Handbook of protocols for standardized measurement of terrestrial invertebrate functional traits

Summary

1. Trait‐based approaches are increasingly being used to test mechanisms underlying species assemblages and biotic interactions across a wide range of organisms including terrestrial arthropods and to investigate consequences for ecosystem processes. Such an approach relies on the standardized measurement of functional traits that can be applied across taxa and regions. Currently, however, unified methods of trait measurements are lacking for terrestrial arthropods and related macroinvertebrates (terrestrial invertebrates hereafter).
2. Here, we present a comprehensive review and detailed protocol for a set of 29 traits known to be sensitive to global stressors and to affect ecosystem processes and services. We give recommendations how to measure these traits under standardized conditions across various terrestrial invertebrate taxonomic groups.
3. We provide considerations and approaches that apply to almost all traits described, such as the selection of species and individuals needed for the measurements, the importance of intraspecific trait variability, how many populations or communities to sample and over which spatial scales.
4. The approaches outlined here provide a means to improve the reliability and predictive power of functional traits to explain community assembly, species diversity patterns and ecosystem processes and services within and across taxa and trophic levels, allowing comparison of studies and running meta‐analyses across regions and ecosystems.
5. This handbook is a crucial first step towards standardizing trait methodology across the most studied terrestrial invertebrate groups, and the protocols are aimed to balance general applicability and requirements for special cases or particular taxa. Therefore, we envision this handbook as a common platform to which researchers can further provide methodological input for additional special cases.

A lay summary is available for this article.

Introduction

Over the last decade, strong calls have been made to shift the research focus of community ecology from purely species‐based approaches to trait‐based ones (among others Lavorel & Garnier 2002; McGill et al. 2006; Diaz et al. 2007b; Suding et al. 2008; Webb et al. 2010; Chown 2012; Mouillot et al. 2013). Despite early work (e.g. Shelford 1911), this call is driven by an increasing awareness that trait‐based approaches can significantly enhance our mechanistic understanding and predictive capabilities of the processes that play a major role in community ecology. Moving from a taxonomic approach to a functional trait approach reduces context dependency and therefore enables generalization across communities and ecosystems that is needed to address macro‐ecological questions (McGill et al. 2006; Suding et al. 2008; Hortal et al. 2015; Kunstler et al. 2016). For example, traits can help explain the effects of climate change on species distribution and range shift (e.g. Kaustuv, Jablonski & Valentine 2001; Berg et al. 2010; Diamond et al. 2011), environmental gradients and stressors on the distribution of species and community (dis)assembly (e.g. Dias et al. 2013; Astor et al. 2014; Woodcock et al. 2014), as well as the effect of community composition on ecosystem processes and the provision of ecosystem services across ecological scales (Naeem & Wright 2003; Messier, McGill & Lechowicz 2010; Luck et al. 2012; Brittain et al. 2013; Deraison et al. 2015). Trait‐based approaches have recently also been advocated as promising tools also in ecotoxicology and environmental risk assessment of chemical substances (Rubach et al. 2011; Van den Brink et al. 2013).

Recent developments in trait‐based ecology have been led by plant ecologists, as plant traits have become effective predictors of community assembly (Götzenberger et al. 2012; HilleRisLambers et al. 2012) and ecosystem processes (Lavorel 2013), and are now widely used. The prime utilization of plant functional traits is to identify abiotic and biotic mechanisms that determine species composition, ecosystem processes and service delivery (Lavorel & Garnier 2002; Diaz et al. 2007a; Luck et al. 2009; de Bello et al. 2010; Lavorel et al. 2013). Plant ecologists have been able to scale up successfully from individual plant physiological traits to vegetation processes, such as competition and environmental filtering, as well as ecosystem processes such as decomposition, across a wide range of plant communities (Diaz et al. 2004; Cornwell et al. 2008; Kunstler et al. 2016), and link trait variability to global carbon cycle and climate models (Atkin et al. 2015). The early success of the plant‐trait approach has fuelled the discussion about which traits need to be measured and how they should be quantified in a standardized way. The development of large online trait data bases in plant ecology, such as LEDA (Kleyer et al. 2008) and TRY (Kattge et al. 2011), now provides quick access to plant‐trait values, allowing comparisons even between ecosystems and biomes. Despite potential limitations of using these data bases (Cordlandwehr et al. 2013), such success in plant ecology has fostered and increasing interest ecologists to adopt a similar trait‐based approach in other taxonomic groups (e.g. Poff et al. 2006; Vandewalle et al. 2010; Aubin et al. 2013; Pakeman & Stockan 2014; Pey, Laporte & Hedde 2014; Fournier et al. 2015; Schmera et al. 2015). Particularly for terrestrial invertebrates, attempts to develop trait frameworks for specific taxa, for example Fountain‐Jones, Baker & Jordan (2015) for beetles, or to construct trait data bases for snails (Falkner et al. (2001), Bouget, Brustel & Zagatti (2008) for saproxylic beetles, Speight & Castella (2010) for hoverflies, Bertelsmeier et al. (2013) for ants (see also Yates et al. 2014), Homburg et al. (2014) for carabid beetles, and Pey, Laporte & Hedde (2014) for soil invertebrates), as well as new statistical developments (e.g. Brown et al. 2014) have been published.

Invertebrates have crucial roles as consumers of primary producers (e.g. herbivores, fungivores, granivores) and the afterlife products of animals and plants (i.e. detritivores, such as feeding on leaf‐litter, dead wood, dung and carrion), they provide a staple food for higher trophic levels (e.g. for predators, parasites and parasitoids) and are recognized as both facilitators of primary production (i.e. pollinators and detritivores) and as ecosystem engineers (e.g. soil bioturbators; see Gagic et al. 2015 for an overview). Hence, knowledge of invertebrate traits is key to understanding multi‐trophic processes and ecosystem functioning (e.g. Lavorel et al. 2013; Schmitz et al. 2015). Current terrestrial invertebrate trait data bases are often built around a set of basic traits from a mixture of studies and observations, which are obtained without uniform methodology and with little consistency in which traits were chosen for measurements. In addition, functional traits, such as species temperature tolerance and drought resistance, are often missing or inferred from the abiotic conditions at the (micro)habitats where they have been observed and not measured directly on individuals. However, (micro)habitat selection of species and realized niche in general might result from interactions between species rather than physiological and phenological characteristics of single individuals and populations (Colwell & Fuentes 1975; Ellers, Dias & Berg 2010; Araujo et al. 2013; Colas et al. 2014; He & Bertness 2014), but see also Warren, Giladi & Bradford (2010). The use of such inferred traits as predictors of community and ecosystem processes has been strongly discouraged (Violle et al. 2007), advocating for traits to be measured on individual organisms. The arguments above raise the urgent need for reliable and unified methods to measure functional traits that are directly linked to species performance. A coherent, unified and standardized trait approach for various types of terrestrial invertebrates requires consensus on (i) what the basic set of functional traits would be and, particularly, on (ii) how they should be measured. A key element in the advance of plant trait‐based approaches has been the provision of a handbook of standardized functional traits that detail the methods and definitions of key traits world‐wide (Cornelissen et al. 2003), and its recent update with additional traits and measuring techniques (Pérez‐Harguindeguy et al. 2013). Such an effort is therefore required in other key organisms such as terrestrial invertebrates. The present work aims to provide such incentive to trait‐based approaches for this broad and diversified group of species, by describing a set of standardized trait measurements to improve the reliability and general applicability of functional traits.

Overall approach to the handbook

This handbook aims to provide a set of protocols for trait measurements that can be used across a wide range of terrestrial invertebrate species, including the major taxonomic groups of Insecta, Collembola, Aranea, Crustacea, Myriapoda, Gastropoda and Oligochaeta. We selected the terrestrial environment as a circumscribed habitat that differs in key features from aquatic ones – rate of temperature change, threat of desiccation, very different osmoregulatory challenges, much greater temperature variability on average and over the short term. We chose these groups of organisms because they are similar enough in lifestyle to apply our protocols to. The handbook does not include specific methods for measuring traits of nematodes, parasites and (semi‐)aquatic invertebrates, although some of the protocols may be used for these groups too.

We recognize that a wide variety of life forms encompassed by the present handbook make it a challenging undertaking. In general, invertebrate traits, overall, may incorporate greater complexity than plant traits, because animals can respond to environmental changes by movement and behaviour. Therefore, the trait protocols contain recommendations for adjustments to accommodate the biology of particular taxonomic groups, while maintaining comparability and standardization across taxa.

The handbook is meant as a first step to advance the trait‐based approach to groups other than plants and vertebrates and to stimulate discussion about additional traits that should be included in the handbook for terrestrial invertebrates. We foresee that this set of traits might be expanded in the future as the functional approach becomes increasingly used among animal ecologists. Moreover, the trait protocols are designed for standardized measurement of traits to facilitate a widespread use and to allow high‐throughput phenotyping to enable measurements on large numbers of species. For this reason, some of the most advanced technological methods that are currently used by specialized research groups only and for few specific taxonomic groups are not part of the standardized methods, but included as special cases in the protocols. We would like to emphasize that the handbook's main purpose is to maximize comparability of measurements across a wide range of taxa. Below, we first provide an overview of the criteria and concepts used for selecting the set of traits, and subsequently we describe the standard format of the protocols, followed by several general recommendations. The protocols themselves are provided as Appendix S1 (Supporting Information).

Trait selection

We reviewed the literature on ecology of terrestrial invertebrates and selected the 29 traits (see Table 1) for which we found clear evidence that they directly link organism performance with environmental conditions or ecosystem processes. These traits have been then further discussed among a group of specialist scientists working on the ecology, ecophysiology and evolutionary aspects of predominantly terrestrial invertebrate fauna at different trophic levels with the aim to standardize the methods for their unambiguous use in any terrestrial biome and for the majority of its constituents.

Table 1. List of the terrestrial invertebrate traits selected for the handbook and considered to be key in responding to the environment and/or effecting ecosystem processes and services at various scales from local plots, to landscapes and biomes. The protocols themselves are provided as Appendix S1 (Supporting Information)

Trait type trait	Definition	Comment
Morphology
Body size	Size of the body. It includes body length, body width, body mass, and body volume	Environmental conditions affect body size which will influence amount and composition of resources used
Eye morphology	Form of the eye. It includes eye number, eye size, eyesight	Eye morphology can be filtered by environmental conditions which will reflect prey and/or predator recognition
Respiration system	Structures developed to perform gas exchange	Type of respiration mode directly affect drought tolerance and desiccation resistance
Hairiness	Degree of hair coverage. It includes hair length and hair density	Abiotic condition and biotic interactions (pollination) affect hairiness providing fitness and performance
Colour	Body coloration. It includes colour, intensity, contrast	Abiotic condition and biotic interactions (e.g. predation) affect pigmentation providing fitness and performance
Feeding
Feeding guild	Food type, upon which species feed. It informs about ‘who eats what or whom’	Feeding guild is a good surrogate for trophic level and position in the food web. It determines the quality of resources, which influences a species growth, reproduction and survival
Ingestion rate	Quantity of food consumed in a given period	The rate of food ingested by an organism reflects its nutritional and energetic requirements and is related to species responses to food quality
Biting force	Biomechanical force exerted on food items by the tip of the mouth parts, claws or forelegs	Biting force mainly determines the effect on trophic network interactions and thus on ecosystem function
Life history
Ontogeny	Developmental history. It includes type and number of developmental stages	Response to environmental stressors and effects on the ecosystem can change significantly across an organism's life history. Changes in environmental conditions can affect ontogeny and ecosystem processes
Clutch size	Number of eggs or juveniles produced in one reproductive event	Clutch size respond significantly to environmental conditions which affect number of offspring and their impact on the ecosystems
Egg size	Size dimension or mass of an egg	Resistance to environmental and particularly climatic conditions increase with egg size, which indirectly determines impact on the ecosystem via changes in population sizes
Life span	Amount of time an adult individual lives, from emergence from last instar until death	Stressors can heavily affect life span which is reflected in different ecosystem functions
Age at maturity	Age at first reproductive event	Time of first reproductive event can be changed under environmental stress, with consequences for population size and ecosystem processes
Parity	The number of times a female lays eggs or gives birth	The spreading of reproductive events over a lifetime has fitness consequences that are related to the trade‐off between current and future reproduction
Reproduction mode	Mode by which new offspring are produced (sexual or asexual)	Mode of reproduction can be changed under environmental stress, with consequences for population sizes and ecosystem processes
Voltinism	The number of generations an organism completes in a single year.	Voltinism is under genetic and environmental control, being mostly influenced by the photoperiod, the local climatic conditions.
Physiology
Resting metabolic rate	Amount of energy expended by an organism at rest	Metabolic rate is related to several organism features such as behaviour, longevity and reproduction output and its reaction norm with temperature can indicate how organisms differ in their response to environmental changes
Relative growth rate	Increase in mass of an organism per unit of time	Relative growth rate is related to other several life‐history traits, such as body size and age at maturity. Therefore, growth rate can influence different fitness components such as fecundity and survival
Desiccation resistance	Ability to withstand dry conditions	Physiological capacity to resist dry conditions is related to species distribution along water availability gradients and to species response to changes in water availability
Inundation resistance	Ability of terrestrial organisms to survive under water	Flooding and increased frequency and intensity of extreme precipitation can impose strong restrictions on survival
Salinity resistance	Ability to withstand conditions of high salinity	Ability to withstand conditions of high salinity determines species survival under high salt stress and will influence growth and reproduction via trade‐offs
Temperature tolerance	Ability to survive at any temperature. It includes hot and cold	Toleration of hot and cold temperatures determines species survival under stress and will influence growth and reproduction via trade‐offs
pH resistance	Ability to withstand acidic or alkaline conditions	Ability to withstand acidic or alkaline conditions determines species survival under acidity stress and will influence growth and reproduction via trade‐offs
Behaviour
Activity time	Activity period of a species within 24 h	Environmental conditions, for example climatic conditions, determine the activity time. This can affect ecosystem function through asynchrony, for example spatiotemporal mismatch in biotic interactions
Aggregation	Clustering of individuals	Clustering of individual reduces microclimatic stress, especially overcoming cold and drought, and can locally result in enhanced ecosystem process rates via high population sizes
Dispersal mode	The form of self‐directed movements an animal uses to move from one place to another	Dispersal mode influences access to new habitat, resources and suitable environments, mates and shelters, and opportunities to escape adverse environmental conditions
Locomotion speed	The pace of self‐propelled movement of an organism	Habitat conditions and biotic interactions influence locomotion speed, which reflect behaviours critical for survival, including efficient use of resources, foraging, predator avoidance, fitness and survival
Sociality	Degree of interactive behaviour with other members of its species to the point of having a recognizable and distinct society	Disturbance and land use changes are expected to affect sociality. High levels of sociality are expected to have a bigger impact on ecosystem function
Annual activity time	Period in an organism's life cycle when growth, development and physical activity are temporarily stopped	Offers the possibility to overcome unfavourable environmental conditions in a resting stage

Overall, the selected set of traits largely covers the primary functions related to species performance, assembly processes and interactions between trophic levels at various spatial scales from plots to landscapes and even biomes. For this first step in generalizing traits across taxonomic groups, we excluded traits that are specific to single groups (e.g. pollen transport mode in bees, web construction strategy in spiders, or chemical and physical defences in ants or some caterpillars) and cannot be standardized across taxa. Selected traits can be considered either response traits (i.e. determining the response of the species to an environmental change or to an interaction with another organism from the same or different trophic level) or effect traits (i.e. contributing to the effect of the species on an ecosystem function or the interaction with the another trophic level) or both (Lavorel & Garnier 2002; Naeem & Wright 2003; Suding & Goldstein 2008; Lavorel et al. 2013). We focus on several traits which, based on the existing literature, are among the most widely used or are in urgent need of standardized measurement protocols that can be applied across taxa. From the user perspective, trait selection is often one of the crucial aspects in trait‐based approaches and it has to be based clearly bearing the research question being asked (Rosado, Dias & de Mattos 2013; Shipley et al. 2016). We do refer to the known functionality of traits considered in our protocols.

Most of the selected traits are quantitative and directly measurable on an individual under standardized conditions; others are categorical (e.g. activity time and feeding guild) or ordinal (e.g. ontogeny and respiration system). Broadly, the selected traits can be grouped into five categories, i.e. morphology, feeding, life history, physiology and behaviour. Morphological traits such as eye morphology, body pigmentation or body size are important features of an organism's interaction with the abiotic and biotic environment. For example, body size across different taxonomic groups is a predictor of multiple ecological processes, such as decomposition and mineralization by soil macro‐detritivores, pollination by bees or water regulation by earthworms (de Bello et al. 2010), and strongly correlated with an individual's metabolic rate (Chown et al. 2007). Body size also scales with many other life‐history traits (Ellers & Jervis 2003) and determines the structure and function of ecological networks (Peters 1983; Brown et al. 2004; Woodward et al. 2005). Feeding traits are related to the trophic position of a species and describe aspects of the morphology and behaviour associated with their diet. Feeding‐related traits can therefore be important for understanding niche partitioning, trophic interactions and the way the structure of ecological networks is shaped (Stang et al. 2009; Ibanez 2012; Ibanez et al. 2013).

Life‐history traits describe the age schedule of reproduction of an organism, including key reproductive aspects such as age at maturity, clutch size, voltinism and life span (Stearns 1992). These traits have strong links to fitness and are expected to be among the most sensitive to environmental stress, making them useful to assess the vulnerability of species to global change. For instance, egg size varies enormously between species (Fox & Czesak 2000) and affects hatching success (Fischer et al. 2006) and resistance to desiccation (Fischer et al. 2006) and heat (Liefting et al. 2010). Moreover, trade‐offs exist between reproductive traits and dispersal (Guerra 2011), leading to a reduced reproductive investment in some insects with strong range expansion under the influence of global warming (Hughes, Hill & Dytham 2003).

Physiological traits refer to features that allow species to tolerate variations in abiotic conditions (resistance adaptations), as well as biochemical modifications that adjust the rate of metabolic function (capacity adaptations) in response to environmental changes (Cossins & Bowler 1987; Somero 1992). Physiological tolerance traits, such as heat tolerance and desiccation resistance, have been successfully applied in predicting species distribution patterns along temperature and humidity gradients (Dias et al. 2013), while growth rate can determine an individuals’ susceptibility to predation (Denno et al. 2002; Coley, Bateman & Kursar 2006) and temperature fluctuations (Fordyce & Shapiro 2003). Further, physiological tolerances can be affected by changes in diet (Verdu et al. 2010).

Finally, Behavioural traits enable flexible, rapid responses to environmental change without any associated changes to physiological or morphological phenotypes. Traits such as activity time, aggregation and locomotion enable organisms to seek out preferred microhabitats and to avoid (a)biotic stress. Behavioural strategies can also increase tolerance to abiotic stresses, for instance through adopting flight strategies that maximize heat dissipation (Verdu, Alba‐Tercedor & Jimenez‐Manrique 2012) or by choosing specific microhabitats to achieve nutritional homeostasis (Clissold, Coggan & Simpson 2013) or escape adverse climatic conditions. Yet in soil fauna species, stratification in soil interacts with other traits, such as physiological traits, thus modifying the individual response to changes in environmental conditions (Cloudsley‐Thompson 1962) and vulnerability to extreme temperature events (van Dooremalen et al. 2012).

The handbook protocols

The trait protocols are described using a standard format aimed to facilitate comparisons among traits. The protocols are provided as Appendix S1 to this study. Each protocol includes four main sections. The section Definition and relevance provides a formal definition and a short, non‐exhaustive justification why that particular trait is of ecological significance based on its role in responding to stressors and/or effecting trophic interactions or ecosystem processes. This section also describes the main approaches to measure a particular trait. The section What and how to measure describes the standardized method and provides the units of expression and, if applicable, mathematical formulas for trait value calculations. The section Additional notes contains, if available, alternative techniques, often more expensive and challenging, and mainly used by more specialized research groups to answer deeper questions. This section may also list modifications of the methods for specific taxonomic groups and draws attention to potential caveats and improvements. Finally, the References list a number of key papers which are cited in the protocol.

Standardization of measurements and acclimation of animals

Organisms respond to a multitude of external environmental factors, leading to differences in trait values due to trait plasticity, learning and shifts in physiological status. As a consequence, trait values may depend on the immediate conditions an organism is subjected to at the place or time of collection. To achieve standardized trait measurements, it is necessary to provide the comparable conditions for all individuals measured, which for many traits requires an acclimation period in order to minimize the effect of local conditions (Cornelissen et al. 2003). By doing this, the trait variability within species will more tightly reflect genetic rather than environmental effect and information about intraspecific trait variability can become valuable (see below). Therefore, the handbook starts off with a standardization protocol that describes recommendations for pre‐treating and acclimating animals to obtain comparable values within and among species for all taxonomic groups. Here, the importance of static conditions relative to fluctuating ones (e.g. Colinet et al. 2015), which reflect the natural environment more closely, is discussed. The matter is not a straightforward one (Chown & Gaston 2016) because the introduction of variable conditions in a standard protocol setting implies that assessments, and subsequent comparisons, have to be made across regimes that differ in mean values, and variation that is described by amplitude, frequency and predictability of a condition (see Angilletta et al. 2006; Chown & Terblanche 2007).

For traits which are expressed in terms of survival time as the unit of measurement, such as inundation resistance, all individuals should have the same nutritional status at the start of the measurements and should either be fully fed or subjected to a short starvation period to empty their gut prior to trait measurements. When measuring feeding traits (e.g. biting force, ingestion rate), it is necessary that all individuals are acquainted with the food items used during the feeding assays. For traits that are strongly temperature dependent such as metabolic rate, food ingestion rate and locomotion speed, thermal acclimation is absolutely necessary, although the acclimation time depends on the organisms and specific life cycles, as well as on the trait and ontogenetic stage of interest. As trait plasticity can occur during an organism's ontogeny (e.g. Wilson & Franklin 2002), it might be sometimes necessary to raise animals under controlled conditions (controlled environmental rooms) and measure traits in individuals born into these rooms. Obviously, in cases where the research interest is focused on the actual survival time when animals are exposed to drought in their habitat, the actual diet composition in the field, or the dispersal distance under natural conditions, then standardized measurements will not need to be imposed, except perhaps for serving as a baseline to measure the extent by which field conditions depart from basal adaptations.

Selection of specimens and number of individuals per species

A key consideration is selecting the appropriate specimens for trait measurements. Aiming to compare standardized trait measurements across studies and taxa of any developmental stage and sex, we recommend selecting healthy, well‐shaped and fully developed individuals of the ontogenetic stage of interest, without any signs of damage and diseases, an approach already suggested in plant‐trait analyses (Cornelissen et al. 2003). The use of interception trapping devices, such as pitfall traps, windowpane traps and Malaise traps to collect species for trait measurements should be regarded with caution as the quality of the captured individuals depends on construction, location, time of day, season or year, weather and trap clearance frequency (Gibb & Oseto 2006), and, importantly, they might be selective for specimen with certain traits. We recommend therefore that the sampling methods should be reported in detail and that additional information on trapping efficiency should be provided together with the trait measurements.

When laboratory strains are used for measurements, care should be taken as laboratory adaptation may cause spurious changes in life‐history and physiological traits of species (Sgro & Partridge 2001; Griffiths, Schiffer & Hoffmann 2005). The type of culturing method, the size of the stock population and the length of the period of laboratory culture are all factors that determine the magnitude of selection response in laboratory population, and therefore, these factors need to be reported meticulously with the trait measurements.

Sample size is a general issue in trait‐based approaches and has already been covered in other publications, although mainly on plants (e.g. Pakeman & Quested 2007; de Bello et al. 2011; Bolnick et al. 2011; Fu et al. 2013; Pérez‐Harguindeguy et al. 2013). If one would like to capture the full spatiotemporal variability of a species trait mean, a proportional number of individuals should be measured from different populations, seasons, communities and ecosystems (Pakeman & Quested 2007; de Bello et al. 2011; Violle et al. 2012). This number will further increase if other sources of intraspecific variation will be included, for example polymorphism, sexual dimorphism and ontogenetic stages (Yang & Rudolf 2010; Violle et al. 2012), which are all particularly important among invertebrates. In general, the minimal number of individuals to be measured for a given species will depend on the variation of the trait values. The higher the variation, for example, in case of behavioural traits, the higher the numbers of individuals to be measured for reliable estimates of the species mean trait value.

Future perspectives

This handbook is a first step towards standardizing trait methodology across some of the most well‐investigated terrestrial invertebrate groups. We are aware that its protocols do not cover all special cases and may miss information for particular taxa. Below we highlight three fields that we hope will be developed further with the aid of this handbook and offer a perspective on these fields of trait research.

Incorporating intraspecific trait variability

Evidence is increasing that intraspecific trait variability plays a significant role in demography and community assembly (de Bello et al. 2011; Bolnick et al. 2011; Violle et al. 2012; Siefert et al. 2015). Within‐species variability may originate from spatial variability in trait values within a species range, or may be due to genetic or environmental variation within a population at a single site. Information on both types of variability is extremely valuable, e.g. for understanding the mechanisms underlying community assembly or as input for models on functional consequences of global drivers (Gaston, Chown & Evans 2008; Yang & Rudolf 2010). Until now, the lack of standardized measurements for invertebrate traits, as well as the tiny sample size for many traits, has prohibited a clear indication of the trait variability beyond the single species level. We believe that the use of the standardized protocols can overcome this gap and we recommend not to report only species trait means for the traits measured, but also measures such a standard deviation (Carmona et al. 2016).

Definition and validation of effect traits

Quantifying community functional trait structure such as the variation in response traits, the diversity and redundancy among species sharing similar effect traits, and the overlap between response and effect traits is important for enhancing predictability of ecosystem functioning under environmental change (Folke, Holling & Perrings 1996; Elmquist et al. 2003; Mori, Furukawa & Sasaki 2013). While our knowledge on response traits of terrestrial invertebrates is relatively good, information on the extent to which response traits and effect traits can be linked within taxa, either via trait correlations or trait trade‐offs, is still largely lacking. Even less is known about response‐to‐effect models across trophic levels (Schmitz 2008; Lavorel et al. 2013; Moretti et al. 2013; Pakeman & Stockan 2014; Deraison et al. 2015), although the degree of overlap between the two types of traits will determine our ability to predict changes in key ecosystem processes under variable environmental conditions. The current definition of response and effect traits in terrestrial invertebrates is based on the literature and expert knowledge, but validation based on controlled experiments is urgently needed.

Construction of a trait data base for terrestrial invertebrates

The benefits of standardized trait measurements to the research community can be amplified if this information is compiled in a communal data base. Following the successful example of the world‐wide TRY initiative (Kattge et al. 2011), we propose that increased access to trait information collected with standardized protocols will promote the interest to use this data. For many research questions, traits obtained from trait data bases can be used as a first step to test hypotheses (Cordlandwehr et al. 2013) and for analyses at broad spatial scales (Hortal et al. 2015). In plant ecology, this has been a very successful approach, sometimes leading to additional trait measurements at different spatial scales (de Bello et al. 2009) or with a stronger focus on intraspecific trait variability (Bolnick et al. 2011). However, the construction and maintenance of such a large data base is a major undertaking that likely requires a dedicated staff and long‐term funding. We hope that an enthusiastic and regular use of this first handbook of protocols for standardized measurement of terrestrial invertebrate functional traits will encourage researchers and funding agencies alike to taking this crucial long‐term option.

Acknowledgements

This work was made possible by support from different agencies funding the following authors: F. de Bello: Czech Academy of Science, GACR grant 16‐15012S; F. Altermatt: Swiss National Science Foundation Grant PP00P3_150698; S.L. Chown: Australian Research Council DP140101240; F.M. Azcárate: Spanish MINECO, Project CGL2014‐53789‐R and Madrid Regional Government, Project REMEDINAL‐3; J.R. Bell: BBSRC strategic grant ‘Delivering Sustainable Systems’ ISP; J. Hortal: Spanish DGCyT project CGL2011‐29317; E. Öckinger: Swedish Research Council, contract 621‐2010‐5589; J.P. Sousa: FCT through the project BiodivERsA/001/2014, project ECOSERVE ‐ Sustainable provisioning of multiple ecosystem services in agricultural landscapes; J. Eller: The Netherlands Organization for Scientific Research, VICI grant 865.12.003; M.P. Berg: the Uyttenboogaart‐Eliasen Society for an Affiliate Professor grant. We also thank I. de Castro, B. Martín‐López and J. Nahmani for their valuable advice and three anonymous referees for their comments on the manuscript.

Data accessibility

This manuscript does not use data. The protocols are provided as Appendix S1.