Biol 498D - Spring 1999

Essay Assignment 2
Links between contemporary and classical papers in ecology.

Farrell, Terence M. 1991. Models and mechanisms of succession: an example from a rocky intertidal community. Ecological Monographs 61 ( 1):95-113.

Summary

The concept of succession has a long history in ecology. Many scientists have described the apparently predictable changes of community composition over time, beginning in un- or de-populated sites. Farrell begins this paper by citing a concise, modem, and inclusive definition of the phenomenon of ecological succession by Connell and Slayer (1977) as the sequence of species colonization and replacement that follows disturbance. Farrell refers again to the work of Connell and Slayer (1977) in defining three scenarios for the effects that early successional species might have on those later in the sequence. Under the model of facilitation, which comes closest to the classical concept of succession, early colonists speed up the establishment of their successors by various means. Under an inhibition model, early-establishing species retard the establishment of late successional species. In the third model, labeled tolerance, the early colonists have no effect on the rate of establishment of later species in the succession.

Farrell is also interested in the effects of consumer organisms on succession, and he sketches out a scheme for how consumers which selectively reduce the population of various successional species might affect the rate at which succession proceeds. The effect of consumers is expected to differ depending on which model of succession applies and whether there is a disproportionate removal of early or late successional species by the consumers.

The paper draws upon these models to pose two key questions which have prompted some experimental investigations of a natural seashore community. First, what factors shape the normal sequence of succession in the community? Second, what is the influence of consumer organisms on the speed of the succession which takes place?

To address these questions, a field experiment was conducted at several points within a 20 km stretch of the Oregon seacoast. At three locations, experimental plots were laid out on flat or sloping basalt rock shorelines in the upper part of the intertidal zone, 2 to 3 meters above mean lower Iow water. Before manipulation, all of the sites were occupied by a low diversity community consisting principally of a dominant alga and barnacle (Balanus glandula); a second, smaller, barnacle (Chthamalus dalli) and three other species of algae; and various herbivores of which limpets were the most abundant. A number of experimental square plots ranging from 6 to 15 cm on a side were marked and subjected to one of several different randomized treatments. The plots whose treatments required continual intervention were maintained monthly. Data on the abundance and species composition of the macroscopic organisms present was collected for each of the plots approximately every four months over periods which ranged from 24 to 33 months at the three sites.

Some plots were left untouched as controls after being marked. The remaining plots were cleared by scraping existing organisms off with a putty knife. Some of the cleared plots were then left undisturbed, while others were covered with sterilized barnacle shells glued to the rock. Other plots were subjected to continuing regimens which included removal of all barnacles, removal of the dominant barnacle species, and removal of the other barnacle species. One set of plots was ringed with copper-based paint to exclude the limpet herbivores, of these limpet-proof plots some were left undisturbed after the initial clearing while others underwent continued barnacle removal.

Farrell found that over the course of the study the control plots community structure remained the same, on average. The cleared, then undisturbed plots tended to approach the macrofaunal composition of the controls over time. The basic successional sequence was the same in all cases: The smaller barnacle became established first, the larger barnacle became dominant later while at the same time the smaller one declined in abundance. Algae increased after the larger, dominant barnacle had become established.

The barnacle removal treatments showed that removing the smaller barnacle had no effect on establishment of the dominant barnacle, while the smaller barnacle was more abundant in treatments where the dominant barnacle was removed. This is evidence for the tolerance model of succession, since the earlier colonizer has no apparent effect on establishment of the later species which apparently easily out-competes it.

The presence of barnacles, or of artificially fixed, barnacle shells, led to increased abundances of algae. This finding illustrates the facilitation model of succession. The fact that barnacle shells were at least as effective as living barnacles suggests that the facilitation of algae by barnacles might be mediated by alteration of the substrate.
When limpets were excluded from the test plots, the observed rate of succession increased. Both the transition from Chthamalus to Balanus barnacles and algal colonization were accelerated. This is consistent with the hypothetical model in which consumers which reduce the abundance of later successional species will slow succession which is characterized by the tolerance or facilitation relationship.

This research demonstrates that several different kinds of relationships may occur between the various successional species in a simple community. Clearly, succession is a complex process and significant questions remain to be answered about the relative importance of the types of relationships modeled here, and about other variables which influence ecological succession.

Historical Links

Some of the earliest published ecological research is on the topic of succession. Cowles (1899) work on vegetation patterns in Lake Michigan sand dunes is clearly focused on some of the same questions as this contemporary paper. Cowles' conclusions are based on description rather than experimentation, but his pioneering use of the concept that spatial gradients could be used as a substitute for temporal change, given what he knew about the construction and vegetation of sand dunes, allowed him to theorize effectively about the course of succession over time with a significant empirical basis. His work is actually closer to the modem approach than some other, more recent descr/ptive theories which describe stages and sequences in detail, without taking a very close look at the processes required to mediate change from one stage to the next.

Cowles focuses on the processes by which colonization by certain plants creates conditions necessary for the establishment of other plants which then replace them. He is most concerned with how early colonizing species create the physical and chemical conditions suitable for the establishment of later arrivals. This view of succession is the dominant theme of the classical theory. In the modem analysis scheme, Cowles and others are seen to be discussing exclusively the obligatory facilitation model of succession, which is now seen as one of many possible models of successional relationships. The modem research has a prominent focus on direct and indirect inter-specific relationships rather than emphasizing solely the abiotic environment.

Cowles view of succession is largely unidirectional, with succession beginning on uninhabited mobile sand dunes and proceeding to a stable forest community end point. This differs from the modem view in which succession is seen as a recurring cycle which follows periodic disturbance events. However, Cowles theory is flexible enough to allow him to account for his observations of the effects of what would now be called repeated disturbances. He acknowledges, for example, that stable, forested sand dunes may become' mobile and lose their vegetation in some circumstances. Cowles labels these processes as rejuvenation or regression, special categories that modem theory has little need for, since the there is now less of an emphasis on an end-state and a continual disturbance regime has become accepted as a typical feature of many environments.

Although Cowles discusses succession mostly in terms of whole communities, he often goes into detail concerning the conditions and processes leading to the arrival and departure of particular, individual species. This individualistic level of analysis anticipates the modem definition of succession used by Farrell, wherein succession is reduced to a collection of explanations for the appearance and disappearance of each species in a sequence of observed stages.

In Clements (1936) exhaustive discussion of the climax, succession is identified as the process by which the development of vegetation in a particular environment reaches a permanent, final stage. Although Clements creates names for and describes in detail many of the stages leading to a climax formation, he shows little interest in elucidating the processes which effect the change form stage to stage. He notes with disapproval that there is a growing tendency to abandon static concepts, and suggests that this may be in part due to the consequences of the level of disturbance imposed by human activity on the environment. Absent human intervention, the environment would be characterized by stability of its climax formations which are shaped mainly by climate.

This appears quite at odds with the modern view that continual disturbance is typical and that the notion of a climax may not be particularly useful, given these conditions. In the contemporary paper, Clement's climax is not seen as of pre-eminent importance or even particularly stable. Farrell notes that large disturbances altered the community structure of some of the control plots he studied. In his discussion of the results, he points out that the algal growth in the final stage of succession may lead to death or dislodging of the dominant mussel species and consequent re-initiation of the succession cycle. The contemporary work places much more emphasis on the processes of change between stages and how they work, and in approaching these processes experimentally rather than descriptively.

This modern paper clearly has direct roots in the Connell's (1961) classic research on interspecific competition in barnacles. The choice of a relatively simple experimental community which features slow growing organisms has been exploited effectively by both projects. It is a system that permits relatively easy manipulation and allows relatively easy data collection. Although Connell was interested in primarily in competition, the experimental design in his work and in the contemporary investigation of succession remains largely similar.

Interestingly, consumer organisms play a role in both experiments. Connell found that he needed to assess the effects of a predator which fed on the competing barnacles under study by excluding it from some plots. In Farrell's .experiment, the major herbivore limpets were excluded from test plots to measure any influence on the rate of ecological succession. Connell found that predation reduced competition, while Farrell finds that consumers slow succession in the community under study. Both of these findings suggest that consumer organisms may slow down processes that lead to the disappearance of species from a community.

Odum (1969) discusses ecological succession as a directional, developmental process during which a collection of correlated trends may be expected to occur. Almost all of his analysis is at the level of community, using measures such as niche specialization, production, and diversity. He does not focus on species replacement mechanisms. This descriptive, speculative work seems to have more in common

with the classic, unidirectional, progressive theory of succession than the modern process-oriented approach. Many fascinating predictions are set forth, and many provocative questions raised, but it appears that the current research focus is to examine smaller, simpler scale problems for which more robust experimental techniques are available. The scope of Farrell's research, though much reduced from Odum's vision of succession, nonetheless generates complex, multi factor models.

This contemporary investigation of the ecological phenomenon of succession presents us with an interesting combination of features from a historical perspective. On the one hand, the theory underlying succession, as outlined by Farrell, has been enlarged and refined considerably since the concept was introduced. There are now models for several types of possible interactions between early and later participants in a pattern of succession, and the complicating effects of consumer species are included in an scheme that interacts differently with each of the tolerance, inhibition, and facilitation models. Earlier work on succession focused mostly on applying a unified model of succession based on facilitation and did not seek to differentiate the effects of various interspecies relationships such as competition and herbivory. On the other hand, in the contemporary paper Farrell's scope of investigation is much reduced from the typical earlier analysis of complex communities and regional scale effects. The modem research is focused on a very simple community studied on a small physical scale, allowing a clear and careful analysis of successional pathways and supporting conclusions of causality, but perhaps limiting the applicability of the results to broader scales of size and complexity that one might want to understand.

Note: All references, except to Farrell (1991), and Connell and Slayer (1977, listed below) are to papers reprinted in Foundations of Ecology, edited by L. A. Real & J. H. Brown, University of Chicago Press, 1991.

Connell, J. H., and R. O. Slatyer. 1977 Mechanisms of succession in natural communities and their role in community stability and organization. American Naturalist 111:1119-1143.