1.4 a	3.1 ± 0.4 b
Greece	4.2 ± 0.8 b	3.4 ± 0.5 b	6.2 ± 0.7 a	3.5 ± 0.6 b
Hawaii	4.2 ± 0.6 b	4.8 ± 0.8 ab	3.2 ± 0.5 b	6.6 ± 0.7 a
Guatemala	2.5 ± 0.3 a	4.6 ± 1.1 a	3.7 ± 1.3 a	3.0 ± 0.5 a

Averages within rows followed by the same letter are not significantly differed (log-rank test; P > 0.05)

Reproductive parameters (Carey, 1993) of the five medfly population in F₂, F₅, F₇, and F₉ laboratory generations are given in Table 4. Among all five populations tested females from Kenya exhibited the lowest variation/fluctuation in both net and gross fecundity rates while adapting to laboratory conditions (Table 4). Analysis of variance revealed a significant effect of population (F=28.2; df=4; P<0.01) and a non-significant effect of adaptation to laboratory (F=2.2; df=3; P=0.08) on female fecundity. In the same analysis the interaction between population and laboratory adaptation was significant (F=4.1; df=12; P<0.01), suggesting that adaptation to laboratory conditions had a different impact on oviposition rates of the five populations tested.

Table 4.

Reproductive parameters of five medfly population (along with 95% confidence intervals produced using standard bootstrap methodology for gross fecundity) in F₂, F₅, F₇ and F₉ laboratory generation. Females were ovipositing in hollow, plastic hemispheres.

Origin	Fecundity
	F2		F5		F7		F9
	Net	Gross	Net	Gross	Net	Gross	Net	Gross
Kenya	433.6 (366.8, 500.3)	534.3 (504.7, 563.9)	525.3 (437, 613.5)	750.4 (709.2, 791.5)	472.7 (434.3, 551.1)	546.9 (518.2, 575.5)	486.1 (406.3, 565.8)	602.6 (570.5, 634.8)
Portugal	371.4 (295.2, 447.6)	584.9 (547.9, 621.9)	272.8 (189.4, 356.1)	615.7 (570.2, 661.3)	240.7 (166.7, 314.6)	507.8 (472.8, 542.7)	440.4 (349.9, 530.8)	728.4 (683.4, 773.3)
Greece	382.2 (299.9, 464.3)	555.4 (520.8, 590)	495.6 (398.2, 592.9)	699.8 (661.6, 738)	426.4 (360.1, 492.6)	489.1 (465.9, 512.2)	377.4 (296.8, 457.9)	627.2 (598.6, 655.8)
Hawaii	180.8 (116.9, 244.5)	338.6 (301.5, 375.7)	131.7 (92, 171.1)	162.8 (146.4, 179.2)	404.8 (320.9, 488.6)	609.9 (572, 647.7)	221.6 (149.6, 293.6)	319.6 (283.7, 355.6)
Guatemala	230.3 (174.4, 286.2)	487.6 (462.2, 513)	210.7 (137.7, 283.6)	385 (352.6, 417.5)	318.1 (235.3, 400.9)	551.8 (516.7, 586.9)	269.3 (191.9, 346.6)	639.3 (596.7, 681.9)

Net fecundity $(\text{eggs})={\displaystyle \sum _{\mathrm{x}=\mathrm{\alpha }}^{\mathrm{\beta }}{\mathrm{l}}_{\mathrm{x}}{\mathrm{M}}_{\mathrm{x}}}$, Gross fecundity $(\text{eggs})=({\displaystyle \sum _{\mathrm{x}=\mathrm{\alpha }}^{\mathrm{\beta }}{\mathrm{M}}_{\mathrm{x}})}$, x = age interval in days, a = age at start of reproduction, b = age at end of reproduction, l_x = Proportion of individuals surviving to age x, M_x = Total numbers of eggs laid by the average female at age x (Carey, 1993)

An event-history diagram (Carey et al., 1998) showing age-specific reproductive patterns and heterogeneity in the five medfly populations in F₂, F₅, F₇, and F₉ laboratory generations is given in Fig. 1. Two specific aspects of this figure merit comment. First, striking differences exist within and among the five populations tested in their age-specific reproduction schedule while adapting to laboratory conditions. For example, females from Hawaii in F₇ and F₉ laboratory generations and those from Portugal in F₉ generation, compressed the period of high egg-laying days (> 20 eggs) in a limited reproductive period compared to F₂ generation. In addition, a big proportion of Hawaiian females died without laying a single egg in F₅ and F₉ laboratory generations, a trend also observed for females from Guatemala in F₂ and F₅ generation. Second, a consistent pattern of age-specific reproduction schedule was observed throughout all generations only for the population of Kenya. This pattern involves zero egg production up to day 10, followed by high egg-laying activity (> 20 eggs) between days 10 and 30.

Figure 1.

Event-history diagram (Carey et al., 1998) of the five medfly populations in F₂, F₅, F₇, and F₉ laboratory generations. Each horizontal line represents the longevity of a single female and different colors designate the level of reproduction for each age. Green 0 eggs, yellow 1–20 eggs, and red > 20 eggs. Fifty to 100 individuals were tested for each medfly population.

Adaptation to captivity had also a differential effect on per capita egg production of the five medfly populations tested (Fig. 2). For example, a substantial decline in oviposition levels was observed in F₅ generation from Hawaii, whereas a large increase in F₉ generation from Portugal compared to the other three generations. Again, females from Kenya exhibited uniform oviposition dynamics during domestication (Fig. 2). Kenyan females reached maximum oviposition levels around day 20 with the average female laying 25–30 eggs per day. Peak egg production showed a gentle decrease after day 20 and reached zero around day 60 in all four generations tested for the Kenyan cohorts. These results were further supported by the use of ROC curve analysis for the pair wise comparisons of oviposition distributions of the different generations within populations. Specifically, oviposition distributions of the different generations of the Kenyan flies do not differ significantly, while oviposition distributions within the other four medfly populations exhibit inconsistent patterns (Appendix 2).

Figure 2.

Age-specific reproduction schedule for five medfly populations in F₂, F₅, F₇, and F₉ laboratory generations. Females were ovipositing in hollow, plastic hemispheres. Fifty to 100 individuals were tested for each medfly population.

Discussion

Three important findings emerge from our study. First, flies from Kenya maintained their life-history traits stable while adapting to the laboratory compared to the other four populations. Second, domestication differentially affected both mortality and oviposition rates of the different populations. Third, a tendency of shortened life span combined with early reproduction was observed for most medfly populations during domestication. In a previous work we demonstrated that medfly populations from six global regions have evolved different life history strategies as a result of selection in ecologically diverse habitats (Diamantidis et al., 2009). Despite the fact that domestication created convergence among populations in the evolution of two demographic traits, longevity and age of first reproduction, notable differences in life span among populations still exist after adaptation in a constant environment.

Taking into consideration the longevity of both sexes combined with net reproductive rates, the age-specific reproductive schedules, and the reproductive periods the population from Kenya was the most stable among all populations tested during domestication. The establishment of an insect colony to the laboratory resembles to the early phases of an invasion, during which a significant loss of genetic variability, compared to the source population, occurs (Dlugosch & Parker, 2008). Similarly if a large number of individuals are wiped out due to poor adaptation in the initial stages of artificial rearing in the laboratory, then bottleneck effects may drastically decrease the genetic variability of the population (Hoffmann et al., 2001; Miyatake, 1998). Despite the fact that adaptation may be accomplished even with very low genetic variability (Zayed, Constantin & Packer, 2007), it is generally thought that high within-population genetic variability, which is related to both presence of multiple alleles and different allele combinations per locus, enhances the adaptive ability of populations to new environments (Ciosi et al., 2008). Mean expected heterozygosity of a population may also be crucial to adaptation since it affects the dynamics of individuals to produce descendants with high genetic variability (Futuyma, 2005). A wealth of studies dealing with medfly worldwide genetic variation, have demonstrated that the populations derived from Kenya are the most highly polymorphic in both terms of mean number of alleles per locus and mean expected heterozygosity (Bonizzoni et al., 2000; Gasperi et al., 2002; Malacrida et al., 1998). Therefore it seems plausible that the sufficient genetic load of the ancestral Kenyan population provides plasticity for better adaptation to captivity, as expressed by a stable adult demographic profile over rearing for ten generations in artificial conditions.

Data analyses have shown significant interaction effects between domestication and population on both mortality and oviposition rates. Again this result could be attributed to the fact that medfly populations inhabiting ecologically distinct habitats are locally adapted and genetically differentiated (Gasperi et al., 2002). Newly introduced medfly populations of a given genetic profile may face a series of adaptational challenges during the establishment of a colony to a novel environment (e.g. laboratory). These challenges mainly include oviposition on artificial substrates, feeding of larvae on an artificial diet rather than fruit, high fly densities that may cause changes in courtship and mating behaviour, and finally constant laboratory conditions (temperature, relative humidity, and photoperiod) that vary greatly from those prevailing in the area of origin (Cayol, 2000). Hence, genetically differentiated medfly populations subjected to the selection pressures mentioned above during adaptation to a constant environment may produce a differential demographic response as observed in our study.

In general, adult longevity followed a decreasing trend for most medfly populations tested during the colonization to the laboratory process. Preoviposition period also showed a tendency to diminish throughout colonization for all populations tested but Kenyan. Previous studies also demonstrate that domesticated medfly populations are rather short lived and initiate egg-laying activity at very young ages compared to wild strains (Vargas & Carey, 1989; Vargas, Miyashita & Nishida, 1984). In a comparative study between one laboratory and four wild medfly populations from the Hawaiian Islands, Vargas and Carey (1989) suggest that during adaptation in a sheltered environment, where extrinsic mortality factors reach almost zero and food is not a limiting factor, selection probably favours individuals with high growth rates. However, these high growth rates incur a longevity cost since laboratory adapted strains are usually short lived. Therefore, it seems plausible that the removal of selective pressures such as climatic variability and uncertainty of host fruits during domestication, selects against extreme long lived individuals with abnormal reproductive patterns. Shortened longevity combined with a short preoviposition period during adaptation to the laboratory had also been the case in previous studies dealing with other members of the Tephritidae family, such as the melon fly Bactrocera cucurbitae (Miyatake, 1998) and the West Indian fruit fly Anastrepha obliqua (Hernandez et al., 2009). This latter result could be attributed to either a genetic trade-off between early fecundity and longevity or the accumulation of deleterious mutations late in life (Zwaan, Bijlsma & Hoekstra, 1995). The negative genetic correlation between early fecundity and longevity could have been selected for through generations of artificial rearing and probably represents a component of an optimal life history of the population (Miyatake, 1998). Although an effort had been made to avoid such effects by mixing the descendants of different age classes in each generation, such trade-offs may also exist for medfly during adaptation to captivity and explain the decrease in longevity and preoviposition period observed in our study for most of the populations tested.

During the last few decades, the growing need for environmental friendly and target-specific methods against medfly has lead to the increased implementation of an area-wide approach using the Sterile Insect Technique (SIT) (Dyck, Hendrichs & Robinson, 2005; Hendrichs, Franz & Rendon, 1995; Hendrichs et al., 2002). In mass rearing facilities where space is a limited factor and millions of individuals should be produced under finite financial resources and hours of labor, beneficial traits such as stable longevity and egg production are essential (Hernandez et al., 2009). The results of our study indicate that wild medfly populations originating from the sub-Saharan East Africa (Kenya) perform better in captivity, and therefore, may be used as source material in order to develop a mass-reared laboratory strain with desirable life history traits.

In conclusion, our study shows that the ancestral Kenyan population adapts better to a novel environment (e.g. laboratory) as expressed by keeping stable several important adult life history traits during domestication. This fact is most likely due to the high levels of heritable variation in fitness traits that endows this population with sufficient plasticity for adaptive evolutionary change (Carvalho, 1993). Furthermore, we found that domestication to constant conditions in the laboratory had a different effect on the mortality and oviposition rates of the different populations. Despite this fact, a clear tendency of decreasing life span and age of first reproduction was observed for most medfly populations tested, suggesting that this is a universal property for the evolution of medfly life-history in captivity. Our findings provide important insights in the life-history evolution of this model species in captivity and suggest that ancestral medfly populations provide ideal candidates for widespread use in respective studies.