Nitrogen limitation and life history adaptation in the grasshopper "Omocestus viridulus"

The role of plant-mediated constraints to herbivore populations in terrestrial ecosystems remains relatively poorly understood. One aim of this study was therefore to explore the effects of low host plant nitrogen (N) content on herbivore performance and feeding behaviour, and thereby to evaluate the utility of the N limitation and nutrient balance concepts. The grass-feeding grasshopper Omocestus viridulus (Orthoptera: Acrididae) provided a well-suited model system as the species exhibits high tissue N demands (11.3 % N dw) compared to other herbivores, and uses a relatively poor food source in the wild (median N content: 2.2 % dw). Using N-poor soil and fertilizer, natural host grasses of contrasting N contents (high vs. low N) were grown in pots in the greenhouse. Juvenile performance experiments were then carried out with first generation (F1) offspring of grasshoppers caught in the field at three sites in Switzerland. The experiments were conducted both in a climate chamber and in outdoor cages under natural climatic conditions. I consistently found lower growth rates in grasshoppers reared on low N grasses, leading to somewhat prolonged development (13 % in the laboratory, 7 % outdoors) and slightly (4 – 5 %) smaller adult size. Juvenile mortality was low (always < 20 %) and similar among the food N treatments. As N contents differed strongly between the experimental grass treatments (e.g. 1.6 % vs. 4.3 % N dw in the climate chamber trial), my results suggest that natural food quality imposes no relevant nutritional constraint on grasshopper performance. Separate laboratory experiments on feeding behaviour served to investigate how grasshoppers cope with low quality food. First, female last-instar nymphs were allowed to feed singly on either high or low N grass (non-choice setting) over a 16 h period. The food’s initial fresh weight was known and converted to dry weight using constants. Individual food consumption was subsequently calculated by subtracting the dry weight of the food leftovers from the initial dry mass (gravimetrical method). This experiment showed that grasshoppers facing N-poor grass displayed substantially elevated food consumption (82 % higher on average) compared to conspecifics fed N-rich food. The animals thus compensated for lower nutritional quality by eating more. In a choice experiment, single grasshopper nymphs were offered high and low N grass simultaneously during a 28 h period. Individual consumption of both foods was calculated gravimetrically and revealed a striking preference for fertilized grass. On average, N-poor grass only accounted for some 20 % of the total ingested plant dry mass. It is argued that the proportion of consumed high N (protein-rich) and low N (energy-rich) grass likely represents the outcome of nutrient balancing. Proximately, the observed food selection can be explained by the modulation of taste receptor responsiveness via metabolic feedback, and/or associative learning of the spatial location of the different food sources. Overall, my results disagree with simple bottom-up predictions of herbivore population control, such as N limitation. Rather, this investigation provides
evidence for a high ability of arthropod herbivores to balance their nutrient intake, and thus corroborates a body of work performed with synthetic diets.
Another set of experiments using the same study organism was aimed at exploring the hypothesis that life history differentiation along a climatic gradient may have allowed a species to extend its geographic range. To this end, a total of eleven O. viridulus populations from eastern Switzerland were selected to span the species’ entire altitudinal range (410 – 2440 m). Temperature measurements and phenological surveys repeated in regular intervals at four field sites allowed quantifying the decline of available heat sums with increasing elevation, and its immediate effect on development rates. During the relatively cool season of 2002, embryonic development at 2215 m altitude started nearly three month later than at 410 m, and nymphal hatching and adult emergence were delayed by roughly two month. This implied a strong truncation of the reproductive period in high altitude animals, a situation predicted by life history theory to promote the evolution of shortened time to maturity (faster juvenile development). During the following, climatically favourable year, the phenological delay in the field of high altitude relative to lowland populations was less pronounced. This suggested that the variance in available season length increases with altitude, and that the strength of natural selection owing to time constraints at high elevation fluctuates considerably. In order to investigate adaptive (genetic) divergence in grasshopper life history, several traits determining development time (or time to maturity) were compared across the eleven study populations under common garden conditions, using offspring of field-caught animals. Embryonic development rates were found to increase according to the population’s altitude of provenance, with embryos from highest altitudes completing development some ten percent faster than lowland animals. This was found both at 27 °C and 19 °C incubation temperature, genotype by temperature interactions thus proved unimportant. Similarly, nymphal development accelerated with elevation of origin, the maximal difference between high and low altitude populations again amounting to roughly ten percent. In contrast, the stage of embryonic overwintering diapause, a pivotal determinant of developmental timing in annual organisms, was constant across six inspected populations from different altitudes. Embryonic development was always arrested just before the onset of embryonic rotation. This result agrees with data from other members of the same subfamily (Gomphocerinae), but disagrees with life history responses to seasonal time constraints reported for other Orthoptera. The documented lack of flexibility (genetic variation) in the diapause stage is therefore best understood as a phylogenetic constraint. This investigation offers strong support for the hypothesis of life history adaptation to local climates, although the degree of differentiation in intrinsic development rate among the populations is relatively modest and essentially limited by the conserved diapause stage. As a result, the immediate influence in the wild of decreasing temperatures along the altitudinal gradient overwhelms the genetic response observed in the laboratory, thus providing an example of cryptic evolution.
Life history theory generally predicts a trade-off between short juvenile development and large adult size, assuming invariant growth rates within species. This basic assumption has been explicitly tested in few organisms. The relationship between growth rate, juvenile development time and body size was examined in O. viridulus grasshoppers from 13 populations from different altitudes (including the populations mentioned above). All experiments were performed with first generation offspring at 32 °C in the climate chamber. I hypothesized that time constraints imposed by the altitudinal decline in season length, and arising from the advantage of earlier emergence of males relative to females (protandry), should have favoured increased intrinsic growth rates in high altitude animals and in the male sex. However, growth rates were similar across the populations. Instead, accelerated development with increasing elevation resulted in an altitudinal body size cline, with animals from the highest sites exhibiting roughly 12 % smaller adult size on average compared to lowland conspecifics. This size pattern was also observed in field animals. As I found a positive correlation between female body size and the number of eggs per clutch (but not offspring size), adaptation of O. viridulus to alpine climates involves a life history trade-off between time to maturity (development time) and fecundity, mediated by body size. An additional (fifth) nymphal stage, inserted before the penultimate nymphal stage and leading to an extended (by 20 %) developmental period and 12 % larger adult size, occurred in some females from low altitude. Individuals with five as opposed to four nymphal stages displayed significantly (5 %) smaller hatchling size, indicating that the insertion of the additional stage is contingent on maternal investment in the egg. It is argued that the absence of the five-stage developmental pathway from high elevation populations likely represents a side effect of the evolution of lower critical size thresholds in high altitude grasshoppers to accelerate development. Contrary to expectations, males were found to grow at substantially lower rates than females. This finding suggests that O. viridulus males not only experience protandry selection, but simultaneous selection for small size associated with reduced energetic requirements during mate search. Relatively slow growth in males probably allows the independent optimization of development time and body size among the sexes. My data further revealed that within populations, large individuals consistently developed faster than small individuals, resulting in a negative correlation between development time and size. This pattern could neither be ascribed to sib competition, as it was expressed in individuals both reared in groups and individually, nor to small hatchling size, a maternal effect. Hence, genetic variation in growth rate among individuals within populations seems probable. This study illustrates that the evolutionary response of intrinsic growth rate to different types of time constraints can greatly differ. I found stasis in growth rate among grasshopper populations in relation to seasonality, but highly divergent growth between the sexes owing to protandry selection. Further, ontogenetic evolution at the among-population level cannot necessarily be predicted based on within-population trait associations. The sign of the time-
size correlation was positive among populations, but negative among individuals within populations. Finally, I illustrate that detailed ontogenetic comparisons can shed light on the developmental cause (here a shift in critical size thresholds) underlying phenotypic evolution.