Food Webs Prelim Arthropod Data

Morphospecies and environmental data for macroarthropod and microarthropod collections from leaf litter in the Liko Na Pilina experiment. In February 2019, we collected single leaf litter samples under canopies of 19 studied plant species to examine macroarthropod populations. This collection was the first of eight collections to be conducted on a quarterly basis and rotated across the four blocks twice. Collections per species were standardized by volume (i.e., one approximately 27 cm x 27 cm bag per species) and distributed spatially as evenly as possible across individuals planted per species in the four plots per block; small collections were made from beneath several individuals per species. We identified 74 macroarthropod morphospecies, ranging between 17 and 35 per plant species. We also performed microarthropod collections as part of a June-July 2017 pilot study focusing on oribatid mites. Litter was collected from 11 tree species. We used four 25-cm2 quadrats placed on the ground at the underneath the plant canopies and found 13 morphospecies of mites. We investigated relationships among the arthropod community (both micro- and macro-arthropods) and the plant functional traits using Redundancy Analyses (RDA) in R (R Development Core, 2018). RDA ordinates invertebrate species using axes that are constrained to be linear combinations of the plant functional traits, in such a way that the relationship between the invertebrate species and the plant functional traits can be clearly seen. We started with the 15 functional traits data that were used to design the experimental treatments (Ostertag et al., 2015) and then ran Variance Inflation Factor (VIF) analyses to look for collinearity. We eliminated all variables with a VIF>10. Functional trait variables included in the RDA were foliar C and P concentrations, foliar δ13C (a measure of integrated water-use efficiency), maximum plant height, seed mass, altitudinal range (a measure of habitat breadth), and shade ratio (a measure of the canopy architecture). While the taxonomic details of the arthropod data sets are still being worked out, RDA shows that the macroarthropod (Figure 5a) and microarthropod (Figure 5b) communities depend on the functional traits of the plant species and thus that primary consumers' functional traits affect the arthropod community. For macroarthropods, C concentration was the only significant variable (P = 0.002), which was tested by the distribution-free Monte Carlo test (999 permutations), in which the distribution of the test statistics under the null hypothesis is generated by random permutations of cases. For microarthropods, only height was significant (P = 0.026) and shade ratio was marginally not significant but potentially important (P = 0.098).本数据集收录Liko Na Pilina实验中枯落物样本的大型节肢动物与小型节肢动物标本的形态种及配套环境数据。2019年2月，我们针对19种研究植物的冠层下方枯落物开展单样本采集，以调查大型节肢动物种群。本次采集为8次季度性采集的首次采样，该系列采集将跨4个样地区块轮换实施，每个区块将完成两轮采样。各植物物种的采集量均按体积标准化（即每个物种对应1个约27cm×27cm的采样袋），并尽可能均匀地分布于每个区块4个样方内各种植株间；每个物种均从多株植株下方分别采集少量样本。本次调查共鉴定出74个大型节肢动物形态种，单种植物对应的形态种数量介于17至35之间。作为2017年6-7月一项以甲螨为核心的预实验的一部分，我们同时开展了小型节肢动物采集工作。本次采样从11种树木的冠层下方收集枯落物，采用4个25cm²的样方置于植物冠层下方地面，共鉴定出13种螨类形态种。我们借助R语言（R Development Core团队，2018）中的冗余分析（Redundancy Analyses, RDA），探究节肢动物群落（含大型与小型节肢动物）与植物功能性状之间的关联。RDA通过将无脊椎动物物种排序轴约束为植物功能性状的线性组合，实现节肢动物物种与植物功能性状间关联的清晰呈现。我们首先采用Ostertag等人（2015）用于设计实验处理的15项功能性状数据集，随后通过方差膨胀因子（Variance Inflation Factor, VIF）分析检测共线性问题，剔除所有VIF值大于10的变量。最终纳入RDA分析的功能性状变量包括：叶片碳（C）与磷（P）浓度、叶片δ¹³C值（表征水分利用效率的综合指标）、植株最大高度、种子质量、海拔分布范围（表征生境广度）以及遮荫率（表征冠层结构）。尽管节肢动物数据集的分类学细节仍有待完善，但RDA分析结果显示，大型节肢动物（图5a）与小型节肢动物（图5b）群落均与植物物种的功能性状存在关联，即初级消费者的功能性状会对节肢动物群落产生影响。对于大型节肢动物而言，叶片碳浓度是唯一具有显著影响的变量（P = 0.002），该结果通过无分布假设的蒙特卡洛检验（Monte Carlo test，999次置换）验证，该检验通过对样本进行随机置换生成零假设下的检验统计量分布。对于小型节肢动物，仅植株高度具有显著影响（P = 0.026），遮荫率虽未达到显著水平（P = 0.098），但仍具有潜在的重要性。