Different nitrogen saturation thresholds for above-, below-, and total net primary productivity in a temperate steppe.

Identifying the thresholds for the positive responses of total net primary productivity (NPP) to nitrogen (N) enrichment is an essential prerequisite for predicting the benefits of N deposition on ecosystem carbon sequestration. However, the responses of below-ground NPP (BNPP) to N enrichment are unknown in many ecosystems, which limits our ability to understand the carbon cycling under the scenario of increasing N availability. We examined the changes in above-ground NPP (ANPP), BNPP, and NPP of a temperate meadow steppe across a wide-ranging N addition gradient (0, 2, 5, 10, 20, and 50 g N m-2 year-1 ) during 5 years. Both ANPP and NPP increased nonlinearly with N addition rates. The N saturation threshold for ANPP (TA ) and NPP (TN ) was at the rate of 13.11 and 6.70 g N m-2 year-1 , respectively. BNPP decreased with increasing N addition when N addition rates ˃5 g N m-2 year-1 , resulting in much lower TN than TA . Soil N enrichment played a key role in driving the negative impacts of high N addition rates on BNPP, and consequently on the earlier occurrence of N saturation threshold for NPP. Our results highlight the negative effects of soil N enrichment on NPP in natural grasslands super-saturated with N. Furthermore, by considering ANPP and BNPP simultaneously, our results indicate that previous findings from above-ground might have over-estimated the positive effects of N deposition on primary productivity.

Assessing community assembly controls over community-scale nutrient resorption responses to nitrogen deposition.

Nutrient resorption is a fundamental physiological process in plants, with important ecological controls over numerous ecosystem functions. However, the role of community assembly in driving responses of nutrient resorption to perturbation remains largely unknown. Following the Price equation framework and the Community Assembly and Ecosystem Function framework, we quantified the contribution of species loss, species gain, and shared species to the reduction of community-level nutrient resorption efficiency in response to multi-level nitrogen (N) addition in a temperate steppe, after continuous N addition for seven years. Reductions of both N and phosphorus (P) resorption efficiency (NRE and PRE, respectively) were positively correlated with N addition levels. The dissimilarities in species composition between N-enriched and control communities increased with N addition levels, and N-enriched plots showed substantial species losses and gains. Interestingly, the reduction of community-scale NRE and PRE mostly resulted from N-induced decreases in resorption efficiency for the shared species in the control and N-enriched communities. There were negative correlations between the contributions of species richness effect and species identity effect and between the number and identity of species gained for the changes in both NRE and PRE following N enrichment. By simultaneously considering N-induced changes in species composition and in species-level resorption, our work presents a more complete picture of how different community assembly processes contribute to N-induced changes in community-level resorption.

Functional structure mediates the responses of productivity to addition of three nitrogen compounds in a meadow steppe.

Atmospheric nitrogen (N) deposition is altering grassland productivity and community structure worldwide. Deposited N comes in different forms, which can have different consequences for productivity due to differences in their fertilization and acidification effects. We hypothesize that these effects may be mediated by changes in plant functional traits. We investigated the responses of aboveground primary productivity and community functional composition to addition of three nitrogen compounds (NH4NO3, [NH4]2SO4, and CO[NH2]2) at the rates of 0, 5, 10, 20 g N m-2 yr-1. We used structural equation modeling (SEM) to evaluate how functional structure influences the responses of productivity to the three N compounds. Nitrogen addition increased community-level leaf chlorophyll content but decreased leaf dry matter content and phosphorus concentration. These changes were mainly due to intra-specific variation. Functional dispersion of traits was reduced by N addition through changes in species composition. SEM revealed that fertilization effects were more important than soil acidification for the responses of productivity to CO(NH2)2 addition, which enhanced productivity by decreasing functional trait dispersion. In contrast, the effects of (NH4)2SO4 and NH4NO3 were primarily due to soil acidification, influencing productivity via community-weighted means of functional traits. Our results suggest that N forms with different fertilizing and acidifying effects influence productivity via different functional traits pathways. Our study also emphasizes the need for in situ experiments with the relevant N compounds to accurately understand and predict the ecological effects of atmospheric N deposition on ecosystems.

Decoupled responses of above- and below-ground stability of productivity to nitrogen addition at the local and larger spatial scale.

Temporal stability of net primary productivity (NPP) is important for predicting the reliable provisioning of ecosystem services under global changes. Although nitrogen (N) addition is known to affect the temporal stability of aboveground net primary productivity (ANPP), it is unclear how it impacts that of belowground net primary productivity (BNPP) and NPP, and whether such effects are scale dependent. Here, using experimental N addition in a grassland, we found different responses of ANPP and BNPP stability to N addition at the local scale and that these responses propagated to the larger spatial scale. That is, N addition significantly decreased the stability of ANPP but did not affect the stability of BNPP and NPP at the two scales investigated. Additionally, spatial asynchrony of both ANPP and BNPP among communities provided greater stability at the larger scale and was not affected by N addition. Our findings challenge the traditional view that N addition would reduce ecosystem stability based on results from aboveground dynamics, thus highlighting the importance of viewing ecosystem stability from a whole system perspective.

Increasing rates of long-term nitrogen deposition consistently increased litter decomposition in a semi-arid grassland.

The continuing nitrogen (N) deposition observed worldwide alters ecosystem nutrient cycling and ecosystem functioning. Litter decomposition is a key process contributing to these changes, but the numerous mechanisms for altered decomposition remain poorly identified. We assessed these different mechanisms with a decomposition experiment using litter from four abundant species (Achnatherum sibiricum, Agropyron cristatum, Leymus chinensis and Stipa grandis) and litter mixtures representing treatment-specific community composition in a semi-arid grassland under long-term simulation of six different rates of N deposition. Decomposition increased consistently with increasing rates of N addition in all litter types. Higher soil manganese (Mn) availability, which apparently was a consequence of N addition-induced lower soil pH, was the most important factor for faster decomposition. Soil C : N ratios were lower with N addition that subsequently led to markedly higher bacterial to fungal ratios, which also stimulated litter decomposition. Several factors contributed jointly to higher rates of litter decomposition in response to N deposition. Shifts in plant species composition and litter quality played a minor role compared to N-driven reductions in soil pH and C : N, which increased soil Mn availability and altered microbial community structure. The soil-driven effect on decomposition reported here may have long-lasting impacts on nutrient cycling, soil organic matter dynamics and ecosystem functioning.

Mowing mitigates the negative impacts of N addition on plant species diversity.

Increasing availability of reactive nitrogen (N) threatens plant diversity in diverse ecosystems. While there is mounting evidence for the negative impacts of N deposition on one component of diversity, species richness, we know little about its effects on another one, species evenness. It is suspected that ecosystem management practice that removes nitrogen from the ecosystem, such as hay-harvesting by mowing in grasslands, would mitigate the negative impacts of N deposition on plant diversity. However, empirical evidence is scarce. Here, we reported the main and interactive effects of N deposition and mowing on plant diversity in a temperate meadow steppe with 4-year data from a field experiment within which multi-level N addition rates and multiple N compounds are considered. Across all the types of N compounds, species richness and evenness significantly decreased with the increases of N addition rate, which was mainly caused by the growth of a tall rhizomatous grass, Leymus chinensis. Such negative impacts of N addition were accumulating with time. Mowing significantly reduced the dominance of L. chinensis, and mitigated the negative impacts of N deposition on species evenness. We present robust evidence that N deposition threatened biodiversity by reducing both species richness and evenness, a process which could be alleviated by mowing. Our results highlight the changes of species evenness in driving the negative impacts of N deposition on plant diversity and the role of mowing in mediating such negative impacts of N deposition.

Correction to: Mowing mitigates the negative impacts of N addition on plant species diversity.

Unfortunately, the panels of (f) in Figures 1, 2, and 4.

Methane emissions from the trunks of living trees on upland soils.

Upland forests are traditionally thought to be net sinks for atmospheric methane (CH4 ). In such forests, in situ CH4 fluxes on tree trunks have been neglected relative to soil and canopy fluxes. We measured in situ CH4 fluxes from the trunks of living trees and other surfaces, such as twigs and soils, using a static closed-chamber method, and estimated the CH4 budget in a temperate upland forest in Beijing. We found that the trunks of Populus davidiana emitted large quantities of CH4 during July 2014-July 2015, amounting to mean annual emissions of 85.3 and 103.1 µg m(-2)  h(-1) on a trunk surface area basis on two replicate plots. The emission rates were similar in magnitude to those from tree trunks in wetland forests. The emitted CH4 was derived from the heartwood of trunks. On a plot or ecosystem scale, trunk CH4 emissions were equivalent to c. 30-90% of the amount of CH4 consumed by soils throughout the year, with an annual average of 63%. Our findings suggest that wet heartwoods, regardless of rot or not, occur widely in living trees on various habitats, where CH4 can be produced.

Impacts of leguminous shrub encroachment on neighboring grasses include transfer of fixed nitrogen.

Shrub encroachment induced by global change and human disturbance strongly affects ecosystem structure and function. In this study, we explore the degree to which invading leguminous shrubs affected neighboring grasses, including via the transfer of fixed nitrogen (N). We measured N concentrations and natural abundance (15)N of shoot tissues from three dominant grasses from different plant functional groups across seven distances along a local transect (up to 500 cm) to the leguminous shrub, Caragana microphylla. C. microphylla did transfer fixed N to neighboring grasses, but the amount and distance of N transferred were strongly species-specific. Shoot N concentrations decreased significantly with distance from C. microphylla, for a rhizomatous grass, Leymus chinensis, and a bunchgrass, Achnatherum sibiricum. However, N concentrations of another bunchgrass, Stipa grandis, were higher only directly underneath the shrub canopy. Shoot δ(15)N values of L. chinensis were enriched up to 500 cm from the shrub, but for S. grandis were enriched only below the shrub canopy. In contrast, δ(15)N of A. sibiricum did not change along the 500-cm transect. Our results indicated the rhizomatous grass transferred fixed N over long distances while bunchgrasses did not. The presence of C. microphylla increased the shoot biomass of L. chinensis but decreased that of S. grandis and A. sibiricum. These findings highlight the potential role of nutrient-acquisition strategies of neighboring grasses in moderating the interspecific variation of fixed N transfer from the leguminous shrub. Overall, leguminous shrubs have either positive or negative effects on the neighboring grasses and dramatically affect plant community composition and structure.

ß-diversity in temperate grasslands is driven by stronger environmental filtering of plant species with large genomes.

Elucidating mechanisms underlying community assembly and biodiversity patterns is central to ecology and evolution. Genome size (GS) has long been hypothesized to potentially affect species' capacity to tolerate environmental stress and might therefore help drive community assembly. However, its role in driving ß-diversity (i.e., spatial variability in species composition) remains unclear. We measured GS for 161 plant species and community composition across 52 sites spanning a 3200-km transect in the temperate grasslands of China. By correlating the turnover of species composition with environmental dissimilarity, we found that resource filtering (i.e., environmental dissimilarity that includes precipitation, and soil nitrogen and phosphorus concentrations) affected ß-diversity patterns of large-GS species more than small-GS species. By contrast, geographical distance explained more variation of ß-diversity for small-GS than for large-GS species. In a 10-year experiment manipulating levels of water, nitrogen, and phosphorus, adding resources increased plant biomass in species with large GS, suggesting that large-GS species are more sensitive to the changes in resource availability. These findings highlight the role of GS in driving community assembly and predicting species responses to global change.

Nitrogen and water availability interact to affect leaf stoichiometry in a semi-arid grassland.

The effects of global change factors on the stoichiometric composition of green and senesced plant tissues are critical determinants of ecosystem feedbacks to anthropogenic-driven global change. So far, little is known about species stoichiometric responses to these changes. We conducted a manipulative field experiment with nitrogen (N; 17.5 g m(-2) year(-1)) and water addition (180 mm per growing season) in a temperate steppe of northern China that is potentially highly vulnerable to global change. A unique and important outcome of our study is that water availability modulated plant nutritional and stoichiometric responses to increased N availability. N addition significantly reduced C:N ratios and increased N:P ratios but only under ambient water conditions. Under increased water supply, N addition had no effect on C:N ratios in green and senesced leaves and N:P ratios in senesced leaves, and significantly decreased C:P ratios in both green and senesced leaves and N:P ratios in green leaves. Stoichiometric ratios varied greatly among species. Our results suggest that N and water addition and species identity can affect stoichiometric ratios of both green and senesced tissues through direct and interactive means. Our findings highlight the importance of water availability in modulating stoichiometric responses of plants to potentially increased N availability in semi-arid grasslands.

Nitrogen effects on grassland biomass production and biodiversity are stronger than those of phosphorus.

Human-induced nitrogen (N) and phosphorus (P) enrichment have profound effects on grassland net primary production (NPP) and species richness. However, a comprehensive understanding of the relative contribution of N vs. P addition and their interaction on grassland NPP increase and species loss remains elusive. We compiled data from 80 field manipulative studies and conducted a meta-analysis (2107 observations world-wide) to evaluate the individual and combined effects of N and P addition on grassland NPP and species richness. We found that both N addition and P addition significantly enhanced grassland above-ground NPP (ANPP; 33.2% and 14.2%, respectively), but did not affect total NPP, below-ground NPP (BNPP), and species evenness. Species richness significantly decreased with N addition (11.7%; by decreasing forbs) probably due to strong decreased soil pH, but not with P addition. The combined effects of N and P addition were generally stronger than the individual effects of N or P addition, and we found the synergistic effects on ANPP, and additive effects on total NPP, BNPP, species richness, and evenness within the combinations of N and P addition. In addition, N and P addition effects were strongly affected by moderator variables (e.g. climate and fertilization type, duration and amount of fertilizer addition). These results demonstrate a higher relative contribution of N than P addition to grassland NPP increase and species loss, although the effects varied across climate and fertilization types. The existing data also reveals that more long-term (≥5 years) experimental studies that combine N and P and test multifactor effects in different climate zones (particularly in boreal grasslands) are needed to provide a more solid basis for forecasting grassland community response and C sequestration response to nutrient enrichment at the global scale.

Extreme drought exacerbates plant nitrogen­phosphorus imbalance in nitrogen enriched grassland.

The nitrogen­phosphorus (N-P) imbalance induced by N enrichment has received increasing concerns, because N:P ratios play a critical role in driving many fundamental ecological processes. Given the simultaneous occurrence of different global change drivers, it is important to understand whether and how would such N-induced N-P imbalance would be mediated by other global change factors. We examined the interactive effects of N addition (10 g N m-2 yr-1) and extreme drought (-66 % rainfall during the growing season) on species- and community-level N:P ratios in both green and senesced leaves in a temperate grassland of northern China. Extreme drought did not alter soil available N:P ratio under ambient N conditions, but increased that under N enriched conditions. Further, extreme drought did not alter the community-level N:P in both green and senesced leaves under ambient N conditions but significantly enhanced that under N enriched conditions. The drought-induced species turnover made a significant positive contribution to the changes in the community-level N:P ratio under N enriched conditions, but not under ambient N conditions. Our results suggest that the N-induced ecosystem N-P imbalance would be exacerbated by extreme drought event, the frequency of which is predicted to increase across global drylands. Such N-P imbalance would have consequences on litter decomposition, nutrient cycling, and the structures of above- and below-ground food webs. Our findings highlighted the complexity in predicting ecosystem N-P imbalance given the interactions between different global change drivers.

Diversity and composition of understory vegetation in the tropical seasonal rain forest of Xishuangbanna, SW China.

Tropical forests vegetation and community research have tended to focus on the tree component, and limited attention has been paid to understory vegetation. Species diversity and composition of the understory of tropical seasonal rain forest were inventoried in a 625 m2 area (for sapling layer) and a 100 m2 area (for herb/seedling layer) in three 1 ha plots. We found 3068 individuals belonging to 309 species, 192 genera and 89 families. The most important family as determined by the Family Importance Value (FIV) was Rubiaceae in both sapling and herb/seedling layers. In terms of Importance Value Index (IVI), the shrub Mycetia gracilis (Rubiaceae) was the most important species in the sapling layer and the pteridophyte Selaginella delicatula (Selaginellaceae) was the most ecological significant species in the herb/seedling layer. Much more vascular plant species were registered in the understory than in the tree layer totaled among the three plots. The species diversity did not differ significantly among the tree layer, sapling layer and herb/seedling layer. Given that we still know little about the understory plant community for growth forms other than trees, the results from the present study indicate that more attention should be paid to the understory vegetation during the decision-making process for biodiversity conservation in the tropical forests.

Effects of nitrogen addition on plant-soil micronutrients vary with nitrogen form and mowing management in a meadow steppe.

Nitrogen (N) addition and mowing can significantly influence micronutrient cycling in grassland ecosystems. It remains largely unknown about how different forms of added N affect micronutrient status in plant-soil systems. We examined the effects of different N compounds of (NH4)2SO4, NH4NO3, and urea with and without mowing on micronutrient Fe, Mn, Cu, and Zn in soil-plant systems in a meadow steppe. The results showed that (NH4)2SO4 addition had a stronger negative effect on soil pH compared with NH4NO3 and urea, resulting in higher increases in soil available Fe and Mn herein. Nitrogen addition decreased plant community-level biomass weighted (hereafter referred to as community-level) Fe concentration but increased Mn concentration, with a greater effect under (NH4)2SO4 addition. Community-level Cu concentration increased with (NH4)2SO4 and NH4NO3 addition only under mowing treatment. Mowing synergistically interacted with urea addition to increase community-level Mn and Zn concentrations even with decreased soil organic matter, possibly because of compensatory plant growth and thus higher plant nutrient uptake intensity under mowing treatment. Overall, responses of plant-soil micronutrients to N addition varied with mowing and different N compounds, which were mainly regulated by soil physicochemical properties and plant growth. Different magnitude of micronutrient responses in plants and soils shed light on the necessity to consider the role of various N compounds in biogeochemical models when projecting the effects of N enrichment on grassland ecosystems.

Diversity and aboveground biomass of lianas in the tropical seasonal rain forests of Xishuangbanna, SW China.

Lianas are important components of tropical forests and have significant impacts on the diversity, structure and dynamics of tropical forests. The present study documented the liana flora in a Chinese tropical region. Species richness, abundance, size-class distribution and spatial patterns of lianas were investigated in three 1-ha plots in tropical seasonal rain forests in Xishuangbanna, SW China. All lianas with > or = 2 cm diameter at breast height (dbh) were measured, tagged and identified. A total of 458 liana stems belonging to 95 species (ranging from 38 to 50 species/ha), 59 genera and 32 families were recorded in the three plots. The most well-represented families were Loganiaceae, Annonceae, Papilionaceae, Apocynaceae and Rhamnaceae. Papilionaceae (14 species recorded) was the most important family in the study forests. The population density, basal area and importance value index (IVI) varied greatly across the three plots. Strychnos cathayensis, Byttneria grandifolia and Bousigonia mekongensis were the dominant species in terms of IVI across the three plots. The mean aboveground biomass of lianas (3 396 kg/ha) accounted for 1.4% of the total community above-ground biomass. The abundance, diversity and biomass of lianas in Xishuangbanna tropical seasonal rain forests are lower than those in tropical moist and wet forests, but higher than those in tropical dry forests. This study provides new data on lianas from a geographical region that has been little-studied. Our findings emphasize that other factors beyond the amount and seasonality of precipitation should be included when considering the liana abundance patterns across scales.

[Effects of nitrogen and phosphorus inputs on soil nematode community in a degraded grassland.] / æ°®ç£·è¾“å ¥å¯¹è¿‡åº¦æ”¾ç‰§é€€åŒ–è‰åŽŸåœŸå£¤çº¿è™«ç¾¤è½çš„å½±å“.

Nematodes, occupying multiple trophic levels in the food web, play important roles in energy flow and nutrient cycling. Most of Chinese natural grasslands have been degraded due to long-term unreasonable utilization, such as over-grazing. External nutrient input is an important way to restore the ecological function of degraded grasslands. The main and intertative effects of nitrogen and phosphorus inputs on soil nematode abundance, trophic group composition and community structure were studied in the grasslands in Xilingol League of Inner Mongolia. Totally, 38 genera of nematodes were recorded. Tylencholaimus, Aphelenchoides, Thonus, and Scutylenchus were dominant genera in this degraded grassland. Nitrogen input decreased total abundances of soil nematodes, and that of omnivores-carnivorous nematodes and plant-feeding nematodes. Phosphorus input increased total abundances of soil nematodes, and that of fungal-feeding nematodes, omnivores-carnivorous nematodes, and plant-feeding nematodes. Nitrogen input inhibited the positive effects of phosphorus input on the abundances of total nematodes, omnivores-carnivorous nematodes and plant-feeding nematodes. Nutrient inputs had no effect on nematode diversity, which would be resulted from the stable plant community. Nitrogen input significantly increased nematode maturity index, decreased plant parasitic nematode maturity index (PPI), and greatly alleviated the negative effects of phosphorus input on PPI and Wasilewska index, indicating that nitrogen input could improve soil health condition and the stability of nematodes community. Our results would help improve our understanding of the effects of nutrient inputs on degraded grassland ecosystem from a soil biotic perspective.

Divergent responses to water and nitrogen addition of three perennial bunchgrass species from variously degraded typical steppe in Inner Mongolia.

Water and nitrogen (N) availability to plants are spatially and temporally variable in arid and semi-arid grasslands. We aimed to investigate the eco-physiological responses of three bunchgrass species to water and N addition along a gradient of habitat degradation in the Inner Mongolian typical grasslands. The effects of water and N addition on aboveground and belowground growth and biomass allocation and water- and nitrogen-use efficiency (WUE and NUE) of Stipa grandis, Agropyron cristatum and Cleistogenes squarrosa from non-degraded, moderately-degraded and heavily-degraded grasslands, respectively, were compared. Stipa grandis had higher specific root length and WUE than C. squarrosa, while C. squarrosa had higher NUE than S. grandis in water- and N-limited conditions. Responses of A. cristatum were intermediate between those of S. grandis and C. squarrosa. Water and N addition did not have a significant effect on growth and biomass allocation of S. grandis, but it increased growth and leaf biomass allocation of A. cristatum and growth and stem biomass allocation of C. squarrosa. The three species differ in WUE, NUE, biomass allocation and responses to water and N addition, and these differences are adaptive to their respective habitats. The degraded grasslands can be restored by an increase in water and N availability such as is expected to occur via climatic change, but S. grandis will not benefit from the increases.

Environmental and spatial variables determine the taxonomic but not functional structure patterns of microbial communities in alpine grasslands.

There is considerable debate regarding how the taxonomic diversity of microbial communities relates to the functional diversity across space while similar questions have been explored in macro-organism communities. Here, we investigated the taxonomic and functional diversity patterns of soil microbial communities by coupling the data obtained from marker genes sequencing and functional gene surveys. Meanwhile, we evaluated the relative effects of environment and geographic distance on shaping these patterns in alpine grasslands of northern China. Although the taxonomic diversity and composition of microbial communities varied across sites, we found no consistent changes in the functional structure. Both the environmental factors and geographic distance concurrently affected the taxonomic diversity patterns but they had no effects on the spatial variations in functional genes. The functional alpha diversity was weakly correlated to the taxonomic alpha diversity across sites. Moreover, we found no significant relationship between the taxonomic and functional composition similarity among microbial communities. Together, our results provide evidence that spatial variation in microbial functions could be independent of their variations in taxonomic diversity. Even the drivers of spatial variations in the functional structure could be totally different from those of taxonomic variations such as environmental differences and dispersal limitation. Our findings suggest that spatial variations of microbial function structure within a community would not follow the variations of taxonomic structures due to different drivers between both of them over space.

The relative contributions of intra- and inter-specific variation in driving community stoichiometric responses to nitrogen deposition and mowing in a grassland.

AIMS: The stoichiometric characteristics of plant communities are important controller for several fundamental ecological processes. The effects of environmental changes on community stoichiometric characteristics are driven by intra- and inter-specific variation. However, the relative importance of both pathways has seldom been empirically examined. METHODS: We quantified the relative contribution of intra- and inter-specific variation to the changes of community nitrogen (N) and phosphorus (P) concentrations after seven-year factorial N addition and mowing treatments in a semi-arid grassland of northern China. RESULTS: Nitrogen addition significantly increased community N and P concentrations and N:P ratio. Mowing significantly increased community N concentration and N:P. Intra-specific variation contributed more than inter-specific variation to the total variability of all the nutritional and stoichiometric characteristics, with intra-specific variation accounting for 68%, 70%, and 75% of the total variation in community-level N, P, and N:P, respectively. Negative covariations between the contribution of intra- and inter-specific variation occurred for community N and P concentrations. Further, N addition and mowing interacted to affect the impacts of intra- and inter-specific variation on community N concentration and N:P stoichiometry. CONCLUSIONS: Our results highlight different ways of trait selection for N addition and mowing treatments. Interactions between those two factors make it more difficult to accurately predict the responses of plant-mediated biogeochemical cycles under co-occurrence of environmental changes.