2Natural ecosystems typically consist of many small and few large organisms 1-4 . The 1 scaling of this negative relationship between body mass and abundance has important 2 implications for resource partitioning and energy usage [5][6][7] . Global warming over the 3 next century is predicted to favour smaller organisms [8][9][10][11][12] , producing steeper mass-4 abundance scaling 13 and a less efficient transfer of biomass through the food web 5 . 5Here, we show that the opposite effect occurs in a natural warming experiment 6 involving 13 whole-stream ecosystems within the same catchment, which span a 7 temperature gradient of 5-25 C. We introduce a mechanistic model that shows how the 8 temperature dependence of basal resource carrying capacity can account for these 9 previously unexpected results. If nutrient supply increases with temperature to offset 10 the rising metabolic demand of primary producers, there will be sufficient resources to 11 sustain larger consumers at higher trophic levels. These new data and the model that 12 explains them highlight important exceptions to some commonly assumed "rules" about 13 responses to warming in natural ecosystems. 14 Body mass (M) is a key determinant of many ecological phenomena 6,7,14 (e.g. growth, 15 metabolism, feeding) and its relationship with abundance (N) at either the individual or 16 species level is well described by a simple power law, b NM  (hereafter "MN-scaling"). 17The exponent b and its controlling factors have generated considerable interest in community 18 ecology for decades 4,6 , with widespread recognition that b is related to energy flow through 19 food webs [5][6][7] . Many studies have found that MN-scaling is conserved in the face of 20 biodiversity loss or species turnover and so may be a relatively stable property of 21 ecosystems 1-3 . Thus, a change in MN-scaling may highlight a fundamental disruption to the 22 processes that govern energy flow through an ecosystem by environmental or anthropogenic 23 stressors. For example, steepening of size-spectra (i.e. a more negative exponent b) following 24 fisheries exploitation is indicative of widespread losses at higher trophic levels 5,15 . warming favoured smaller phytoplankton and led to steeper size-spectra 13 . 36We tested the generality of this predicted temperature effect on MN-scaling across 13 37Icelandic streams that span a natural temperature gradient of 5-25 °C (Fig. 1a), but are 38 otherwise very similar in their physical and chemical properties [20][21][22][23][24] . Natural experiments and 39 space-for-time substitutions have some limitations (e.g. non-random allocation of 40 temperature "treatments", no observation of the warming process but rather its end point; see 41Supplementary Methods for discussion of these limitations), however, the streams occur in 42 the same catchment and thus are free of the usual confounding effects of biogeographical 43 differences or other environmental gradients 23,25 . The constituent species are a subset of those 44 commonly fou...

Climate warming has been linked to an apparent general decrease in body sizes of ectotherms, both across and within taxa, especially in aquatic systems. Smaller body size in warmer geographical regions has also been widely observed. Since body size is a fundamental determinant of many biological attributes, climate-warming-related changes in size could ripple across multiple levels of ecological organization. Some recent studies have questioned the ubiquity of temperature–size rules, however, and certain widespread and abundant taxa, such as diatoms, may be important exceptions. We tested the hypothesis that diatoms are smaller at warmer temperatures using a system of geothermally heated streams. There was no consistent relationship between size and temperature at either the population or community level. These field data provide important counterexamples to both James’ and Bergmann's temperature–size rules, respectively, undermining the widely held assumption that warming favours the small. This study provides compelling new evidence that diatoms are an important exception to temperature–size rules for three reasons: (i) we use many more species than prior work; (ii) we examine both community and species levels of organization simultaneously; (iii) we work in a natural system with a wide temperature gradient but minimal variation in other factors, to achieve robust tests of hypotheses without relying on laboratory setups, which have limited realism. In addition, we show that interspecific effects were a bigger contributor to whole-community size differences, and are probably more ecologically important than more commonly studied intraspecific effects. These findings highlight the need for multispecies approaches in future studies of climate warming and body size.

Riparian zones are complex, dynamic habitats that play a critical role in river ecosystem functioning. Terrestrial invertebrates comprise much of the diversity found in riparian habitats and facilitate the transfer of energy between aquatic and terrestrial systems. However, the consequences for terrestrial invertebrates of invasion of riparian zones by invasive non-native plants (INNP) remain poorly understood. Responses of terrestrial macroinvertebrate morphospecies to invasion by two common INNP, Fallopia japonica (Japanese knotweed) and Impatiens glandulifera (Himalayan balsam) were assessed, relative to local environmental factors. Terrestrial invertebrates were collected from 20 sites on low order streams in June and August alongside data on physical attributes and land use. Greater cover of F. japonica and I. glandulifera cover reduced total invertebrate abundance and morphospecies diversity at the individual sample scale, whilst increasing spatial heterogeneity of invertebrates at the site scale. Impatiens glandulifera reduced morphospecies diversity at the site scale with increasing cover, but this was not observed for F. japonica. INNP affected terrestrial invertebrate morphospecies abundance and diversity, to a greater extent than prevailing environmental conditions. Our findings therefore offer support for managing riparian plant invasions to improve habitat heterogeneity, restore terrestrial invertebrate diversity and repair aquatic-terrestrial linkages.

Invasion of riparian zones by non‐native plants is a global issue and commonly perceived as a challenge for river and fishery managers, but the type and extent of ecological changes induced by such invasions remain poorly understood. Established effects on sediment delivery, allochthonous inputs, and channel shading could potentially alter aquatic macroinvertebrate assemblages, with implications for in‐stream ecological quality. We assessed responses in the diversity, quality, and heterogeneity of stream macroinvertebrate communities to riparian invasion by non‐native plants. Macroinvertebrates were collected from 24 sites on low order streams in central and southern Scotland during spring and autumn. The effect of invasive non‐native plants (INNP) on macroinvertebrates was assessed relative to that of local physical and chemical factors. Invasive non‐native plants cover was associated with stronger effects than other factors on local diversity of macroinvertebrates (33% reduction at the highest INNP cover) but also increased macroinvertebrate abundance across sites. Invaded sites were also associated with lower macroinvertebrate biomonitoring scores. Community composition differed between invaded and uninvaded sites in autumn, but not in spring. However, INNP influence on macroinvertebrate composition was generally secondary to that of physicochemical variables (e.g. channel shade, substrate diversity). We demonstrate that the influence of INNP extends beyond well‐known impacts on plant communities to reductions mainly in stream macroinvertebrate diversity. Combined with the negative impact on pollution‐sensitive macroinvertebrate taxa this raises concerns over the ecological health of streams with heavily invaded riparian zones. Our findings suggest that efforts to improve low order streams by actively managing severe riparian invasions are merited, but the size and uncertainty of the likely ecological gains must also be evaluated against the effort involved.