Size-specific mortality patterns in two species of forest floor Collembola

Summary Forest floor populations of Orchesella cincta (L) and Tomocerus minor (Lubbock) (Collembola; Entomobryidae) have been sampled over two generations and from length frequency distributions survivorship curves were constructed for each generation. Size-specific mortality rates were computed from size-specific survivorship data. The results reveal an asymmetrically U-shaped mortality pattern, with disproportionally high juvenile mortality. The pattern changes from one generation to the next, in particular the post-hatching mortality. Evidence from behavioural ecological research on a Collembola-specialized ground beetle is cited in favour of the hypothesis that size-specific mortality patterns are moulded by size-specific predation.     

Hurricane disturbance in a temperate deciduous forest: patch dynamics, tree mortality, and coarse woody detritus

Patch dynamics, tree injury and mortality, and coarse woody detritus were quantified to examine the ecological impacts of Hurricane Fran on an oak-hickory-pine forest near Chapel Hill, NC. Data from long-term vegetation plots (1990–1997) and aerial photographs (1998) indicated that this 1996 storm caused patchy disturbance of intermediate severity (10–50% tree mortality; Woods, J Ecol 92:464–476, 2004). The area in large disturbance patches (>0.1 ha) increased from <1% to approximately 4% of the forested landscape. Of the forty-two 0.1-ha plots that were studied, 23 were damaged by the storm and lost 1–66% of their original live basal area. Although the remaining 19 plots gained basal area (1–15% increase), across all 42 stands basal area decreased by 17% because of storm impacts. Overall mortality of trees >10 cm dbh was 18%. The basal area of standing dead trees after the storm was 0.9 m²/ha, which was not substantially different from the original value of 0.7 m²/ha. In contrast, the volume and mass of fallen dead trees after the storm (129 m³/ha; 55 Mg/ha) were 6.1 and 7.9 times greater than the original levels (21 m³/ha; 7 Mg/ha), respectively. Uprooting was the most frequent type of damage, and it increased with tree size. However, two other forms of injury, severe canopy breakage and toppling by other trees, decreased with increasing tree size. Two dominant oak species of intermediate shade-tolerance suffered the largest losses in basal area (30–41% lost). Before the storm they comprised almost half of the total basal area in a forest of 13% shade-tolerant, 69% intermediate, and 18% shade-intolerant trees. Recovery is expected to differ with respect to vegetation (e.g., species composition and diversity) and ecosystem properties (e.g., biomass, detritus mass, and carbon balance). Vegetation may not revert to its former composition; however, reversion of biomass, detritus mass, and carbon balance to pre-storm conditions is projected to occur within a few decades. For example, the net change in ecosystem carbon balance may initially be negative from losses to decomposition, but it is expected to be positive within a decade after the storm. Repeated intermediate-disturbance events of this nature would likely have cumulative effects, particularly on vegetation properties.     

Environmental and neighbourhood effects on tree fern distributions in a neotropical lowland rain forest

Questions: To what extent are the distributions of tropical rain forest tree ferns (Cyatheaceae) related to environmental variation, and is habitat specialization likely to play a role in their local coexistence? Location: Lowland rain forest at La Selva Biological Station, Costa Rica. Methods: Generalized linear (GLM) and generalized additive (GAM) logistic regression were used to model the incidence of four tree fern species in relation to environmental and neighbourhood variables in 1154 inventory plots regularly distributed across 6 km² of old‐growth forest. Small and large size classes of the two most abundant species were modelled separately to see whether habitat associations change with ontogeny. Results: GLM and GAM model results were similar. All species had significant distributional biases with respect to micro‐habitat. Environmental variables describing soil variation were included in the models most often, followed by topographic and forest structural variables. The distributions of small individuals were more strongly related to environmental variation than those of larger individuals. Significant neighbourhood effects (spatial autocorrelation in intraspecific distributions and non‐random overlaps in the distributions of certain species pairs) were also identified. Overlaps between congeners did not differ from random, but there was a highly significant overlap in the distributions of the two most common species. Conclusions: Our results support the view that habitat specialization is an important determinant of where on the rain forest landscape tree ferns grow, especially for juvenile plants. However, other factors, such as dispersal limitation, may also contribute to their local coexistence.     

Habitat disturbance reduces seed dispersal of a forest interior tree in a fragmented African cloud forest

Habitat fragmentation and disturbance are known to impact animals and plants in different ways, depending on species' characteristics and the type and scale of habitat modification involved. In contrast, direct or indirect ramifications on mutualistic relationships between plants and animals are less clear, possibly because general patterns are confounded by the diffuse nature of many of these interactions. Here, we examine how fragment size and/or severe disturbance of a Kenyan mountain cloud forest affects the frugivore community and seed removal of a large-seeded, bird-dispersed tree of the forest interior, for three consecutive years. Forest deterioration reduced avian visitation and seed removal rates independent of fragment size, consistently so despite strong temporal variation in fruit production over the three-year study. In disturbed forest fragments, seed removal rates were on average 3.5 times lower than in more intact ones. Strong differences in both visitation and seed removal rates were largely attributable to shifts in frugivore assemblages, characterized by loss or reduced abundance of the most effective seed dispersers, most of which were forest specialists. Although some disturbed fragments benefited from visits of non-forest dependent seed dispersers, such 'resilience' was not predictable or reliable in time or space. We conclude that disruption of seed disperser-seed interactions in highly fragmented and disturbed tropical forests may be persistent in time when resiliency is inadequate, possibly posing long-term effects on tree communities.     

Community composition and tree population structure in a sub-tropical broad-leaved forest along a disturbance gradient

Floristic composition, diversity, dominance and distribution pattern of species and tree population structure were studied in three stands of a sub-tropical wet hill broad-leaved forest of Meghalaya, India, along a disturbance gradient. Tree species diversity declined with increasing disturbance. Disturbed stands showed low equitability or high dominance and the undisturbed stand exhibited high equitability or low dominance. Contagious distribution among the tree species increased with increasing intensity of disturbance. Species showing regular distribution were restricted only to the undisturbed stand. Effect of disturbance on tree population structure was analysed using density-diameter curves. In the disturbed stands tree species showed reverse J-shaped and/or negative exponential curves, while those in the undisturbed stand exhibited sigmoid to bimodal mound shaped curves.     

Response of tree biomass and wood litter to disturbance in a Central Amazon forest

We developed an individual-based stochastic-empirical model to simulate the carbon dynamics of live and dead trees in a Central Amazon forest near Manaus, Brazil. The model is based on analyses of extensive field studies carried out on permanent forest inventory plots, and syntheses of published studies. New analyses included: (1) growth suppression of small trees, (2) maximum size (trunk base diameter) for 220 tree species, (3) the relationship between growth rate and wood density, and (4) the growth response of surviving trees to catastrophic mortality (from logging). The model simulates a forest inventory plot, and tracks recruitment, growth, and mortality of live trees, decomposition of dead trees (coarse litter), and how these processes vary with changing environmental conditions. Model predictions were tested against aggregated field data, and also compared with independent measurements including maximum tree age and coarse litter standing stocks. Spatial analyses demonstrated that a plot size of ~10 ha was required to accurately measure wood (live and dead) carbon balance. With the model accurately predicting relevant pools and fluxes, a number of model experiments were performed to predict forest carbon balance response to perturbations including: (1) increased productivity due to CO₂ fertilization, (2) a single semi-catastrophic (10%) mortality event, (3) increased recruitment and mortality (turnover) rates, and (4) the combined effects of increased turnover, increased tree growth rates, and decreased mean wood density of new recruits. Results demonstrated that carbon accumulation over the past few decades observed on tropical forest inventory plots (~0.5 Mg C ha^–1 year^–1) is not likely caused by CO₂ fertilization. A maximum 25% increase in woody tissue productivity with a doubling of atmospheric CO₂ only resulted in an accumulation rate of 0.05 Mg C ha^–1 year^–1 for the period 1980–2020 for a Central Amazon forest, or an order of magnitude less than observed on the inventory plots. In contrast, model parameterization based on extensive data from a logging experiment demonstrated a rapid increase in tree growth following disturbance, which could be misinterpreted as carbon sequestration if changes in coarse litter stocks were not considered. Combined results demonstrated that predictions of changes in forest carbon balance during the twenty-first century are highly dependent on assumptions of tree response to various perturbations, and underscores the importance of a close coupling of model and field investigations.     

Impacts of drought on tree mortality and growth in a mixed hardwood forest

Abstract. The tree and shrub species on a 16-ha watershed in the Coweeta Basin were sampled in 1984 and again in 1991 to determine the effects of drought on tree species composition and basal area growth. Mortality and radial growth were determined for tree species within three community types that represent a moisture gradient from moist to dry: cove-hardwoods > mixed-oak > oak-pine. Tree mortality from 1984 to 1991 was 20% and 23% in the cove-hardwoods and mixed-oak communities, respectively, compared to only 12% in the oak-pine type. With the exception of Oxydendrum arboreum and Robinia pseudoacacia, the oaks had higher percentage mortality than any other genus; Quercus velutina had 29%, 37%, and 20% mortality in the cove-hardwoods, mixed-oak, and oak-pine types, respectively; Quercus prinus had 23% mortality in the mixed-oak type; Quercus coccinea had 36% mortality in the mixed-oak type; and Quercus marilandica had 27% mortality in the oak-pine type. Mortality occurred mostly in the small-size class individuals (< 10 cm in diameter) for all species, suggesting that thinning was still an important process contributing to mortality 29 yr after clearcutting. Although growth of Liriodendron tulipifera was much higher than growth of either Quercus prinus or Quercus coccinea, growth in Liriodendron was significantly reduced by the 1985–88 drought and no growth reduction was observed for these two dominant Quercus species during the same time period.     

Apparent climatically induced increase of tree mortality rates in a temperate forest

We provide a first detailed analysis of long-term, annual-resolution demographic trends in a temperate forest. After tracking the fates of 21 338 trees in a network of old-growth forest plots in the Sierra Nevada of California, we found that mortality rate, but not the recruitment rate, increased significantly over the 22 years of measurement (1983–2004). Mortality rates increased in both of two dominant taxonomic groups ( Abies and Pinus ) and in different forest types (different elevational zones). The increase in overall mortality rate resulted from an increase in tree deaths attributed to stress and biotic causes, and coincided with a temperature-driven increase in an index of drought. Our findings suggest that these forests (and by implication, other water-limited forests) may be sensitive to temperature-driven drought stress, and may be poised for die-back if future climates continue to feature rising temperatures without compensating increases in precipitation.     

Intermediary disturbance increases tree diversity in riverine forest of southern Brazil

Floods are frequently associated with disturbance in structuring riverine forests and they lead to environmental heterogeneity over space and time. We evaluated the distribution of tree species, ecological groups, species richness and diversity from the point bar to the slope of a riverside forest in southern Brazil (Lat. 30°01′S, Long. 52°47′W) to analyze the effects of flooding on soil properties and forest structure. A plot of 50 × 200 m divided in five contiguous transects of 10 × 200 m parallel to the river was installed, where we measured all the individual trees with pbh ≥ 15 cm. A detailed topographical and soil survey was carried out across the plot and indicated significant differences in organic matter and most mineral nutrients through the topographical gradient. The 1,229 surveyed individuals belonged to 72 species and 35 families. We used Partial CCA and Species Indicator Analysis to observe the spatial distribution of species. Both analyses showed that species distribution was strongly related to the flooding gradient, soil properties and also by space and pure spatial structuring of species and environmental variables (spatial autocorrelation), although a large part of variation remains unexplained. The ecological groups of forest stratification, plant dispersal and requirements for germination indicated slight differences among frequently, occasional and non-flooded transects. Species richness and diversity were higher at intermediate elevations and were associated to the increased spatial–temporal environmental heterogeneity. Across the plot, the direct influence of flooding on tree species distribution created a vegetation zonation that is determined by predicted ecological traits.     

树木年龄和断面积对加拿大北方林树木死亡率的影响

以加拿大北部的杨树(Populus spp.)、斑克松(Pinus banksiana)、黑云杉(Picea mariana)为对象,采用长期定位试验,对134块固定样地的活立木及枯死木进行调查,并运用线性回归的方法研究树木年龄、断面积和林分类型对3种树木死亡率的影响.结果表明:随着树龄和断面积的增加,林木的死亡率呈上升趋势.杨树在斑克松林中的死亡率较高,而在黑云杉林中死亡率较低.在黑云杉林中,树龄是影响斑克松死亡率的主要因子;而在杨树林中,断面积是影响斑克松死亡率的重要因子;不同林分类型中树龄对黑云杉死亡率的影响显著.树种组成对树种的死亡率有显著影响;树木年龄、断面积和林分类型之间的交互效应对各树种的死亡率均有显著影响;不同林分类型中同一树种的死亡状况有明显差异.     

基于广义线性混合效应模型的森林树木死亡研究

基于计数模型方法,同时考虑样地的随机效应,构建林分水平死亡模型,探究影响树木死亡的因素,以期为森林资源的监测与管理提供参考依据。以美国德州东部森林连续清查的样地数据为数据源,按4∶1的比例将其进行随机抽样,划分为训练集和验证集数据,将立地因子、林分因子和气候因子作为模型的自变量,林木死亡株数则作为模型的因变量,运用计数模型和混合效应模型方法进行模型的构建,并分析影响林木死亡株数的因子。使用赤池信息准则(AIC)、贝叶斯信息准则(BIC)和-2倍对数似然函数值(-2logL) 3种模型评价指标评估各模型间的拟合效果;采用平均绝对误差(MAE)和均方根误差(RMSE) 2种评价指标评估其预测效果,以便筛选出最佳的林分水平死亡模型。结果表明：立地因子方面,林木死亡株数与海拔(P<0.01)呈显著的负效应,与坡度(P<0.05)呈显著的正效应,说明林木死亡株数随海拔的升高而减少,随坡度的增加而增多;林分因子方面,林木死亡株数与林分年龄(P<0.001)和树木基面积(P<0.001)呈显著的正效应,与林分平方平均胸径(P<0.001)和林分密度(P<0.05)...     

Fruit production is influenced by tree size and size‐asymmetric crowding in a wet tropical forest

In tropical forest communities, seedling recruitment can be limited by the number of fruit produced by adults. Fruit production tends to be highly unequal among trees of the same species, which may be due to environmental factors. We observed fruit production for ~2,000 trees of 17 species across 3 years in a wet tropical forest in Costa Rica. Fruit production was modeled as a function of tree size, nutrient availability, and neighborhood crowding. Following model selection, tree size and neighborhood crowding predicted both the probability of reproduction and the number of fruit produced. Nutrient availability only predicted only the probability of reproduction. In all species, larger trees were more likely to be reproductive and produce more fruit. In addition, number of fruit was strongly negatively related to presence of larger neighboring trees in 13 species; presence of all neighboring trees had a weak‐to‐moderate negative influence on reproductive status in 16 species. Among various metrics of soil nutrient availability, only sum of base cations was positively associated with reproductive status, and for only four species. Synthesis Overall, these results suggest that direct influences on fruit production tend to be mediated through tree size and crowding from neighboring trees, rather than soil nutrients. However, we found variation in the effects of neighbors and nutrients among species; mechanistic studies of allocation to fruit production are needed to explain these differences.     

Drought induces lagged tree mortality in a subalpine forest in the Rocky Mountains

Extreme climatic events are key factors in initiating gradual or sudden changes in forest ecosystems through the promotion of severe, tree-killing disturbances such as fire, blowdown, and widespread insect outbreaks. In contrast to these climatically-incited disturbances, little is known about the more direct effect of drought on tree mortality, especially in high-elevation forests. Therefore projections of drought-induced mortality under future climatic conditions remain uncertain. For a subalpine forest landscape in the Rocky Mountains of northern Colorado (USA), we quantified lag effects of drought on mortality of Engelmann spruce Picea engelmannii , subalpine fir Abies lasiocarpa , and lodgepole pine Pinus contorta . For the period 1910–2004, we related death dates of 164 crossdated dead trees to early-season and late-season droughts. Following early-season droughts, spruce mortality increased over five years and fir mortality increased sharply over 11 years. Following late-season droughts, spruce showed a small increase in mortality within one year, whereas fir showed a consistent period of increased mortality over two years. Pine mortality was not affected by drought. Low pre-drought radial growth rates predisposed spruce and fir to drought-related mortality. Spruce and fir trees that died during a recent drought (2000–2004) had significantly lower pre-drought growth rates than live neighbour trees. Overall, we found large interspecific differences in drought-related mortality with fir showing the strongest effect followed by spruce and pine. This direct influence of climatic variability on differential tree mortality has the potential for driving large-scale changes in subalpine forests of the Rocky Mountains.     

Secondary poisoning of mustelids in a New Zealand Nothofagus forest

Eleven radio-tagged stoats ( Mustela erminea ) and one weasel ( M. nivalis ) died of secondary poisoning following Talon 20 P^TM (20 ppm brodifacoum) poisoning operations which killed mice ( Mus musculus ), ship rats ( Rattus rattus ) and probably brushtail possums ( Trichosurus vulpecula ) in a New Zealand beech ( Nothofagus ) forest. This poisoning method could be an especially useful way of restoring New Zealand native bird populations because it kills several predator species in one operation. Potential unwanted side-effects must be researched before its routine use. This research also demonstrates the potential hazards of second-generation anticoagulant rodenticides to conservation of rodent predators in Europe.     

鼎湖山南亚热带常绿阔叶林不同树种存活对邻体组成的响应差异

理解群落结构和动态的主导机制是生态学研究的基本目标之一。群落内树种的存活受到其邻近树木的显著影响。为探究不同树种的存活对邻体组成的响应差异, 本研究基于鼎湖山南亚热带阔叶林20 ha森林动态监测样地中常见的90个树种的存活监测数据和功能性状数据, 建立了一系列关于邻体效应的树种存活模型。结果表明: 约58%的树种存活对邻体组成有敏感的响应, 共存树种间的功能性状差异影响着50%的树种存活动态。不同树种对邻体组成的响应差异与其耐阴性相关, 耐阴能力较弱的树种更倾向于表现出对邻体的敏感性。低比叶面积、高叶干物质含量、木材密度和最大胸径意味着较强的耐阴能力, 与光资源利用策略有关的生态位分化可能是邻域尺度上物种共存的原因。本研究为量化邻体间的相互作用和解释局域群落的物种共存提供了新的视角。     

The response of forest floor vegetation and tree regeneration to deer exclusion and disturbance in a riparian deciduous forest,central Japan

The response of forest floor vegetation and regeneration of major treespecies to deer exclusion in a riparian forest were studied for 3 years withtheinteractive effects of natural disturbances. At the start of this study, deerdensity had quickly increased to a fairly high level (29–31 individualskm^–2) during the last decade and had severely reduced theamount of forest floor vegetation in the study area. Dwarf bamboos, which weredominant before, had almost diminished from the forest floor. During the studyperiod, aboveground biomass increased steadily but species diversity did notchange much in the exclosures. Outside the exclosures, the seedlings of alltreespecies were damaged greatly by deer browsing, especially the taller ones. Deerbrowsing had little effect on the emergence of tree seedlings, but deertrampling might have accelerated emergence indirectly by disturbing the soilsurface for some species. Differences in plant responses were also observed fordeer browsing and the presence of dwarf bamboo that strongly inhibits therecruitment of tree seedlings. The nine major species were classified intothreegroups according to the response of seedlings to deer browsing and disturbance.Group 1 consists of the species whose seedling survival is affected bybrowsing,but seldom by disturbances (Phellodendron amurense,Kalopanax pictus, Quercus crispulaandMalus toringo). Groups 2 and 3 consist of species adaptedto tree-fall disturbance (Betula spp.) and ripariandisturbance (Alnus hirsuta var.sibirica, Ulmus davidiana var.japonica, Populus maximowiczii andSalix sachalinensis), respectively, and seedling survivalof these two groups is principally affected by light conditions. The effect ofdeer browsing on seedling survival and growth was greater for Group 2 than forGroup 3.     

Dynamic steady state of patch-mosaic tree size structure of a mixed dipterocarp forest regulated by local crowding

A patch age- and tree size-structured simulator was applied to demonstrate the landscape dynamics of a lowland mixed dipterocarp forest, using census data over a 3 year interval from two 1 ha plots in northern West Kalimantan, Indonesia (Western Borneo). Tree growth rate and recruitment rate were estimated as functions of tree size and local crowding. The effect of local crowding was assumed to be one-sided through light competition, where the basal area for all trees larger than a target tree inside the circle of 10 m radius around the target was employed as the index of crowding. Estimated parameters were similar between the two plots. Tree mortality was expressed by descending function of tree size with asymptotic mortality for large trees corresponding to the gap formation rate. One parameter specifying the survival of trees at gap formation, which was required for the landscape-level simulation of a shifting-gap mosaic, was left undetermined from plot census data. Through simulation, this parameter was estimated so as to best fit the observed among-patch variation in terms of local basal area. The overall time course of simulation and tree size structure were not sensitive to this parameter, suggesting that one-sided competition along the vertical forest profile is a stronger determinant of average forest structure than among-patch horizontal heterogeneity in this forest. Simulated dynamic steady state successfully reproduced the observed forest architecture in the gap-dynamic landscape. It took about 400 years for a vacant landscape to be replaced by a steady-state architecture of forest. Sensitivity analysis suggests that steady-state basal area and biomass are most sensitive to changing gap formation rate and intrinsic size growth rate.     

Species diversity,susceptibility to disturbance and tree population dynamics in tropical rain forest

Abstract. In 1964 a census of all trees > 9.7 cm diameter at breast height (DBH) was conducted on 22 plots totalling 13.2 ha in lowland tropical evergreen rain forest on Kolombangara, Solomon Islands. Over the following 30 yr (1964–1994), populations of all individuals > 4.85 cm DBH of the 12 most common tree species and amounts of disturbance have been monitored on a declining number of these plots (in 1994, nine plots totalling 5.4 ha were still being recorded). Between November 1967 and April 1970, Kolombangara was struck by four cyclones, although only two of these caused substantial amounts of damage to the canopy structure. Multivariate analysis has identified six forest types on Kolombangara (Greig-Smith et al. 1967). The species richness and diversity of trees in the 1964 census, turnover rates of the populations monitored over 1964–1975, and the amount of disturbance sustained during a cyclone in 1970, were all positively correlated across five of the forest types. The sixth forest type was a consistent outlier in these analyses and is believed to have been seriously disturbed by humans about a century ago. The floristics, turnover and disturbance data support Connell's intermediate disturbance hypothesis. The most species-rich forest types contained a higher proportion of fast-growing individuals and species that are early-successional and which have low density timber. Properties of these species rendered them more susceptible to damage when struck by the 1970 cyclone. They showed higher turnover rates because disturbance-dependent species are also characterised by higher mortality and recruitment rates. Thus, periodic cyclones appear to favour the maintenance of differences in species diversity and composition between forest types.     

Variation in tree growth,mortality and recruitment among topographic positions in a warm temperate forest

Questions: Do the population dynamics of trees differ among topographic positions and, if so, how does topographic position affect the population dynamics of species that are distributed in a topography‐specific manner? Which is the most important life stage in determining vegetation patterns? Location: Primary and secondary warm temperate evergreen broad‐leaved forest (40 ‐ 280 m a.s.l.) on the western part of Yakushima Island, Japan. Methods: Mortality, recruitment, DBH growth and distribution of stems (= 5 cm DBH) in a 2.62‐ha plot were surveyed in 1992 and 2002 to determine the relationships between population parameters and (1) topography and (2) distribution patterns of 17 common tree species. Results: Common species (n = 17) were classified into three distribution pattern groups: group A, distributed mainly on convex slopes; group B, on concave slopes, and group C, not aggregated with respect to topographic position. Stem mortality, recruitment and DBH growth were greater in group A than in group B within each topographic class. The hierarchy of stem mortality among topographic classes for groups A and B was convex > planar > concave. Stem recruitment density was relatively high on the convex and concave slopes, respectively, for groups A and B. Conclusions The topographical positions of adult trees were not always most suited for adult survival and growth. For group A, the distribution pattern of adults was determined in the juvenile stage, while this was not the case for group B. Studies of juvenile stages are important for understanding the demographic basis of vegetation distribution patterns.     

Maternal testosterone in tree swallow eggs varies with female aggression

Variation in the investment of maternal resources in eggs, such as proteins and lipids, can have a profound influence on the growth and development of young. Maternal resources transferred to eggs also include androgens found in the yolk. In several species of birds the concentration of testosterone in the yolk either increases or decreases with laying order. Yolk testosterone has been shown to have various effects on the young including enhanced growth and dominance as well as reduced survival. Previous work suggested that the concentration of testosterone in the yolk may be influenced by the female's social conditions, specifically the frequency of aggressive interactions. In tree swallows, Tachycineta bicolor, we found that yolk testosterone was correlated with the aggressive interactions of the female before and during egg laying. In contrast to other species, yolk testosterone did not vary with laying order in tree swallows. Thus, patterns of yolk testosterone are more variable than thought previously and may be influenced by the social conditions experienced by the female during laying.