Life‐history traits and physiological limits of the alpine fly Drosophila nigrosparsa (Diptera: Drosophilidae): A comparative study

Abstract Interspecific variation in life‐history traits and physiological limits can be linked to the environmental conditions species experience, including climatic conditions. As alpine environments are particularly vulnerable under climate change, we focus on the montane‐alpine fly Drosophila nigrosparsa. Here, we characterized some of its life‐history traits and physiological limits and compared these with those of other drosophilids, namely Drosophila hydei, Drosophila melanogaster, and Drosophila obscura. We assayed oviposition rate, longevity, productivity, development time, larval competitiveness, starvation resistance, and heat and cold tolerance. Compared with the other species assayed, D. nigrosparsa is less fecund, relatively long‐living, starvation susceptible, cold adapted, and surprisingly well heat adapted. These life‐history characteristics provide insights into invertebrate adaptations to alpine conditions which may evolve under ongoing climate change.

At higher altitudes, species are adapted to the harsh environmental conditions including short growing seasons, resource shortage, and extreme low and high temperatures (Franz, 1979;Hodkinson, 2005). Due to low average temperatures, oviposition and growth rate are low, the latter also resulting in long development times (Mani, 1962;Schnebel & Grossfield, 1986). Alpine species are expected to be relatively cold and heat tolerant due to the temperature extremes they encounter both during the day and throughout the year (Gaston & Chown, 1999). To our knowledge, the influence of environmental factors on life-history strategies of alpine insects has rarely been investigated.
The genus Drosophila contains some of the best studied animal species, whereby Drosophila melanogaster is in major focus of research (Hoffmann, Sørensen, & Loeschcke, 2003). So far, climate change research on Drosophilidae has rarely focused on thermal evolution of alpine species, although mountain ecosystems are predicted to be particularly vulnerable to changing temperatures (IPCC, 2007). For this reason, we are currently establishing Drosophila nigrosparsa ( Figure 1) as an alpine study system (Austrian Science Fund, project number P 23949). Drosophila nigrosparsa is native to montane-alpine regions, occurring around the timber line at about 2,000 m above sea level (a.s.l.) in Central Europe (Bächli & Burla, 1985). It uses fungal fruiting bodies as natural oviposition substrate (Kinzner et al., 2016); a specific food source of adults is not known. The exact position of this species in the Drosophila phylogeny is not clear, but it seems that D. nigrosparsa is related to the Hawaiian species (Cicconardi, Marcatili, Arthofer, Schlick-Steiner, & Steiner, 2017). In contrast to Drosophila alpina (Bächli & Burla, 1985), D. nigrosparsa is culturable in the laboratory without an obligatory diapause, making it an ideal alpine study organism for laboratory experiments.
The main aim of this study was to test whether selected lifehistory traits and physiological limits of the novel alpine model organism Drosophila nigrosparsa reflect the alpine environment. Thus, we characterized multiple traits of D. nigrosparsa under laboratory conditions, namely oviposition rate, productivity, development time, longevity, larval competition performance, starvation resistance, and heat and cold resistance. We focus on thermal limits because of increasing temperatures due to climate change. It was not our aim to examine the adaptability to climate change but to provide a baseline for future investigations. To benchmark D. nigrosparsa's performance, additional Drosophila species with different ecological niches were tested in a subset of assays. Our goal was to directly compare our focal species with species both well characterized in former studies and common as study organisms in ecological and evolutionary experiments. Drosophila obscura, a species with wide temperate distribution, served as a link between the specialist, alpine-montane D. nigrosparsa, the cosmopolitan D. melanogaster, and the tropical Drosophila hydei (Gibert, Moreteau, Pétavy, Karan, & David, 2001).

| Origin and maintenance of flies
Fly strains are named according to the first letter of the species name, and the year when the assays were carried out. All species were identified following Bächli and Burla (1985) and were kept at a 16L:8D photoperiod and ca. 70% relative humidity; rearing temperatures varied as described below. No specific permissions were required because none of the species used in this study is endangered or protected.
In 2007, Drosophila melanogaster (M 2015  In a large-scale field survey in 2012, Drosophila nigrosparsa (N 2013 ) was captured again at Kaserstattalm (K) but also at Pfitscherjoch F I G U R E 1 The alpine fly Drosophila nigrosparsa (P;South Tyrol,Italy,46.98°N,11.68°E,2,000 m a.s.l.). From each of the populations, 100 males and 100 females were used (N 2013 Table S1) to provide laboratory temperature conditions as near as possible to natural conditions. The 5th generation of N 2013 K and N 2013 P was randomly split into eight replicate lines each.
To adapt flies to laboratory conditions and to check for parasites and disease (Ashburner & Roote, 2007) To allow enough space for mating, batches of up to 100 emerged adults were transferred to perforated transparent 300-ml plastic cups on 90-mm Petri dishes ( Figure 2) containing 40 ml grape-juice agar (24% grape juice, 2.5% sucrose, 2% agar, and 0.2% methyl-4-hydroxybenzoate in deionized water; modified from Sullivan, Ashburner, & Hawley, 2000). Approximately 0.10 g of dried yeast and 0.30 g of medium (malt for D. nigrosparsa, cornmeal for the other species) were added.

| Oviposition rate and longevity
For N 2013 , 50 females and 50 males (4-8 days old) were transferred to culturing cups with grape-juice agar. New agar plates were provided twice a week after 72 or 96 hr, when also the oviposition rate, the number and sex of dead flies, and the dead flies' age were determined.

| Productivity and development time
For N 2013 , five females and five males per line of the same age were allowed to oviposit for 48 hr on malt medium with additional 0.10 g dry yeast as a protein source. The number, sex, and development time of emerged flies were recorded once a week, and the mean number F I G U R E 2 Schematic illustration of a culturing cup. Petri dish with grape-juice agar medium, dried yeast, and malt medium to enhance oviposition, topped with a perforated plastic cup. Yeast and malt media were covered with a plastic roof that prevented flies from falling into the media from above but allowed access to the media from the side of emerged flies per parental female was calculated. The assay was replicated 11 times for each line.

| Larval competition
First-instar larvae of D. nigrosparsa (N 2013 ) and D. subobscura (ratio 1:2) were transferred into 36-ml vials containing 5 ml plain agar (2% agar and 0.2% methyl-4-hydroxybenzoate in deionized water) at the bottom to provide moisture covered with 3 ml malt medium. Drosophila subobscura was used as a competition standard for D. nigrosparsa due to its co-occurrence at the timber line sharing mushrooms as oviposition substrate (Bächli, 1977;Kinzner et al., 2016). Three different densities of D. nigrosparsa : D. subobscura (12:24, 24:48, 48:96) were replicated five times per line. Larva-to-adult viability, sex, and development time of emerged flies were recorded three times a week, and adults were removed to avoid additional oviposition.

| Acute critical maximum and minimum temperature
Procedures were similar to those of Overgaard, Kristensen, Mitchell, & Hoffmann (2011). Briefly, for acute critical maximum temperature, For acute critical minimum temperature, the same protocol as described above was applied; flies were at most 11 (N 2013 ) or 13 days (H 2015 , M 2015

| Statistical analyses
The experiments longevity, productivity, development time, larval competition, starvation resistance, and heat knockdown were analyzed using analysis of variance (ANOVA). The experiments oviposition rate over lifespan and acute critical maximum and minimum temperature were analyzed using analysis of covariance (ANCOVA).
The acute critical maximum and minimum temperatures at which 50% of the flies were knocked down (CT max and CT min , respectively) were calculated using linear regressions. All analyses were performed in R version 3.1.1 (R Core Team, 2014).
To gain information on the reproducibility of the oviposition rate, the concordance correlation coefficient (CCC) was calculated describing the egg counter's measurement error (Castañeda, Calabria, Betancourt, Rezende, & Santos, 2012). This coefficient is meaningful, easily measured, and similar to intra-class correlation coefficient, which evaluates a researcher's accurateness and precision. Due to overly high oviposition rates of H 2015 , surpassing 2,000 eggs per replicate per 24 hr, eggs on corn media were analyzed via manual image analysis using Fiji ImageJ 2.0. (Schindelin et al., 2012). Also, for accurate visual counting, a counting grid was placed on the H 2015 and M 2015 media after oviposition.

| Oviposition rate
Counts were accurate with a mean error below 0.4% (CCC analysis, Table S2). Mean oviposition rate over lifespan (averaged per week) was lowest for N 2013 K and N 2013 P ( Figure 3a, Data S1), without a significant difference between the two populations (Table 1). Mean oviposition rate was highest for H 2015 with more than seven times higher rates than for the N 2013 populations. Mean oviposition rate differed significantly among species at the constant temperature regime (Table 1). Oviposition rate increased steeply at an early age for all spe- Week 2 before decreasing gradually.  Table 1).

| Productivity and development time
For N 2013 K, 1.63 ± 0.13 females and 1.13 ± 0.10 males emerged per parental female (Figure 5a, Data S1). For N 2013 P, 1.79 ± 0.15 females and 1.34 ± 0.88 males emerged per parental female. Productivity did not differ significantly between the two populations (Table 1).
Sex ratio of emerged adults (males : females) was 0.68 (N 2013 K) and 0.77 (N 2013 P), that is, significantly more females than males emerged for both populations and within the populations (Table 1) (Table 1), with females taking 60.14 ± 0.42 and males 60.18 ± 0.47 days.

| Larval competition
At density 12:24 N 2013 : D. subobscura, larva-to-adult viability for N 2013 females was about 8% and for males between 8% and 9% (Figure 6a, T A B L E 1 Results of the analysis of variance for life-history and physiological-limit assays

| Starvation resistance
Mean starvation resistance was highest for H 2015 (125.95 ± 3.82 hr, Figure 7, Data S1) and lowest for O 2015 (83.00 ± 3.14 hr). Mean starvation resistance differed significantly among the species of the constant regime but not between the populations N 2013 K and N 2013 P nor between the sexes, except for O 2015 where females lived significantly longer than males (Table 1). N 2013 reached the most extreme starvation resistance values with maxima of more than 320 hr and minima of about 30 hr (Figure 7).  (Table 1).

| Acute critical maximum temperature
Linear regression revealed a significant relation between increasing temperature and acute heat knockdown for all species and populations (Table 2). At the lowest temperature (36.50°C for the variable regime and 37.00°C for the constant regime), no fly was in coma (Table 2). Fifty percent knock-down temperature (CT max ) was at 37.55°C (N 2013 (Table 1).

| Acute critical minimum temperature
Linear regression revealed a significant relation between decreasing temperature and acute cold knockdown for all species and populations (Table 2). At the highest temperature (9.00 C), no fly was in coma.  (Table 1).

| DISCUSSION
Life-history traits are among the most basic and essential characteristics describing a species (Pease & Bull, 1988) and contribute to a broader understanding of an organism. Under climate change, it is necessary to understand a species' ecology, its possible responses to resulting selection pressures, and its potential for rapid adaptation to environmental alterations (Angilletta & Dunham, 2003;Hoffmann & Sgrò, 2011). Life-history traits include those involved in responding to extreme conditions, such as heat and cold tolerance and starvation resistance, as well as patterns of reproduction and longevity that can all provide insights into evolutionary Analysis of covariance (ANCOVA) was performed where noted in parentheses. Each line represents a single analysis. For oviposition rate, ramping heat knockdown, and acute critical maximum and minimum temperature, only female flies were used, and thus, no sex-dependent analysis was conducted. egg-laying (Ashburner, Golic, & Hawley, 2005). In our study, the constant temperature regime of 19°C (N 2015 ) enhanced oviposition compared with a more natural, fluctuating regime (N 2013 ). Nevertheless, in both tests, the number of eggs laid by D. nigrosparsa was relatively low compared with the other assayed species, except D. obscura. The low oviposition rate might be explained by cold adaptation (Carbonell et al., 2017;Kubrak, Nylin, Flatt, Nässel, & Leimar, 2017;Schnebel & Grossfield, 1986). Drosophila nigrosparsa has a reproductive diapause depending on photopheriodism similar to D. littoralis (Lankinen, 1985). Longevity can contribute to the reproductive success of an organism, although senescence is also apparent in Drosophila and reduces overall productivity (Figure 3) (Nuzhdin, Pasyukova, Dilda, Zeng, & Mackay, 1997). Generally, males lived longer than females in our study, which might reflect female energy investment in egg production and mating behavior (Chapman, Liddle, Kalb, Wolfner, & Partridge, 1995;Fowler & Partridge, 1989). Similarly, D. buzzatii males lived longer in another study (Scannapieco, Sambucetti, & Norry, 2009). However, this is not a universal pattern given that M 2015 females outlived the males. Differences in experimental conditions can shorten or extend the lifespan of an organism (Aigaki & Ohba, 1984;Helfand & Rogina, 2003), which no doubt accounts for the difference between the N 2013 F I G U R E 6 Larval competition between Drosophila nigrosparsa and Drosophila subobscura.  (Figure 3). This difference probably reflects the influence of the temperature regime: A constant but warm temperature results in a continuous and relatively high metabolic rate, whereas this is not the case with a fluctuating temperature regime.
The latter is more appropriate for alpine species such as D. nigrosparsa, as they experience strong daily temperature fluctuations in nature. For D. montana, another cold-adapted species, males lived over 130 days on average (Hoikkala, Saarikettu, Kotiaho, & Liimatainen, 2008), but this assay was at 4°C and therefore not comparable with our study.
In any case, species differences were discovered, but no species came close to the known maximum lifespan of some Hawaiian drosophilids which can exceed 9 months (Carson, Hardy, Spieth, & Stone, 1970).
Besides the low oviposition rate, D. nigrosparsa (N 2013 ) also had a low productivity ( Figure 5) when compared with D. melanogaster (Ochando & Ayala, 1999) and D. pseudoobscura (Gowaty, Kim, Rawlings, & Anderson, 2010). The specific numbers of eclosed flies strongly depend on the study design (Barker, 1973). Furthermore, rearing temperature and mating frequency influence productivity (Barker, 1973;Gowaty et al., 2010). Reproductive output might be negatively correlated with cold resistance (Jenkins & Hoffmann, 1999) and, as D. nigrosparsa is cold adapted (Table 1), this may contribute to the low productivity of the N 2013 populations. Finally, it is imaginable that the culture medium we used is suboptimal for D. nigrosparsa, possibly leading to its low productivity in the laboratory. However, we have tested a range of media types for D. nigrosparsa, and none of these lead to a high productivity.
Short development time is expected to enhance larval survival and reproductive success (Roff, 2000). Moreover, development is expected to be fast in alpine environments due to a short seasonal growth period (Fischer & Fiedler, 2002). In grasshoppers, there is no difference of development time between low-elevation and high-elevation populations (Carron, 1996). However, D. nigrosparsa displayed long development times compared with other Drosophila species. Egg-to-adult development time was, on average, 61 days ( Figure 6), while it takes D. montana, another cold-adapted species, about 27 days (Salminen, Vesala, & Hoikkala, 2012) and D. melanogaster only 8.5 days (Ashburner et al., 2005) to develop under optimum conditions. At 24°C, D. hydei develops within 14 days and D. virilis within 18 days (Ashburner et al., 2005). Development time is influenced by environmental conditions such as photoperiod (Salminen et al., 2012) and temperature as well as genetic variation (Norry, Bubliy, & Loeschcke, 2001;Zwaan, Bijlsma, & Hoekstra, 1992). Drosophila birchii and D. serrata have a longer development time with increasing latitude (Griffiths, Schiffer, & Hoffmann, 2005;Sgrò, Blows, & Noor, 2003), and cold adapted strains of D. subobscura have a longer development time than warm-adapted strains (Santos, Brites, & Laayouni, 2006). If cold adaptation is generally linked to longer development times, this might explain the increased development time of D. nigrosparsa in relation to other species.
In natural populations, competition might play a role in the abundance of Drosophila species (Grimaldi & Jaenike, 1984). Larval competition influences growth rate, development time, and pre-adult survival (James & Partridge, 1998), and strong larval competitiveness might have positive effects on these traits. Here, larval competitiveness of D. nigrosparsa (N 2013 ) was assayed using the sympatric and resourcesharing fly D. subobscura (Kinzner et al., 2016). The percentage of eclosed N 2013 adults decreased by more than half from the first to the second density ratio, but did not decrease further ( Figure 5, Table 1).
The media used in our assays may have influenced the hatching rate, but as both species use fungal fruiting bodies as rearing substrate in nature (Kinzner et al., 2016), the influence of the artificial media used here may be similar for the two species. In our experiment, we cannot distinguish between the influence of intra-and interspecific competition. Thus, the decreasing emergence success might not only reflect the interspecific competitiveness but also the combined effect of intra-and interspecific competition. Shorrocks, Rosewell, Edwards, & Atkinson (1984) assumed that, although there may be competition among different drosophilids in nature, they will probably not oust one another even on a shared resource. It seems that D. nigrosparsa larvae are not strong competitors because the percentage of emerged adults at the lowest density was lower than without interspecific competition but with higher intraspecific larval density (80 eggs, ca. 20% hatching success, data not shown). Moreover, D. nigrosparsa (N 2013 ) does not seem to be competitive in response to an early occupation of resources given its extended development time. Concerning the species' competitive ability, further research is needed not only on larvae but also on adults.
Ephemeral and fragmented food resources, such as mushrooms, are limited patches (Krijger, Peters, & Sevenster, 2001). Thus, and due to shortage or suboptimal quality of food, starvation might often be experienced in Drosophila species, and limited access to food might play an important role in high altitudes (Goenaga, Fanara, & Hasson, 2013). However, in concordance with former studies assuming high tolerance for tropical species (van Herrewege & David, 1997;Karan et al., 1998;Parsons, 1983;Sisodia & Singh, 2010) Both species had much lower starvation resistance compared with our results. Reasons could be the maintenance of the flies (24°C with 35% relative humidity versus 19°C with 70% relative humidity in our study), the experimental setup, but also the origin of flies.
Temperature affects life-history traits, and knowledge of basic temperature performance is essential for understanding a species' biology (Angilletta, 2009;Chown & Terblanche, 2006). Ramping heat knockdown evaluates the impact of increasing temperature on flies, not measuring the lethal temperature but an ecologically more realistic limit, the heat coma (Hoffmann, Dagher, Hercus, & Berrigan, 1997).
Generally, heat knockdown performance also depends on the experimental design (Santos, Castañeda, & Rezende, 2011;Terblanche, Deere, Clusella-Trullas, Janion, & Chown, 2007). In other studies, the mean heat knockdown temperature for D. melanogaster was higher than in our assay (Kellermann et al., 2012;Overgaard, Kristensen, Mitchell, & Hoffmann, 2011;Sgrò et al., 2010), possibly because of our rearing regime. However, the heat resistance of D. hydei seemed unaffected by temperature regime (Santos et al., 2011), but see Kellermann et al. (2012). Interestingly, D. obscura reached a considerably higher value than observed for this species by Kellermann et al. (2012). The results of this study are more similar to those of a preliminary study in our laboratory (Eberl, 2016). A reason for these divergent results could be the different origins of flies. The strain used in the current study originated from a population at the alpine timber line (same location as N 2013 and N 2015 ), whereas the other populations were from lower altitudes in Denmark (Kellermann et al., 2012) and Germany (Eberl, 2016). Sisodia and Singh (Sisodia & Singh, 2010) proposed that intraspecific variation strongly depends on the populations' environment.
For example, D. melanogaster populations from a subtropical habitat were less temperature-stress resistant than populations from temperate regions (David, Allemand, van Herrewege, & Cohet, 1983;Sisodia & Singh, 2010;Stanley & Parsons, 1981). Beppu, Yoshida, & Kimura (1996)  37.44 ± 0.05°C N 2013 P, and 38.11 ± 0.07°C N 2015 ) were surprisingly high for an alpine-montane species relative to tropical or widely spread temperate species in our study as well as in others (e.g., Overgaard, Kristensen, Mitchell, & Hoffmann, 2011). Kellermann et al. (2012) showed that the upper thermal limits of drosophilids are linked to the phylogeny rather than to environmental factors at the species' main distribution range. However, D. nigrosparsa was not part of their study, and we have only limited information about this species' exact position in the phylogeny of Drosophilidae (Cicconardi et al., 2017), as molecular information on many drosophilids is lacking. The variation among individuals was high for N 2013 K and N 2013 P relative to the four other treatments tested (Figure 7). This might be explained by higher genetic variation in the former, which were founded by 100 females and males per population, whereas all others were founded by single pairs of flies. However, we have no genetic data to confirm this assumption.
After all, the relatively high heat resistance of D. nigrosparsa could be adaptive-it may allow the species to stay in the sun and microhabitats that can heat up beyond 35°C (Franz, 1979), which could be relevant because of, for example, biotic interactions or competition (Gilman, Urban, Tewksbury, Gilchrist, & Holt, 2010).
Physiological adaption to extreme temperatures is an important limiting factor to an ectotherm's distribution (Andersen et al., 2015;Overgaard, Kearney, & Hoffmann, 2014), especially affecting individuals living in extreme habitats like the Alps. The acute critical maximum temperature, where 50% of females fell in coma, was lower for D. nigrosparsa (N 2013 and N 2015 ) than for habitat generalist Drosophila species such as D. pseudoananassae and D. simulans (Overgaard, Kristensen, Mitchell, & Hoffmann, 2011). Concerning the generalist species, both D. hydei and D. melanogaster were more sensitive to heat in this than in another study (Overgaard, Kristensen, Mitchell, & Hoffmann, 2011). Our results of the acute critical maximum temperatures reflect the ramping heat knockdown results (Figure 8). The acute critical minimum temperature for D. nigrosparsa was similarly low at the two culturing temperatures (N 2013 , N 2015 ; Tables 1 and 2).
Drosophila hydei and D. melanogaster had considerably weaker cold resistance. Overgaard, Kristensen, Mitchell, & Hoffmann, (2011) observed lower values for both species. Other generalist drosophilids, such as D. pseudoananassae, were more vulnerable to cold temperatures, and D. simulans was close to D. melanogaster (Overgaard, Kristensen, Mitchell, & Hoffmann, 2011 T A B L E 2 Linear regression models for acute critical maximum and minimum temperature experiments increasing altitude and rapid variations in temperature and distinctive seasonality (Barry, 1992), might enforce better cold resistance, so that alpine species can withstand colder temperatures than tropical or lowland temperate species. Moreover, development under variable thermal conditions might widen the thermal limits .
Drosophila nigrosparsa seems well adapted to the harsh conditions at high altitudes as the range between upper and lower thermal limit is larger than that of any other species tested here (N 2013 2.70-37.55°C, N 2015 2.83-37.68°C), enabling it to greatly withstand variable temperatures in higher or northern regions (Calosi, Bilton, Spicer, Votier, & Atfield, 2010;Goto & Kimura, 1998). The tropical and cosmopolitan species were less robust.
The challenging alpine climate has most likely selected for a wide range of temperature tolerance (Gaston & Chown, 1999;Goto, Yoshida, Beppu, & Kimura, 1999). Seasonality also plays a major role in evolution, and cold tolerance is key to selection in temperate or arctic (Goto et al., 1999) and most likely also alpine regions. In conclusion, the alpine D. nigrosparsa can withstand relatively high temperatures compared with drosophilids from warmer origins like D. melanogaster and D. hydei. Climate change research predicts warmer winters and warmer conditions in general (IPCC, 2007). Nevertheless, what extent of temperature rise species can withstand does not necessarily lead to any conclusion about long-term effects or impacts on productivity or food resources. Moreover, biological interactions strongly influence species' distributions, which are also changed by climatic alteration (Davis, Jenkinson, Lawton, Shorrocks, & Wood, 1998). Kellermann et al. (2012) suggested that drosophilids cannot easily evolve to increase upper thermal tolerance. Further, the study suggested quite small temperature safety margins (deviation of temperature maxima at place of origin and thermal limit) for tropical species, which thus might be particularly vulnerable to climate change (Kellermann et al., 2012). Temperate species might be least affected (Kellermann et al., 2012). The vulnerability of alpine species is still unknown but alpine environments are thought to be particularly susceptible (IPCC, 2007). Moreover, a high extinction rate for species adapted to cool habitats was observed (Bernardo & Spotila, 2006;Sinervo et al., 2010). Thus, D. nigrosparsa and other alpine species might be among the most affected organisms with ongoing climate warming.
Referring to our main question, the life-history traits and physiological limits of the alpine fly Drosophila nigrosparsa seem to be, at least partly, a result of the harsh mountain environment. On the one hand, the low oviposition rate and productivity as well as elongated