Small rodent population cycles and plants - after 70 years, where do we go?

Small rodent population cycles characterise northern ecosystems, and the cause of these cycles has been a long-lasting central topic in ecology, with trophic interactions currently considered the most plausible cause. While some researchers have rejected plant-herbivore interactions as a cause of rodent cycles, others have continued to research their potential roles. Here, we present an overview of whether plants can cause rodent population cycles, dividing this idea into four different hypotheses with different pathways of plant impacts and related assumptions. Our systematic review of the existing literature identified 238 studies from 150 publications. This evidence base covered studies from the temperate biome to the tundra, but the studies were scattered across study systems and only a few specific topics were addressed in a replicated manner. Quantitative effects of rodents on vegetation was the best studied topic, and our evidence base suggests such that such effects may be most pronounced in winter. However, the regrowth of vegetation appears to take place too rapidly to maintain low rodent population densities over several years. The lack of studies prevented assessment of time lags in the qualitative responses of vegetation to rodent herbivory. We conclude that the literature is currently insufficient to discard with confidence any of the four potential hypotheses for plant-rodent cycles discussed herein. While new methods allow analyses of plant quality across more herbivore-relevant spatial scales than previously possible, we argue that the best way forward to rejecting any of the rodent-plant hypotheses is testing specific predictions of dietary variation. Indeed, all identified hypotheses make explicit assumptions on how rodent diet taxonomic composition and quality will change across the cycle. Passing this bottleneck could help pinpoint where, when, and how plant-herbivore interactions have - or do not have - plausible effects on rodent population dynamics.

Bank vole alarm pheromone chemistry and effects in the field

Chemical communication plays an important role in mammalian life history decisions. Animals send and receive information based on body odour secretions. Odour cues provide important social information on identity, kinship, sex, group membership or genetic quality. Recent findings show, that rodents alarm their conspecifics with danger-dependent body odours after encountering a predator. In this study, we aim to identify the chemistry of alarm pheromones (AP) in the bank vole, a common boreal rodent. Furthermore, the vole foraging efficiency under perceived fear was measured in a set of field experiments in large outdoor enclosures. During the analysis of bank vole odour by gas chromatography–mass spectrometry, we identified that 1-octanol, 2-octanone, and one unknown compound as the most likely candidates to function as alarm signals. These compounds were independent of the vole’s sex. In a field experiment, voles were foraging less, i.e. they were more afraid in the AP odour foraging trays during the first day, as the odour was fresh, than in the second day. This verified the short lasting effect of volatile APs. Our results clarified the chemistry of alarming body odour compounds in mammals, and enhanced our understanding of the ecological role of AP and chemical communication in mammals.

Population cycles and outbreaks of small rodents: ten essential questions we still need to solve

Most small rodent populations in the world have fascinating population dynamics. In the northern hemisphere, voles and lemmings tend to show population cycles with regular fluctuations in numbers. In the southern hemisphere, small rodents tend to have large amplitude outbreaks with less regular intervals. In the light of vast research and debate over almost a century, we here discuss the driving forces of these different rodent population dynamics. We highlight ten questions directly related to the various characteristics of relevant populations and ecosystems that still need to be answered. This overview is not intended as a complete list of questions but rather focuses on the most important issues that are essential for understanding the generality of small rodent population dynamics.

In utero behavioral imprinting to predation risk in pups of the bank vole

Abstract

In the predator–prey arms race, survival-enhancing adaptive behaviors are essential. Prey can perceive predator presence directly from visual, auditory, or chemical cues. Non-lethal encounters with a predator may trigger prey to produce special body odors, alarm pheromones, informing conspecifics about predation risks. Recent studies suggest that parental exposure to predation risk during reproduction affects offspring behavior cross-generationally. We compared behaviors of bank vole (Myodes glareolus) pups produced by parents exposed to one of three treatments: predator scent from the least weasel (Mustela nivalis nivalis); scent from weasel-exposed voles, i.e., alarm pheromones; or a control treatment without added scents. Parents were treated in semi-natural field enclosures, but pups were born in the lab and assayed in an open-field arena. Before each behavioral test, one of the three scent treatments was spread throughout the test arena. The tests followed a full factorial design (3 parental treatments × 3 area treatments). Regardless of the parents’ treatment, pups exposed to predator odor in the arena moved more. Additionally, pups spend more time in the center of the arena when presented with predator odor or alarm pheromone compared with the control. Pups from predator odor–exposed parents avoided the center of the arena under control conditions, but they spent more time in the center when either predator odor or alarm pheromone was present. Our experiment shows that cross-generational effects are context-sensitive, depending on the perceived risk. Future studies should examine cross-generational behavioral effects in ecologically meaningful environments instead of only neutral ones.

Significance statement

We exposed bank voles to odors signaling predation risk to assess the effects parental predation exposure on the behavior of their offspring. Besides predator odor, we also assessed the role of a conspecific alarm cue as a novel way of spreading the predation risk information. Pup behaviors were assessed in the open-field arena, a standard way of assessing animal behavior in a wide range of contexts. We found that also alarm pheromone increased the time pups spend in the center of the arena similarly to predator odor. While previous studies suggested that offspring would be more fearful, our results indicate that the cross-generational effects are very context-dependent; i.e., they differ significantly depending on which scent cue is presented in the open-field arena. This shows the need for better tools or measurements to translate laboratory results into ecologically meaningful frameworks.

Interactions between rodents and weeds in a lowland rice agro-ecosystem: the need for an integrated approach to management

Rodents and weeds are important pests to rice crops in Southeast Asia. The interaction between these 2 major pests is poorly documented. In temperate cereal systems, seeds of grass weeds can be an important food source for rodents and weed cover along crop margins provides important refuge for rodents. In 2012 and 2013, a replicated study (n = 4) in Bago, Myanmar compared 4 treatments (rodents and weeds; no rodents and weeds; rodents and no weeds; no rodents and no weeds) each of 0.25 ha in transplanted rice. Weeds were managed with hand weeding in the wet season, and hand weeding and herbicides in the dry season. Plastic fences were installed to exclude rodents. We examined the weed cover and relative abundance of weed species, rodent damage, rodent population dynamics and rice yield loss caused by rodents and weeds. The dominant rodent species was Bandicota bengalensis. In the dry season, Cyperus difformis was dominant at the tillering stage and Echinochloa crus‐galli was the dominant weed species at the booting stage. In the wet season E. crus‐galli was a dominant weed throughout the season. Damage by rodents was higher in the dry season. There were larger economic benefits for best weed management and effective rodent control in the dry season (258 US$/ha) than in the wet season (30 US$/ha). Concurrent control of weeds in and around rice fields combined with coordinated community trapping of rodents during the early tillering stage and ripening stage of rice are recommended management options.

The need to implement the landscape of fear within rodent pest management strategies

Current reactive pest management methods have serious drawbacks such as the heavy reliance on chemicals, emerging genetic rodenticide resistance and high secondary exposure risks. Rodent control needs to be based on pest species ecology and ethology to facilitate the development of ecologically based rodent management (EBRM). An important aspect of EBRM is a strong understanding of rodent pest species ecology, behaviour and spatiotemporal factors. Gaining insight into the behaviour of pest species is a key aspect of EBRM. The landscape of fear (LOF) is a mapping of the spatial variation in the foraging cost arising from the risk of predation, and reflects the levels of fear a prey species perceives at different locations within its home range. In practice, the LOF maps habitat use as a result of perceived fear, which shows where bait or traps are most likely to be encountered and used by rodents. Several studies have linked perceived predation risk of foraging animals with quitting-harvest rates or giving-up densities (GUDs). GUDs have been used to reflect foraging behaviour strategies of predator avoidance, but to our knowledge very few papers have directly used GUDs in relation to pest management strategies. An opportunity for rodent control strategies lies in the integration of the LOF of rodents in EBRM methodologies. Rodent management could be more efficient and effective by concentrating on those areas where rodents perceive the least levels of predation risk. © 2017 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Foraging decisions in wild versus domestic Mus musculus: What does life in the lab select for?

What does domestication select for in terms of foraging and anti-predator behaviors? We applied principles of patch use and foraging theory to test foraging strategies and fear responses of three strains of Mus musculus: wild-caught, control laboratory, and genetically modified strains. Foraging choices were quantified using giving-up densities (GUDs) under three foraging scenarios: (1) patches varying in microhabitat (covered versus open), and initial resource density (low versus high); (2) daily variation in auditory cues (aerial predators and control calls); (3) patches with varying seed aggregations. Overall, both domestic strains harvested significantly more food than wild mice. Each strain revealed a significant preference for foraging under cover compared to the open, and predator calls had no detectable effects on foraging. Both domestic strains biased their harvest toward high quality patches; wild mice did not. In terms of exploiting favorable and avoiding unfavorable distributions of seeds within patches, the lab strain performed best, the wild strain worst, and the mutant strain in between. Our study provides support for hypothesis that domestic animals have more energy-efficient foraging strategies than their wild counterparts, but retain residual fear responses. Furthermore, patch-use studies can reveal the aptitudes and priorities of both domestic and wild animals.

Divergent foraging behaviour of a desert rodent, Notomys fuscus, in covered and open microhabitats revealed using giving up densities and video analysis

We used a combination of giving up densities (GUD) and behavioural analysis from video footage to test the response of an Australian desert rodent, Notomys fuscus, to the experimental provision of cover microhabitat in the Strzelecki Desert, Australia. In many ecosystems, cover microhabitats are considered to be safe foraging locations for rodents. The response of bipedal desert rodents to cover microhabitats is less certain, with varied findings, due to their use of open habitats for fast travel. Notomys fuscus returned lower GUDs in cover than in open microhabitats and moved slowly for a greater amount of time under cover than in the open. These results suggest that N. fuscus has a preference for foraging under cover, where predation risk is lower, and under these conditions took longer to assess the cover microhabitat before foraging. This was distinctly different to the ‘get in, get out’ behaviour associated with foraging in open habitats. We advocate for the combined measurement of GUD and behavioural analysis using video footage as a way to improve understanding of rodents’ foraging behaviour.

Habitat manipulation in lowland rice-coconut cropping systems of the Philippines--an effective rodent pest management strategy?

BACKGROUND

Reduction of vegetation height is recommended as a management strategy for controlling rodent pests of rice in South-east Asia, but there are limited field data to assess its effectiveness. The breeding biology of the main pest species of rodent in the Philippines, Rattus tanezumi, suggests that habitat manipulation in irrigated rice-coconut cropping systems may be an effective strategy to limit the quality and availability of their nesting habitat. The authors imposed a replicated manipulation of vegetation cover in adjacent coconut groves during a single rice-cropping season, and added artificial nest sites to facilitate capture and culling of young.

RESULTS

Three trapping sessions in four rice fields (two treatments, two controls) adjacent to coconut groves led to the capture of 176 R. tanezumi, 12 Rattus exulans and seven Chrotomys mindorensis individuals. There was no significant difference in overall abundance between crop stages or between treatments, and there was no treatment effect on damage to tillers or rice yield. Only two R. tanezumi were caught at the artificial nest sites.

CONCLUSION

Habitat manipulation to reduce the quality of R. tanezumi nesting habitat adjacent to rice fields is not effective as a lone rodent management tool in rice-coconut cropping systems.

Mammalian predator-prey interaction in a fragmented landscape: weasels and voles

The relationship between predators and prey is thought to change due to habitat loss and fragmentation, but patterns regarding the direction of the effect are lacking. The common prediction is that specialized predators, often more dependent on a certain habitat type, should be more vulnerable to habitat loss compared to generalist predators, but actual fragmentation effects are unknown. If a predator is small and vulnerable to predation by other larger predators through intra-guild predation, habitat fragmentation will similarly affect both the prey and the small predator. In this case, the predator is predicted to behave similarly to the prey and avoid open and risky areas. We studied a specialist predator's, the least weasel, Mustela nivalis nivalis, spacing behavior and hunting efficiency on bank voles, Myodes glareolus, in an experimentally fragmented habitat. The habitat consisted of either one large habitat patch (non-fragmented) or four small habitat patches (fragmented) with the same total area. The study was replicated in summer and autumn during a year with high avian predation risk for both voles and weasels. As predicted, weasels under radio-surveillance killed more voles in the non-fragmented habitat which also provided cover from avian predators during their prey search. However, this was only during autumn, when the killing rate was also generally high due to cold weather. The movement areas were the same for both sexes and both fragmentation treatments, but weasels of both sexes were more prone to take risks in crossing the open matrix in the fragmented treatment. Our results support the hypothesis that habitat fragmentation may increase the persistence of specialist predator and prey populations if predators are limited in the same habitat as their prey and they share the same risk from avian predation.

Can rodent outbreaks be driven by major climatic events? Evidence from cyclone Nargis in the Ayeyawady Delta, Myanmar

BACKGROUND

Massive rodent population outbreaks occurred in the Ayeyarwady Delta, Myanmar, in July 2009, 15 months after cyclone Nargis. Satellite imagery with high temporal frequency was used to identify the area and planting time of rice at a landscape scale of > 80 000 ha, and household surveys of farmers were conducted to validate the mapping and to quantify losses.

RESULTS

Farmers did not have problems with rodents in 2007-2008; rodents were the principal problem in the 2009 summer and monsoon rice crops. The landscape scale modeling indicated that high rodent densities in 2009 were associated with extended or delayed cropping and harvesting time because of asynchronous planting, and with an increase in the amount of abandoned agricultural land after cyclone Nargis.

CONCLUSION

Asynchronous planting following cyclone Nargis provided abundant high-quality food for an extended period, which in turn led to a lengthened breeding season of rodents. The outbreak of populations 15 months after cyclone Nargis is consistent with the time it would take rodent populations to build from a low base after a major flooding event. To prevent rodent outbreaks effectively, synchronous planting, use of rice varieties with a similar maturation date and good field sanitation are important actions for subsequent rice crops after a major weather event.

Response of small rodents to manipulations of vegetation height in agro-ecosystems

Some small mammal populations require human interference to conserve rare or threatened species or to minimize adverse effects in plant production. Without a thorough understanding about how small rodents behave in their environment and consideration of how they react to management efforts, management will not be optimal. Social behavior, spatial and temporal activity patterns, predator avoidance and other behavioral responses can affect pest rodent management. Some of these behavioral patterns and their causes have been well studied. However, their impact on pest rodent management, especially for novel management approaches, is not always clear. Habitat manipulation occurs necessarily through land use and intentionally to reduce shelter and food availability and to increase predation pressure on rodents. Rodents often respond to decreased vegetation height with reduced movements and increased risk sensitivity in their feeding behavior. This seems to result mainly from an elevated perceived predation risk. Behavioral responses may lessen the efficacy of the management because the desired effects of predators might be mediated. It remains largely unknown to what extent such responses can compensate at the population level for the expected consequences of habitat manipulation and how population size and crop damage are affected. It is advantageous to understand how target and non-target species react to habitat manipulation to maximize the management effects by appropriate techniques, timing and spatial scale without causing unwanted effects at the system level.

Managing crop damage caused by house mice (Mus domesticus) in Australia

A large-scale outbreak of the house mouse populations occurs in grain growing in Australia on average once every four years. High densities of mice cause major yield losses to cereal crops, and low to moderate densities of mice also cause some losses. Several predictive models based on rainfall patterns have been developed to forecast mouse density. These models carry some uncertainty and the economic value of basing management actions on these models is not clear. Baiting is the most commonly used method and zinc phosphide and other rodenticide bait are effective in reducing up to 90% of mouse populations. Ecologically-based best farming practice for controlling mice has recently been developed on the basis of long-term field studies of mouse populations. No effective biological control method has been developed for mice. However, grain growers still cannot make economically rational decisions to implement control because they do not know the pest threshold density (D(T)) above which the economic benefits of control exceed the economic costs of control. Applied predator-prey theory suggests that understanding the relationship between mouse density and damage is the basis for determining D(T). Understanding this relationship is the first research priority for managing mouse damage. The other research priority is to develop a reliable method to estimate unbiased mouse density.

Population characteristics of house mice (Mus musculus) on southern Yorke Peninsula, South Australia

Seasonal population characteristics of house mice (Mus musculus), including the effect of season on body mass, were studied at Innes National Park, southern Yorke Peninsula. Mice were caught with Elliott traps, ear-notched, and released. Over 1550 trap-nights (January to December 2006, excluding May), 202 mice were caught. The overall capture success rate was 13.03 mice per 100 trap-nights. The recapture rate was 42.57%. Body mass of adult house mice varied significantly among seasons (P = 0.009). In particular, mouse body mass varied between autumn and winter (P = 0.018), and spring and winter (P = 0.023). The body mass of mice captured in autumn and then recaptured in winter was also significantly different (P = 0.006). This study is the first published for M. musculus population characteristics on Yorke Peninsula and adds to the relatively limited information available on house mouse populations in non-agricultural habitats.

Foraging patterns of voles at heterogeneous avian and uniform mustelid predation risk

Temporal variation of antipredatory behavior and a uniform distribution of predation risk over refuges and foraging sites may create foraging patterns different from those anticipated from risk in heterogenous habitats. We studied the temporal variation in foraging behavior of voles exposed to uniform mustelid predation risk and heterogeneous avian predation risk of different levels induced by vegetation types in eight outdoor enclosures (0.25 ha). We manipulated mustelid predation risk with weasel presence or absence and avian predation risk by reducing or providing local cover at experimental food patches. Foraging at food patches was monitored by collecting giving-up densities at artificial food patches, overall activity was automatically monitored, and mortality of voles was monitored by live-trapping and radiotracking. Voles depleted the food to lower levels in the sheltered patches than in the exposed ones. In enclosures with higher avian predation risk caused by lower vegetation height, trays were depleted to lower levels. Unexpectedly, voles foraged in more trays and depleted trays to lower levels in the presence of weasels than in the absence. Weasels match their prey's body size and locomotive abilities and therefore increase predation risk uniformly over both foraging sites and refuge sites that can both be entered by the predator. This reduces the costs of missing opportunities other than foraging. Voles changed their foraging strategy accordingly by specializing on the experimental food patches with predictable returns and probably reduced their foraging in the matrix of natural food source with unpredictable returns and high risk to encounter the weasel. Moreover, after 1 day of weasel presence, voles shifted their main foraging activities to avoid the diurnal weasel. This behavior facilitated bird predation, probably by nocturnal owls, and more voles were killed by birds than by weasels. Food patch use of voles in weasel enclosures increased with time. Voles had to balance the previously missed feeding opportunities by progressively concentrating on artificial food patches.

Detecting predation of a burrow-nesting seabird by two introduced predators, using stable isotopes, dietary analysis and experimental removals

Burrowing seabirds are vulnerable to extirpation by introduced predators such as rats, but much evidence of predation is circumstantial. On Taukihepa, an island off southern New Zealand, two possible predators exist with sooty shearwaters (Puffinus griseus): the weka (Gallirallus australis), a large rail, and the ship rat (Rattus rattus), both introduced to the island. It was expected that chick predation would be principally by weka, the much larger of the two predators. To measure losses of sooty shearwater chicks to weka or rats, nests were monitored with burrow-scopes at six sites in the summers of 2003–04 and 2004–05. In three of the sites rats were removed on 4-ha grids by trapping. In the other three sites rats were not trapped. In addition, weka were removed from all six sites in 2005. Concurrent diet analysis of weka and rat stomachs was undertaken as well as stable isotopic analysis (δ13C, δ15N) of samples from rats and weka. These were compared with possible prey items including sooty shearwaters. Additional stable isotope samples were taken from Pacific rats (Rattus exulans), a small rat species present with weka and sooty shearwaters on nearby Moginui Island. Weka diet comprised ~40% of bird remains by volume and calculations using Isosource, an isotopic source portioning model, estimated sooty shearwaters contributed 59% (range: 15–71%) of weka diet during the sooty shearwater chick-raising period. Ship rats, in contrast, had very depleted δ13C isotope signatures compared with sooty shearwaters and bird remains contributed <9% of diet by volume, with Isosource calculations suggesting that ship rats consumed more passerine birds (mean: 30%; range 5–51%) than sooty shearwaters (mean 24%; range: 0–44%). In both summers, more chicks were lost on sites from which rats had been removed than on control sites. When weka were removed in 2005, fewer chicks were lost than in 2004 and significantly fewer weka-killed chicks were found on weka-removal sites than on non-removal sites. Weka were the principal predator of sooty shearwater chicks, depredating an estimated 9.9% of nests. Combining several techniques quantified the loss and identified the principal predator of a seabird in decline.

Ecologically based management of rodents in the real world: applied to a mixed agroecosystem in Vietnam

Rodents cause significant damage to lowland irrigated rice crops in the Red River Delta of Vietnam. A four-year study was conducted in 1999-2002 to examine the effectiveness of applying rodent control practices using the principles of ecologically based pest management. Four 100-150 ha study sites adjacent to villages were selected and farmers on two treated sites were asked to follow a set of rodent management practices, while farmers on the untreated sites were asked not to change their typical practices. Farmers on the treated sites were encouraged to use trap-barrier systems (TBS's; 0.065-ha early planted crop surrounded by a plastic fence with multiple capture traps; one TBS for every 10-15 ha), to work together over large areas by destroying burrows in refuge habitats soon after planting (before the rats reestablish in the fields and before the onset of breeding), synchronizing planting and harvesting of the their rice crops, cleaning up weeds and piles of straw, and keeping bund (embankment) size small (<30 cm) to prevent burrowing. A 75% reduction in the use of rodenticides and plastic barrier fences (without traps or an early crop) was achieved on treated sites. The abundance of rodents was low after implementation of the management practices across all sites. There was no evidence for an effect of treatment on the abundance of rodents captured each month using live-capture traps, and no difference in damage between treatments or in yields obtained from the rice crops. Therefore, ecologically based rodent management was equally effective as typical practices for rodent management. Farmers on the treated sites spent considerably less money applying rodent control practices, which was reflected in the comparative increase in the partial benefit:cost of applying ecologically based rodent management from 3:1 on treated sites and untreated sites prior to the implementation of treatments to 17:1 on treated sites in the final year of the project.

Short- and long-term demographic changes in house mouse populations after control in dryland farming systems in Australia

In Australia, outbreaks of house mice (Mus domesticus) cause significant damage to agricultural crops. Rodenticides are used to reduce damage to crops, but the demographic consequences of applying rodenticides are poorly understood. Furthermore, it is not known whether the reduction induced by rodenticides would be similar to that of a natural crash in abundance at the end of mouse outbreaks. I compared the demographic responses of populations of mice to broad-scale field application of fast-acting, acute rodenticides (strychnine and zinc phosphide) in three grain-growing regions of Australia on baited and unbaited sites through live-trapping of mouse populations before baiting and up to four months after baiting. The reductions in population density in each region immediately after baiting were <40%, 92% and 98%. There were few consistent changes in demographic responses across the three regions for bodyweight (no change, increased or decreased), proportion of juveniles (increased or decreased), sex ratio (no change or bias towards females), survival (no change or decreased) and relative body condition (no change or increased). The differences in demographic responses appeared to be related to differences in the efficacy of the rodenticide. A natural crash in densities occurred over a 2–4-week period after baiting and induced a >85% decline in population densities across all regions on baited and unbaited sites. The natural crash caused increases and decreases in bodyweights, a reduction in the proportion of juveniles, male bias, poor survival and poor relative body condition. Poor survival was the only demographic parameter that was consistent for baiting and the natural crash. Five of seven demographic responses for mice during the natural crash were similar to those found in the literature for the decline phase of cyclic vole and lemming populations in the Northern Hemisphere. These results raise the question of whether mouse populations should be baited if a natural crash would occur anyway, but the timing of the natural crash is always uncertain and rodenticides are inexpensive.

Skuas at penguin carcass: patch use and state-dependent leaving decisions in a top-predator

Foraging decisions depend not only on simple maximization of energy intake but also on parallel fitness-relevant activities that change the forager's 'state'. We characterized patch use and patch leaving rules of a top-predatory seabird, the Brown Skua (Catharacta antarctica lonnbergi), which during its reproductive period in the Antarctic establishes feeding territories in penguin colonies. In feeding trials, we observed how skuas foraged at penguin carcass patches and analysed patch leaving decisions by incorporating the estimated state of foraging birds and patch availability. Patches were exploited in a characteristic temporal pattern with exponentially decreasing remaining patch sizes (RPSs) and intake rates. Patch size decreased particularly fast in small compared to large patches and exploitation ended at a mean RPS of 47.6% irrespective of initial size. We failed to identify a measure which those birds equalized upon patch departure from raw data. However, when accounting for the birds' state, we ascertained remaining patch size and intake rates to have the lowest variance at departure whereas food amount and feeding time remained variable. Statistical correction for territory size only and combined with state had lower effects, but remaining patch size remained the measure with lowest coefficient of variation. Thus, we could clearly reject a fixed-time or fixed-amount strategy for territorial skuas and rather suggest a state-dependent strategy that equalizes remaining patch size. Thus our results provide evidence that under natural conditions, territorial skuas adjust their foraging decision on actual energy requirements, i.e. offspring number and age.

Can farm-management practices reduce the impact of house mouse populations on crops in an irrigated farming system?

The impacts of a range of farm-management practices on house mouse (Mus domesticus) populations were tested in a large replicated field study in a complex irrigated farming system in southern New South Wales, Australia. An advisory panel, made up of farmers, extension officers, industry representatives and scientists developed a series of best-practice farm-management actions to minimise the impact of mice. Twelve experimental sites were split into six treated sites, where farmers were encouraged to conduct the recommended practices, and six untreated sites, where farmers conducted their normal farming practices. Mouse abundance was generally low to moderate for the 4-year project (5–60% adjusted trap success). We found significant reductions in population abundance of mice on treated sites when densities were moderate, but no differences when densities were low. Biomass of weeds and grasses around the perimeter of crops were significantly lower on treated sites because of applications of herbicide sprays and grazing by sheep. We could not detect any significant difference in mouse damage to crops between treated and untreated sites; however, levels of damage were low (<5%). Yields of winter cereals and rice crops were significantly higher on treated sites by up to 40%. An analysis of benefits and costs of conducting farming practices on treated sites compared with untreated sites showed a 2 : 1 benefit to cost ratio for winter cereals, 9 : 1 for rice and 4 : 1 for soybeans.