Establishing social cooperation: The role of hubs and community structure

Cooperation is essential for successful human societies. Thus, understanding how cooperative and selfish behaviors spread from person to person is a topic of theoretical and practical importance. Previous laboratory experiments provide clear evidence of social contagion in the domain of cooperation, both in fixed networks and in randomly shuffled networks, but leave open the possibility of asymmetries in the spread of cooperative and selfish behaviors. Additionally, many real human interaction structures are dynamic: we often have control over whom we interact with. Dynamic networks may differ importantly in the goals and strategic considerations they promote, and thus the question of how cooperative and selfish behaviors spread in dynamic networks remains open. Here, we address these questions with data from a social dilemma laboratory experiment. We measure the contagion of both cooperative and selfish behavior over time across three different network structures that vary in the extent to which they afford individuals control over their network ties. We find that in relatively fixed networks, both cooperative and selfish behaviors are contagious. In contrast, in more dynamic networks, selfish behavior is contagious, but cooperative behavior is not: subjects are fairly likely to switch to cooperation regardless of the behavior of their neighbors. We hypothesize that this insensitivity to the behavior of neighbors in dynamic networks is the result of subjects’ desire to attract new cooperative partners: even if many of one’s current neighbors are defectors, it may still make sense to switch to cooperation. We further hypothesize that selfishness remains contagious in dynamic networks because of the well-documented willingness of cooperators to retaliate against selfishness, even when doing so is costly. These results shed light on the contagion of cooperative behavior in fixed and fluid networks, and have implications for influence-based interventions aiming at increasing cooperative behavior.

Background: The measurement and identification of changes in the social structure in response to an exceptional event like COVID-19 can facilitate a more informed public response to the pandemic and provide fundamental insights on how collective social processes respond to extreme events.Objective: In this study, we built a generalized framework for applying social media data to understand public behavioral and emotional changes in response to COVID-19.Methods: Utilizing a complete dataset of Sina Weibo posts published by users in Wuhan from December 2019 to March 2020, we constructed a time-varying social network of 3.5 million users. In combination with community detection, text analysis, and sentiment analysis, we comprehensively analyzed the evolution of the social network structure, as well as the behavioral and emotional changes across four main stages of Wuhan's experience with the epidemic.Results: The empirical results indicate that almost all network indicators related to the network's size and the frequency of social interactions increased during the outbreak. The number of unique recipients, average degree, and transitivity increased by 24, 23, and 19% during the severe stage than before the outbreak, respectively. Additionally, the similarity of topics discussed on Weibo increased during the local peak of the epidemic. Most people began discussing the epidemic instead of the more varied cultural topics that dominated early conversations. The number of communities focused on COVID-19 increased by nearly 40 percent of the total number of communities. Finally, we find a statistically significant “rebound effect” by exploring the emotional content of the users' posts through paired sample t-test (P = 0.003).Conclusions: Following the evolution of the network and community structure can explain how collective social processes changed during the pandemic. These results can provide data-driven insights into the development of public attention during extreme events.

Evolutionary game of cooperative behavior among social capitals in PPP projects: A complex network perspective

We consider a spatial generalization of evolutionary game theory in which strategies are distributed over a spatial array of sites. We assume that the strategy corresponding to a given site has local interactions with the strategies sitting on neighbouring sites, and that the strategies change if neighbouring strategies are doing better. After briefly setting the stage with a formal definition of spatial evolutionary game theory, we consider the spatial extension of the Hawk-Dove game, and we show that the results are qualitatively different from those obtained from classical evolutionary game theory. For example, the proportion of Hawks in the population is in general lower in the spatial game than in the classical one. We also consider spatial generalizations of the extensions of the Hawk-Dove game obtained by including strategies such as Retaliator and Bully. Here, too, the results from the spatial game are very different from the classical results. In particular, with space Retaliator is a much more successful strategy than one would expect from classical considerations. This suggests that, in general, spatial structure may facilitate the evolution of strategies such as Retaliator, which do not necessarily prosper classically, and which are reminiscent of the \`nice', \`provokable' and `forgiving' strategies which seem to play a central role in the evolution of cooperation. The results indicate that including spatial structure in evolutionary game theory is a fruitful extension.

BackgroundThe topological structure of the wiring of the mammalian brain cortex plays an important role in shaping the functional dynamics of large-scale neural activity. Due to their central embedding in the network, high degree hub regions and their connections (often referred to as the ‘rich club’) have been hypothesized to facilitate intermodular neural communication and global integration of information by means of synchronization. Here, we examined the theoretical role of anatomical hubs and their wiring in brain dynamics. The Kuramoto model was used to simulate interaction of cortical brain areas by means of coupled phase oscillators—with anatomical connections between regions derived from diffusion weighted imaging and module assignment of brain regions based on empirically determined resting-state data.ResultsOur findings show that synchrony among hub nodes was higher than any module’s intramodular synchrony (p < 10−4, for cortical coupling strengths, λ, in the range 0.02 < λ < 0.05), suggesting that hub nodes lead the functional modules in the process of synchronization. Furthermore, suppressing structural connectivity among hub nodes resulted in an elevated modular state (p < 4.1 × l0−3, 0.015 < λ < 0.04), indicating that hub-to-hub connections are critical in intermodular synchronization. Finally, perturbing the oscillatory behavior of hub nodes prevented functional modules from synchronizing, implying that synchronization of functional modules is dependent on the hub nodes’ behavior.ConclusionOur results converge on anatomical hubs having a leading role in intermodular synchronization and integration in the human brain.Electronic supplementary materialThe online version of this article (doi:10.1186/s12868-015-0193-z) contains supplementary material, which is available to authorized users.

Adaptive foragers and community ecology: linking individuals to communities and ecosystems

In the social and biological networks,each agent experiences a birth-and-death process.These evolving networks may exhibit some unique characteristics.Recently,the birth-and-death networks have gradually caught attention,and thus far,most of these studies on birth-and-death networks have focused on the calculations of the degree distributions and their properties.In this paper,a kind of random birth-and-death network (RBDN) with reducing network size is discussed,in which at each time step,with probability p(0pq=1-p.Unlike the existing literature,this study is to calculate the average degrees of the proposed networks under different network sizes.First,for the reducing RBDN,the steady state equations for each node's degree are given by using the Markov chain method based on stochastic process rule,and then the recursive equations of average degree for different network sizes are obtained according to these steady state equations.Second,by means of the recursive equations,we explore four basic properties of average degrees as follows:1) the average degrees are limited,2) the average degrees are strictly monotonically increasing,3) the average degrees are convergent to 2mq,and 4) the sum of each difference between the average degree and 2mq is a bounded number.Theoretical proofs for these four properties are also provided in this paper.Finally,on the basis of these properties,a generation function approach is employed to obtain the exact solutions of the average degrees for various network sizes.In addition to the theoretical derivations to the average degrees,computer simulation is also used to verify the correctness of exact solutions of the average degrees and their properties.Furthermore,we use numerical simulation to study the relationship between the average degree and node increasing probability p.Our simulation results show as follows:1) with the increasing of p,the convergent speed of the average degree to 2mq is increasing;2) with the increasing of m,the convergent speed of the average degree to 2mq is decreasing.In conclusion,for the proposed RBDN model,the main contributions of this study include 1) providing the recursive equations of the average degrees under different network sizes,2) investigating the basic properties for the average degrees,and 3) obtaining the exact solutions of the average degrees.

The effect of hub nodes on the community structure in scale-free networks

Identifying influencers is an important step towards understanding how information spreads within a network. Social networks follow a power-law degree distribution of nodes, with a few hub nodes and a long tail of peripheral nodes. While there exist consolidated approaches supporting the identification and characterization of hub nodes, research on the analysis of the multi-layered distribution of peripheral nodes is limited. In social media, hub nodes represent social influencers. However, the literature provides evidence of the multi-layered structure of influence networks, emphasizing the distinction between influencers and influence. Information seems to spread following multi-hop paths across nodes in peripheral network layers. This paper proposes a visual approach to the graphical representation and exploration of peripheral layers and clusters by exploiting the theory of k-shell decomposition analysis. The core concept of the proposed approach is to partition the node set of a graph into pre-defined hub and peripheral nodes. Then, a power-law based modified force-directed method is applied to clearly display local multi-layered neighborhood clusters around hub nodes based on a characterization of the content of message that we refer to as content specificity. We put forward three hypotheses that allow the graphical identification of the peripheral nodes that are more likely to be influential and contribute to the spread of information. Hypotheses are tested on a large sample of tweets from the tourism domain.

This paper starts from the observation that social networks follow a power-law degree distribution of nodes, with a few hub nodes and a long tail of peripheral nodes. While there exist consolidated approaches supporting the identification and characterization of hub nodes, research on the analysis of the multi-layered distribution of peripheral nodes is limited. In social media, hub nodes represent social influencers. However, the literature provides evidence of the multi-layered structure of influence networks, emphasizing the distinction between influencers and influence. The latter seems to spread following multi-hop paths across nodes in peripheral network layers. This paper proposes a visual approach to the graphical representation and exploration of peripheral layers and clusters to exploit underlying concept of k-shell decomposition analysis. The core concept of our approach is to partition the node set of a graph into hub and peripheral nodes. Then, a power-law based modified force-directed method is applied to clearly display local multi-layered neighbourhood clusters around hub nodes. Our approach is tested on a large sample of tweets from the tourism domain. Empirical results indicate that peripheral nodes have a greater probability of being retweeted and, thus, play a critical role in determining the influence of content. Our visualization technique helps us highlight peripheral nodes and, thus, seems an interesting tool to the visual analysis of social influence.

Identifying influential users and predicting their network impact on social networks have attracted tremendous interest from both academia and industry. Social networks follow a power-law degree distribution of nodes, with a few hub nodes and a long tail of peripheral nodes. While there exist consolidated approaches supporting the identification and characterization of hub nodes, research on the analysis of the multi-layered distribution of peripheral nodes is limited. In social media, hub nodes represent social influencers. However, the literature provides evidence of the multi-layered structure of influence networks, emphasizing the distinction between influencers and influence. Information seems to spread following multi-hop paths across nodes in peripheral network layers. This paper proposes a visual approach to the graphical representation and exploration of peripheral layers and clusters by exploiting the theory of k-shell decomposition analysis. The core concept of the proposed approach is to partition the node set of a graph into pre-defined hub and peripheral nodes. Then, a power-law based modified force-directed method is applied to clearly display local multi-layered neighborhood clusters around hub nodes based on a characterization of the content of message that we refer to as content specificity. We put forward three hypotheses that allow the graphical identification of the peripheral nodes that are more likely to be influential and contribute to the spread of information. Hypotheses are tested on a large sample of tweets from the tourism domain.

This paper focuses on semantic networks that represent the user opinions expressed by social media users on a given set of topics. These networks are found to follow a power-law degree distribution of nodes, with a few hub nodes and a long tail of peripheral nodes. While there exist consolidated approaches supporting the identification and characterization of hu b nodes, research on the analysis of the multi-layered distribution of periph eral nodes is limited. In social media, hub nodes represent social influencers. However, the literature provides evidence of the multi-layered structure of influence networks, emphasizing the disti nction between influencers and influence. The latter seems to spread followi ng multi-hop paths across nodes in peripheral network layers. This paper proposes a visual approach to the graphical representation of peripheral layers. The core concept of our approach is to partition the node set of a graph into hub and peripheral nodes. Then, a modified force-directed method is applied to c learly display local multi-layered neighborhood clusters around hub nodes. Our approach is tested on a large sample of tweets from tourism domain. Our algorithm is visually compared with state-of-the-art network drawing techniques.

Due to the complex nature of mega construction projects (MCPs), technological innovation risks have significantly increased. Cooperation is widely accepted as a proactive approach to resolving these risks. An in-depth study of technological innovation cooperative behavior (TICB) helps understand the underlying reasons, but studies need to pay more attention to it. This study explored the factors affecting TICB for MCPs and developed a conceptual model based on the Theory of planned behavior (TPB). It established a structural equation model to verify the relationship between influencing factors. An example verified the feasibility of the model. The results show that cooperative attitude, subjective cooperative norm, perceived cooperative behavior control, and cooperative scenarios positively affect cooperative behavior through cooperative intention. Cooperative attitude plays a mediating role between cooperative scenarios and cooperative intention. Perceived cooperative behavior control has no direct effect on cooperative behavior. This study provides a theoretical reference to guide future empirical studies and enriches the knowledge of TICB for MCPs.

Recently, the demands of big data combining with cloud computing have enabled dramatically growing of intelligent data in various fields. Data security issues have been considered one of the most critical aspects for the implementations of big data and cloud computing. Big Data Security and Intelligent Data in Clouds (BDS-IDC) have been immerse in people's daily life and are promised to have a great growth in the future, such as smart data analysis, smart big data, sensitive-oriented information protection, and cloud-computing security. Multiple protection approaches and combined security schemes as well as intercrossed cloud platforms are becoming common requirements for many big data and cloud users. Considering the next generation of big data executions, it is expected to integrate big data with approaches of intelligent data in cloud systems. The main challenges of achieving this aim will go across most emerging technologies from cyber security and high performances including effective security mechanism against old and new threats, complex system architectures, computing resources allocation optimizations, distributed heterogeneous computing, energy-saving issues, and using smart data in cyber physical systems. This special issue presents a number of outstanding papers researching in the latest explorations and studies in Big Data Security and Intelligent Data in Clouds. We encourage audiences who are interested in high-performance pattern recognition to read following work. Li et al. 1 developed algorithms and investigated various classifiers to determine the authenticity of short social network postings, an average of 20.6 words, from Facebook. This work presented and discussed several experiments using a variety of classifiers. The research was one of the first works that focused on authorship authentication in short messages, such as posting on social network sites. Moreover, Yibin Li et al. 2 designed an approach to minimize the necessary calculations and memory space. The authors developed Field-Programmable Gate Array (FPGA)–based parallel architecture to apply Partial Dynamic Reconfiguration (PDR), which enables the remote configurations for the number of processing units. According to the experimental results, the detection precision could reach a miss rate less than 1.97% and a false positive rate of 1%. A reliable broadcast scheme is also a crucial issue in the target field. Xia et al. 3 investigate a novel reliable broadcast scheme that explored the advantages of lost data piggybacking. The scheme allowed all the vehicles to piggyback some received packets cooperatively to help other vehicles to recover the lost packets. The simulation results showed that the proposed cooperative offloading scheme could achieve much higher broadcast reliability and lower propagation delay, in comparison with existing solutions. Furthermore, the implementation of cloud computing is broad in multiple domains. Dou et al. 4 introduce an energy-aware dynamic VM scheduling method for QoS enhancement in clouds over big data to address the above challenge. This method consisted of 2 main VM migration phases to reduce service prices and execution time, which had been evaluated by experiments. In addition, Leng et al. 5 presented a balanced RDF graph partitioning algorithm for storing massive RDF data on cloud. The authors first devise a modularity-based multilevel label propagation algorithm to partition RDF graph roughly and then use a balanced K-mediods clustering algorithm for final k-way partitioning. This special issue also addresses the aspect of combining video with social networks. Cui et al. 6 proposed novel video recommendation algorithm based on the combination of video content and social network. The proposed algorithm consisted of the trust friends computing model and video's quality evaluation model. The trust friends computing method took into account similarity between users, interaction between users, and the active degree of a user. Son et al. 7 introduced a new conditional proxy re-encryption (CPRE) scheme, namely, the CPRE for Mobile Cloud (CPRE-M), which utilizes the back-end cloud to the extreme extent so that the overhead of terminals is drastically reduced. The scheme outsources a significant amount of computation overhead caused by re-encryption key generation, condition value change, and decryption, to the cloud. The proposed scheme allows users to verify the correctness of outsourced computation under refereed delegation of computation (RDoC) model. The simulation results show CPRE-M outperforms its existing alternatives. Cahyani et al. 8 examine the extent to which data acquisition from Windows phone is supported by 3 popular mobile forensics tools. The effect of device settings modification and alternative acquisition processes on the extraction results is also examined. The results show that current mobile forensic tool support for Windows Phone 8 remains limited. In this sense, the logical acquisition support was more complete in comparison to the physical one. Furthermore, enabling flight mode and disabling location services are recommended to limit data alteration during the acquisition process. Existing information security incident handling strategies may not be adequate as cloud data are generally virtualized, geographically distributed, and ephemeral, presenting both technical and jurisdictional challenges. Ab Rahman et al. 9 present an integrated cloud incident handling and forensic-by-design model. Afterwards, the authors validate the model using a set of controlled experiments on a cloud-related incident. Three popular cloud storage applications were deployed, namely, Dropbox, Google Drive, and OneDrive. This study demonstrates the utility of the model for organizational cloud users to undertake incident investigations. To improve the security of the microblog network, Liu et al. 10 proposed a defending scheme against malicious URL spreading in microblog networks with Hub nodes. After a node found a new malicious URL, it edits a warning message about the malicious URL. If the normal node gets malicious message, it sends private message to the Hub node and update its blog article. Malicious URL messages spread rapidly in the networks due to the influence of Hub nodes. So the security of entire microblog networks can be improved against malicious URL without increasing the network load. Experiments show that the scheme can effectively defend against malicious URL in the any scale of microblog networks. Man et al. 11 introduced a new social network access control model using logical authorization language, named as RuleSN, which can be efficiently used in cloud systems. The model provides high performance of authorization expressiveness and flexibility that can effectively describe relations of User to User, User to Resource, Resource to Resource, and attributes of users and resources. The paper elaborates the formal definitions of the RuleSN model, then the model's authorization specification and verification policies are described, and finally, the implementation, application, and expressiveness of the model are discussed. Recommender systems have shown great potential to address information overload problems, namely, to help users find interesting and relevant objects within a huge information space. To achieve more accurate recommendation, Jianxun et al. 12 investigate a recommendation algorithm INBIw, which improves on the original weighted network-based inference (NBIw) by introducing a tunable parameter β to depress the influence of high-degree nodes. To evaluate the recommendation performance of INBIw, ranking position rate and hitting rate are calculated. The results of experiments based on MovieLens data set show that the INBIw outperforms previous methods, including the global ranking method, collaborative filtering, network-based inference, and NBIw. The high-energy–consuming petroleum equipment in oilfield represents a high proportion in the total energy consumption of petroleum industry. Zhao et al. 13 analyzes the energy consumption constitutions in different production processes, focusing on the running state and energy-saving countermeasure of the high-energy–consuming mechanical equipment. The proposed model provides concurrent data processing for shortening the production cycle and the Fuzzy Weighting Subspace Clustering algorithm to implement the equipment comparability, applying the big data on the running status statistics and applying cloud computation analysis on the corresponding energy-saving countermeasure. Wang et al. 14 present a novel virtual network embedding (VNE) algorithm to increase revenue and utilization of substrate network as well as improve acceptance fairness of virtual networks. To this end, they present a virtual topology pretransformation mechanism leveraging reusable technology. Then they model the problem as an integer linear programming problem and solve the VNE problem with a discrete particle swarm optimization–based algorithm. An incentive convergence mechanism is proposed to reduce mapping complexity, which is used to accelerate convergence and to save more bandwidth. The results prove that the proposed method improves existing algorithms in terms of physical resource utilization, acceptance fairness, revenue/cost ratio, and searching efficiency. Li et al. 15 model the collaboration relations in the regional enterprise cluster as a generalized social collaboration network and explore the dynamic growth process of the Facilities Collaboration Network (FCN) for different strategies of facilities selection, including the random selection and the balanced selection. Using performance indexes such as network size, the distribution of node degree and act degree, clustering coefficient, the average shortest distance, and the number of n-cliques, they analyze the characteristics of these strategies for cloud manufacturing. Based on these characteristics, they propose 2 mechanisms for self-optimization, including the dynamic weighing of facilities and the concentrated processing of successive subtasks in the process. Finally, they analyze the mechanisms' effects on the characteristics of FCN and the performance in manufacturing. Anomaly detection plays a crucial part in identifying unforeseen attacks for network and information security. To address this issue, Li et al. 16 propose a Multilayer Anomaly Detection approach, which extracts and combines features from different network layers. To reduce redundancy and noise, an algorithm called RanPF is designed by applying principal components analysis (PCA) into random forest (RF) algorithm. RanPF uses features selected by PCA to decide the height of every tree in RF and provides a method to select which features for tree nodes to use. To obtain high-quality features, they adopt an attribute learning mechanism. Naive Bayes is used to characterize the attribute information. A series of experiments conducted on 2 real-life datasets demonstrate that the approach outperforms the state-of-the-art methods in terms of detection rate and false alarm rate. Tao et al. 17 investigated the cloud resource allocation for data access with noncooperative game based on a Gauss–Seidel iteration method. They proposed a repeated game scheme, which can approximately achieve the near Pareto-optimal flow allocation among the vehicles. Considering the vehicles' irrational behavior, a punishment strategy was designed to prevent the vehicles from behavior deviation. The analysis based on these models lays a theoretical method foundation on cloud resource allocation process. The validity of the modeling and the accuracy of the analysis were verified through extensive simulations. In the past decade, there has been a surge of interest in exploring the properties of scale-free networks. Two interesting properties have attracted much attention: assortativity and community structure. However, these properties have been studied separately in either theoretical models or real-world networks. Jiang et al. 18 show that both properties are highly related in scale-free networks. First, due to the positive degree of assortativity, assortative networks present high embeddedness. This leads to the high density of communities. Second, because of the negative degree of assortativity, disassortative networks exhibit great hubs in communities, which results in the high compactness of communities that nodes can reach each other via a few hops. Finally, as a result of the equal attractiveness of nodes in neutral networks, a big portion of links act as community bridges. Thus, neutral networks display sparse and less compact communities. Li et al. 19 investigate the selfish load balancing problem in mobile distributed crowdsourcing networks. This work leverages the d-choice method based on Ball and Bin theory for effective balancing under limited information and the proportional allocation scheme for selfish load balancing, maintaining good load balancing property among selfish users. The results show that, chance-choice outperforms several existing algorithms, comparing with proportional allocation scheme, it could decrease the load gap between the maximum and the minimum in system by 50% to 80% and reduce the overhead complexity from O(n) to O(1) comparing with the Max-weight Best Response algorithm. This special issue is composed of a collection of articles, which provide recent advances in some fields related to Big Data Security and Intelligent Data in Clouds. More precisely, the manuscripts investigate different topics including forensic in cloud-of-things devices, encryption in mobile cloud environments, defense against malicious URL spreading in microblog networks, social network access control model in cloud computing, dynamic resource sharing algorithms, load balancing for cloud computing, cloud-based audio/video streaming techniques, energy-aware scheduling in clouds, sensor network security in cloud computing, as well as real world applications related to industry, manufacturing, and vanets. We hope that the readers could benefit from the articles presented in this special issue and would contribute to these exciting research areas. The guest editors of this special issue wish to express their sincere gratitude to all of the authors who contributed to this issue. We extend our thanks to the reviewers for their hard work and their feedbacks. Likewise, we wish to express our gratitude to the Editor-in-Chief Geoffrey C. Fox for his valuable help in arranging this special issue and for giving the authors the opportunity to publish their work in Concurrency and Computation: Practice and Experience journal. Lastly, we wish to thank the journal's staff for their assistance.

During 1970s and 1980s, a large area of mountains in the Czech Republic was influenced by long-term industrial air pollution. Among the most degraded areas were the peaks of the Moravskoslezske Beskydy Mts, where vast clearings resulted from emissions and subsequent forest destruction. This study is aimed at determining the degree of deforestation that is necessary to cause changes in structure and species diversity of small mammal communities that were observed previously. Communities of rodents and insectivores were monitored for a minimum of 3 years at two mountain ranges of the Moravskoslezske Beskydy Mts (Czech Republic) by standard mouse snap-traps. The localities (Smrk and Kněhyně) differ by the degree of human disturbance. Clearings on Smrk Mt are very large (> 30 ha) with no remaining original forest growth as a result of intensive air pollution, unlike the same habitat type at Kněhyně Mt, where the clearings are minor (< 3 ha) and contain living solitary trees. Structure and diversity of small mammal communities in clearings were compared with those from original forests and other mountain habitats. Communities of small mammals at clearings in Smrk Mt (with dominatingMicrotus agrestis) are structurally very different from all other habitats, while structure of communities at Kněhyně clearings are very similar to those of original mountain forest (Complete linkage clustering based on Renkonen index). The community of the original mountain spruce forest at Kněhyně had the highest species diversity (according to Shannon-Weaver, Brillouin, and Simpson indices, Shannon evenness, and rarefaction), while species diversity at clearings of Smrk was the lowest. Shannon diversity of community at Kněhyně primeval forest is similar to that of Kněhyně clearings, while at Smrk Mt the forest diversity is higher than that of clearings. The species diversity of mountain forest and clearings at Kněhyně Mt was significantly higher than that in the same habitats at Smrk Mt. Our results obtained in disturbed habitats at Kněhyně and Smrk Mts suggest that the degree of deforestation may influence the presence and/or the degree of community changes. If the forest destruction is relatively small (clearings < 3 ha), the structure and diversity of small mammal communities do not differ from those of original forest.