1Despite the reinterpretation of the importance of spatial and temporal scales in geomorphological studies by Schumm and Lichty (1965) and the subsequent orientation of geomorphology towards process studies, the Davisian cycle of erosion still holds a special place in the education of most British and North American university geography students. Textbooks either give a passing nod to Davis and the cycle of erosion (e.g., Christopherson, 2006) or illustrate the concept in detail (e.g., Holden, 2005. McKnight and Hess, 2005). The figure of W. M. Davis is also almost certain to make an appearance in any unit or textbook looking at the history of geographic thought (e.g., Gregory, 1985. Johnston, 1989. Livingstone, 1992. Inkpen, 2005). The perceived importance of Davis as a thinker in geomorphology is cemented by the devotion of the hefty, possibly definitive historical text on his work (Chorley et al., 1973) in which the whole volume is devoted to Davis’s ideas. Despite the extensive treatment of Davis in textbooks, attempts to contextualise Davisian thought have been rare (see Livingstone, 1992). Chorley, Beckinsale and Dunn (1973), for example, recognize that Davis was not using a purely Darwinian view of evolution in his ideas about landscape change, but their discussion of the nature of Davis’s evolutionary thought is couched in general and vague terms. “ The main problem arising from this association is due to the fact that in the later nineteenth century the highly attractive label of ‘evolution’ had practically become a synonym for any ‘change’, and often ‘history’ in general. This identification has tended to obscure the special character of the concept of evolution, and in the same way the concept of the cycle of erosion has been identified with all types of change in landforms and with landform development in general. Thus it is only possible to grasp the special and restrictive characteristics of the cyclic idea by understanding some of the contemporary implications of the term ‘evolution’. The late nineteenth century view of evolution, particularly in its popular, non-biological conation, implied an inevitable, continuous and irreversible process of change producing an orderly sequence of transformations, wherein earlier forms could be considered as stages leading to later forms ” (Chorley et al., 1973, p. 193). Their statement only begins to place the Davisian cycle of erosion into the wider contemporary debates concerning evolutionary theory.
2Although it may seem arcane and pointless to current thinking about landscape development to contextualise Davis and his evolutionary thought, developing such an understanding may be helpful in judging why Davis’s idea remained so influential in Anglo-American physical geography and how his ideas may still have relevance, or not, for understanding long-term landscape change. The Neo-Lamarckian “baggage” that the “cycle” of erosion carries is at odds with much current thinking about biological evolution and landscape evolution. Davis’s cycle of erosion remains a useful educational tool, but it is difficult to apply in full to current studies of landscape development. However, while it needs to be recognised that Neo-Lamarckian thinking has little or no place in modern evolutionary biology, the directionality, or broadly predictable trajectory, of the development of fluvial landscapes, in the absence of climatic change or renewed uplift, is in line with Neo-Lamarckian ideas. The general terms of the cycle of erosion, such as developmental sequences, may also be relevant to current thinking.
3In this paper we contend specifically that the evolutionary flavour of the Davisian cycle of erosion cannot be divorced from the complex threads of evolutionary thought that existed at the turn of the century in North America. In Davis’s work, as Young (1985) noted in Darwin’s evolutionary thought, there are echoes of the contemporary society. The Davisian cycle of erosion, however, cannot be reduced or conceived as a simple reflection of a category of evolutionary thought defined from the vantage point of the twenty first century. Rather, Davis’s ideas reflect a mix of the general threads of evolutionary thought, and his own unique contribution to these ideas within physical geography.
4Within the cycle of erosion, a key element in its uniqueness lies in the Davisian use of time. Simplified, two contradictory views of time have been identified in geological thought (Gould, 1987): time as a cycle and time as an arrow. Time as a cycle reflects the repetition of the same cycle of events over and over again. Time as an arrow reflects the irreversible alteration of forms over time. Kennedy (1992) extends this use of time in her analysis of six different geologists and their use of time. She distinguishes time as a repeating sequence of events, always returning to and implying a normality and predictability to the geological system. Opposed to this view is time as a sequence of non-repeating, irreversible, unique events providing an unpredictable trajectory to changes in the geological system. Davis’s ideas can be regarded as containing elements of both views of time in discussing surface landforms, an important element in assessing Davis’s significance to current theories of long-term landscape change. Attempts to trace the development of geomorphology are rare in other languages, but as Giusti (2004) noted the influence of Davis’s views was not limited to Anglo-American geomorphology. As he notes, it was the adoption of Davis’s ideas by French geomorphologists, and their rejection by geologists, that led to an ideological split between the two disciplines.
5This paper provides an overview of the major themes of the debate over evolutionary ideas at the turn of this century within which Davis developed his cycle of erosion. Summarizing, or rather caricaturing, the different threads of this debate, the Davisian cycle of erosion is analysed to assess which, if any, of the threads it follows. The uniqueness of Davis’s idea is illustrated by analysing his application of the cycle to the development of the Appalachians. From this assessment it is possible to illustrate how the Davisian cycle of erosion carries its contextual baggage into the future. The potential links of such a context to Davis is provided by the general academic environment of the period, and also by his tutor/tutee relationship with Shaler, a key player in the development and acceptance of evolution with a Neo-Lamarckian flavour in North America (Livingstone, 1989). The French geomorphologist Baulig (1950) wrote that he saw tears in Davis’s eyes on the day he announced the death of Shaler, to whom is dedicated The Coral Reef Problem (Davis, 1928): “To the memory of Nathaniel Southgate Shaler who, as my teacher at Harvard sixty years ago, first led me to know in outline Darwin’s theory of coral reefs, and who, ten years later, inviting me to return to Harvard as his assistant in field geology, opened the door of opportunity through which my life’s work was entered upon…”.
6Reconstruction of past ideas, despite the problems of “whiggish” histories, requires some framework for analysis and so some simplification and categorization is required. This is not to imply that any individual worker, particularly one outside the biological sciences, such as Davis, was only influenced by one, pure category of thought concerning evolutionary ideas. Elements of each thread of evolutionary thought could be discerned within any one concept as general characteristics, such as change over time and competition that are common to a number of such threads. Use may or may not reflect a coherent use of one thread of evolutionary thought exclusively.
7Given the reservations expressed above, three main threads can be identified: Darwinian evolution, Neo-Lamarckian, and orthogenesis, evolutionary directionality along channels of internal constraint (Gould, 2002, p. 66). Each has a distinct set of characteristics and a set of common elements that are interpreted in different manners. In a paper of this length it would be impossible to deal in detail with the three identified threads. Fortunately, Gould’s The Structure of Evolutionary Theory provides comprehensive discussions of the three threads, and in particular, the roles of Cope and Hyatt in the development of Neo-Lamarckianism (see, in particular, Gould, 2002, p. 366-370 for Cope and Neo-Lamarckianism; p. 370-383 for Hyatt and Neo-Lamarckianism; and p. 366-367 for the orthogenetic version of Neo-Lamarckianism)
8Darwinian evolution was concerned with the alteration of the form of organisms by random variation. Variations that enhanced reproductive success of individuals were preserved and passed on to offspring. The central element to the theory was the idea of reproductive success and, hence, competition between organisms and the random nature of variation. This implied that evolution followed no predetermined path and that the organism itself could not direct evolution. Although changes in form and function were often expressed as gradual, no explicit exclusion exists of rapid changes in form either (Kennedy, 1992). Evolution was usually gradual, non-directional, irreversible, and unique.
9Neo-Lamarckianism was not a unified movement (Bowler, 1989), but can be characterised as assuming that changes in the organisms were a result of the activity of the organism itself, i.e., use inheritance. In this manner evolution was directed and, therefore, potentially purposeful. The element of purposefulness, Livingstone (1984) suggests, made some forms of Neo-Lamarckianism open to interpretation from the perspective of natural theology. Design and design fulfilment were teleologically logical outcomes of this position. Adaptation could also be seen as the response of the organism to the environment and a strong element of environmental determinism emerged within the movement. Within this view change can be more rapid because random variation is not required for a beneficial change in form or function. Change is, however, directional, progressive, goal-orientated, and irreversible or reversible.
10Within Neo-Lamarckianism an important set of concepts were concerned with the relationship between the changes visible in a species and those visible in any individual of that species. Recapitulation theory stated that the growth of an embryo was repeated in the way an organism evolved. An organism would retrace its evolutionary development during its embryonic development. Development sequences could, therefore, be found in the fossil record illustrating how life had developed following the same pattern as revealed in the embryonic development of the currently most advanced species. Organisms developed along a predetermined path of progressive change (Gould, 1977). A key figure in the development and application of recapitulation theory was E. Haeckel (Gould, 1977). He viewed the biological realm as composed of entities that were aggregates of individuals from lower strata. Six levels of biological reality occur with the lowest forms being plastids, cells, and other basic constituents. Individuals of species were further up the hierarchy, but their actions were governed by the same biological processes as those which governed lower and higher levels. Instead, there was the recognition of an all-pervading developmental process across different strata, including phytogenetic and ontogenetic development. These were simply the expressions of the same developmental process on different entities.
11Evolutionary change within recapitulation, as envisioned by Haeckel (1866), was achieved by two means. The first simply involved additions to the stages of development of the embryo, with no shortening of the individual stages. The second involved the acceleration of the early stages of development, or condensation, to permit completion of the developmental sequence in a fixed time period, but with the addition of new evolutionary stages. Cope (1887) later added the idea that acceleration could be so rapid as to result in the deletion of some stages of the progression. The embryo need not, therefore, go through all stages of development. Haeckel (1866) also noted that temporal and spatial dislocation could occur in the order of stages. Within the development of an embryo, for example, the eyes develop long before expected if the evolutionary sequence were strictly adhered to. This means that development need not proceed at the same pace through the evolutionary stages throughout the embryo.
12A particularly North American brand of Neo-Lamarckianism developed under the influence of Louis Agassiz and his pupils which highlighted similar ties to embryology and regularity of development (Bowler, 1989). The views of Cope (1887) and Hyatt (1866), in particular, illustrate the close conceptual ties between Neo-Lamarckianism and orthogenesis (Gould, 2002). Cope and Hyatt used recapitulation theory to understand the evolution of organisms (Bowler, 1977, 1983, 1989; Gould, 1977). Cope (1887) distinguished between two types of change. Firstly, new species arose by modification of existing structures and, hence, produced deflections of evolutionary development. New genera arose by addition or subtraction from the sequence of orthogenetic changes. The latter were responsible for progressive evolution. Deflections of evolution he considered as following Darwinian law, whilst “real” evolution was by addition “to the end of ancestral ontogeny” (Gould, 1977, p. 86). Within Cope’s evolutionary view, therefore, it was possible to have development that was not progressive, but which operated within the structures of progressive evolution, and progressive change.
13Hyatt (1866) extended the idea of regularity to evolutionary pathways by suggesting that each group of organisms could develop from simple to advanced forms before degenerating back to forms close to the original level. Hyatt’s concept of “racial senility” saw evolution as a cycle of advance and decay, like a life cycle applied to a whole species. This re-emphasised the commonality of process between strata of reality and the inter link between individual and species. Within these views evolution is a directional, gradual process of accumulation of pre-determined characteristics by the organism. Evolution followed a predetermined, staged pathway be it progressive or towards eventual decay.
14Orthogenesis, closely associated with this last view of evolution, envisaged evolution as change in a regular manner determined by directional forces within the organism. Within this mode of thought selection and adaptation had no role to play. Organisms developed as predetermined no matter what the environmental or competitive stresses. Species extinction was, therefore, a function of internal directional forces not of “blind” selective processes external to the organism. The species thus contained the seeds of its own destruction within itself — a very moralistic view of evolutionary success or failure.
15These three related threads of thought provided the context within which Davis developed his cycle of erosion. His contact with Neo-Lamarckian ideas at least is implicit in his teaching by Shaler. Similarly, as a practising geologist he is unlikely not to have come across the works of Cope and Hyatt. Lastly, his general Darwinian traits have often been assumed (e.g. Chorley et al., 1973). Within the above diverse threads and the varying, often contradictory views of time and change, the Davisian cycle of erosion and its application to the Appalachians will be assessed.
16Before considering evidence for a Neo-Lamarckian basis for Davis’s ideas, a brief outline of his basic concept, the cycle of erosion may prove useful. Providing a simple description of the cycle of erosion, however, is problematic. A precise knowledge of Davis intent is impossible. Only by reading Davis’s work, and that of his predecessors, without any knowledge of the last hundred years of geomorphology could we hope to understand his work without our interpretation being clouded by the ideas of his successors.
17Although Davis identifies three important determinants of landscape, structure, process and stage, the central concept is that of the stage of development. Each stage is associated with a particular suite of landforms and a suite of processes and rates of operation. Initial uplift produces a landscape with height difference or relative relief AB (fig. 1).
Fig. 1 - Davis’s schematic diagram of the change in relief through time of the idealised cycle of erosion.
Fig. 1 – Diagramme schématique de l’évolution du relief selon le cycle de l’érosion idéalisé par Davis.
The horizontal scale, α to ω, indicates the passage of time (from Davis, 1899).
L’échelle horizontale, de α à , indique le temps (d’après Davis, 1899).
18Initial, consequent streams start at height A and reduce the altitude of channels to C, whilst the divides remain high at D. This means that relative relief increases during the initial stage of development, youth. River channels continue to erode and decrease in altitude, but at a slower rate for the rest of the cycle. Divides meanwhile are reduced more rapidly by headwater dissection. Relative relief is at a maximum in the period when landform diversity is at its greatest. This is the stage of landscape maturity. Finally, a long stage occurs of progressive landscape lowering to a very low altitude and relative low rates of process operation, the stage of old age.
19According to Thorn (1988) Davis’s cycle of erosion encompasses three models: a cycle of landscape development; a cycle of river development at the level of the drainage network and the individual river; and lastly, a cycle of slope development. Each cycle went through predetermined stages, which defined various attributes to each stage. The important points in the development of landscape were that relative relief altered as the landscape developed and its changes provided the impetus for variations in the rates of process operation. This was matched by an ability of the landscape to produce a variety of landforms and, hence, the diversity of landforms was greatest when the relative relief was greatest during landscape maturity. Drainage development followed a similar, predefined cycle of change. Initially, the drainage network was disorganized with a few major, consequent streams draining the landscape. With the onset of maturity the drainage network became increasing efficient and organized with the development of moderate floodplains in valleys. By old age, extensive floodplains and meanders occurred within broad, gently sloping valleys. Slope development underwent a similar cyclical change. Davis suggested that slopes progressively declined over time in response to the generation and movement of sediment across the slope surface. Coupled with these cycles was the concept of grade being a goal within the fluvial and slope systems. Davis viewed the landscape as achieving balance in maturity between the forces available to do work and the amount of work to be done. In youth, this balance was absent and the inequality resulted in the presence of indicative forms such as waterfalls. Once grade or equality was attained in maturity, landforms reflected this. Rivers, for example, had smooth, concave longitudinal profiles. Slopes attained a convexo-concave form, which gradually declined in angle.
20Given that, as Chorley et al. (1973) recognize, Davis’s view and description of the cycle underwent its own evolution throughout his life, any general analysis would do an injustice to Davis’s own intellectual development of the cycle. Instead of attempting a definitive assessment of the cycle of erosion in all its shades and the relationship to different evolutionary strands, this paper concentrates on a number of papers published by Davis between 1889 and 1905, in particular, but not exclusively, his 1899 outline of the cycle of erosion, his 1905 discussion of the arid cycle and his 1889 application of this concept to the development of the Atlantic seaboard. These papers and others cited were part of a volume of collected works published in 1954, and all page references refer to that edition, unless otherwise stated. Concentrating on this one period helps to focus on particular phrases and interpretations that could be seen as evidence of the influence of the particular North American brand of Neo-Lamarckianism. Using these papers, Davis’s views on the nature of evolution, the rate and processes of change, and his view and use of time, particularly in explaining the operation of the cycle of erosion in the field, are considered.
21As Campbell and Livingstone (1983) noted, Davis’s recapitulationist tendencies made him favourably disposed towards Neo-Lamarckianism. Davis viewed evolution as a primary fact of all development, whether organic or inorganic. In this, his thinking was very much in line with the intellectual climate of the age, in which concepts such as “social Darwinism” were also developed. Davis states (1904, p. 679): “ So wonderful is organization of these land and water forms in physiographic maturity and old age, so perfect is their systematic interdependence, that one must grudge the monopoly of the term ‘organism’ for plants and animals, to the exclusion of well-organized forms of land and water. By good fortune, ‘evolution’ is a term of broader meaning; we must share its use with the biologists; and we are glad to replace the violent revolutions of our predecessors with the quiet processes that evolution suggests ”. This quote illustrates Davis’s acceptance of evolution as a broad term, as a metaphor, applicable to organic and inorganic realms. Importantly, he does not see the organic in isolation from the inorganic, but rather conceives them as a functional whole: “ The evolution of the earth and the evolution of organic forms are doctrines that have reinforced each other; the full meaning of both is gained only when one is seen to furnish the inorganic environment, and the other to exemplify the organic response ” (ibid., p. 675). The inorganic provides the basis for an organic response immediately bringing forth a Lamarckian sense of directionality by the physical environment of organic development: “ The responses of one kind are brought down as inheritances from beginnings in an earlier time, maintained to-day because their physiographic controls are persistent... The most important inherited responses are those determined by long persistent conditions of the environment, such as are common to the physiographies of all ages... ” (Davis, 1902, in Davis, 1954, p. 51-52). Within his geographical cycle of erosion explanation and evolution went hand in hand: (i) “ Any mass of land constituting a single geographical individual or a natural group of such individuals, must, as soon as it is exposed to the destructive forces of the atmosphere begin its long sequence of development ” (Davis, 1889, in Davis, 1954, p. 195); (ii) “ A whole series of forms must be in this way evolved in the life history of a single region, and all forms of such a series, however, unlikely they may seem at first, should be associated under the element of time, as merely expressing different stages of development of a single structure ” (Davis, 1899, in Davis, 1954, p. 253-254).
22For Davis the geographical region itself was an individual capable of ageing and developing and, hence, possessing an organic quality of its own. The complex inter-related nature of the region and its landscape were the only requirements needed for Davis to permit it to be classed as capable of evolution. Central to the development process in any landscape was the river. Davis viewed the river, again, as an individual following a predictable life history and acting as the focus for geomorphological activity in the region; the sculpting force of evolution. The image of the landscape and river as individuals were essential for Davis. As organic objects, they could be expected to have a definite life cycle and a definite beginning and end point.
23The entities of landscape and river were individuals tied together by the developmental process. As with the concepts expounded in recapitulation, the entities were but strata of reality and, therefore, both subject to the same laws as the lower strata and linked by a pervading developmental process. In other words, as individuals the development could be predicted and each stage characterised by certain properties. In this manner a general model of development could be proposed for the landscape and the sculpting force. Importantly, however, the rivers were not subject to progressive evolution without changing the stage. Within the stage Davis invoked Darwinian evolution to explain processes such as consequent river capture. Rivers developed by competition for catchment within the stage, but the characteristics changed as the stage changed.
24The complex, interconnected nature of the landscape, essential in enabling it to be defined as an individual, also meant that each part was intimately connected with every other part of the landscape. Identification of the forms characteristic of certain stages in development, therefore, was tantamount to explanation of the form. Davisian analysis was based upon a strategy of classification. The stage model is, therefore, an explanatory approach dependent upon a view of the landscape and its sculpting agent as organic individuals following accepted life cycles. The forms, which are transitory, represent an expression of the different stages in the life of an individual. Reading Davis in this manner highlights his allegiances to the orthogenetic flavour of Neo-Lamarckianism put forward by Cope (1887) and Hyatt (1866). Neo-Lamarckian concepts of directed development remain, but the force or goal for evolution is not provided by the physical environment, but by a quasi-organic life of individuals, the geographical region and the river: “ For this reason I have endeavoured to examine the forms of the land surface in detail, and to arrange them in their generic relations, in order to come to a closer appreciation of the meaning of the form of the earth and its development “ (Davis, 1889b, in Davis, 1954, p. 194).
25The application of the ideal cycle and the description of each stage were not without value judgements related to the use of the life cycle and individual. Davis tends to couch his descriptions in terms that suggest a hierarchy of activity, vigour, and importance for each stage of landscape development, terms parallelling the progress to species senility of Hyatt (1866). Davis viewed and compared maturity to growth in an organism with this stage representing the most advanced level of development and organisation. At maturity destructive processes were the most powerful and drainage adjustment, facilitated by tributary competition, was prominent. This view is the basis for his observations concerning the development of an individual geographical region: “ When this work begins, with every mark of immaturity in its small accomplishment, we may regard the individual as young, that is, but little advanced in the long cycle of systematic change through which it is destined to pass. When much more work has been accomplished, and the variety of form resulting is at its greatest, the individual may be called mature; and finally, when the features of maturity weaken as the relief is reduced and the intensity of form is lost, we find a resemblance to organic decay, and are warranted in the use of such a term as old age ” (Davis, 1889b, in Davis, 1954, p. 195).
26From the above it is clear that Davis’s view of evolution was based on Neo-Lamarckian ideas as interpreted within the North American context. In particular, he viewed the inorganic as being as subject to evolution as the organic. The inorganic realm possessed individuals in the form of geographical units and rivers that followed a predictable and differentiated life cycle. Within this life cycle change was predictable and, as with the organic, subject to highs and lows of activity and alteration of form. Development was directional, towards the effective lowering of the landscape, but also within each stage directed to the completion of various stage-specific tasks.
27Davis highlights the gradual nature of evolution in bringing about change by contrasting rates of processes from catastrophic geology with slow acting evolution. He makes this point in relation to the Welsh uplands and consequent river capture which he describes as illustrating “ the patient progress of inorganic natural selection ” of subsequent extension (Davis, 1895). He used the same phrase to describe rivers along the Blue Ridge in North Carolina (Davis, 1903). He clearly liked the description, as he used a variant, “ a most beautiful process of inorganic natural selection ” in a journal aimed at a more general audience (Davis, 1900). Davis viewed gradual change as the cornerstone for understanding the development of the inorganic and organic. Moreover, it was gradual change brought about by the operation of “normal” (i.e. humid temperate climates) processes observable today that produced every geomorphological form. Without the continuation of the operation of these “normal” processes, no possibility exists to link the past to the present in Davis’s view. Continuity of process suites was, therefore, central to the Davisian view of both geological and geomorphological explanation, as it was to geological and geomorphological thought more generally. This view resulted in a key concern of Davis and others, that geomorphology was only the visible expression today of the longer-term geological development of the landscape. Tying geology and geomorphology, and by implication human geography, together meant that Davis viewed concepts from one as easily transferable to the other. Once again, the idea of an all-pervading developmental process between different strata of reality seems to guide his thinking. Concepts of uniformitarianism and developmental evolution could, therefore, be equally applied to explanations of geological data on change and to explain current forms in the landscape: (i) “ The real foundation of the broad consideration of earth sciences rests on the continuity of ordinary processes through the long periods of recorded earth-history ” (Davis, 1904, p. 678); (ii) “ The conception of the continuity of processes, without extra-natural interference, has been forced to fight its way against opposition; now it has gained at least a very general verbal acceptance among us, and is quietly drifting into popular belief. To realize its full meaning is an arduous task, not only because of the opposition of inherited prejudices, but even more because of the inherent difficulty of the problem. To think that processes such as those of today have done all the work of the past is appalling; yet we are constrained to believe it ” (Davis, 1904, p. 678); (iii) “ It is the assurance of orderly continuity that binds the past to the present in the endless sequence of events, and shows us that geography is only today’s issue of a perpetual journal, whose complete files constitute geology ” Davis, 1904, p. 679).
28Davis identifies destructive processes associated with fluvial activity as the normal set of processes doing work in the landscape: “ But as most of the land surface of the earth is acted on chiefly by weather changes and running water, these will be treated as forming a normal group of destructive processes, while the wind of arid deserts and the ice of frigid deserts will be considered as climatic modifications of the norm, and set apart for particular discussion ” (Davis, 1899, in Davis, 1954, p. 250). Despite the emphasis on continuation of process suites, Davis differentiated between process suites in different stages. The intensity and composition of processes varies with the stage of development: “ The amount of change caused by destructive processes increases with the passage of time, but neither the amount nor the rate of change is a simple function of time. ... The rate of change under normal processes, which alone will be considered for the present, is at the very first relatively moderate; it then advances rather rapidly to a maximum, and net slowly decreases to an indefinitely postponed minimum ” (Davis, 1899, in Davis, 1954 p. 251). What constitutes a normal set of processes and the normal rate of operation is, therefore, temporally differentiated. More specifically, process suites can only change when the stage changes. Stages are, therefore, given a special status in the development of the landscape; they are equivalent to the progressive changes as opposed to deflections, in evolutionary development noted by Cope (1887).
29Although Davis appears, as with Lyell, Dana and Horton (Kennedy, 1992), to have a vision of a norm for the physical landscape, it is dynamic rather than static. Within any landscape the set of processes at work will vary within and between stages as will the precise set of forms produced. The processes change rate and composition, however, in a predictable and regular manner and the forms produced correspond and converge to the suite of forms associated with each stage of development. Davis’s vision is also spatially distinct because he views different parts of his idealised landscape achieving each stage at different times and competing. This view echoes the potential for spatial and temporal dislocation of recapitulation expressed by Haeckel (1866).
30Within the idealised cycle of erosion, the whole landscape does not undergo an equal amount of erosion. The development of subsequent drainage results in competition between streams that affects the original consequent drainage patterns: “ The growth of subsequent streams and drainage areas must be at the expense of the original consequent streams and consequent drainage areas. In other words, the simple consequent drainage of youth is modified by the development of subsequent drainage lines, so as to bring about an increasing adjustment of streams to structures, which is highly characteristic of the mature stage of the geographical cycle ” (Davis, 1899, original italics in bold in Davis, 1954, p. 262).
31Davis is using his vision of a river as an individual as the basis for permitting competition between individuals. The rate of development of each river or river branch is determined by its underlying geology. The relative rate of downcutting is a function of the geology and its resistance. Less resistance strata are worn away faster and headward growth is more rapid. This underlying inequality in the development of rivers results in the elimination of the less efficient rivers and the capture of the headwaters by successful rivers. In this manner Davis introduces a method of river selection to the development of the landscape that is based on a key evolutionary principle, whether Darwinian or Lamarckian: competition between individuals. By introducing this concept, Davis highlights the key role of underlying geological structure in determining the forms produced in a cycle. Landscape development is, therefore, not just directed by the imperatives of the cycle of erosion, but also by the imperatives of underlying geology. Viewed in this light a Davisian landscape has two determinants of its form, its stages of development and the underlying geology. The former determines the suite of forms associated with development, and the latter determines location.
32Within the Davisian cycle of erosion the development of rivers and slopes tends towards a “graded” condition. Davis viewed grade as a goal of the landscape at the scale of individual slopes and rivers. Grade, simplified, represented the equality between ability to do work and work to do. Both are concepts with very subjective definitions. Davis viewed the river as being overcompetent in its youth, having an excess of ability over work. Valley entrenchment resulted. This excess is eventually brought into balance as load increases with the onset of river maturity. Once grade is established, the graded slope of the river can only be attained by altering volume and load relationships, a very slow form of alteration. Such a graded condition does not occur everywhere over the landscape at the same time. Grade conditions move from the trunk stream in early maturity to the smaller headwaters by late maturity until by old age even the wet-weather rills are graded.
33Graded valley sides occur when divide migration and valley floor adjustment are completed, that is, during maturity. Davis is able to extend the concept of grade to the valley sides by extending the definition of the river to include the waste streams from the valley. In this manner the definition of the individual undergoing development is extended to include the whole river catchment. Grade is achieved once the agencies of removal are just sufficient to cope with the weathering products at any point plus the weathering products from further upslope. Graded slopes are initially steep, but decline as material is removed and the supply of material declines as the agencies of removal decrease in intensity. Eventually, in old age a thick cover of weathered debris covers the declining slopes and reduces weathering to zero with no new forms being produced. As with river slope a graded condition is not achieved everywhere at once, but spreads upslope from the base of the slope.
34Even when dealing with the development of individual sections of the catchment, the importance of viewing that portion in relation to the development of the whole catchment is a key concern for Davis. This view of a nested expression of general concepts is discernible in his paper on tornadoes (1891) as an expression of a general tendency for air and matter to move in the same direction at a range of scales from a local tornado to a regional, cyclonic storm to the “polar whirl” and even to the scale of the rings of Saturn or nebula masses: (i) “ It is, then, from the very ancient times when the planets were rings, all turning one way around the sun, that they inherit the common impulse that gives them all the same direction of axial rotation. ... they think that the rings were only annular segregations from a vague, irregularly scattered nebular mass, that, as a whole, turned one way in spiral courses; and that this slow turning of the primeval nebula determined the direction in which the rings revolved, and all the rest from this ” (Davis, 1891, p. 76); (ii) “ When we speak of inheritance, we generally think of the inheritance of property from a parent; and this means that we live in a country of established laws. Established laws give us security in the transmission of inheritance. But how local, short-lived, how vacillating, are our human laws of inheritance compared with these eternal laws of physical inheritance, persistently in operation since the first segregation of chaos ” (Davis, 1891, p. 77-78). Similarly, even at the level of individual slopes and rivers the importance of a goal towards which the form is developing is essential for understanding the nature of and explaining the form. In both cases, the goal represents a form, which is subject to slow, gradual change, although its initial development is marked by rapid adjustment. In keeping with with Dana, Horton, and Lyell, a tendency exists to view the physical landscape as conforming to a predictable form of behaviour. Unlike Dana and Lyell, however, Davis carries this view through different scales, from the whole catchment to individual slopes. Critically, individual elements cannot be understood if divorced from the whole.
35Previous materials might suggest that Davis can be categorized with Dana, Horton, and Lyell in Kennedy’s (1992) terminology as a figure whose work centred on the present state of the world being the norm. Davis, in applying his cycle of erosion, however, did not view the world in the same manner as these writers. The Davisian cycle of erosion, although derived from observations of the current state of the world, does not give it a privileged status. Davis’s focus was rather upon the developmental sequence of which the current state was part. Within that cycle grade occurred. Periods of slow, gradual change occurred, but normal behaviour was only normal for particular stages of the cycle.
36When applying the stage model to the arid environment, Davis (1905) insists on using “normal” processes of fluvial activity to initiate landscape development. Significantly, however, the landscape described is one where time or energy is not sufficient for the “normal” features to develop. Within each stage the description is of a landscape that is following the normal sequence but at a reduced level and intensity. The landscape appears to be missing several key elements that would ensure its “normal” development: “ The processes of drainage adjustment are, however, at the best, of less importance here than in the normal cycle, because of the absence of main valleys, deep-cut by trunk rivers, and the resulting deficient development of deep- set subsequent streams, as has already been suggested ” (Davis, 1905, in Davis, 1954, p. 302). Parallels occur with the ideas of deletion and retrogressive development of Cope (1887) and Hyatt (1866). Within each stage the “normal” process suites are absent and unable to produce the landscape elements characteristic of the same stage in the “normal” cycle. The resulting landscape is empty of a range of forms expected of the normal cycle and the subsequent stage cannot develop progressively because the previous stage was incomplete.
37Davis, it could be argued, still had a very constrained view of landscape development because the cycle predicted stages through which the landscape had to develop. In this manner, his view of time was, as with Lyell and Hutton, that of a cycle. This cycle contained development with the stage model, but the development was not progressive in nature. The stage model did not indicate continuous increased complexity of organisation over time, rather it predicted a decline in indicators such as relief, work, and landscape complexity with the onset of old age. The model was one in which progress was followed by decay unless the cycle was renewed.
38Renewal of the cycle meant that the predictability of landscape forms, as indicated by the idealized cycle, had disappeared. Understanding a landscape, such as the Appalachians, required the application of the principles of the cycle of erosion, but to a landscape that bore the marks of previous cycles. Depending on the stage of the cycle before renewal, the nature of these marks could vary. Importantly, the development of the landscape during the previous cycle was initiated and carried out, however, by individual rivers whose development in turn was determined by geological structure. Interpretation was, therefore, complex and relied on the application of general principles to complex particular circumstances. Although the cycle was predictive, it was used to explain the development and presence of current forms rather than the future landscape. Davis used the model as a tool to explain landscape individuality rather than to present a common future for all landscapes. From this point of view Davis’s views were not concerned with optimality of current conditions, but rather with explaining the unique sequence of changes that resulted in the current set of landforms. Within this model, however, the importance of normal processes and the attainment of graded conditions were central themes. These were, however, coupled with a competitive view of river development and recognition of spatially and temporally differentiated attainment of graded conditions and different developmental levels. Combined, these meant that Davis was more attuned to Darwin, Hutton, and Gilbert in recognising that the sequence of change was a key explanatory element in the landscape.
39The use of the cycle of erosion can be illustrated by reference to Davis’s study of the Appalachians (Davis, 1889b). The original topography of the geographical region was developed in the Permian. Davis suggested that the middle Susquehanna, its branches and upper portions of the Schuykill and Lehigh, are descendants of the original consequent drainage. Although descended from the drainage of the original topography, they have not come through time unscathed or unmodified. In terms of the general course, however, they reflect the topography of that period. The development of the Newark depression reversed the flow of some streams. This, coupled with the adjustments resulting from the several erosional cycles experienced by the region, has destroyed most of the original drainage in the rest of the region. In the lower courses all rivers have altered the original Permian or pre-Permian courses largely as a result of the Newark and Cretaceous superimposition and more recent downward deformation of estuaries. All these geological changes have had a resultant response in the river course and form. Each change has been associated with a range of adjustments by the rivers, such as river capture, which have been instrumental in determining the location of each in its maturity.
40The complexity and skill with which the cycle of erosion needs to be applied and understood to any one case is recognized by Davis. Explaining the development of the course of a real river requires a high degree of conceptualization and piecing together of evidence. He is sure that such complicated cases are not the focus of ordinary study. Instead he believes that this should involve only the simplest case of river development, presumably the cycle of erosion itself. Importantly, he also compares his reconstruction of the development of these Pennsylvanian rivers with the search by biologists for missing links in their developmental chain. Nowhere does he suggest the extension of his cycle to predict future courses. His focus is on using the cycle to fill in the gaps in the past, not to extend the present into the future: (i) “ If this theory of the history of our rivers is correct, it follows that any one river as it now exists is of so complicated an origin that its development cannot become a matter of general study and must remain unhappily only a subject for special investigation for some time to come. It was my hope on beginning this essay to find some teachable sequence of facts that would serve to relieve the usual routine of statistical and descriptive geography, but this is not the result that has been attained. The history of the Susquehanna, the Juniata or the Schuylkill is too involved with complex changes, if not enshrouded in mystery, to become intelligible to any but advanced students; only the simplest cases of river development can be introduced into the narrow limits of ordinary instruction ” (Davis, 1889, in Davis, 1954, p. 482); (ii) “ …as long as we are ignorant of the whole history of our rivers, through which alone present features can be rightfully understood, we must continue to search after the natural processes of their development as carefully and thoroughly as the biologist searches for the links missing from his scheme of classification ” (Davis, 1889, in Davis, 1954, p. 483).
41The Davisian cycle of erosion still retains an important place as an introduction to geomorphological thinking about long-term development of the landscape. Although it is recognized that Davis was not using a strictly Darwinian view of evolution in his analysis, few authors have identified the North American Neo-Lamarckian flavour of his thinking. His conception of the geographical unit and the river as individuals, his extension of that concept to the whole slope and his orthogenetic view of landscape change all point to a Neo-Lamarckian flavour in his view of evolution. His application of this view at different scales in the drainage basin and the frequently explicit progressive, goal-directed view of change he advocated are at odds with the Darwinian image of evolutionary change as random and goalless. Importantly, however, despite the cyclical view of time espoused, Davis manages to give uniqueness in his application of his theory to landscape development. He used his theory not as an ideal the landscape has to follow, but as a tool for understanding how individual, unique landscapes had been produced by the repeated and complicated application of the cycle of erosion and rejuvenation. In this manner the cycle of erosion combined elements of a cyclical view of time with the irreversible and unique image of the arrow of time. The more complicated version of Davis’s cycle of erosion, the cycle of erosion as a descriptive practice, is one that is rarely discussed in textbooks. This more complicated view of Davis’s cycle of erosion is one that meshes more firmly with current thinking in landscape evolution where complexity, developmental pathways and process suites are important concepts. Perhaps it is time for a less whiggish revisiting of the legacy of Davis than that presented by Chorley, Beckinsale and Dunn (1973), and even a recognition that the ideas of Davis may still have something to offer current practitioners of geomorphology.