Kin competition accelerates experimental range expansion in an arthropod herbivore

With ongoing global change, life is continuously forced to move to novel areas, which leads to dynamically changing species ranges. As dispersal is central to range dynamics, factors promoting fast and distant dispersal are key to understanding and predicting species ranges. During range expansions, genetic variation is depleted at the expanding front. Such conditions should reduce evolutionary potential, while increasing kin competition. Organisms able to recognise relatives may be able to assess increased levels of relatedness at expanding range margins and to increase their dispersal in a plastic manner. Using individual-based simulations and experimental range expansions of a spider mite, we demonstrate that plastic responses to kin structure can be at least as important as evolution in driving range expansion speed. Because recognition of kin or kind is increasingly documented across the tree of life, we anticipate it to be a highly important but neglected driver of range expansions.


Introduction
Range expansions and biological invasions have traditionally been studied from an ecological and conservation biological perspective, primarily in relation to climate change and invasive species (Keane & Smith 2002). The speed and extent of a range expansion can only be affected through a change in the following underlying life history traits: dispersal and reproduction (Fisher 1937  Paradoxically, spatial sorting of genotypes during invasion is tightly associated with a successive loss of genetic variation due to subsequent founder effects. These founder effects render genetic drift 5 important, and have the potential to further affect evolutionary change (Hallatschek et al. 2007). Loss of genetic variation may not only constrain evolutionary change but also increase local levels of genetic relatedness (Newman & Pilson 1997;Kubisch et al. 2013;Nadell et al. 2016). In many arthropods, for instance, single female colonisers found highly related populations (Dingle 1978).
Increased relatedness has a strong impact on dispersal, both in terms of evolutionary and plastic mechanisms (e.g. Bowler & Benton 2005, Ronce 2007). In general, dispersal is a spatial process leading to fitness maximisation . It is therefore strongly context-and condition-dependent (Clobert et al. 2009) with high local densities typically boosting emigration rates, thereby enabling individuals to increase their short term performance by avoiding resource competition (Bowler & Benton 2005). Short-term performance is, however, only an incomplete measure of fitness as the latter also depends on aspects of relatedness, eventually determining the spread of genes within the population (Hamilton 1964). With increasing relatedness, competition among kin will become one of the major interactions, even in highly cooperative or social species ). If populations are tightly kin-structured, emigration of individuals reduces local resource competition among kin while also providing a chance of colonising new habitat, even if individual dispersal costs are high. Kin competition (i.e., competition between genetic relatives) is therefore expected to be a strong driver of dispersal evolution by maximising inclusive fitness (Hamilton & May 1977).
Evidently, plastic adjustments of dispersal, conditional to the local level of relatedness, may be even more adaptive (Bitume et al. 2013). A major prerequisite for relatedness-dependent dispersal to be effective is the presence of kin recognition mechanisms that lead to kin discrimination (e.g., Waldman 1987; Blaustein et al. 1988;Waldman et al. 1988;Tang-Martinez 2001), that is, some sort of association and phenotype matching. This process involves the discrimination of traits of kin or self from traits of any other individual, either by learning or by means of recognition alleles. While kin recognition strategies based on spatial association and learning are widely documented, evidence is accumulating 6 that kin recognition in the absence of social interactions and learning is neither uncommon in animals  (Dudley et al. 2013). Assuming that relatives (kin by descent) at range expansion fronts will be identical-by-state (kind by sharing identical traits), plastic increases of dispersal are anticipated to be a key driver of range expansions and may explain the paradox of fast expansions despite severe genetic diversity loss (Estoup et al. 2016). Relatedness-dependent dispersal is mechanistically driven by selection of kind rather than of kin (Queller 2011). We therefore use here the term relatedness-dependent dispersal as a special form of relatedness-dependent dispersal to refer to the recognition based on identity-by-state (IBS) rather than by Hamilton's identity-by-descent (IBD) mechanisms. In organisms in which variation in dispersal and/or reproduction is primarily environmentally driven, relatedness-dependent dispersal following assessment of IBS may even be the primary driver of fast range expansions.  (Williams et al. 2016b). These studies included reshuffling or replacing experiments to quantify the eco-evolutionary loop, that is how evolution feeds back on the range expansion dynamics. In this experimental procedure, individuals are systematically replaced by individuals from a source population or from a random patch in the experimental range expansion to avoid the evolution of traits related to spatial sorting or local adaptation while maintaining population densities, age, and sex structure constant. Such approaches have, however, the major drawback that patterns in relatedness and phenotype (state) are destroyed as well. Given the presumed relevance of kin competition for dispersal, and the central role of dispersal for range expansions, we expect that observed differences in expansion speed previously attributed to spatial sorting and selection could equally likely result from changes in relatedness and subsequent changes in relatedness-dependent dispersal.
We here set out to test the relative importance of plastic relatedness (IBS)-dependent dispersal compared to spatial sorting and selection for the dynamics of range expansions. We use the two-spotted spider mite Tetranychus urticae Koch (Acari, Tetranychidae) as a model organism because the impact of relatedness on dispersal kernels has been extensively studied in this species, which has been developed Yano 2014). Importantly, kin recognition has been shown to play an important role in conditiondependent dispersal (Bitume et al. 2013). We firstly developed a highly parameterized yet simple simulation model based on spider mite life histories and relatedness-by-IBS-dependent dispersal reaction norms to formalise our hypotheses and predictions. In order to provide empirical proof of principle, we subsequently conducted two experiments in which genetic diversity (i.e., evolutionary potential) and relatedness were manipulated to infer whether range expansion dynamics are jointly affected by spatial sorting and kin competition. This was accomplished by contrasting evolved trait divergence and the rate of range expansion in two sets of experimental range expansions that differed in the level of genetic variation and spatial structure. While a first experiment followed the earlier used replacement manipulations that eliminate both kin structure and evolution, a second experimental range expansion prevented evolution while maintaining kin structure.

General model algorithm
The model is individual-based and simulates demographic and evolutionary processes along a onedimensional array of patches (metapopulation). Patches contain resources, which are consumed by individuals at different rates depending on their life stage (juvenile or adult). Resources are refreshed on a weekly basis. We parameterised relatedness-dependent dispersal according to earlier research (partly published in (Bitume et al. 2013). Relatedness-dependent dispersal was here studied under average densities, that is, no further density-dependence was implemented. We thus assume the relatedness-dependent dispersal kernels to be relevant for the average population densities during range expansion. A detailed model description and additional results on in silico trait evolution are available in APPENDIX 1 in the Supporting Information.
Males and females of Tetranychus urticae differ in a number of aspects. Firstly, males are smaller when reaching the adult life stage, and hence contribute less to resource consumption. Secondly, dispersal behaviour differs between the two sexes, with adult females being the dominant dispersers, whereas juveniles and males disperse very little. Lastly, the species is characterized by a haplodiploid life cycle, where non-mated females only produce haploid male offspring, and mated females can produce both haploid male and diploid female eggs. The sex ratio of spider mites is female-biased, with approximately 0.66 males to females. For these reasons and for the sake of simplicity, we designed the model to only include female mites, where the genotype of the individual is passed on from mother to daughter.
Individuals carry the following genetic traits: age at maturity, fecundity, longevity, and a locus

Experimental range expansion
An experimental range expansion consisted of a linear system of populations: bean leaf squares (2  2 cm) connected by parafilm bridges [8  1 cm], placed on top of moist cotton. A metapopulation was initialised by placing ten freshly mated one-day-old adult females on the first patch (population) of this system. At this point, the metapopulation comprised only four patches. The initial population of ten females was subsequently left to settle, grow, and progressively colonise the next patch(es) in the linear array through ambulatory dispersal. Three times a week, all patches were checked and one/two new patches were added to the system if mites had reached the second-to-last/last patch. Mites were therefore not hindered in their dispersal attempts, allowing for a continuous expansion of the range. A regular food supply was secured for all populations by renewing all leaf squares in the metapopulation once every week; all one-week-old leaf squares were shifted aside, replacing the two-week-old squares that were put there the week before, and in their turn replaced by fresh patches. As the old patches slightly overlapped the new, mites could freely move to these new patches. Mites were left in this experimental metapopulation for approximately ten generations (80 days) during which they could gradually expand their range.

Treatments
We performed two experiments, each of which contrasted two types of experimental metapopulations (see Figure 1). In the first experiment, we contrasted unmanipulated control LS-VL strains, further abbreviated as CONTROL, with a treatment where females in the metapopulations were replaced on a weekly basis by randomly chosen, but similarly aged, females from the LS-VL stock. This Replacement From Stock treatment is further abbreviated as REPLACEMENT. The metapopulations within the CONTROL treatment thus started with a high enough amount of standing genetic variation for evolution to act on. Kin structure was not manipulated in this treatment and kin competition was therefore expected to increase towards the range edge (see introduction). The REPLACEMENT treatment maintains age and population structure (i.e., if x females were on a patch before the replacement, they were replaced by x females from the stock) but prevented genetic sorting, and destroyed local relatedness, thus preventing both spatial sorting and kin competition. In this experiment, we thus In addition to monitoring range expansions along the linear system, we quantified life history trait variation and genetic variation in gene expression between core and edge populations at an unprecedented level of detail (APPENDIX 3, 4 in Supporting Information). All data were analysed using general(ized) linear models with proper error structure; the individual traits longevity, fecundity, sex ratio and survival were integrated into a simulated growth rate measure by means of bootstrapping within replicates. A detailed overview of all used statistical models can be found in APPENDIX 3.

Results
The eventually reached range size is a measure of range expansion speed. Despite the incorporation of uncertainties regarding condition-dependent dispersal thresholds, our model (APPENDIX 1 in Supporting Information) predicted range expansions to proceed at a 28.7% slower mean rate when signatures of both kin competition and spatial sorting were removed, while expansion rates were only 4% slower when only spatial sorting was prevented, but kin competition was present (Fig. 2).
In the experiments, we detected a 28% lower rate of range expansion in the treatment with mites replaced from stock, in which kin competition and evolution were constrained, versus the nonmanipulated control treatment (CONTROL-REPLACEMENT contrast: GLM, day × treatment interaction for range size, F1,54.8=7.62; P=0.007; Fig. 3). However, no statistically significant differences were found in the experiment that contrasted the evolving with non-evolving kin lines that inhibited spatial sorting Bootstrapping growth rates did not follow the empirically determined higher growth rates at the leading edges (see APPENDIX  Our conclusion that increased relatedness at expanding range margins may further boost expansion dynamics is based on the assumption that organisms are able to assess local genetic relatedness via kin recognition abilities, and to exhibit an according conditional dispersal response. Kin recognition strategies can be categorised into (i) mechanisms based on predictable kin overlap in space and time (for example parent birds treating hatchlings in their nest as their offspring or larvae interacting with presumed relatives because eggs laid by parents are spatially concentrated), (ii) kin discrimination following initial interactions and learning (offspring getting habituated to cues from nestlings or individuals living nearby and learning to associate specific cues with presumed kinship), and (iii) kin discrimination based on innate, typically genetic cues that enable recognition of relatives (when kinds indicate kin; Queller 2011) under conditions that are not predictable in space or time (Waldman 1988).
The first two mechanisms are obviously not relevant within our framework as they are independent of any spatial signature of genetic diversity loss and increased relatedness at expanding range margins. For relatedness-dependent dispersal to occur within a metapopulation or during range expansions, individuals need to be able to reliably assess the level of kin competition within a larger spatial neighbourhood, typically the local patch. Because gene flow renders levels of local relatedness dynamic, phenotype matching based on self-learning or recognition alleles is anticipated to be the main relevant

2012)
, support for the latter mechanism is much more controversial than for mechanisms related to learning or direct spatial associations . The current debate is strongly altruismcentred, fuelled by the absence of genetic kin-recognition strategies in many cooperating and/or groupliving species. Within this context, theoretical developments demonstrate that genetic kin recognition mechanisms can only be stable when mutation rates are high, or when extrinsic mechanisms, such as parasite-host dynamics, maintain the diversity of alleles linked to cues involved in kin through fluctuating selection ).
Dispersal decisions are taken hierarchically in response to multiple ultimate and proximate drivers (Matthysen 2012;Legrand et al. 2015). Because dispersal is theoretically leading to an ideal free distribution of fitness expectations in a metapopulation, it is a fitness-homogenising process (Bonte & Dahirel 2017). In response to local selection pressures, such as density, sex ratio, disturbance and relatedness (Bowler & Benton 2005), dispersal leads to the spread of genotypes in spatially structured systems. It is thus key to the maintenance of genetic diversity at the metapopulation level. Given that most likely all organisms inhabit spatially structured environments, kin(d)-recognition strategies might thus have primarily evolved to enable relatedness-dependent dispersal in order to avoid kin 19 competition. This perspective would not only explain the existence of genetic kin recognition across a wide range of non-cooperative organisms, it also suggests that its link with range expansions is relevant far beyond our studied system. It is for instance not improbable that mobile vertebrates engage in exploratory behaviour to assess the level of relatedness with conspecifics in areas distant from the breeding territory (Delgado et al. 2014). Less mobile or sedentary sexual species (insects, plants) might by contrast use genetically based cues following in-and outbreeding to assess the level of relatedness at larger spatiotemporal scales from a set of close interactions with conspecifics.
If future research confirms the omnipresence of genetic kin(d)-recognition strategies (including in humans Wedekind et al. 1995;Jacob et al. 2002), spread mechanisms based on kin(d) recognition might be widespread, and, in organism with predominantly plastic reproduction and dispersal, potentially more relevant than the earlier proposed eco-evolutionary feedbacks following spatial sorting. From an even broader perspective, our work calls for a further integration of hitherto rather isolated disciplines related to the evolution of sociality and spatial dynamics to increase our understanding of patterns of biodiversity from local to regional scales.