Aardvarck to Zyzyxia– functional groups across kingdoms


Author for correspondence(tel +33 4 67 61 32 68; fax +33 4 67 41 21 38;emaillavorel@cefe.cnrs-mop.fr)

Functional groups: concepts and applications Montpellier, France, September 2000

Though the desire to group species according to their ecological role is not a new one (Theophrastus classified plants into trees, shrubs and herbs in about 300 bc), the concept of the functional group has received renewed attention in relation to research on the effects of global change on terrestrial ecosystems (Smith et al., 1997) and on relationships between biodiversity and ecosystem function (Körner, 1993; Wardle et al., 1999). A large amount of activity and numerous publications have resulted, mostly in the area of plant functional types (Woodward & Cramer, 1996; Lavorel & Cramer, 1999), but also for other organisms including soil biota (Lavelle, 1997; Mikola & Setälä, 1998), vertebrates (Simberloff & Dayan, 1991), and aquatic vertebrates and invertebrates (Covich et al., 1999). The objective of the symposium ‘Functional Groups: concepts and applications’ was to build bridges, finding common concepts and approaches between those working in these hitherto ill-connected areas of investigation.


Though the variety of views on functional groupings for different organisms and ecosystems cannot be ignored, scientists in the different disciplines are faced with common motivations, approaches and difficulties. Participants identified some common principles and conceptual or methodological problems. The first striking and sobering issue is that of concepts and definitions. In spite of continued efforts by proponents of the functional group approach to publish in widely read journals (Simberloff & Dayan, 1991; Lavelle, 1997; Lavorel et al., 1997) terms continue to be used with little consistency. A frequent confusion concerns functional groups in their stricter definition (groups of species with a similar role or effect for ecosystem functioning) and response groups (groups of species with a similar response to an ecological factor). The two types of functional classifications also correspond with different approaches and lists of traits measured. Detailed physiological studies at the individual level (e.g. T.S. Barigah and co-workers, INRA Nancy, France; C. Descolas-Gros, Université Montpellier II, France) are more commonly used for functional groups, whereas response groups are identified through community-level studies of response to a diversity of abiotic, disturbance and biotic factors (e.g. G. Bornette, Laboratoire d’Ecologie des Hydrosystèmes Fluviaux, Villeurbanne, France; T. Dutoit, Université d’Aix-Marseille, France; S. Jauffret, CEFE-CNRS, Montpellier, France). The use of often nonoverlapping traits between the two types of studies makes it difficult to reconcile the two types of classifications.

Across fields, ‘complete’ functional classifications that are relevant to ecosystem dynamics and functioning would need to include aspects of the following:

  • • Demography (life history characteristics and adaptive strategies)
  •  Feeding strategy (i.e. interspecific interactions)
  •  Effects on ecosystem function through biogeochemical cycles or changes in ecosystem structure.

One striking difference that appeared between disciplines concerns the basis on which species are grouped. In plants, virtually all classifications are based on the use of traits, while in animals (vertebrates, soil macro-fauna and aquatic organisms), species tend to be grouped in relation to their action in the ecosystem (Bouché, 1977)

Animals – the notion of the guild

For animals, and especially vertebrates (J. Blondel, CEFE-CNRS, Montpellier, France), the notion of guilds has long pervaded thoughts on community structure (Diamond, 1975). Guilds are defined as groups of species that exploit the same class of environmental resources in a similar way (Root, 1967). They do so using a diversity of strategies, such as differentiation in beak size in birds. Guilds have also recently been applied to plants (Wilson, 1999) to examine the structure of communities from the point of view of groups with strong intragroup, but weak intergroup, competition. Although the notion of guilds does have common points with the idea of functional classification, the former emphasizes community structure as opposed to ecosystem dynamics and functioning in the latter.

Working from the point of view of guilds and trophic networks, classifications of soil fauna and aquatic invertebrates have traditionally focused on the role played by species in trophic relationships. However, recent work has demonstrated the need to consider other functions, such as mechanical effects of organisms on ecosystems (bioturbation and ecosystem engineering at large) in addition to effects on biogeochemical cycles through trophic relationships (P. Lavelle, Ecologie des Sols Tropicaux – IRD Bondy, France for the soil macro-fauna; M. Gérino, CESAC-CNRS, Toulouse, France and G. Stora, Centre d’Océanologie de Marseille, Marseille, France for aquatic invertebrates).

Plants – one functional group?

In plants, research on functional types has examined a very large number of responses and functions. However, from the trophic perspective adopted for animals, one could argue that plants should belong to a single functional group. From a whole ecosystem perspective, plants are indeed a single group whose main function is the production of organic matter from the (mostly) inorganic resources of the environment. This perspective, however, ignores other trophic interactions such as those involved in selective herbivory of different organs or animal-mediated dispersal. Following this perspective, interspecific differences concern the intensity at which the autotrophic function is performed, as the result of a fundamental trade-off between traits favouring resource acquisition and those relevant to resource conservation (Grime et al., 1997; Poorter & Garnier, 1999). The intensity can also be regulated by response to other environmental constraints such as disturbances or biotic interactions. The difficulty in relating this intensity with responses to diverse and simultaneous environmental factors then stems from the frequent lack of overlap between classifications developed for different factors.

In plants, a wealth of studies about response to disturbance, resource gradients and competition highlight the importance of a few central traits which appear relevant to both response and function (H. Gondard, CEFE-CNRS, Montpellier, France; T. Dutoit, Université d’Aix-Marseille, France). These include Raunkiaer’s (1934) life forms or life-form schemes based on a similar philosophy (Box, 1981). The three traits included in Westoby’s (1998) plant ecology strategy scheme (seed size, plant height and specific leaf area) have also been repeatedly identified as related to plant functioning in different environmental conditions, response to disturbance and biotic interactions (Garnier et al., CEFE-CNRS, Montpellier, France). Interestingly, body size also plays a key role in the ecology of animals (Peters, 1983; West et al., 1997). More specifically, the body area : volume ratio – an index that may be compared to the specific leaf area for plants – has been related to metabolic rates in particular.

Looking ahead

Future challenges lie in linking response groups with functional effects, especially through the use of appropriate (rather than ad hoc) trait lists. For soil invertebrates, existing classifications still rely heavily on taxonomy, since many of them concern only one particular taxonomic group (e.g. earthworms, nematodes and collembola). This should be reconsidered, because aspects of soil functioning, such as modifications of soil structure, depend on a variety of organisms that may not be taxonomically close. For plants, the way forward is to use a common list of key traits obtained from a standardized methodology (Westoby, 1998; Weiher et al., 1999). This should lead to the constitution of databases of plant traits (S. Gachet et al., Université d’Aix-Marseille) (Fitter & Peat, 1994) that can be used to test hypotheses related both to the distribution of plant species and to the impact of changes in species distribution on ecosystem functioning.

Important common basic and methodological issues concern the concept of functional group, as opposed to functional trait. Indeed, except in cases with distinct qualitative variation in function (e.g. nitrogen-fixing in plants, different trophic levels in the soil or aquatic ecosystems), variation in function is quantitative and continuously related to one or a few traits (e.g. specific leaf area in plants, body size in animals). One may then be justified in challenging the need to establish groups, whereas the interesting question lies in the identification of the determining traits, trade-offs among them, and quantitative relationships between these traits and specific ecosystem functions. For example, plant specific leaf area has been shown to correlate with ecosystem productivity (Reich et al., 1997) and rate of litter decomposition (Cornelissen et al., 1999). In the same way, animal body size relates to metabolic rates and population properties such as fecundity. Few studies to date have focused systematically on the types of relationships that may be found between traits and functions, and especially on the structure of the inter- and intraspecific variance around these relationships.

When groups may be needed, as in the case of vegetation mapping, modelling, or management of diverse ecosystems, fuzzy classifications such as those used in palaeo-ecology (S. Gachet et al. Université d’Aix-Marseille) (Sykes et al., 1996) are likely to be more relevant than classical hierarchical classification methods (Pillar, 1999). In any case, formal studies of the effects of methods for aggregating species on predicted ecosystem function are essential (N. Picard, CIRAD Forêts, Montpellier, France).

The usefulness of functional groups has been demonstrated, at least intuitively, for forest management (J.-C. Rameau, ENGREF, Nancy, France), and managed grasslands or rangelands (Sinclair & Arcese, 1995; Díaz et al., 2001). However, the difficulty of obtaining classifications relevant to multiple functions makes it difficult to manage for integrated functionality of the ecosystem. Part of this problem may be solved by changing the focus from indicator groups to indicator traits, acknowledging that several traits will probably be needed for assessing the status or dynamics of managed ecosystems.


The concept of species functional group has developed at a different pace and with different perspectives in various biological disciplines. A substantial effort should be made to harmonize terminology, thereby improving communication across and within fields. For all organisms, one of the most exciting challenges lies in the identification of key traits of species related both to their response to environmental factors and to ecosystem functioning.


The meeting, part of the the French Ecophysiological Society and the Global Change and Terrestrial Ecosystems (IGBP) programme, was organized by Marie-Laure Navas (Ecole Nationale Supérieure Agronomique de Montpellier (ENSA-M, Montpellier, France), Eric Garnier, Sandra Lavorel and Catherine Roumet (CEFE-CNRS, Montpellier, France). Many thanks to Maryse Gautier, Gérard Laurent and Jean Richarte for their help in the organization, and to the ENSA-M communication unit for providing the exceptional Philippe Lamour conference room. Summaries of communications will be published as issue 25 of the Bulletin de la Société d’Ecophysiologie (2000): see http://www.ese.u-psud.fr/.