Body size is intrinsically linked to metabolic rate and life-history traits, and is a crucial determinant of food webs and community dynamics. The increased temperatures associated with the urban-heat-island effect result in increased metabolic costs and are expected to drive shifts to smaller body sizes . Urban environments are, however, also characterized by substantial habitat fragmentation , which favours mobile species. Here, using a replicated, spatially nested sampling design across ten animal taxonomic groups, we show that urban communities generally consist of smaller species. In addition, although we show urban warming for three habitat types and associated reduced community-weighted mean body sizes for four taxa, three taxa display a shift to larger species along the urbanization gradients. Our results show that the general trend towards smaller-sized species is overruled by filtering for larger species when there is positive covariation between size and dispersal, a process that can mitigate the low connectivity of ecological resources in urban settings . We thus demonstrate that the urban-heat-island effect and urban habitat fragmentation are associated with contrasting community-level shifts in body size that critically depend on the association between body size and dispersal. Because body size determines the structure and dynamics of ecological networks , such shifts may affect urban ecosystem function.

The increasing urbanization process is hypothesized to drastically alter (semi‐)natural environments with a concomitant major decline in species abundance and diversity. Yet, studies on this effect of urbanization, and the spatial scale at which it acts, are at present inconclusive due to the large heterogeneity in taxonomic groups and spatial scales at which this relationship has been investigated among studies. Comprehensive studies analysing this relationship across multiple animal groups and at multiple spatial scales are rare, hampering the assessment of how biodiversity generally responds to urbanization. We studied aquatic (cladocerans), limno‐terrestrial (bdelloid rotifers) and terrestrial (butterflies, ground beetles, ground‐ and web spiders, macro‐moths, orthopterans and snails) invertebrate groups using a hierarchical spatial design, wherein three local‐scale (200 m × 200 m) urbanization levels were repeatedly sampled across three landscape‐scale (3 km × 3 km) urbanization levels. We tested for local and landscape urbanization effects on abundance and species richness of each group, whereby total richness was partitioned into the average richness of local communities and the richness due to variation among local communities. Abundances of the terrestrial active dispersers declined in response to local urbanization, with reductions up to 85% for butterflies, while passive dispersers did not show any clear trend. Species richness also declined with increasing levels of urbanization, but responses were highly heterogeneous among the different groups with respect to the richness component and the spatial scale at which urbanization impacts richness. Depending on the group, species richness declined due to biotic homogenization and/or local species loss. This resulted in an overall decrease in total richness across groups in urban areas. These results provide strong support to the general negative impact of urbanization on abundance and species richness within habitat patches and highlight the importance of considering multiple spatial scales and taxa to assess the impacts of urbanization on biodiversity.

Although strict asexuality is supposed to be an evolutionary dead end, morphological, cytogenetic, and genomic data suggest that bdelloid rotifers, a clade of microscopic animals, have persisted and diversified for more than 60 Myr in an ameiotic fashion. Moreover, the genome of bdelloids of the genus Adineta comprises 8%-10% of genes of putative non-metazoan origin, indicating that horizontal gene transfers are frequent within this group and suggesting that this mechanism may also promote genetic exchanges among bdelloids as well. To test this hypothesis, we used five independent sequence markers to study the genetic diversity of 576 Adineta vaga individuals from a park in Belgium. Haplowebs and GMYC analyses revealed the existence of six species among our sampled A. vaga individuals, with strong evidence of both intra- and interspecific recombination. Comparison of genomic regions of three allele-sharing individuals further revealed signatures of genetic exchanges scattered among regions evolving asexually. Our findings suggest that bdelloids evolve asexually but exchange DNA horizontally both within and between species.

Occasional" sexuality occurs when a species combines clonal reproduction and genetic mixing. This strategy is predicted to combine the advantages of both asexuality and sexuality, but its actual consequences on the genetic diversity and species longevity are poorly understood. Androgenesis, a reproductive mode in which the offspring inherits its entire nuclear genome from the father, is often reported as a strictly clonal reproductive mode. Androgenesis is the predominant reproductive mode within the hermaphroditic, invasive lineages of the mollusk genus Corbicula. Their ability to reproduce clonally through androgenesis has been determinant in their invasive success, having colonized during the 20th century American and European freshwater systems, where they became notorious invaders with a widespread, global distribution. However, in androgenetic Corbicula clams, occasional genetic mixing between distinct lineages has also been observed when the sperm of one lineage fertilizes the oocyte of another one. Because of these occasional introgressions, the genetic relationships between Corbicula species remained unclear, and the biogeographic origins of the invasive androgenetic lineages have been challenging to identify. To address these issues, we analyzed the patterns of allele sharing for several nuclear and mitochondrial molecular markers among Corbicula individuals collected across both the native and invasive range. Our results show the occurrence of an allelic pool encompassing all Corbicula freshwater species worldwide, including sexual and androgenetic ones, which highlights the substantial genetic mixing within this genus. However, the differences in allele sharing patterns between invasive lineages, and the low diversity within each lineage, suggest recent, distinct biogeographic origins of invasive Corbicula androgenetic lineages. Finally, the polyploidy, high heterozygosity, and hybrid phenotypes and genotypes found in our study probably originated from hybridization events following egg parasitism between distinct Corbicula lineages. This extensive cross-lineage mixing found in Corbicula may generate nuclear diversity in an otherwise asexually reproducing species.

The genus Mentha is taxonomically and phylogenetically challenging due to complex genomes, polyploidization and an extensive historical nomenclature, potentially hiding cryptic taxa. A straightforward interpretation of phylogenetic relationships within the section Mentha is further hindered by dominant but outdated concepts on historically identified hybrid taxa. Mentha spicata is traditionally considered to be of hybrid origin, but the evidence for this is weak. Here, we aim to understand the phylogenetic relationships within the section Mentha using large sample sizes and to revisit the hybrid status and identity of M. spicata. We show that two of three traditional species in the subsection Spicatae are polyphyletic, as is the subsection as a whole, while the real number of cryptic species was underestimated. Compared to previous studies we present a fundamentally different phylogeny, with a basal split between M. spicata s.s. and M. longifolia s.s. Cluster analyses of morphological and genotypic data demonstrate that there is a dissociation between morphologically and genotypically defined groups of samples. We did not find any evidence that M. spicata is of hybrid origin, and we conclude its taxonomic status should be revised. The combination of genetic and phenotypic information is essential when evaluating hyperdiverse taxonomic groups.

18Corbicula clams were introduced during the 20 th century into America and Europe, where they became 19 notoriously successful invaders with a widespread, global distribution. Their ability to reproduce clonally through 20 androgenesis ("all-male asexuality") has been determinant in their invasive success, with only four invasive clonal 21 lineages detected across Europe and America, one of which is very abundant and widespread on both continents. 22Due to their "all-male asexuality" and egg parasitism between distinct lineages, the evolutionary and geographic 23 origins of the invasive androgenetic lineages have been challenging to identify. We analyzed here the patterns of 24 allele sharing for different molecular markers among Corbicula individuals collected worldwide. We identify three 25 distinct genetic pools containing androgenetic Corbicula lineages. While one sexual Corbicula species forms a 26 distinct fourth genetic pool, the other sexual lineages cluster with the androgenetic ones based on shared alleles. 27One genetic pool contains most androgenetic lineages and sexual C. sandai from Lake Biwa in Japan, pointing to 28 this lake as a likely origin of androgenetic Corbicula lineages. Although three distinct biogeographic origins of 29Corbicula androgenetic lineages have been identified, their recent radiation and cross-lineage genetic mixing 30 hamper classical species delimitation within this clam genus. 31 32

Signorovitch et al.[1] comment that an Oenothera-like meiosis [2] could produce a pattern similar to what we observed in our study of natural isolates of the bdelloid rotifer Adineta vaga, which we attributed to horizontal gene transfers (HGTs) [3]. Indeed, our HGT hypothesis appears at first sight difficult to conciliate with their observation of a congruent pattern of allele sharing at four large loci possibly located on different chromosomes [4]. However, one might imagine conditions under which massive horizontal gene transfer between bdelloid individuals could produce such a pattern, notably if the individuals involved had previously lost most of their heterozygosity because of their exposure to frequent desiccation (which produces DNA double-strand breaks [5]). In the published A. vaga genome the loss of heterozygosity due to large-scale gene conversion events or break-induced replication covers only about 10% of the genome [6], but this percentage may be much higher in environmental isolates that often experience dessication. Besides, if an Oenothera-like mode of meiosis occurs in bdelloids frequently enough to be detected in a single sampling of 29 individuals (as in [4]), one would expect males and meiosis to be observed at least occasionally, and instances of congruent allele sharing across loci should turn up frequently in genetic surveys. This was not the case in [3]: among the 82 A. vaga individuals sequenced for four nuclear markers, no trio of individuals presented congruent patterns of shared sequences at different loci. For these reasons, and in the absence of any direct evidence for an Oenothera-like meiosis in bdelloids, we still consider inter-bdelloid HGTs a more parsimonious explanation for our results.

We thank Wilson et al. (2018) for their thorough re-analysis of our data and for their constructive criticisms that led our groups to exchange many stimulating emails over the last two years. Although we agree that inter-individual contamination can yield patterns suggestive of inter-individual recombination, we are not fully convinced by their criticisms of our 2016 dataset and would like to point here briefly to some inaccuracies and likely errors in their interpretation of our chromatograms (available at https://github.com/jflot/Debortoli2016CurrentBiology).Wilson et al.'s criticism of our results rests on two main arguments: based on their ConTAMPR analysis of the pattern of minor peaks in some of our chromatograms (often barely distinguishable from background noise), they interpret our proposed inter-specific recombination patterns as merely the result of cross-contamination, namely the inadvertent coextraction of several individuals in the same tube; and based on their finding of triple peaks in some other chromatograms they conclude that our proposed intra-specific recombination patterns are also caused by cross-individual contamination.If we inadvertently co-extracted several individuals at once as suggested by Wilson et al., we should observe superposition of several sequences of approximately equal intensities in at least some of our COI chromatograms (the sequencing of which did not involve any whole-genome amplification step), especially given that rotifers are eutelic (hence juveniles and adults contain strictly the same amount of DNA). This was never the case: the minor peaks were always much smaller than the main ones. This appears more compatible with the hypothesis of minute amounts of carry-over or post-PCR contamination than with co-extraction of several individuals. It is well known that there is tiny amounts of DNA 'floating around' in any laboratory (Gruber et al. 2015); background-level contamination may also occur post-PCR, during cycle-sequencing and capillary electrophoresis. Alternatively, contamination by the gut content of the rotifers we sequenced (which might contain traces of ingested DNA from dead rotifers) cannot be ruled out and could explain some of the minor peaks observed. These possible explanations for the minor peaks incriminated by Wilson et al. were not considered in their article.To see whether the minor peaks in the COI chromatograms of our 2016 article were unusually abundant, we compared them with those of another published COI dataset of groundwater amphipods (Flot et al. 2012). These amphipods are several millimetres long so the chances of