INTRODUCTION
A major challenge facing all organisms is to adapt to environments that vary within their lifespan. A route to responding to and surviving such variation is phenotypic plasticity, the ability of individual genotypes to change phenotype when exposed to different environments during their life-cycle . One pervasive, natural source of environmental variation is the risk of mortality from predation, known as predation risk. Predation risk induces a suite of changes in the behaviour, life history and morphology of many plants and animals . Of particular interest are morphological responses to predation risk which range from the production of spines to changes in the shape of a portion of the body or the entire body plan of an organism.
Many studies of predator-induced shape change have focussed on linear assessments of shape, which measure the distance between two points. Key examples include changes in defensive dorsal spine length , body depth that affects vulnerability to gape-limited predators and morphological features associated with behavioural swimming escape responses . However, this type of analysis only captures a subset of the overall shape variation.
Measurement of overall shape is a multivariate analysis (i.e. it involves multiple different variables). A well-established method for assessing multivariate plasticity in shape is geometric morphometrics , which uses anatomical coordinates as shape variables to measure relative differences in shape. This approach has been used extensively to measure predator-induced changes in shape for a wide range of organisms, such as fish , amphibians and snails .
Geometric morphometrics not only allows the measurement of shape overall, but also the extent of modularity and integration between different aspects of shape (Klingenberg, 2014). Modularity refers to the level of covariation between different traits within morphological structures, or modules, relative to the level of covariation between these structures. Integration refers to the level of co-variation between different traits throughout a morphological structure or even a whole organism (Klingenberg, 2014). Therefore, modularity exists if the level of integration within modules is strong compared to the integration between modules (Klingenberg, 2009).
In addition to considering multivariate shape, there has been a shift from the standard ”two-environment” approach for assessing phenotypic plasticity and estimating reaction norms (Roff 1999) to analysing changes along a gradient. More recently, it has become standard to evaluate plasticity in multiple traits and along environmental gradients of more than two environments .
Despite such work, shape has rarely been assessed as a plastic trait in water fleas (Daphnia species), an iconic organism for the study of size-selective, predator-induced phenotypic change. Instead, research has largely focused on assessing the production of inducible morphological defences, such as the head spikes of Daphnia pulex , called ‘neckteeth’, which develop in response to predator cues (kairomones) released from their midge larvae predators . Although there has been some research into the dorsal pattern of induced morphology inD. pulex , the question of how overall shape changes in response to predation risk remains unanswered.
Other examples of predator-induced changes in Daphnia suggest that overall shape may change in response to predation risk. Considerable variation exists in many features of the Daphniaspp. body plan including body width , alignment , shoulder shape and carapace strength . Furthermore, changes in body size are well-documented in the context of size selective predation theory and empirical assessment of life history responses to predation risk . Together with associated research on fish, this suggests that overall shape might increase survival and therefore provide important fitness benefits to plasticity.
In this study, we evaluate shape plasticity along a gradient of increasing predation risk in three genotypes of D. pulex which differ in their sensitivity to predator cues. We apply morphometric landmark-based analysis to photographs of Daphnia taken by , in which D. pulex were exposed to six levels of predation risk from their midge larvae predator, Chaoborus flavicans . We combine geometric morphometrics with phenotypic trajectory analysis to formally evaluate the multivariate change in shape and estimate measures of both modularity and integration to evaluate if there are coherent units of the body plan that respond to predation risk, and whether these units change independently or together .
In advance of the analysis, we predict narrower bodies and bigger heads will form part of the predator-induced response in D. pulex . Narrower bodies enhance the predator escape response in fish and amphibians and we expect that bigger heads are more likely to interfere with predation linked to the neckteeth defence. Given these two predictions, we might also predict modularity and / or integration of these responses. From a modularity perspective, we predict that the head region and lower body are separated in their response to predation due to the nature of the induced morphological change in the head region. This means that the neckteeth defence and associated shape changes of the head would be localised to that part of the animal and relatively independent of changes to the body. Furthermore, under the modularity hypothesis, we might also predict that changes in the dorsal region, where the neckteeth form, respond independently to the ventral portion of the daphnid. As for the integration hypothesis, we predict some level of integration between the head and lower body regions driven by a negative developmental correlation, where the head gets larger and the body gets narrower and longer.