Experimental Gerontology 48 (2013) 1189–1195

Contents lists available at ScienceDirect

Experimental Gerontology
journal homepage: www.elsevier.com/locate/expgero

Laboratory selection for increased longevity in Drosophila melanogaster

reduces field performance
Janneke Wit a,⁎, Torsten Nygaard Kristensen b,1, Pernille Sarup a,2, Jane Frydenberg a, Volker Loeschcke a
a
Department of Bioscience, Integrative Ecology and Evolution, Aarhus University, Ny Munkegade 114-116, DK-8000 Aarhus C, Denmark
b
Department of Molecular Biology and Genetics, Center for Quantitative Genetics and Genomics, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark

a r t i c l e i n f o a b s t r a c t

Article history: Drosophila melanogaster is frequently used in ageing studies to elucidate which mechanisms determine the onset
Received 17 May 2013 and progress of senescence. Lines selected for increased longevity have often been shown to perform as well as or
Received in revised form 19 July 2013 superior to control lines in life history, stress resistance and behavioural traits when tested in the laboratory.
Accepted 25 July 2013
Functional senescence in longevity selected lines has also been shown to occur at a slower rate. However, it is
Available online 2 August 2013
known that performance in a controlled laboratory setting is not necessarily representative of performance
Section Editor: T.E. Johnson in nature. In this study the effect of ageing, environmental temperature and longevity selection on performance
in the field was tested. Flies from longevity selected and control lines of different ages (2, 5, 10 and 15 days) were
Keywords: released in an environment free of natural food sources. Control flies were tested at low, intermediate and high
Ageing temperatures, while longevity selected flies were tested at the intermediate temperature only. The ability of flies
Release–capture to locate and reach a food source was tested. Flies of intermediate age were generally better at locating resources
Functional senescence than both younger and older flies, where hot and cold environments accelerate the senescent decline in perfor-
Environmental temperature mance. Control lines were better able to locate a resource compared to longevity selected lines of the same age,
Longevity selection
suggesting that longevity comes at a cost in early life field fitness, supporting the antagonistic pleiotropy theory
Fitness in the field
of ageing.
© 2013 Elsevier Inc. All rights reserved.

1. Introduction
For decades scientists have investigated ageing in Drosophila
melanogaster and other model organisms in laboratory settings to elu-
cidate the underlying mechanisms, as well as to understand how ageing
has evolved. Ageing is defined by Zwaan (1999) as “the total effect of
those intrinsic changes in an organism that adversely affect its vitality
and that renders it more susceptible to the many factors that can
cause death”. Many life history and physiological traits are correlated
to the ageing process, illustrating its quantitative nature.
Life history traits are expected to trade off with one another as a re-
sult of the restriction of adaptation potential due to constraints posed by
the evolutionary history as well as the physiology of organisms
(Stearns, 1989, 1992). Many of the studies on trade-offs and correlated
traits in D. melanogaster have focussed on lines selected for increased
longevity (Bubliy and Loeschcke, 2005; Luckinbill et al., 1984; Rose,
⁎ Corresponding author. Tel.: +45 87156547.
E-mail addresses: janneke.wit@biology.au.dk (J. Wit), torsten.nygaard@nordgen.org
(T.N. Kristensen), pernille.sarup@biology.au.dk (P. Sarup), jane.frydenberg@biology.au.dk
(J. Frydenberg), volker.loeschcke@biology.au.dk (V. Loeschcke).
1
Current address: NordGen — Nordic Genetic Resource Center, Raveien 9, 1430 Å̊s,
Norway.
2
Current address: Department of Molecular Biology and Genetics, Center for
Quantitative Genetics and Genomics, Aarhus University, Blichers Allé 20, DK-8830 Tjele,
Denmark.
1984) but the trade-off between longevity, reproduction and stress re-
sistance has been studied in other lines, too (Chapman, 2001; Cordts
and Partridge, 1996; Fowler and Partridge, 1989; Harshman et al.,
1999; Hoffmann and Parsons, 1993a,b; Prowse and Partridge, 1997).
Trade-off patterns have been shown to be conflicting both between
(Chippindale et al., 1997; Force et al., 1995; Vermeulen and Bijlsma,
2006; Wit et al., 2013) as well as within (Arking et al., 2002; Leroi
et al., 1994a,b) selection regimes.
In nature, selection for increased longevity could be limited due to the
constraints posed by trade-offs in suboptimal environments (Partridge
and Barton, 1993; Stearns, 1989). These constraints might be less signif-
icant in a laboratory environment with ample food, limited competition
and likely fewer incidences of adverse biological interactions. In line with
suggestions of laboratory selection leading to correlated responses un-
like those expected in nature (Harshman and Hoffmann, 2000; Houle,
1991), lines selected for late life reproduction in our laboratory (Bubliy
and Loeschcke, 2005; Sarup et al., 2011) have previously been shown
to perform similar, if not superior, in a range of traits when compared
to control lines in a standardized laboratory environment (Wit et al.,
2013). Others, too, find that lines selected for increased longevity show
similar or increased performance in a number of correlated traits
(Arking and Wells, 1990; Chippindale et al., 1994, 1997; Force et al.,
1995; Service et al., 1985) The most notable exception to this trend is
early life reproduction (Luckinbill et al., 1984; Service, 1989; Zwaan
et al., 1995, but see Arking et al., 2002; Leroi et al., 1994a; Vermeulen

0531-5565/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.exger.2013.07.012
1190 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195

and Bijlsma, 2006) while developmental time, body size and starvation
resistance have occasionally been shown to be negatively affected by se-
lection for increased life span, too (Buck et al., 2000; Force et al., 1995;
Vermeulen et al., 2006).
Another factor influencing correlated responses in experiments
with laboratory selection lines is experimental setup. Kristensen
et al. (2008a) showed that costs as well as strong benefits associated
with cold acclimation were evident in a field-release setup but not in
the laboratory. Inbreeding effects have also been shown to be stron-
ger in field- rather than laboratory studies (Kristensen et al., 2008b),
while D. melanogaster with a genetically modified heat-stress re-
sponse also show more positive responses when tested in the labora-
tory compared to thermal assays in the field (Sørensen et al., 2008).
Thus, it appears from these studies that negative effects of genetic
manipulation, inbreeding, or thermal stress might be compensated
for in a laboratory setting, while testing in a natural environment un-
covers detrimental effects. Also, assay environment influences per-
formance in laboratory settings. This is highlighted by a strong
genotype-by-environment interaction for early life reproduction in
flies selected for late life reproduction (Leroi et al., 1994b). Overall,
these studies indicate that extrapolating results from different rear-
ing, selection and test environments requires caution, and testing
under natural or semi-natural conditions might help in elucidating
robust and ecologically relevant correlated responses.
Laboratory studies on correlated traits often use relatively young flies,
typically less than a week old (Harshman et al., 1999; Kellermann et al.,
2007; Kristensen et al., 2008a; Sørensen et al., 2008; Vermeulen et al.,
2005). When studying ageing, performance later in life can also
give information about why some flies live longer than others. It
has been shown that longevity selected lines keep performing better
in behavioural traits as they age compared to control lines, i.e., their
rate of functional senescence has decreased (Arking and Wells, 1990;
Graves et al., 1988; Wit et al., 2013). Other studies, in non-selected
lines, show a steady decline in performance over time for multiple
types of stress resistance (Piazza et al., 2009; Sørensen and Loeschcke,
2002; Stratman and Markow, 1998, but see Magwere et al., 2006). Gen-
erally, the younger the fly, the better it performs in these traits (but see
Service, 1987; Service et al., 1985).
To test if the fitness superiority of a set of longevity selected lines
(Bubliy and Loeschcke, 2005) as assessed in Wit et al. (2013) holds in
the field, a release–capture study was undertaken and the ability of
flies to locate, and reach, food in an otherwise resource-free environment
was tested. Further, the effect of temperature and age on this ability in
control lines was assessed, as well as the performance over time of con-
trol versus longevity selected lines of D. melanogaster (Bubliy and
Loeschcke, 2005). Based on laboratory results both control and longevity
selected lines are hypothesized to have a lower performance as they age,
with a less pronounced decline in the longevity selected lines. Longevity
selected lines are expected to start out at similar, if not better perfor-
mance than the control lines based on previous studies in the laboratory
of e.g., stress resistance, and a stressful environment is hypothesized to
enhance the detrimental effects of ageing.
assay the effect of longevity selection on capture patterns all 6 lines
(C1, C2, C4, LS1, LS2, and LS4) were used. Two generations prior to the
experiment the lines were transferred from 25 °C to 20 °C.
Experimental flies for the release assays were reared under standard
laboratory conditions of 20 °C, Leeds medium and 12:12 dark:light, in
200 ml culture bottles. To control density ~10 parental pairs per bottle
were allowed to lay eggs for 24 h. After hatching the flies were kept in
300 ml bottles at a density of 200 flies and tipped 3 times weekly
until the experiment was performed. Flies used for the study on the ef-
fect of temperature on capture rates were 2, 5, 10 and, for releases at low
temperature only, 15 days old. For the study on the effect of longevity
selection flies were 5, 10 and 15 days old. The ages were chosen such
that the time frame would allow for age to affect behaviour. Another
reason for the age classes chosen was that natural life span estimates
vary from a few days to a few weeks (Rosewell and Shorrocks, 1987;
Turelli and Hoffmann, 1995), thus 2 to 15 days are ecologically relevant
ages.
2.2. Longevity assay
The longevity assay was based on the setup used by Bubliy and
Loeschcke (2005) and Wit et al. (2013). In short, experimental flies for
the longevity assay were reared at a controlled density of ~40 larvae
per vial with 7 ml Leeds medium at 20 °C and 12:12 dark:light. Eclosing
flies were collected during an 8-hour window. One day after collecting
they were sexed while sedated using CO2 anaesthetics and placed at a
density of 15 males and 15 females per vial containing 4 ml Leeds me-
dium. Flies were transferred to new vials and mortality was scored
every other day until all flies had died. For each of the six lines 20 repli-
cate vials were assayed.
2.3. Field release — effects of ageing at low and high temperatures
The general setup of the field release–capture is described in Kristensen
et al. (2008b). The releases at high and low temperatures were
performed on different days. Three release sites were set up under sim-
ilar conditions in open woodland for each temperature. These releases
will be referred to as Low 1, 2 and 3 and High 1, 2 and 3 for the releases
at low and high temperatures, respectively (Table 1). Flies were
transported to the release site at a density of 200 flies in 200 ml bottles.
Temperature in the car during transportation was kept at 20 ± 1 °C.
Capture points were set up in a straight line from the release site in
two directions every 5 m until 25 m, each consisting of 1.5 l buckets
with yeasted mashed banana. Perpendicular to the direction in which
the buckets were placed, two more buckets were placed 3 m from the
original bucket on either side. At the release site, in the shade,
Table 1
Summary of the number of flies released (Released (#)) and percentage caught (Captured
(%)) per release as well as temperatures during the releases. R = release. Selection =
releases with control and longevity selected lines. High and Low = releases with
control line at high and low temperatures, respectively. Rel = temperature at time of
release.

2. Materials and methods R Released (#) Captured (%) Temperature (°C)

Rel Max Min Mean

2.1. Maintenance and origin of experimental flies
Selection 1 12,000 30.7 22 26 17 22.6
Selection 2 12,000 16.5 23 24 16 20.1
Two sets of D. melanogaster lines were used for this study; control Selection 3 12,000 15.4 23 26 16 22.5
(C) and longevity selected (LS) lines. The longevity selected lines, LS1, Selection 4 12,000 7.9 20 25 16 21.6
LS2 and LS4, had been selected for increased mated longevity as de- Selection 5 12,000 31.1 20 26 16 20.9
Selection 6 12,000 21.5 20 25 17 21.0
scribed by Bubliy and Loeschcke (2005) for 52 generations, while the
High 1 9000 6.8 28 28 27 27.3
control lines, C1, C2 and C4, were established simultaneously and kept High 2 9000 24.3 28 28 26 26.8
at a standard maintenance regime with a generation time of about High 3 9000 9.6 26.5 28 26 26.8
two weeks. Both types of lines have been kept at 25 °C, 12:12 dark: Low 1 12,000 10.8 14.5 18 15 16.7
light on a yeast–sugar–oatmeal–agar (Leeds) medium. For the study Low 2 12,000 5.6 15 17 16 16.3
Low 3 12,000 11.7 15 19 15 17.6
on the effects of ageing at low and high temperatures C1 was used. To
 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195
temperature was continuously recorded with a data logger (SmartButton,
ACR Systems Inc., Canada).
Immediately prior to the release two bottles with 200 flies of the
same regime were mixed and transferred to a vial with 4.5 mg of fluo-
rescent micronized dust (Radiant Cop., Richmond, CA). The three (at
high temperature) or four (at low temperature) colours were randomly
assigned to the age classes at each release site independently. 3000 flies
were released per age class per release site. Flies were captured from the
bait by means of a net or aspirator.
For the releases at high temperatures flies were collected three times
(four for release High 1) between 12.15 p.m. and 4.45 p.m. At the low
temperature releases flies were collected every hour for five (six for
Low 2) hours between 12.50 p.m. and 6.00 p.m. Vials with captured
flies (1 to N500 per vial) were put on ice immediately to knock them
out and prevent transfer of coloured dust. After transport back to the
laboratory they were stored at −20 °C until they were scored for colour
under ultraviolet light.
2.4. Field release — effects of selection for increased longevity on ageing
The setup was the same as described for the temperature releases.
Six releases were setup on the same day under similar conditions in
open woodland and will be referred to as Selection 1–6 (Table 1). A
day prior to the releases the three lines per selection regime and age
group were mixed at equal ratios to create six release groups: 5, 10
and 15-day-old control (C) groups, and 5, 10 and 15-day-old longevity
selected (LS) groups. Thereafter the six groups were kept in 200 ml
bottles at a density of 200 flies and were transported to the release
site in these bottles.
Prior to the release at 11.00 a.m. two bottles with 200 flies of the
same regime were mixed and transferred to a vial with 4.5 mg of fluo-
rescent micronized dust (Radiant Cop., Richmond, CA). The six colours
were randomly assigned to the different selection and age groups at
each release site independently. Flies were captured from the bait by
means of a net every hour from 12.00 p.m. until 3.00 p.m.
2.5. Statistical analysis
For all statistical analyses JMP® statistical package version 9.0.0. for
Mac was used. Survival curves were analysed by means of a Cox propor-
tional hazards test. Release data was analysed with GLM logit models as
described by Kristensen et al. (2008b). Fixed factors consisted of selec-
tion regime, age and sex. Each of the 12 releases was analysed separate-
ly. For both sets of releases the overall capture rates were analysed,
including all interaction terms. When interactions were significant the
data was split into sexes first, and subsequently by selection regime to
100
80
Survival (%)
60
40
20
Female
0
20 40 60 80 100 120
Age (days)
Fig. 1. Survival curves for longevity selected lines (open symbols) and control lines (closed symbols) for females and males.
1191
assess ageing effects, and by age to assess the effect of selection regime
on capture rates.
3. Results
3.1. Life span of longevity selected flies at 20 °C
As the LS lines were selected at 25 °C but reared at 20 °C prior
to performing the current experiments, survival at 20 °C was tested
to ensure the differential life span persisted at the rearing temper-
ature. Fig. 1 shows the survival curves for females and males of all
lines used for the release experiments. The Cox proportional haz-
ards showed no interaction effect between sex and selection
(p = 0.0637). The effects of selection regime and sex were both
overall significant (selection: χ2(df = 1) = 2016.77; p b 0.001; sex:
χ2(df = 1) = 20.24; p b 0.001). An ANOVA on the average survival
of the individual lines confirmed these findings (results not shown).
3.2. Field releases
Overall capture percentages varied between 5.6 and 31.1% in the 12
releases (Table 1). Temperature was relatively stable within but varied
between release types (Table 1). Between the first two age classes
female capture rates increased from 5 to 10% on average up to 23% in re-
lease Selection 1. For males the increase between the first two ages was
more extreme, ranging from 3 to 31% in release High 2, to on average
from 5 to 19% in the selection releases. At later ages a decrease was
seen for all three types of releases and both sexes (Fig. 2). For capture
rates with individual lines annotated, see Supplementary Fig. 1.
3.2.1. Effects of ageing at different temperatures
In the High and Low temperature releases the pattern of initial in-
crease in capture rate with age was stronger for males than for females.
Overall there was a sex effect in all High and Low temperature releases
(Table 2), though within the two release types neither sex consistently
was caught more often (Fig. 2). Both male and female capture rates are
affected by age (Table 3). All males performed better at five days of age
than they did at 2 days of age. Females showed a less consistent pattern.
In releases Low 1, Low 2 and High 1, the 5-day-old females had a similar
or lower likelihood of being caught than 2-day-old females. Compared
to the C groups in the Selection releases, 5-day-old flies overall got
caught at about the same rate, though males of the High 2 release had
a very high capture rate. At 10 and 15 days of age, the C groups in the
Selection releases were caught relatively more often than C flies in the
High and Low temperature releases for both sexes. As a result, the
100 C1
C1
C2 C2
C4 C4
L1 80 L1
L2 L2
L4 L4
Survival (%)
60
40
20
0 Male
140 160 20 40 60 80 100 120 140 160 180
Age (days)
1192 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195

20 20

15 15

Flies caught (%)

Flies caught (%)

10 10

5 5

0 Female 0 Male

2 5 10 15 2 5 10 15
Age (days) Age (days)
Release: High Low Selection C Selection L

Fig. 2. Average capture rates (%) for females and males. Low and high represent the releases conducted at low and high temperatures, respectively. Selection C and Selection L represent the
capture rates of control and longevity selected groups, respectively, in the Selection releases.

peak in the capture rate was at 5 days of age in the High and Low tem- Age affected capture rates between sexes and selection regimes in all
perature releases, and at 10 days for Selection releases. releases, except for LS females in Selection release 4 and LS males in Se-
lection release 2 (Table 3).

3.2.2. Effects of selection for increased longevity on field performance in

4. Discussion
ageing flies
The ANOVA on overall capture rates revealed significant interaction
This study addressed the effect of age on field performance and the
terms between selection regime, sex and age (data not shown). There-
extent to which this behaviour is affected by ambient temperature or
fore the data was split up into the distinct age classes, and thereafter
longevity selection. Capture rates of flies of different ages originating
by selection regime and sex to look at the effect of ageing on the capture
from an outbred population were assessed at high and low tempera-
rates. Sex but not selection consistently affected capture rate of 5-day-
tures in open woodland deprived of natural resources. In another exper-
old flies (Table 4). For 5-day-old flies selection regime significantly
iment the same setup was used but with control and longevity selected
affected capture rates in Selection releases 2 and 4 (p = 0.034 and
groups to see if selection for delayed reproduction affects capture rates
p = 0.046, respectively). Selection release 6 also showed an effect of
at different ages.
selection at 5 days of age after analysing sexes separately due to a
Capture rate was hypothesized to decrease with age as a result of
significant sex ∗ selection regime interaction term (female χ2 = 9.50,
functional senescence in all environments and lines tested. Hot and
p b 0.01; male χ2 = 35.25, p b 0.001). Because of sex ∗ selection
cold environments might be experienced as more stressful than the in-
regime interaction effects in all releases of both 10- and 15-day old
termediate temperature at which the selection experiments were
flies (Table 4), the data was split up to assess the effect of selection
conducted (Kristensen et al., 2007, 2008a, 2008c; Sørensen et al., 2001,
regime and sex on the different age classes separately (Table 5).
Capture rates of males differed between the two selection regimes
Table 3
when flies were 10 as well as 15 days old. Control males got caught χ2 results of the GLM logit model testing the effect of age on capture rates per sex for all
more often at both ages than LS males (Fig. 2). 10-day-old control releases. Selection = releases with control and longevity selected lines. High and
females were caught significantly more often than LS females, but at Low = releases with control line at high and low temperatures, respectively. R =
15 days of age the response was varied, and only significant in Selection release. Degrees of freedom = 2 for all analyses.

releases 1 (LS N C), 2 (C N LS), 4 (LS N C) and 6 (C N LS). R Control Longevity selected

Female Male Female Male

Selection 1 140.31⁎⁎⁎ 49.85⁎⁎⁎ 12.59⁎⁎ 11.98⁎⁎

Table 2 Selection 2 128.02⁎⁎⁎ 317.24⁎⁎⁎ 14.63⁎⁎⁎ 5.49
Relative capture success of females vs. males for the three releases at high (High 1–3) and Selection 3 94.07⁎⁎⁎ 197.67⁎⁎⁎ 36.12⁎⁎⁎ 8.67⁎
three at low (Low 1–3) temperatures. Significance is based on χ2 results of the GLM logit Selection 4 46.98⁎⁎⁎ 78.58⁎⁎⁎ 2.49 8.01⁎
model. R = release. Degrees of freedom = 1 for all analyses. Selection 5 51.52⁎⁎⁎ 347.49⁎⁎⁎ 12.58⁎⁎ 45.37⁎⁎⁎
Selection 6 49.22⁎⁎⁎ 249.11⁎⁎⁎ 13.56⁎⁎ 53.66⁎⁎⁎
R 2 days 5 days 10 days 15 days
High 1 125.66⁎⁎⁎ 369.85⁎⁎⁎
High 1 1.94⁎⁎⁎ 0.25 0.09⁎⁎⁎ High 2 643.22v 1500.32⁎⁎⁎
High 2 3.73⁎⁎⁎ 0.67⁎⁎⁎ 0.47⁎ High 3 261.82⁎⁎⁎ 715.39⁎⁎⁎
High 3 4.96⁎⁎⁎ 0.46⁎⁎⁎ 0.10⁎⁎⁎ Low 1 795.29⁎⁎⁎ 217.96⁎⁎⁎
Low 1 2.47⁎⁎⁎ 1.05⁎⁎⁎ 0.24⁎⁎⁎ 0.02⁎⁎⁎ Low 2 522.10⁎⁎⁎ 80.82⁎⁎⁎
Low 2 3.38⁎⁎⁎ 2.13⁎⁎⁎ 0.46⁎⁎⁎ 0.22⁎⁎⁎ Low 3 1071.14⁎⁎⁎ 287.73⁎⁎⁎
Low 3 2.03⁎⁎⁎ 1.34⁎⁎⁎ 0.14⁎⁎⁎ 0.02⁎⁎⁎
⁎ Significance (p = 0.05).
⁎ Significance (p = 0.05). ⁎⁎ Significance (p = 0.01).
⁎⁎⁎ Significance (p b 0.001). ⁎⁎⁎ Significance (p b 0.001).
 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195 1193

Table 4 shown to increase with age, while increased desiccation resistance in

Relative capture success of females vs. males (Sex) and of control vs. longevity selected lines selected for this trait is the result of an increase in percentage of
(Sel) for the six selection releases (Selections 1–6) at 5 days-of-age. Significance as
indicated for Sex and Sel is based on the χ2 test. Sex ∗ Sel: χ2 results of the GLM logit
body water content during the first days of adult life (Folk et al.,
model testing the interaction between sex and selection regime at 5, 10 and 15 days-of- 2001). In other words, as Gibert et al. (2001) suggest, flies might not
age. R = release. Degrees of freedom = 1 for all analyses. be physiologically mature at the younger ages, consequences of which
need not always be evident in laboratory assays.
R 5 days 10 days 15 days
As described in Section 3.2.1, the peak in capture rates differs be-
Sex Sel Sex ∗ Sel Sex ∗ Sel Sex ∗ Sel
tween the releases in hot and cold environments compared to that at
Selection 1 1.65⁎⁎⁎ 0.90 0.09 68.29⁎⁎⁎ 136.16⁎⁎⁎ an intermediate temperature. At 5 days-of-age capture rates in all
Selection 2 1.98⁎⁎⁎ 0.79⁎ 0.56 14.19⁎⁎⁎ 29.13⁎⁎⁎ three releases are relatively similar, but as the flies age and performance
Selection 3 2.43⁎⁎⁎ 1.20 0.43 53.33⁎⁎⁎ 18.60⁎⁎⁎
Selection 4 1.78⁎⁎⁎ 0.77⁎ 0.02 4.36⁎ 24.54⁎⁎⁎
at intermediate temperatures still increases, the more stressful environ-
Selection 5 2.10⁎⁎⁎ 1.04 0.02 36.63⁎⁎⁎ 42.86⁎⁎⁎ ments severely limit capture rates of 10 and 15-day-old flies. This
Selection 6 1.75⁎⁎⁎ 1.64⁎⁎⁎ 7.85⁎⁎⁎ 28.27⁎⁎⁎ 40.16⁎⁎⁎ observation fits well with the hypothesis that a stressful environment
⁎ Significance (p = 0.05). more strongly affects behaviour of older compared to younger flies.
⁎⁎⁎ Significance (p b 0.001). Our data supports findings in another model organism, Caenorhabditis
elegans. C. elegans mutant age-1 has been shown to live up to 80% lon-
ger than wild-type nematodes (Friedman and Johnson, 1988). They
2008). As stress resistance decreases with age (Grotewiel et al., 2005; have, like D. melanogaster, also been shown to have essentially simi-
Minois, 2001; Minois and Le Bourg, 1999; Minois et al., 2001) capture lar or superior phenotypes when looking at for example
rates were predicted to decrease faster in those environments. In addi- thermotolerance and activity levels (Duhon and Johnson, 1995;
tion to these general hypotheses the longevity selected lines where hy- Lithgow et al., 1995). However, while these age-1 mutants have
pothesized to be better at locating resources than the control lines of been shown to compete well with wild-type C. elegans at normal lab-
the same age, based on previous assessments of performance and behav- oratory conditions, the relative fitness of age-1 mutants was nega-
iour in the laboratory (Wit et al., 2013). tively affected when subjected to starvation cycles mimicking field
conditions (Walker et al., 2000).

4.1. Capture rates do not progressively decrease with age

The most striking pattern in capture rates was the initial increase in
performance with age, followed by a decrease (Fig. 2). This pattern is
unlike that predicted by previous studies on the effect of age on perfor-
mance in the laboratory where younger flies normally outperform older
flies (Simon et al., 2006; Sørensen and Loeschcke, 2002; Stratman and
Markow, 1998; Wit et al., 2013), though at temperatures reminiscent
of the lowest two temperatures used in these experiments, walking
speed in D. melanogaster has been shown to increase between 2 and
7 days-of-age (Gibert et al., 2001). The pattern is robust across the
three types of releases (high or low temperatures and longevity selected
versus control lines). These were all performed at different times, ren-
dering unlikely a role for adverse rearing conditions in the relatively
poor performance of all young flies in all releases. An explanation for
the increase in performance with age could be that after metamorphosis
adults continue to accumulate resources such as carbohydrates and
lipids (Arrese and Soulages, 2010) which results in increasing field fit-
ness with age initially. Some studies have indeed shown alterations in
lipid content with age, including the lines used here (Minois and Le
Bourg, 1999, Neda Nasiri Moghadam, personal communication). Also,
starvation resistance in females (Service et al., 1985) and activity levels
in females selected for postponed senescence (Service, 1987) have been
4.2. The effect of selection for increased longevity on capture rates
Ageing affected the capture rates in both longevity selected and con-
trol lines (Table 3). Comparing the ageing patterns, control lines show a
greater fluctuation with age than longevity selected lines (Supplemen-
tary Fig. 1). This finding is supported by the significant effect of selection
on capture rates of 10-day-old flies. No natural resources were available
in the habitats where the releases were performed, therefore flies
would starve and die if they did not find the bait. Thus, control lines
appear to be more fit in natural environments than longevity selected
lines at the ages tested in this study. This is in contrast to the
observations in the laboratory and suggests a cost associated with the
ability to live long, as predicted by the antagonistic pleiotropy theory
of ageing (Williams, 1957). This theory suggests that an increased life
span is accompanied by antagonistic responses early in life.
In the longevity selected lines of this study virtually no trade-offs be-
tween fecundity, stress resistance and functional senescence with life
span have been found when the flies are maintained and tested at
25 °C in the laboratory (Wit et al., 2013). Rather, the longevity selected
lines are more fecund throughout life and perform better at nearly all
resistance and behavioural traits assessed (Wit et al., 2013). The ageing
pattern in the current study is, however, suggestive of such a negative

Table 5
Relative capture success of females vs. males (columns C and LS), or of control vs. longevity selected groups (columns Female and Male) for the six releases with control and longevity
selected flies (Selections 1–6) at ages 10 and 15 days. R = release, C = control group, LS = longevity selected group. Significance is based on χ2 results of the GLM logit model.
Degrees of freedom = 1 for all analyses.

R 10 days 15 days

Sex Selection Sex Selection

C LS Female Male C LS Female Male

Selection 1 0.79⁎⁎⁎ 1.55⁎⁎⁎ 1.33⁎⁎⁎ 2.60⁎⁎⁎ 0.57⁎⁎⁎ 1.83⁎⁎⁎ 0.78⁎⁎⁎ 2.53⁎⁎⁎

Selection 2 0.96 1.64⁎⁎⁎ 1.95⁎⁎⁎ 3.34⁎⁎⁎ 0.59⁎⁎⁎ 1.31 1.92⁎⁎⁎ 4.25⁎⁎⁎
Selection 3 0.95 2.61⁎⁎⁎ 1.26⁎⁎⁎ 3.46⁎⁎⁎ 0.73⁎⁎ 1.50⁎ 0.85 1.76⁎⁎⁎
Selection 4 0.96 1.43⁎ 1.39⁎⁎ 2.07⁎⁎⁎ 0.60⁎⁎ 1.95⁎⁎⁎ 0.71⁎ 2.31⁎⁎⁎
Selection 5 0.85⁎⁎⁎ 1.39⁎⁎⁎ 1.22⁎⁎⁎ 2.01⁎⁎⁎ 0.72⁎⁎⁎ 1.34⁎⁎⁎ 1.04 1.94⁎⁎⁎
Selection 6 0.75⁎⁎⁎ 1.31⁎⁎⁎ 1.61⁎⁎⁎ 2.79⁎⁎⁎ 0.76⁎⁎⁎ 1.74⁎⁎⁎ 1.52⁎⁎⁎ 3.48⁎⁎⁎
⁎ Significance (p = 0.05).
⁎⁎ Significance (p = 0.01).
⁎⁎⁎ Significance (p b 0.001).
1194 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195
correlation in the form of decreased field performance relatively early in
life. Extrapolating the slope of the decrease in performance between 10
and 15 days-of-age to later time points indicates that longevity selected
females might outperform control females already at day 20 in 4 of the 6
releases. That could be suggestive of longevity selected flies performing
worse early in life, but maintaining their level of performance for longer,
eventually doing better than control lines. The assumption behind this
hypothesis however is that we actually measure an important fitness
component in nature in these releases. We argue that this is the case be-
cause flies that do not succeed in finding the buckets with banana will
die because no natural resources are available for D. melanogaster at
the field sites.
An explanation for the relatively poor capture rates of the longevity
selected lines could be related to the one trait that has been shown to
correlate negatively with increased life span in laboratory studies. A
negative correlation between life span and cold resistance was found
both in the lines used here (Wit et al., 2013) and in a multi-species com-
parison (Janneke Wit, personal observation). Ambient temperature dur-
ing releases of the selection lines was relatively low (Table 1). Therefore,
the low capture rates of the longevity selected lines might be the result
of their decreased ability to cope with cold stress. Our data seems to
contradict the findings in a previous study on behaviour in field cages
of lines selected for early and late reproduction (Hoffmann et al.,
2003). At colder temperatures than experienced in a laboratory envi-
ronment, these lines maintain their early and late timing of reproducing.
However, in the laboratory the lines selected for delayed reproduction
also have an increased life span (Partridge et al., 1999) while in the
field cages no difference in life span between early and late reproducing
lines was detected. This suggests that selection on reproductive timing,
a trait thought to be ecologically relevant and highly evolutionarily con-
served (Hoffmann et al., 2003; Mitrovski and Hoffmann, 2001), is robust
across environments. Life span and possibly correlated traits are, on the
other hand, affected by the more adverse environment encountered in
the field cages.
Another explanation for the relatively poor field performance of lon-
gevity selected lines does not involve a trade-off between increased life
span and performance. The selection lines were selected at 25 °C in the
laboratory. In this environment the longevity selected flies are equally
good or superior at virtually every trait tested to date in an environment
reminiscent of that during selection (Wit et al., 2013). These lines are in
other words very well adapted to their ‘natural’ habitat as evidenced by
their long life span and decreased rate of functional senescence in com-
parison to the control lines. These flies also live longer at 20 °C (Fig. 1),
the temperature at which the experimental generation was reared.
However, we cannot be sure that the longevity selected lines also live
longer under the circumstances encountered in the field. Therefore,
the possibility of the longevity selected lines being inferior to the con-
trol lines both with respect to survival and capture rates, cannot be
excluded.
5. Conclusions
The results of this study reinforce the notion that caution is needed
when extrapolating empirical findings in the lab to the field. In all re-
leases young–younger–youngest does not equal good–better–best, un-
like the prediction of progressive senescent decline. More studies are
needed to elucidate the mechanisms underlying this observation. The
convincing superiority of longevity selected lines in the laboratory did
not translate into increased field fitness. These observations support
the antagonistic pleiotropy theory of ageing. Increased longevity in a
lab setting comes at the cost of field fitness. However, as life span of
the studied lines was assessed in a laboratory environment, it cannot
be ruled out that longevity selected lines overall perform poorly in the
field environment. This aspect should be accounted for in future studies.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.exger.2013.07.012.
Conflict of interest
The authors have no conflicts of interests.
Acknowledgements
The authors are grateful to Mads Fristrup Schou, Vanessa Kellermann,
Doth Andersen, Neda Nasiri Moghadam, Golshah Ayoubi, Marie
Rosenstand Hansen and Anne Marie Vestergaard Henten for their help
with the execution of the release experiments. We also want to thank
Kuke (R.) Bijlsma and Alexei A. Maklakov for commenting on an earlier
version of this manuscript and Kjeld and Bente Nygaard Kristensen for
allowing us to perform the release–capture experiments at Kistrupgaard.
For financial support we thank the Lundbeck Foundation (R31-A2517)
and The Danish Council for Independent Research | Natural Sciences
(PS: 10-093806/IPD/MAJ; VL: 437369; TNK: 09-064936).
References
Arking, R., Wells, R.A., 1990. Genetic alteration of normal aging processes is responsible
for extended longevity in Drosophila. Dev. Genet. 11, 141–148.
Arking, R., Buck, S., Novoseltev, V.N., Hwangbo, D.S., Lane, M., 2002. Genomic plasticity,
energy allocations, and the extended longevity phenotypes of Drosophila. Ageing
Res. Rev. 1, 209–228.
Arrese, E.L., Soulages, J.L., 2010. Insect fat body: energy, metabolism, and regulation. Annu.
Rev. Entomol. 55, 207–225.
Bubliy, O.A., Loeschcke, V., 2005. Correlated responses to selection for stress resistance
and longevity in a laboratory population of Drosophila melanogaster. J. Evol. Biol. 18,
789–803.
Buck, S., Vettraino, J., Force, A.G., Arking, R., 2000. Extended longevity in Drosophila is consis-
tently associated with a decrease in developmental viability. J. Gerontol. 55, B292–B301.
Chapman, T., 2001. Seminal fluid-mediated fitness traits in Drosophila. Heredity 87, 511–521.
Chippindale, A.K., Hoang, D.T., Service, P.M., Rose, M.R., 1994. The evolution of develop-
ment in Drosophila melanogaster selected for postponed senescence. Evolution 48,
1880–1899.
Chippindale, A.K., Leroi, A.M., Saing, H., Borash, D.J., Rose, M.R., 1997. Phenotypic plasticity
and selection in Drosophila life history evolution. II. Diet, mates and the cost of repro-
duction. J. Evol. Biol. 10, 269–293.
Cordts, R., Partridge, L., 1996. Courtship reduces longevity of male Drosophila melanogaster.
Anim. Behav. 52, 269–278.
Duhon, S.A., Johnson, T.E., 1995. Movement as an index of vitality: comparing wild type
and the age-1 mutant of Caenorhabditis elegans. J. Gerontol. 50, B254–B261.
Folk, D.G., Han, C., Bradley, T.J., 2001. Water acquisition and partitioning in Drosophila
melanogaster: effects of selection for desiccation-resistance. J. Exp. Biol. 204, 3323–3331.
Force, A.G., Staples, T., Soliman, S., Arking, R., 1995. Comparative biochemical and stress
analysis of genetically selected Drosophila strains with different longevities. Dev.
Genet. 17, 340–351.
Fowler, K., Partridge, L., 1989. A cost of mating in female fruit flies. Nature 338, 760–761.
Friedman, D.B., Johnson, T.E., 1988. Three mutants that extend both mean and maximum life
span of the nematode, Caenorhabditis elegans, define the age-1 gene. J. Gerontol. 43,
B102–B109.
Gibert, P., Huey, R.B., Gilchrist, G.W., 2001. Locomotor performance of Drosophila
melanogaster: interactions among developmental and adult temperatures, age, and
geography. Evolution 55, 205–209.
Graves, J.L., Luckinbill, L.S., Nichols, A., 1988. Flight duration and wing beat frequency in
long-and short-lived Drosophila melanogaster. J. Insect Physiol. 34, 1021–1026.
Grotewiel, M.S., Martin, I., Bhandari, P., Cook-Wiens, E., 2005. Functional senescence in
Drosophila melanogaster. Ageing Res. Rev. 4, 372–397.
Harshman, L.G., Hoffmann, A.A., 2000. Laboratory selection experiments using Drosophila:
what do they really tell us? Trends Ecol. Evol. 15, 32–36.
Harshman, L.G., Hoffmann, A.A., Clark, A.G., 1999. Selection for starvation resistance in
Drosophila melanogaster: physiological correlates, enzyme activities and multiple
stress responses. J. Evol. Biol. 12, 370–379.
Hoffmann, A.A., Parsons, P.A., 1993a. Direct and correlated responses to selection for des-
iccation resistance: a comparison of Drosophila melanogaster and D. simulans. J. Evol.
Biol. 6, 643–657.
Hoffmann, A.A., Parsons, P.A., 1993b. Selection for adult desiccation resistance in Drosophila
melanogaster: fitness components, larval resistance and stress correlations. Biol. J. Linn.
Soc. 48, 43–54.
Hoffmann, A.A., Scot, M., Partridge, L., Hallas, R., 2003. Overwintering in Drosophila
melanogaster: outdoor field cage experiments on clinal and laboratory selected pop-
ulations help to elucidate traits under selection. J. Evol. Biol. 16, 614–623.
Houle, D., 1991. Genetic covariance of fitness correlates: what genetic correlations are
made of and why it matters. Evolution 45, 630–648.
Kellermann, V.M., Hoffmann, A.A., Sgrò, C.M., 2007. Hsp90 inhibition and the expression
of phenotypic variability in the rainforest species Drosophila birchii. Biol. J. Linn. Soc.
92, 457–465.
Kristensen, T.N., Loeschcke, V., Hoffmann, A.A., 2007. Can artificially selected phenotypes
influence a component of field fitness? Thermal selection and fly performance under
thermal extremes. Proc. R. Soc. B Biol. Sci. 274, 771–778.
 J. Wit et al. / Experimental Gerontology 48 (2013) 1189–1195
Kristensen, T.N., Hoffmann, A.A., Overgaard, J., Sørensen, J.G., Hallas, R., Loeschcke, V.,
2008a. Costs and benefits of cold acclimation in field-released Drosophila. Proc. Natl.
Acad. Sci. U.S.A. 105, 216–221.
Kristensen, T.N., Loeschcke, V., Hoffmann, A.A., 2008b. Linking inbreeding effects in captive
populations with fitness in the wild: release of replicated Drosophila melanogaster lines
under different temperatures. Conserv. Biol. 22, 189–199.
Kristensen, T.N., Barker, J.S.F., Pedersen, K.S., Loeschcke, V., 2008c. Extreme temperatures
increase the deleterious consequences of inbreeding under laboratory and semi-
natural conditions. Proc. R. Soc. B Biol. Sci. 275, 2055–2061.
Leroi, A.M., Chen, W.R., Rose, M.R., 1994a. Long-term laboratory evolution of a genetic life-
history trade-off in Drosophila melanogaster. 2. Stability of genetic correlations. Evolu-
tion 48, 1258–1268.
Leroi, A.M., Chippindale, A.K., Rose, M.R., 1994b. Long-term laboratory evolution of a ge-
netic life-history trade-off in Drosophila melanogaster. 1. The role of genotype-by-
environment interaction. Evolution 48, 1244–1257.
Lithgow, G.J., White, T.M., Melov, S., Johnson, T.E., 1995. Thermotolerance and extended
life-span conferred by single-gene mutations and induced by thermal stress. Proc.
Natl. Acad. Sci. U.S.A. 92, 7540–7544.
Luckinbill, L.S., Arking, R., Clare, M.J., Cirocco, W.C., Buck, S.A., 1984. Selection for delayed
senescence in Drosophila melanogaster. Evolution 38, 996–1003.
Magwere, T., Pamplona, R., Miwa, S., Martinez-Diaz, P., Portero-Otin, M., Brand, M.D.,
Partridge, L., 2006. Flight activity, mortality rates, and lipoxidative damage in Drosophila.
J. Gerontol. 61, B136–B145.
Minois, N., 2001. Resistance to stress as a function of age in transgenic Drosophila
melanogaster overexpressing Hsp70. J. Insect Physiol. 47, 1007–1012.
Minois, N., Le Bourg, É., 1999. Resistance to stress as a function of age in Drosophila
melanogaster living in hypergravity. Mech. Ageing Dev. 109, 53–64.
Minois, N., Khazaeli, A.A., Curtsinger, J.W., 2001. Locomotor activity as a function of age
and life span in Drosophila melanogaster overexpressing Hsp70. Exp. Gerontol. 36,
1137–1153.
Mitrovski, P., Hoffmann, A.A., 2001. Postponed reproduction as an adaptation to winter
conditions in Drosophila melanogaster: evidence for clinal variation under semi-
natural conditions. Proc. R. Soc. Lond. B Biol. Sci. 268, 2163–2168.
Partridge, L., Barton, N.H., 1993. Optimality, mutation and the evolution of ageing. Nature
362, 305–311.
Partridge, L., Prowse, N., Pignatelli, P., 1999. Another set of responses and correlated re-
sponses to selection on age at reproduction in Drosophila melanogaster. Proc. R. Soc.
Lond. B Biol. Sci. 266, 255–261.
Piazza, N., Gosangi, B., Devilla, S., Arking, R., Wessells, R., 2009. Exercise-training in young
Drosophila melanogaster reduces age-related decline in mobility and cardiac perfor-
mance. PLoS One 4, e5886.
Prowse, N., Partridge, L., 1997. The effects of reproduction on longevity and fertility in
male Drosophila melanogaster. J. Insect Physiol. 43, 501–512.
Rose, M.R., 1984. Laboratory evolution of postponed senescence in Drosophila melanogaster.
Evolution 3, 1004–1010.
Rosewell, J., Shorrocks, B., 1987. The implication of survival rates in natural populations of
Drosophila: capture recapture experiments on domestic species. Biol. J. Linn. Soc. 32,
373–384.
1195
Sarup, P.M., Sørensen, P., Loeschcke, V., 2011. Flies selected for longevity retain a young
gene expression profile. Age 33, 69–80.
Service, P.M., 1987. Physiological mechanisms of increased stress resistance in Drosophila
melanogaster selected for postponed senescence. Physiol. Zool. 60, 321–326.
Service, P.M., 1989. The effect of mating status on lifespan, egg laying, and starvation re-
sistance in Drosophila melanogaster in relation to selection on longevity. J. Insect
Physiol. 35, 447–452.
Service, P.M., Hutchinson, E.W., MacKinley, M.D., Rose, M.R., 1985. Resistance to environ-
mental stress in Drosophila melanogaster selected for postponed senescence. Physiol.
Zool. 58, 380–389.
Simon, A.F., Liang, D.T., Krantz, D.E., 2006. Differential decline in behavioral performance
of Drosophila melanogaster with age. Mech. Ageing Dev. 127, 647–651.
Sørensen, J.G., Loeschcke, V., 2002. Decreased heat-shock resistance and down-regulation
of Hsp70 expression with increasing age in adult Drosophila melanogaster. Funct. Ecol.
16, 379–384.
Sørensen, J.G., Dahlgaard, J., Loeschcke, V., 2001. Genetic variation in thermal tolerance
among natural populations of Drosophila buzzatii: down regulation of Hsp70 expres-
sion and variation in heat stress resistance traits. Funct. Ecol. 15, 289–296.
Sørensen, J.G., Loeschcke, V., Kristensen, T.N., 2008. Lessons from the use of genetically
modified Drosophila melanogaster in ecological studies: Hsf mutant lines show highly
trait-specific performance in field and laboratory thermal assays. Funct. Ecol. 23,
240–247.
Stearns, S.C., 1989. Trade-offs in life-history evolution. Funct. Ecol. 3, 259–268.
Stearns, S.C., 1992. The Evolution of Life Histories. Oxford University Press, Oxford.
Stratman, R., Markow, T.A., 1998. Resistance to thermal stress in desert Drosophila. Funct.
Ecol. 12, 965–970.
Turelli, M., Hoffmann, A.A., 1995. Cytoplasmic incompatibility in Drosophila simulans:
dynamics and parameter estimates from natural populations. Genetics 140,
1319–1338.
Vermeulen, C.J., Bijlsma, R., 2006. Changes in genetic architecture during relaxation
in Drosophila melanogaster selected on divergent virgin life span. J. Evol. Biol. 19,
216–227.
Vermeulen, C.J., Van De Zande, L., Bijlsma, R., 2005. Resistance to oxidative stress induced
by paraquat correlates well with both decreased and increased lifespan in Drosophila
melanogaster. Biogerontology 6, 387–395.
Vermeulen, C.J., Van De Zande, L., Bijlsma, R., 2006. Developmental and age-specific ef-
fects of selection on divergent virgin life span on fat content and starvation resistance
in Drosophila melanogaster. J. Insect Physiol. 52, 910–919.
Walker, D.W., McColl, G., Jenkins, N.L., Harris, J., Lithgow, G.J., 2000. Evolution of life span
in C. elegans. Nature 405, 296–297.
Williams, G.C., 1957. Pleiotropy, natural selection, and the evolution of senescence. Evolution
11, 398–411.
Wit, J., Sarup, P.M., Lupsa, N., Malte, H., Frydenberg, J., Loeschcke, V., 2013. Longevity for
free? Increased reproduction with limited trade-offs in Drosophila melanogaster se-
lected for increased life span. Exp. Gerontol. 48, 349–357.
Zwaan, B.J., 1999. The evolutionary genetics of ageing and longevity. Heredity 82, 589–597.
Zwaan, B.J., Bijlsma, R., Hoekstra, R.F., 1995. Direct selection on life span in Drosophila
melanogaster. Evolution 49, 649–659.