Evol Ecol (2010) 24:115–131 DOI 10.1007/s10682-009-9294-4 ORIGINAL PAPER
Are high-latitude individuals superior competitors? A test with Rana temporaria tadpoles Beatrice Lindgren Æ Anssi Laurila
Received: 16 April 2008 / Accepted: 16 January 2009 / Published online: 1 February 2009 Ó Springer Science+Business Media B.V. 2009
Abstract Species with a wide distribution over latitudinal gradients often exhibit increasing growth and development rates towards higher latitudes. Ecological theory predicts that these fast-growing genotypes are, in the absence of trade-offs with fast growth, better competitors than low-latitude conspecifics. While knowledge on key ecological traits along latitudinal clines is important for understanding how these clines are maintained, the relative competitive ability of high latitude individuals against low latitude conspecifics has not been tested. Growth and development rates of the common frog Rana temporaria increase along the latitudinal gradient across Scandinavia. Here we investigated larval competition over food resources within and between two R. temporaria populations originating from southern and northern Sweden in an outdoor common garden experiment. We used a factorial design, where southern and northern tadpoles were reared either as single populations or as mixes of the two populations at two densities and predator treatments (absence and non-lethal presence of Aeshna dragonfly larvae). Tadpoles from the high latitude population grew and developed faster and in the beginning of the experiment they hid less and were more active than tadpoles from the low latitude population. When raised together with high latitude tadpoles the southern tadpoles had a longer larval period, however, the response of high latitude tadpoles to the competition by low latitude tadpoles did not differ from their response to intra-population competition. This result was not significantly affected by density or predator treatments. Our results support the hypothesis that high latitude populations are better competitors than their low latitude conspecifics, and suggest that in R. temporaria fast growth and development trade off with other fitness components along the latitudinal gradient across Scandinavia. Keywords Activity Growth Intraspecfic competition Latitudinal clines Rana temporaria B. Lindgren A. Laurila (&) Population and Conservation Biology/Department of Ecology and Evolution, Evolutionary Biology Center, Uppsala University, Norbyva¨gen 18D, 75236 Uppsala, Sweden e-mail:
[email protected] B. Lindgren e-mail:
[email protected]
123
116
Evol Ecol (2010) 24:115–131
Introduction Large-scale climatic variation is an important factor causing adaptive intraspecific variation in a wide range of organisms (e.g. Conover and Schultz 1995; Angilletta et al. 2003; Ashton 2004; Blanckenhorn and Dermot 2004). For example, the shorter growth season at higher latitudes often selects for latitudinal clines in growth and development rates in ectotherms, as successful completion of the juvenile stages is crucial for many organisms and increased body size often enhances over winter survival (Roff 1980; Arendt 1997; Altwegg and Reyer 2003; Munch et al. 2003; Sears 2005). Accordingly, increased growth and development rates towards higher latitudes have been found in a wide range of ectotherms, including insects (Roff 1980; Arnett and Gotelli 1999; Gilchrist et al. 2001), fish (Conover and Present 1990; Conover et al. 1997; Imsland et al. 2000) and amphibians (Berven and Gill 1983; Merila¨ et al. 2000). While many morphological, behavioural and physiological traits can affect the competitive ability of individuals in exploitative competition, in mobile organisms body size and activity level are often positively correlated with competitive ability. Large animals are competitively superior because of their greater capacity to forage (Connell 1983; Schoener 1983; but see Persson 1985). High activity levels also increase competitive ability by improving harvesting rate (Werner 1992, 1994; Grill and Juliano 1996). Therefore, the higher growth and activity rates of high latitude populations may make them competitively superior over slower growing low latitude populations. Such a competitive superiority of fast-growing genotypes is supported by studies making use of among-strain variation of hatchery and wild salmonid fishes (e.g. Devlin et al. 1999; Biro et al. 2004, 2006). The relationship between latitudinal variation in growth and development and other important ecological traits, such as competitive ability or predator avoidance, are seldom studied. However, such information would be valuable if we are to understand how phenotypic variation along climatic gradients is maintained or how intra- and inter-specific interactions are modified if the climate changes. In the only study we are aware of that investigated competitive ability along a latitudinal gradient, James and Partridge (1998) found that high latitude Drosophila melanogaster populations were poorer competitors at high temperatures than low latitude populations when tested against a common laboratory strain. To our knowledge, no studies examining the competitive ability of conspecific animal populations along a latitudinal gradient have been conducted. In anuran tadpoles, intraspecific competition can be intense and negatively affect growth and survival, especially when mortality due to predation is low (see Wilbur 1980; Skelly and Kiesecker 2001 for reviews). Density-dependence during the aquatic larval stage can have a major effect on population regulation in amphibians (Wilbur 1980; Smith 1983; Berven 1990; Loman 2004), although more recent studies have also emphasized the importance of the terrestrial stages for population regulation (Hellriegel 2000; Biek et al. 2002; Vonesh and De la Cruz 2002). Growth and development rates of larval common frogs (Rana temporaria) increase along the latitudinal gradient across Scandinavia (e.g. Merila¨ et al. 2000; Laugen et al. 2003; Lindgren and Laurila 2005). Furthermore, high latitude tadpoles exhibit higher activity levels (Laurila et al. 2008), which, together with their larger body size, is likely to increase the competitive ability of high latitude tadpoles as compared to low latitude conspecifics. Importantly, high latitude tadpoles maintain their higher growth and development rates as well as higher activity over a range of temperatures (e.g. Laugen et al. 2003; Lindgren and Laurila 2005; Laurila et al. 2008) strongly suggesting adaptation to season length rather than to prevailing temperature (Conover and Schultz 1995; Laugen
123
Evol Ecol (2010) 24:115–131
117
et al. 2003; Palo et al. 2003). High latitude tadpoles also have higher growth efficiency turning ingested food more effectively into biomass than low latitude conspecifics (Lindgren and Laurila 2005). In R. temporaria, intraspecific competition among tadpoles can be intense, and Loman (2004) found evidence for density-dependent regulation at the larval stage in several ponds in southern Sweden, suggesting that larval competition is an important ecological factor affecting population size in this species. Here we investigated intraspecific competition between and within two populations of R. temporaria tadpoles situated 1,000 km apart along the latitudinal gradient across Sweden. We conducted an outdoor experiment where we addressed two main questions. First, do the fast-growing and active northern tadpoles have a competitive advantage over their southern conspecifics? We predicted that competition has a stronger negative effect on the slow-growing, less active southern tadpoles when raised together with the faster growing, more active northern tadpoles. In mixed-population treatments, we assigned the population origin of the tadpoles by using a diagnostic microsatellite marker with no alleles shared between the populations. As size-dependent competitive ability may depend on the amount of resources available for each individual (Persson 1985), we conducted the experiment at two densities. Second, since predator presence has a strong effect on activity of R. temporaria tadpoles (e.g. Laurila et al. 2004), we conducted the experiment both in the presence and absence of a non-lethal (caged) insect predator and asked: does predation risk affect the competitive outcome between northern and southern tadpoles? We predicted that the northern tadpoles would be less inclined to reduce activity and sacrifice growth due to predator presence, leading to an increase of competitive superiority of the northern over the southern tadpoles in the presence of predators.
Materials and methods R. temporaria has a large distribution area in Europe from northern Spain to northernmost Norway and breeds in a wide range of aquatic habitats (Gasc et al. 1997). Based on our previous studies along the latitudinal gradient (Merila¨ et al. 2004, A. Richter-Boix et al. in preparation), we chose two representative populations from the southern and northern parts of the gradient. The low-latitude population was obtained by collecting 12 pairs of adult R. temporaria in a pond outside Uppsala (Stora Almby, 59°510 N, 17°280 E; hereafter U population) on April 15 2004. This population has growth and development rates typical to an open-canopy pond in this area (A. Richter-Boix et al., manuscript in preparation) The high latitude population was obtained by collecting ca. 500 eggs from each of 12 freshly laid egg clutches near Kiruna (Jukkasja¨rvi, 67°510 N, 21°020 E; hereafter K population) on May 20. Also this population has growth and development rates typical to this geographic area (population N5 in Merila¨ et al. 2004). The K population is located ca. 1,000 km N of the U population. The growth season length (daily mean temperature over 5°C) is ca. 194 and 113 days in U and K populations, respectively (Odin et al. 1983). Both populations breed in permanent, open-canopy ponds. The populations are fairly large (ca. 100 breeding females in U, [50 females in K), and larval competition is likely to occur within both populations. In addition, large populations of R. arvalis and Bufo bufo tadpoles may compete with R. temporaria tadpoles in U, but R. temporaria is the only amphibian species in K. Both ponds have large dragonfly and diving beetle predators, however, predator density is higher in U (Laurila et al. 2008). To be able to conduct a competition study between two phenologically separated populations, the tadpoles from the two populations need to hatch at approximately the
123
118
Evol Ecol (2010) 24:115–131
same time. By collecting adults in U and storing them in the laboratory we delayed their reproduction until the breeding started in K. Our original plan was to collect adults also in K and make artificial crossings in the same manner as in the southern population. Unfortunately, we were unable to capture a sufficient number of adult individuals in K and used field-collected fresh eggs instead. To evaluate the possibly harmful effects of time delay in breeding of the U population, we conducted the crosses in two batches. Five pairs were crossed the following day after being captured (April 16, with methods described in Ra¨sa¨nen et al. 2003) to produce five full-sib families which acted as treatment controls. The remaining seven pairs were kept in the laboratory at 4°C in plastic boxes filled with moist peat moss (Sphagnum) until they were crossed on May 20 to produce seven full-sib families for the experimental U population. For both populations ca. 500 eggs per family were raised in the laboratory (18°C, 18 h light:6 h dark photoperiod) in two 3 l vials (250 eggs per vial) filled with reconstituted soft water (RSW; APHA 1985) until they reached Gosner stage 25 (Gosner 1960, gills fully absorbed). The water in the vials was changed completely every 3 days. To investigate whether the extended storing of the U adults affected the larval life history traits of the offspring, we reared U tadpoles from both the early and late crosses until metamorphosis in temperature-controlled laboratory (18°C, 18 h light:6 h dark photoperiod), corresponding to the conditions in Uppsala area during the late larval period (Orizaola and Laurila 2009; A. Richter-Boix and A. Laurila, unpublished). About 10–11 (early) or 5–8 (late cross) U tadpoles from each family were raised individually in 1 l vials filled with 0.8 l of RSW. The vials were arranged on four shelves (blocks) to account for a known vertical temperature gradient within the room. The tadpoles were fed ad libitum chopped spinach and the water was changed every 4 days. Experimental setup The experiment was conducted outdoors in a fenced field close to Uppsala. Due to logistic limitations we could not replicate the experiment under high latitude conditions, however, our laboratory studies indicate that K individuals maintain higher growth and development rates over a range of relevant temperatures (14–22°C; e.g. Laugen et al. 2003; Lindgren and Laurila 2005; Laurila et al. 2008) suggesting that our results are robust in respect to the location of the experiment. We filled 60 opaque plastic tanks (36 9 40 9 90 cm) with 80 l of tap water 2 weeks before the start of the experiment and added 1 l of pond water as an algal oculum. We added 10 g of dried aspen (Populus tremula) leaves and 6 g of rabbit pellets to act as a nutrient base and to provide food and shelter for the tadpoles. The tanks were covered with mosquito net to prevent colonisation by insects. The photoperiod during the experiment corresponded to late larval period in both populations (day length in June in Uppsala corresponds roughly to August day length in Kiruna), suggesting that the populations were likely to experience the time of season in a similar manner. Moreover, in laboratory the effects of photoperiod on the growth and development in R. temporaria tadpoles are considerably smaller than the genetic differences found between southern and northern populations (Laurila et al. 2001; A. Laurila and S. Pakkasmaa, unpublished data) suggesting that photoperiod cues played only a minor role in the responses. The two populations reached Gosner stage 25 simultaneously, and the experiment was started 2 days later (June 3, day 0 of the experiment). A 2 9 2 9 2 9 2 randomized block design consisting of two populations (U, K), two densities (low and high), two betweenpopulation competition treatments (mixed and alone), two predator treatments (present or
123
Evol Ecol (2010) 24:115–131
119
absent) and five spatial blocks was used. The tanks were placed in five rows (the blocks) 1 m apart, with one randomly placed replicate of each treatment combination within each block giving us five replicates per treatment and a total of 60 tanks (note that the mixedpopulations treatment tanks with U and K tadpoles together were the same for the two populations). For each population we pooled tadpoles in equal amounts from each family to one container before adding them to the experimental tanks. Each tank received either 10 (low density) or 30 (high density) tadpoles and the populations were allocated to these treatments either alone (only U or K tadpoles) or as mixes of the two in equal proportions (i.e. 5 or 15 tadpoles from both populations). As predators we used late-instar dragonfly larvae (Aeshna sp.) collected in ponds near Uppsala. In the experiment, the predators were kept in transparent plastic cages (Ø 80 mm, height 210 mm) with a double net bottom (mesh size 1.5 mm) allowing the tadpoles to receive both visual and chemical cues from the predator. One cage was placed in the middle of each tank, hanging ca. 7 cm from the bottom. Each cage in the predator treatment received one dragonfly larva, whilst in the no-predator treatment the cage was left empty. The predators were fed every second day with ca. 300 mg of R. temporaria tadpoles. Response variables Tadpole behaviour was observed around noon for three consecutive days during two periods: days 4–6 (period 1) and days 14–16 (period 2) of the experiment. On each observation day the number of visible (i.e. not hiding) and the number of active (moving tail) tadpoles was recorded in each tank at five occasions separated by 30 min. After the first individuals metamorphosed (stage 42, Gosner 1960; occurrence of at least one forelimb), the containers were checked daily for metamorphs. Metamorphosed tadpoles were sacrificed using MS-222 and preserved in 70% ethanol. From the preserved individuals we later measured body length (BL; from the tip of the snout to the cloaca) using a digital calliper to the nearest 0.01 mm, and body mass (BM) to the nearest 0.1 mg. Larval period (LP) was measured as days elapsed from the start of the experiment to metamorphosis. Growth rate (GR) was estimated by dividing BM with LP. Tissue samples were taken from the metamorphs in the mixed treatments in order to establish population origin by using molecular methods (see below). Genetic analysis We used the microsatellite marker RtlH (Pidancier et al. 2001; Palo et al. 2003) to assign population origin of the tadpoles in the mixed-population treatment. This locus had no shared alleles in the two populations (Table 1). Before genotyping the experimental tadpoles, we extracted and analysed the DNA from the parents of the Uppsala population and from ten individuals from each Kiruna family (120 individuals in total) in order to determine the alleles occurring in the populations. The DNA extractions were made from tissue samples using a standard salt extraction protocol. We then amplified the DNA using PCR. The PCRs were performed in a total volume of 10 ll with *100 ng nuclear DNA, 125 lM of each dNTP, 109 PCR buffer II (PE Biosystems), 1.5 mM MgCl2, 0.25 U of AmpliTaq DNA polymerase (PE Biosystems) and 400 nM of each primer (one labelled with fluorescent dye). The PCR protocol used were as follows: 94°C for 3 min (initial denaturation), 35 cycles of 94, 50 and 72°C for 30 s each and 72°C for 5 min (elongation step). The amplified samples were analysed using a MegaBACE genotyper and scored
123
120
Evol Ecol (2010) 24:115–131
Table 1 Allele frequencies and sample size (individuals) analyzed for RtlH microsatellite locus in the U parental generation, K tadpoles of known family origin and experimental individuals in the mixed populations-treatment Allele
U, parents
K, family
U, experimental
K, experimental
203
0.607
0
0.537
0
207
0
0.392
0
0.385
209
0
0.608
0
0.615
211
0.393
0
0.463
0
Sample size (N)
24
120
187
187
using the software Fragment Profiler 1.2 (Amersham Biosciences). The genotype of two tadpoles out of 376 failed to be scored and they were excluded from the analyses. Statistical analyses We used mixed model ANOVAs with REML estimation (proc Mixed in SAS 9.1) to test if storing the parental U individuals in the laboratory affected metamorphic body weight, body length, growth rate or larval period of the tadpoles. Cross type and block were treated as fixed effects and family (nested under cross type) as a random effect. The behavioural and life history data from the outdoor competition experiment were analysed using MANOVAs followed by univariate ANOVAs. Our experimental design was 2 9 2 9 2 9 2 9 5 randomized block design. However, since the two populations in the mixed treatment could not be differentiated in the behavioural analyses, separate analyses were run for behavioural and life history data. In the behavioural analyses we had four factors: mix, with three levels [U alone, both populations (mix), and K alone], density, predator and block. For each tank, we calculated the mean proportion of active and visible individuals in relation to initial tadpole density during each 3-day period. The behavioural data were arcsin-squareroot transformed before the analyses. Six tanks were excluded from the behavioural analyses because algal blooming interfered with the observations. Removal of these tanks did not affect the results qualitatively. Tank means were used as response variables in the life history trait analyses. Survival was determined as the number of metamorphosing tadpoles divided by the original number of tadpoles in the experimental unit and analysed with type III general linear models with a logit link function and binomial error structure using GENMOD procedure in SAS. Results Effects of crossing date on U life history Metamorphs from the later crosses grew faster (GR; F1,8.27 = 69.80, P \ 0.001), had longer bodies (BL; F1,9.14 = 16.17, P = 0.003) and larger body mass (BM; F1,8.67 = 69.91, P \ 0.001) at metamorphosis, but did not develop faster (LP; F1,10.2 = 2.56, P = 0.140; Fig. 1). Mixed model ANOVAs showed significant family effects on BL (P = 0.040) and LP (P = 0.032), and close to significant effects on BM (P = 0.063) and GR (P = 0.054; Fig. 1). Block had a significant effect on LP (F3,80.3 = 15.72, P \ 0.001), but not on GR, BL or BM (P = 0.321, 0.931 and 0.548, respectively), suggesting that development rate was the most sensitive variable to temperature variation within the laboratory.
123
Evol Ecol (2010) 24:115–131
121
Fig. 1 Family means for body length, body mass, larval period and growth rate of R. temporaria tadpoles in early and late crosses in the Uppsala population
Behaviour During period 1, K tadpoles in single population tanks were generally more active and hid less than U tadpoles, whilst tadpoles in the mixed treatment were intermediate in these traits (Mix- treatment in Table 2; Fig. 2). Density had a significant main effect on activity, as tadpoles were more active at high than low density (Table 2; Fig. 2). Predator main effect was not significant (Table 2), but tadpoles tended to be less active and hide more at high density in the presence of a predator, bringing about a significant density 9 predator interaction (Table 2; Fig. 2). As shown by the significant density 9 mix 9 predator interaction, this effect was especially clear in the mixed treatment, where low density tadpoles were more active and hid less in the presence of a predator (Table 2; Fig. 2). Density had a strong effect on tadpole behaviour during period 2. In contrast to period 1, tadpoles were now more active and hid less at low density (Table 2; Fig. 2). The main effects of predator and mix were not significant, but there was a significant density 9 mix 9 predator interaction, with tadpoles in the low density-mixed treatment combination being less active and hiding more than tadpoles in the single population treatments in the presence of a predator (Table 2; Fig. 2). Moreover, the opposite, although not as strong, pattern emerged in the absence of a predator, i.e. the tadpoles hid less and were more active in the mixed than in the single population treatment (Fig. 2). Block had no significant effect on any of the behavioural traits (Table 2). Life history and survival Populations differed strongly in larval life history traits, with K tadpoles having shorter LP and higher GR and BL (Table 3; Fig. 3). High density had strong negative impacts on BL,
123
123
3.02
1, 38
2, 38
2, 38
1, 38
2, 38
Predator
Density 9 Mix
Mix 9 Predator
Density 9 Predator
Density 9 Mix 9 Predator
Significant results are given in bold
0.10
2, 38
Mix
4.68
4.80
2.13
5.88
0.25
1, 38
1.07
4, 38
Density
0.015
0.035
0.133
0.061
0.755
0.006
0.621
0.383
3.05
4.14
1.56
2.79
0.02
8.36
4.74
0.56
F
F
P
Active 1
Visible 1
Block
df
0.059
0.049
0.223
0.074
0.656
0.001
0.036
0.696
P
3.00
0.00
1.56
0.14
2.01
0.98
42.25
1.29
F
Visible 2
0.062
0.985
0.224
0.868
0.165
0.386
