Intraspecific phenotypic variation in life history traits of Daphnia galeata populations in response to fish kairomones

Phenotypic plasticity is the ability of a genotype to produce different phenotypes depending on the environment. It has an influence on the adaptive potential to environmental change and the capability to adapt locally. Adaptation to environmental change happens at the population level, thereby contributing to genotypic and phenotypic variation within a species. Predation is an important ecological factor structuring communities and maintaining species diversity. Prey developed different strategies to reduce their vulnerability to predators by changing their behaviour, their morphology or their life history. Predator-induced life history responses in Daphnia have been investigated for decades, but intra-and inter-population variability was rarely addressed explicitly. We addressed this issue by conducting a common garden experiment with 24 clonal lines of European Daphnia galeata originating from four populations, each represented by six clonal lines. We recorded life history traits in the absence and presence of fish kairomones. Additionally, we looked at the shape of experimental individuals by conducting a geometric morphometric analysis, thus assessing predator-induced morphometric changes. Our data revealed high intraspecific phenotypic variation within and between four D. galeata populations, the potential to locally adapt to a vertebrate predator regime as well as an effect of the fish kairomones on morphology of D. galeata.

280 A significant interaction effect of 'Treatment x Population' was revealed mainly for within-281 population differences in the population from Greifensee (popG) and Jordan Reservoir (popJ) as 282 well as among those two populations.
283 In the presence of fish kairomones the age at first reproduction (AFR) differed significantly 284 within popJ (p=<.0001) and within popG (p=0.0023). Additionally, AFR differed significantly 285 between popG and popJ (p=0.0347) in the absence of fish kairomones, meaning that genotypes 286 of popG reproduce later compared to genotypes of popJ regardless of the treatment.
287 The total number of offspring (offspring) differed significantly between popG and popJ in the 288 absence of predator kairomones (p=0.0198) and the presence of predator kairomones 289 (p=0.0023), as well as between treatment groups (popG-fish vs. popJ-control (p=0.0311)).
290 Additionally, the number of offspring differed significantly between treatments within popJ 291 (p=0.0243) resulting in an increase of offspring for popJ exposed to fish.
292 In the presence of predator kairomones the somatic growth rate (SGR) differed significantly 293 within popG (p=<.0001) and popJ (p=0.0135) ( Figure 1E, Figure 2E). The visualization of growth 294 differences between treatments and populations (dSGR, Figure 5) showed that all genotypes 295 from popG had a negative growth rate in fish exposed environment, resulting in a smaller body 296 size. Four out of six genotypes from popJ had a negative growth rate, while genotypes from 297 popLC and popM vary in growth rate across treatments.
305 Effect of predator kairomones on the morphological trait 'shape' 306 A total of 83% of shape variation was explained by the first four Principal Components (PC1= 307 42%, PC2=24%, PC3=11% and PC4= 6%) ( Figure S1). The geometric morphometric analysis 308 showed that 'Treatment' was a meaningful factor for shape variation (Df=454, F=3.4177, 309 Z=3.1515, Pr(>F)=0.001**). Visualization revealed an overall shape change towards a smaller 310 body. In detail, the head area changed to a ventral position, while the tail area changed to a 311 dorsal position ( Figure 6A).  Table 3A). The visualizations showed a homogenous change from all directions to a smaller 316 body form for popG ( Figure 6B). Within popJ the overall shape change towards a smaller body  (Table 3B).
327 There was a significant interaction effect of 'Treatment x Population' on shape (Df=451, 328 F=2.5725, Z=2.3747, Pr(>F)=0.004**). The p-value matrix revealed that there was a statistical 329 significance difference within popLC between treatments (p=0.043*; Table 3C). 342 Detailed experimental information for each genotype can be found in the supporting material 343 (appendix, Figure S2 to Figure S7). 354 effect within as well as among the populations. In popJ, the variation of a trait itself, not the 355 change in the trait median value as a response was extremely reduced for two life history traits, 356 AFR and total number of broods ( Figure 2C). Almost all individuals of popJ started to reproduce 357 at the same age and produce the same amount of broods when exposed to fish kairomones, 358 showing a striking homogeneity under stress. On the contrary, in popM the variation for AFR 359 increased, resulting in a broader range of ages at first reproduction when exposed to 360 kairomones. Overall our study with a total of 24 genotypes revealed a broad spectrum of  . One might argue that the genotypes exposed to fish water 376 had a nutritional benefit due to introduced bacteria which they could have fed on. Since our 377 model organism Daphnia galeata falls into the low efficiency bacteria feeders with a medium 378 sized filter mesh of 1-1.6µm (Geller and Müller 1981), we cannot exclude this effect with 379 certainty. However, we think this potential of nutritional benefit can be ignored, because if 380 there had been a benefit of bacterial uptake in our experiment especially for the fish exposed 381 group, we would have expected to find an increase of relative fitness for all clones exposed to 382 this medium, which we did not observe (Table 2A). 383 We observed that the random factor 'Genotype' was the main driver for the observed 384 phenotypic variation of the two traits total number of offspring and somatic growth rate. The 385 phenotypic variation between genotypes was best visualized by plotting the differences of 405 For three life history traits we found a statistically significant block effect. The difference 406 between experimental rounds for somatic growth rate, total number of offspring in first brood 407 and shape could be attributed to the high clonal variation we observed in all life history traits.
408 Since we did not find a significant 'Treatment' effect for total number of offspring first brood 409 (brood1), we rule out that the block effect was connected to the presence of fish kairomones or    Differences of somatic growth rate (dSGR).
Differences of somatic growth rate (dSGR) as µm per day (mean +/-SD); calculated as: mean of SGR (fish) minus mean SGR (control) equals dSGR per genotype, sorted by populations.   Results of geometric morphometric analysis.