1. Worldwide, increasing numbers of stream restoration projects are being initiated to rehabilitate waterways modified by urbanisation. However, many of these projects have limited success in restoring stream communities. Prompted by this, we investigated previously unrecognised barriers to aquatic insect colonisation in urban streams. 2. To investigate whether the availability of suitable substrata for oviposition limited the longitudinal distribution of caddisflies, large boulders were added to the upstream reaches of one stream. Prior to the addition, more egg masses were observed downstream and this longitudinal pattern persisted subsequently. 3. Malaise trapping revealed that adult caddisfly diversity and abundance was greater downstream than upstream. Furthermore, in a previous study the authors found larval caddisflies reflected the longitudinal distribution of adults. 4. The only obvious potential obstructions between reaches were roads beneath which the stream flowed through culverted crossings. Malaise trapping was used to examine the effect of road culverts and bridges on caddisfly dispersal. Numbers of caddisflies caught declined upstream and about 2.5 · more individuals were taken in traps immediately below than above five culverts. 4. Bridges, which had a more open structure than culverts, had no significant effect on the size of catches made above and below them. 5. Road culverts could act as partial barriers to upstream flight, with consequences for larval recruitment in urban streams. We recommend that urban planners and designers of restoration projects consider possible synergistic effects of poor oviposition habitat and barriers to aquatic insect dispersal, which may be critical for the colonisation of urban headwater streams and for the maintenance of stream insect populations.

Tree-holes provide an important microhabitat that is used for feeding, roosting and breeding by numerous species around the world.Yet despite their ecological importance for many of New Zealand's endangered species, few studies have investigated the abundance or distribution of tree-holes in native forests. We used complementary ground and climbed tree surveys to determine the abundance, distribution and characteristics of tree-holes in undisturbed Nothofagus forest in the Lewis Pass, New Zealand. We found that hole-bearing trees were surprisingly abundant compared with many other studies, including Australian Eucalyptus species and American beech. In fact, we estimated as many as 3906 tree-holes per hectare, of which 963 holes per hectare were potentially large enough to provide roost sites for hole-nesting bats in New Zealand, while only eight holes per hectare were potentially suitable for specialist hole-nesting birds. This was of great interest as primary cavity-excavating animals are absent from New Zealand forests, compared with North America and Australia. Moreover, tree-hole formation in New Zealand is likely to be dominated by abiotic processes, such as branch breakage from windstorms and snow damage. As has been found in many other studies, tree-holes were not uniformly distributed throughout the forest. Tree-holes were significantly more abundant on the least abundant tree species, Nothofagus fusca, than on either N. menziesii or N. solandri. In addition to tree species, tree size was also an important factor influencing the structural characteristics of tree-holes and their abundance in this forest. Moreover, these trends were not fully evident without climbed tree surveys. Our results revealed that ground-based surveys consistently underestimated the number of tree-holes present on Nothofagus trees, and illustrate the importance of using climbed inspections where possible in tree-hole surveys. We compare our results with other studies overseas and discuss how these are linked to the biotic and abiotic processes involved in tree-hole formation. We consider the potential implications of our findings for New Zealand's hole-dwelling fauna and how stand dynamics and past and future forest management practices will influence the structural characteristics of tree-holes and their abundance in remnant forest throughout New Zealand.

Summary1. Species richness is typically positively correlated with ecosystem size, yet there is no general consensus on the proximate mechanisms (resource concentration, disturbance, colonization-extinction dynamics or habitat heterogeneity) driving this relationship. Ecosystem-size effects are often attributed to increasing resource concentration, but the inherent intercorrelation of resource concentration with other potential proximate mechanisms has led to strong debate over its significance as a mediator of ecosystem-size effects. 2. We disentangle the proximate mechanisms underlying ecosystem-size effects on species richness by experimentally reversing resource concentration and enhancing drought disturbance, while holding colonization-extinction dynamics and habitat heterogeneity constant, in field microcosms. 3. Contrary to theory and much empirical evidence, species richness decreased with increasing ecosystem size, due explicitly to experimental manipulation of the resource-concentration gradient. Structural equation modelling revealed that resource concentration was the principal driver of ecosystem-size effects on species richness, while drought disturbance and habitat quality strongly determined the identity and composition of colonizing species. 4. Our results support the logical contention that 'ecosystem size' is not a mechanism of effect in its own right, and that with appropriate experimental manipulation it is possible to tease apart the multiple underlying proximate drivers of ecosystem-size effects on species richness. 5. Our results imply that the universally accepted relationship between ecosystem size and biodiversity can be reversed by nutrient enrichment, an increasingly observed human-induced driver of global environmental change.

No abstract

Okeover Stream, on the University of Canterbury campus in Christchurch, New Zealand, has been the subject of restoration efforts since 1998. Our study focused on quantifying the response of this urban stream to current restoration efforts. Initially, physico-chemical conditions and biological communities at three sites along the Okeover Stream were compared with three physically similar sites on each of nearby Waimairi Stream and Avon River. General physical and chemical parameters were similar in all streams with circum-neutral pH, specific conductivity ranging from 167 to 173 µS/cm, dissolved oxygen ranging from 9.0 to 9.2 mg/litre, low turbidity, and similar hydrological conditions. However, analysis of heavy metals in the sediment showed mean lead (Pb) communities. Taxonomic richness and caddisfly diversity increased downstream, with twice as many taxa at the most downstream site than the uppermost sampling site. In upper reaches, copper (Cu), Pb, and zinc (Zn) concentrations in sediments all exceeded ANZECC ISQG-low trigger values. Despite ongoing restoration efforts in Okeover Stream, sedimentation, the presence of high heavy metal concentrations, intermittent flows in headwaters, and possible barriers to adult recolonisation seem to be having a continuing negative impact on benthic communities, especially in the headwaters.