Published in Selbyana 2001.

Mark W. Moffett, Museum of Vertebrate Zoology, University of California, 3101 Valley Life Sciences, Berkeley, CA 94720, U.S.A., moffett@uclink4.berkeley.edu

The Geometry of Canopy Biology

If we expand our definition of "canopy" beyond what I suggested in Moffett (2000) to encompass all the parts of any community of sessile organisms that project into a medium, on what basis might canopy biology rest on firmer ground as a discipline?

Ecologists have traditionally dealt with plants (or other sessile organisms) in two dimensions, or as points on the earth. While orientation in canopy ants, the conservation of orchids, and foliage uptake of pollutants bear on canopy biology sensu Moffett (2000), much of the research on these topics fall squarely within this tradition. For example, most studies of orientation in canopy ants ignore spatial issues bearing on plant topographies, as well as other issues unique to canopy substrates, such as properties of pheromone diffusion from bark versus from leaves. In a typical experiment, ants are entirely removed from their canopy environment, as when a species normally found on foliage is studied on a flat laboratory surface. While we can learn a great deal about canopy organisms with this kind of approach, none of it has to do specifically with their origins in the canopy.

In contrast, the core of canopy biology as an independent discipline can be characterized in large part as the science of treating communities of plants (or other sessile organisms) as three dimensional. Expressed another way, creating a robust science focused on, for example, aerial plant organs requires us to "put the canopy into canopy biology" through research that contributes directly to understanding the distinctive aspects of life associated with plants. Typically this can be achieved by introducing the "z" axis or other canopy-specific attributes (see studies on ant orientation by Jander & Voss 1963, Beugnon & Fourcassie 1988, Fourcassie & Beugnon 1988, Jander 1990, Klotz & Reid 1992, Wohlgemuth et al. 2001). For example, Diekmann et al. (2000) conforms to the 2D "mainstream tradition" except for some material on three dimensional gap structure, which by my criterion is the only part of their coverage of ecological geometry that represents canopy research sensu stricto. The same philosophy can apply to aquatic systems, which offer unique experimental opportunities: consider the value of artificial reefs of varied structure to understanding productivity and diversity of canopy residents (Bohnsack 1989, Hixon & Beets 1993, Carr & Hixon 1997).

I have outlined six approaches to "putting the canopy into canopy biology," calling these the core issues of the subject (Moffett 1999). By framing questions in regard to one or more of these issues, researchers can contribute fundamentally to canopy biology as a discipline. I review these issues below. To keep my treatment brief, I’ve chosen examples from the forest literature, although there is information on each issue available for other kinds of canopies. While I find that most scientific results can be partitioned along these lines, the issues clearly are not independent, and of course many studies make important contributions in two or more of them. I believe that in time a large part of canopy science may ultimately codify according to how these issues have contributed over ecological and evolutionary time to the diversity of both the host organisms and their occupants.

1) Community ecospace. For any canopy, the quantity and quality of space available to canopy dwellers depends on host structure. How much so is a matter of conjecture. Any small nimble bird or agile climbing animal like a gibbon seems to experience canopies as a volume, although even here not all points in the volume may be accessible, such as vegetation too dense to be navigated by a bird (Cuthill & Guilford 1990) or spaces too wide for primates to cross (Cant, 1992, Cannon & Leighton 1994). At the other extreme, small flightless arthropods such as mites or earthworms are unlikely to register community ecospace as a volume. Ants, for example, are restricted to within millimeters of every surface within their territory. Because of this, and in spite of models to the contrary (Hölldobler & Lumsden 1980), ants experience canopies as something between a 2D and 3D space (Moffett 1994). In essence, a canopy represents for them a highly warped surface. Like a science fiction ship using a wormhole to bridge points normally experienced as distant from each other, Oecophylla (weaver ants) create shortcuts through this space by linking bodies into chains to access new branches (Hölldobler & Wilson 1977), and thereby can bridge whole tree crowns that otherwise could only be reached by way of a long march to and from the ground.

2) Aggregate structural properties of the community. Here I include the nonuniform vertical distribution of canopy structures and the concomitant stratification of other elements of a canopy environment, such as microclimate. Parker & Brown (2000) criticize studies of stratification for their lack of reproducibility, inconsistent terminology, and other weaknesses. Regardless of the difficulties, understanding stratification is at the core of much of canopy science (Moffett 2000). Whereas many studies of terrestrial nutrient interception treat the canopy as a single "black box" with overall inputs and outputs (lc. Coxson & Nadkarni 1995), some researchers have uncovered a complex internal pattern within canopies, which act as atmospheric filters (Wiman et al. 1985, Meyers et al. 1989, Lovett & Lindberg 1992) and nutrient transfer systems (Pike 1978, Reiners & Olson 1984, Coxson et al. 1992). Defining strata or gradients requires broad spatial averaging, while explaining any patterns that emerge necessitates research at a finer spatial scale. For example, bark pH can vary with height in trees (Hyvärinen et al. 1992). If the height distribution of an epiphyte corresponds to that of bark pH, we could propose that the plants prefer a particular pH. Testing this hypothesis requires determining the detailed distribution of bark pH's at the actual locations occupied by the plants (e.g., Gauslaa 1995) followed by manipulations of pH in the field or laboratory (e.g., Hallingbäck 1990).

3) Host distribution. In all canopies, the substrate changes from place to place, for example from one host to the next. The likelihood that many canopy residents are specialists on one or a few plant taxa (e.g., Erwin 1982) suggests the usefulness of considering hosts as islands over evolutionary time (Janzen 1968, 1973), both in explaining levels of resident diversity, and potentially even in modeling processes occuring within and between plants that generate such bounty. But rather than treating hosts as islands in a uniform ocean to conform to the perspective of MacArthur & Wilson (1967), "patchwork" biogeography models could treat communities as a continuum of host islands of varied acceptability as sources of food, retreats, or transit routes to each canopy-dwelling species. In addition, diversity is organized at scales both smaller and larger than that of a host. Any stable canopy element that qualifies as an island sensu Haila (1990) could harbor a distinct community, such as a flower, phytotelmatum, and leaf (to a microbe) (Seifert 1975, Andrews et al. 1987, Jenkins & Kitching 1990, Richardson 1999), and potentially also certain microclimatic features (Herwitz & Slye 1992). Even ant territories could form habitat islands for residents within tropical canopies (Moffett 1994). Territories of different ant species exist as mosaics that overlay, but are partially independent of, the mosaic of the trees themselves (e.g., Dejean et al. 1999). Ants scour their territories to drive off intruders and kill prey while promoting survival of species-specific assemblages of associates (Hölldobler & Wilson 1990). Like other island-like canopy features, then, ants may add to the potential for species to pack into vegetable space.

4) Host architecture. All canopies from redwood forests to biofilms (Lawrence et al. 1991) have varied architectural parts. In forest ecology, there is a burgeoning literature on this topic, that is, the size, angles, distributions, development, and spatial relations of aerial plant parts. Classically, architecture is described for trees by the models of Hallé et al. (1978). The "Hallé-Oldeman architectural model" classification has been described by one key practitioner to be "comparable to the development of the binary system of nomenclature by Linnaeus" (Tomlinson 1983). Nonetheless, the system has been little used by nonmorphologists, arguably due in part to an overall neglect of the potential importance of substrate architecture on canopy organisms (but see Hallé 1990). There are studies of the effects of simple architectural attributes, especially branch angle and width, on the growth of epiphytes (e.g., Rasmussen 1975, Ingram & Nadkarni 1993, Moe & Botnen 1997) and on animal locomotion, particularly in reptiles (Plummer 1981, Irschick & Losos 1999, Beuttell & Losos 1999) and primates (Demes et al. 1995, Povinelli & Cant 1995, Warren 1997, Dagosto & Yamashita 1998, Hamrick 1998, McGraw 1998, Garber & Rehg 1999, just to mention recent citations on positional behavior). Yet there have been no broadly scaled studies of how plant architecture influences canopy life. Consider that many tropical canopies have well-beaten vertebrate highways which in some cases can be detected by the epiphytes that spread to each side of a branch like hair from a part (Perry 1978, Sillett et al. 1995). Perry (1978) found evidence of multispecies use and active pruning, but as yet no one has mapped such a trail in relation to the tree architectures available locally, or documented how the trail originates, how long it lasts, and how its usage shifts with changes in canopy physical structure and resource availability.

5) Open space. Not even "biofilms" are a continuous matrix of organisms: all canopies consist of a framework occupying a dynamic fluid matrix that has open communication and exchange with the adjacent atmosphere or hydrosphere. Open space (air or water) within canopies merits special consideration because of its potential effects on microclimate and on the locomotion or dispersal of organisms, and because many aspects of the subject remain virtually ignored. Space within or between forest trees is commonly distinguished into two categories, with "gaps" being the result of plant death and shyness-related space often being the result of plant growth and reconfiguration, that is, plant foraging (Hutchings & de Kroon 1994; this distinction is somewhat of a simplification because physical abrasion can also be involved in shyness patterns: Franco 1986). Gaps are the most commonly studied spaces because of their role in forest succession and species diversity patterns (Lieberman et al. 1989). Spaces represent barriers to some species and pathways ("flyways") to others (Brady et al. 1989, Aluja et al. 1989, Cuthill & Guilford 1990, Cannon & Leighton 1994, Brigham et al. 1997, Aylor 1999), but little information exists on how open space might be involved in structuring the population and species distributions of residents within canopies. There is a tendency to think in terms of canopy structures such as trunks and branches when the space between structures could be the resource used, as might be the case among gliding animals (Moffett 2000). Within the open spaces, boundary layers – the regions of lowered fluid velocity that exist around any surface in a flowing medium – are a general feature of attached communities. Their presence partially isolates canopies from the surrounding medium, and thereby can increase community reliance on efficient and potentially autogenically-controlled internal (within-canopy) nutrient cycling. This isolation may be particularly important in flowing water (Mulholland 1996), where canopy physiognomy can substantially ameliorate the downstream displacement of chemicals or of any organisms that are moving within a canopy or that have a poorly developed capacity to attach to a substrate.

6) Properties of structural elements. In all canopies, the sessile hosts present associated species with a variety of surfaces, both between host individuals or species and within each host (such as wood versus leaves in a tree). These structural elements vary in their physical and chemical properties, such as the capacity for insulation or water absorption, tendency to leach nutrients, efficiency at transmitting vibration, and their texture, stability, density, hardness, compliance, stiffness, strength, pH, and so on. How do such variables effect life on or in a host? One of the oldest areas of canopy investigation in terrestrial biology is the question of substrate choice by epiphytes, especially cryptogams (e.g., Barkman 1958). Another area of intensive study has been herbivory as it relates to secondary compounds, nutrient content, and the mechanical difficulties of feeding (Scholwater 2000). Outside these research focal points, the literature is widely scattered, with many potential research avenues of enormous prospects.


Continue reading this paper:

Abstract

Seeing the Forest for the Herbs

More to Pond Scum Than Meets the Eye

The Geometry of Canopy Biology

Getting to the Root of the Matter

Conclusions

© Mark W. Moffett, please e-mail naturalist@erols.com to obtain a complete reprint.