versión On-line ISSN 1996-7489
versión impresa ISSN 0038-2353
S. Afr. j. sci. vol.110 no.11-12 Pretoria nov./dic. 2014
Daan F. ToerienI; Maitland T. SeamanI, II
ICentre for Environmental Management, University of the Free State, Bloemfontein, South Africa
IIStrategic Cluster: Water Management in Water-scarce Areas, University of the Free State, Bloemfontein, South Africa
Towards the end of the 20th century there were almost 500 small towns of fewer than 50 000 persons in South Africa, accommodating about one tenth of the country's population. Little was known or said in national debates about the future of these places. A decade later this situation had changed and many studies have been or are being undertaken on small towns. For instance, the South African Government recognised that to stem the continued migration from rural to urban areas, a different approach was needed to economic development in rural municipalities and a 'Small Towns Regeneration Project' was initiated. Concerns about a perceived decline of rural towns also stimulated a quest to develop or find methods and/or measures to monitor the well-being of towns. Elsewhere in the world, small and medium enterprise 'observatories' were established to study and report on all aspects of small and medium enterprises, an approach recently followed in South Africa. New ways are needed to improve our understanding of the enterprise dynamics of South African towns. In this contribution, we examine the potential utility of the enterprise richness (i.e. the number of enterprise types) of South African towns and show that enterprise richness has a strong and fully quantifiable relationship with the total number of enterprises in the towns. This contribution adds a new dimension to the capability to make predictions about the enterprise structures of South African towns.
Keywords: enterprise diversity; enterprise architecture; enterprise proportionality
The role of lower-order urban centres in regional development constitutes an important subject of debate1 both internationally and in South Africa, which has seen a dramatic rise in scholarship on small town geographies since 2000.2 A large number of empirical case studies form part of this scholarship. Two broad strands of enquiry related to small towns in South Africa have dominated recent writings.3 The first focused on small town growth and development potential, particularly local economic development, and the second strand focused on various new rural and small town activities relative to debates on post-productivist landscapes.3
A third strand of investigation originated from ideas about the similarities between economic wealth and biological wealth. Beinhocker4 is adamant that economic wealth and biological wealth are thermodynamically the same sort of phenomena. He stated that both are systems of low entropy and patterns of order that have evolved over time under constraint of fitness functions. Like each living organism, each individual enterprise is also in constant competition for survival and only the fittest survive. The similarities between living organisms and enterprises offer the opportunity to transfer lessons learnt from natural ecology to enterprise development in towns. The third-strand studies built on these possibilities.
The latter studies followed two broad themes. Firstly, they explored the interface between natural ecology and regional geography. It was concluded on the basis of norms set for natural ecosystems that towns are 'enterprise ecosystems'.5 Based on the similarities between biological species and enterprises and consideration of the well-known species equilibrium model6 of natural ecology, it was concluded that towns in the Karoo can conceptually be viewed as islands in a sea of farms, and the extent to which entrepreneurs successfully crossed the sea and established enterprises in available entrepreneurial spaces in the towns, determined the enterprise architectures of the latter.7
Secondly, a large number of business sector proportionalities (i.e. statistically significant correlations between the enterprise numbers in specific business sectors and total enterprise numbers of towns) were reported for a large group of South African towns8 as well as for Karoo towns9. Such proportionalities also occurred in the tourism sector of towns in arid and semi-arid South Africa.10 These studies revealed additional ways in which the enterprise dynamics of South African towns could be understood and predicted.
A logical conclusion of Beinhocker's4 views is that the organisms that survive the competitive forces in natural ecosystems determine their biological diversity. Likewise, surviving enterprises determine the enterprise diversity, and hence the enterprise architecture, of towns. Examination of the utility of diversity concepts used in natural ecology could, therefore, be useful in developing an understanding of the enterprise diversity of South African towns.
The biological diversity of the earth has, ever since Wallace and Darwin, been a source of amazement and curiosity, and an area of formal inquiry. Tilman11 remarked:
We still do not know, for example, how hundreds of plant species and thousands of insect species coexist on a hectare of rainforest or prairie, or how millions of species coexist on earth.
This lack of understanding is also broadly true for the coexistence of enterprises in towns.
There are basically three reasons for natural ecologists' interest in ecological diversity and its measurement: (1) it is a central theme in ecology, (2) measures of diversity are often used as indicators of the well-being of ecological systems and (3) considerable debate surrounds the measurement of diversity.12 Magurran12 stated that biological diversity is like an optical illusion: the more it is looked at, the less clearly defined it appears to be.
There is, however, a simple explanation for why biological diversity is so hard to define, which consists of two components: (1) the variety of and (2) the relative abundance of species.12 Biological diversity is, therefore, measured by recording the number of species, by describing their relative abundances or by using a measure that combines the two components.13 The term 'species richness' should be used to refer to the number of natural species in a given area or in a given sample.14 The equivalent term in economic geography is 'enterprise richness', which is used in this contribution.
In the context of enterprise dynamics, questions can be raised about how enterprises coexist in towns. In pursuing this issue we can build on the knowledge and experience of natural ecology. A number of natural ecological diversity indices have been derived using some combination of the number of biological species recorded in natural ecosystems (= species richness) and the total number of individuals summed over all species.12 A widely used diversity index is the Margalef (DMg) Index13:
where S is the number of different species observed and N is the total number of individuals recorded.
The Margalef Index simply states that in natural ecosystems the species richness is a function of the natural log of the total number of individuals present. An ecosystem that is able to carry more individuals will have a greater species richness than ecosystems that can carry fewer individuals and the differences are totally quantifiable. If the same is true for the enterprise richness of South African towns and the relationship can be quantified, a powerful predictor of the link between the total number of enterprises of South African towns and their enterprise richness could be gained.
It is important to note that functional diversity in natural ecosystems differs from species diversity.15 Functional diversity reflects the extent of functional differences among the species in a natural community. An enterprise diversity index, a functional index based on the relative presence of 19 different business sectors in South African towns, was inter alia used to examine functional differences between towns in water abundant and water poor areas of South Africa.16 However, mathematical modelling of the enterprise richness (i.e. enterprise types) of South African towns has not been studied before - a situation which is addressed in this contribution.
The primary purpose of this study was to examine if there is a statistically significant relationship between the number of enterprise types (i.e. enterprise richness and not just functional types) and the total number of formal enterprises present in South African towns.
Over the past three decades, one of the more controversial issues in natural ecology has been the hypothesis that greater diversity and trophic complexity in natural ecosystems increase population and ecosystem stability.17 This concept has been repeatedly challenged1819 but still remains a possibility.11
Are towns with a greater enterprise richness more stable than towns that are less enterprise rich? A possible answer could be derived from comparing different types of towns. Groups of South African towns have been identified based on their enterprise structures.8 A secondary purpose of the study was, therefore, to examine the enterprise richness-enterprise number relationships of different groups of South African towns in order to determine whether such relationships, if they exist, could be used as indicators of town (i.e. enterprise ecosystem) stability.
Methods and results
Enterprise architectures and enterprise richness of towns
In total, 134 towns (Table 1) were examined in this study. There was no logic to the choice of the towns apart from the fact that they were, at the time of writing, part of a database of the enterprise architecture of South African towns. The selected towns were from eight South African provinces (Table 1) and represented a large subset of South African towns.
Town data was obtained as follows. A list of all enterprises in a specific town was prepared from telephone directories as described by Toerien and Seaman5. Each enterprise was allocated to one of 19 business sectors5 (Table 2) by examining and/or 'Googling' its name. If it was impossible to determine the sector in which the enterprise functioned it was ignored in further analyses.5 The enterprise type was then identified. For example, an enterprise with a name such as Union Wholesale Traders would be allocated to the trade services sector and the enterprise type would be recorded as a wholesaler. Salon Isabel would be allocated to the personal services sector and be identified as a beauty parlour. We previously identified more than 500 different enterprise types in our database of South African towns and these were used to allocate the enterprise types. For each town, we recorded: (1) a business profile consisting of the number of enterprises in each of 19 business sectors and the enterprises expressed as percentages of the total number of enterprises per town and (2) the number of different enterprise types (Table 1).
Statistical distribution of enterprise numbers
The selection of 134 towns contained many more small than large towns (Table 1). Statistica 12® software was used to examine the statistical distribution of the enterprise numbers of the selected towns. The median number (73 enterprises per town) was appreciably lower than the average number (138 enterprises per town) and the enterprise numbers were clearly not normally distributed. Further analysis showed that the enterprise numbers of the 134 towns had a log-normal distribution. This distribution suggested that, in order to test a hypothesis that enterprise richness is statistically related to enterprise numbers, an approach similar to that of Margalef could be followed.
Enterprise richness of a large group of South African towns
We related the number of enterprise types observed (as the dependent variable) to the log10 values of the total number of enterprises observed in towns (the independent variable) and then tested the statistical significance of the resulting equation(s). Enterprise richness increased as the enterprise numbers of towns increased (Figure 1); the best-fit relationship was statistically highly significant (p<0.01 and n=134) and 97% of the variance was explained:
Bigger towns not only have more enterprises than smaller towns but also more enterprise types. The increase in enterprise richness is, however, not linear but moderates at higher enterprise numbers (Equation 2), similar to the Margalef Index in natural ecology. As towns grow, their 'entrepreneurial spaces'9 increase in a predictable fashion (or vice versa).
For example, should (for whatever reason) the total number of enterprises of a town increase from 100 to 200, Equation 2 predicts that the enterprise richness will increase from 49 to 81, an increase of 32 types. Similarly, if a town degenerates and its total enterprises reduce from 200 to 100, the enterprise types present will decrease from 81 to 49 types. However, if a town's total enterprises rose from 200 to 300, the change in enterprise types would be from 81 to 108, a mere increase of 27 types (or vice versa). As towns get bigger, the enterprise richness grows at a slower rate than the total enterprises. In other words, the 'entrepreneurial space' for similar enterprise types increases, increasing the likelihood of heightened competition between peer group enterprises. The contrary is also true - if a town degenerates, it sheds entrepreneurial space and enterprises in a predictable fashion.
Close examination of Figure 1 indicates that at higher enterprise numbers the best-fit line does not fit the data very closely. Because smaller towns dominate the group of selected towns (Table 1), it is possible that their influence might have introduced a degree of spuriousness in the enterprise number-enterprise richness response of Figure 1. To test this possibility we divided the selected towns into two groups: Group 1 - small towns (88 towns with fewer than 115 enterprises per town) and, Group 2 - large towns (46 towns with 115 or more enterprises). The relationship between enterprise richness and the enterprise numbers of each of these groups was determined as described earlier and the results are presented in Figures 2 and 3.
The statistical relationships of the two groups, i.e.:
were statistically highly significant (p<0.01) and in excess of 90% of the variance of both groups was explained.
To understand the differences among Equations 2, 3 and 4, we used the equations to predict the enterprise richness of towns of different sizes (Table 3). For lower enterprise numbers, the predictions of the 'all town' (Equation 2) and 'small town' (Equation 3) models were very similar. From 100 to 200 enterprises, the 'all town' model lagged behind the other two models but at higher values it exceeded the 'large town' model considerably. From about 200 enterprises, the 'small town' model started exceeding the 'large town' (Equation 4) model. Below 100 enterprises, the predictions of the 'large town' model exceeded those of the two other models. Taken together, it is clear that town size has some effect and it is advisable to use the small town model for predictions of the enterprise richness of towns with up to 150 enterprises. For predictions for towns with more than 150 enterprises, the 'large town' model should be used.
Clusters of towns and enterprise richness
Although the small town data in Figure 2 clearly shows a strong relationship between enterprise numbers and enterprise types, it is also evident that there is quite a bit of data variation around the best-fit line. The earlier detection of so-called 'proportionality-in-proportionality' phenomena in the enterprise architectures of towns in arid and semiarid South Africa10 raised the question of whether the above variability could be a result of differences in the enterprise architectures of different groups of towns. The proportionality-in-proportionality term refers to a phenomenon in which a selection of South African towns exhibited an overall statistically significant correlation between town size (measured by the number of enterprises) and the number of enterprises in specific business sectors.10 In these cases, specific regression coefficients defined the proportion of the total number of enterprises that sector enterprises constitute. Yet if the larger group was separated into sub-groups of towns with similar enterprise architectures, the above relationships were more clearly defined and distinct differences between the regression coefficients were observed. There was not only proportionality at an overhead level but also proportionality at lower levels; thus proportionality-in-proportionality.
To detect the possible presence of a proportionality-in-proportionality phenomenon in the overall enterprise numbers-enterprise richness relationship established here (Equation 2), the 134 towns were clustered on the basis of their enterprise architectures as described elsewhere.5 Use was made of PRIMER v.6 software20 and a complete linkage clustering strategy was used. Eight clusters (groups) with at least three member towns were identified at a correlation level of 0.55 (Figure 4 and Table 4). To identify differences between the clusters, the enterprise architecture of each cluster was calculated from the total enterprises for each of the 19 business sectors expressed as a percentage of the total enterprises of each cluster (Table 5).
Each cluster received a designation based on the identity of towns clustered in it. 'Highway' towns (Cluster 1) are located on major routes and have strength in the tourism and hospitality, trade and vehicle services sectors. They differ from 'Tourism' towns (Cluster 5) mainly by way of a weaker construction sector but stronger tourism and hospitality and trade sectors. Three clusters have strength in the agricultural products and services sector. 'Small Agri' towns (Cluster 2) differ from 'Large Agri' towns (Cluster 4) and 'Agri Tourism' towns (Cluster 6) in terms of greater strength in the agricultural products and services and vehicle sectors. Large Agri towns are weaker in the agricultural products and services sector and in the tourism and hospitality sector, but have strength in the financial services, trade and vehicle sectors. Agri Tourism towns (Cluster 7) are strong in agricultural products and services and the tourism and hospitality sectors, but weaker in the trade, financial services and vehicle services sectors. Cluster 3 contains a large group of towns that are weak in the agricultural products and services sector, strong in the trade sector and have mostly well-balanced enterprise architectures - they are called 'Trade' towns. Cluster 8 towns are exceptionally strong in the tourism and hospitality sector as well as the real estate services sector but weaker in the trade sector; they are called 'Gentry' towns because gentrification is a significant factor in these towns (e.g. Clarens21 and Prince Albert22).
The clusters differed greatly in their enterprise architectures (Table 5). As a consequence, the clusters met the requirement that different groups of towns could be used to test for proportionality-in-proportionality phenomena in enterprise number-enterprise richness relationships. These relationships and other characteristics were determined for each cluster through the use of Microsoft Excel (Table 6).
With the exception of Cluster 2, which has only small towns, all clusters contained a range of different-sized towns (Table 6). The composition of the enterprise architectures (Table 5) rather than the magnitude of enterprise development in towns defined specific clusters of towns. The statistical significance of the relationships between enterprise numbers and enterprise richness of the different clusters was highly significant in all cases and in no case was less than 95% of the variance explained. Figure 5 shows the enterprise number-enterprise richness relationship of Cluster 4 as a visual example of the goodness of fit obtained when cluster-level analyses were done. There is clearly some proportionality-in-proportionality within the enterprise number-enterprise richness relationship of the 134 towns (Equation 2), which could be discerned by examination at cluster level.
To test if the proportionality-in-proportionality phenomenon would introduce distortions that would have to be considered, the predictive powers of the cluster equations were examined by comparing predictions for a range of town sizes (expressed as the total number of enterprises in towns) and what was limited to the range of town sizes included in each cluster (Table 7). This exercise showed that the predictions stemming from the different cluster equations did not differ by much, particularly for larger towns (see Clusters 3 and 5 in Table 7). The relationship between enterprise richness and total enterprise number seems to hold for whatever group of towns is considered and seems to represent a general property of the enterprise dynamics of South African towns. Although there are subtle differences between clusters, at this time, Equations 3 and 4 should rather be used for predictive purposes.
Discussion and conclusions
There clearly is a quantitative link among the factors that control the growth or decline of towns, the total number of enterprises and enterprise richness. On the basis of the results presented, the hypothesis that the relative abundance of enterprises (= enterprise richness) in South African towns offers no or little potential to serve as a potential indicator of business well-being of towns can be rejected. Enterprise richness is one of the tools that may be useful in assessing the economic health of towns.
The fact that the enterprise richness of South African towns increases with increased town size is not unexpected and could be inferred from earlier studies of South African towns.2324 In the 1960s and 1970s, the fashion (largely driven by a belief in central place theory23) was to construct town classification hierarchies based on the levels of services provided. For instance, Davies and Cook24 identified eight different orders among 601 settlements in South Africa. However, Figures 1 to 3 and Figure 5 indicate that the increase in enterprise richness, and hence the level of business services provided in South African towns, is a continuous rather than a stepwise phenomenon (which the hierarchy approach of Davies and Cook seemed to infer). Our results caution against simply categorising towns on the basis of service levels.
Toerien and Seaman25 applied systems thinking to enterprise dynamics in South African towns. They provided proof that the gross domestic products (GDPs) of towns drive the available money, which drives the number of people in the towns, which drives the total number of enterprises in the towns. Toerien and Seaman8 showed that there are strong proportionalities between the total number of enterprises and the enterprise numbers of a wide range of business sectors in South African towns. This study now adds the fact that the enterprise richness of towns is distinctly and quantitatively linked to the total number of enterprises, which increases our predictive capabilities. The underlying reasons for the observed regularities in the numbers of enterprise types in South African towns are still obscure and deserve further research.
To illustrate the potential value of our results we turn to the burning issue of the potential use of hydraulic fracturing ('fracking') of shales in the Karoo for the exploitation of shale gas.26 The potential benefits or detriments of such exploitation are being strongly debated,27 but without any indication of what Karoo towns may gain or lose in terms of enterprise development. Toerien and Seaman25 quantified some of the expected impacts and Table 3 provides further guidance. For instance, should fracking activities in the vicinity of a Karoo town with 100 existing enterprises result in an increase of 50 new enterprises, the enterprise richness would increase by 19 additional enterprise types. In other words, 31 of the new enterprises would enter business sectors that are already served by one or more existing enterprises. In addition to expanding the business services in the town there would be increased business competition in some business sectors. However, should pollution of groundwater as a result of fracking activities result in a loss of economic activities and 50 enterprises from a town of 150 enterprises, the contrary picture would emerge. Apart from 50 enterprises, 19 enterprise types would disappear, probably to the detriment of the residents. This predictive capability should be factored into considerations of the application of fracking in South Africa.
What views have guided the thinking about the developmental roles of small towns? In 1950 to 1970, small towns were seen as centres for innovation and modernisation in rural areas.28 The concept of 'urban functions in rural development' suggested that a rational rural spatial strategy is to develop a well-articulated, integrated and balanced urban hierarchy.29 Rural development would be promoted by locating more service supply points for a variety of services, agricultural inputs and consumer goods to the rural areas.29 However, this approach was criticised on the grounds that low rural consumption is caused by social inequality and low incomes rather than by access difficulties.30 Southall31 suggested that small towns contribute to rural impoverishment in Africa because they are 'vanguards of exploitation' of the rural poor by external forces which, depending on the case, may be colonial powers, multinational enterprises, central national government, local administrators and élites and, in some cases, international donors. However, Hardoy and Satterthwaite30 cautioned against universal generalisations and prescriptions and suggested that attention should be given to the social dimensions of small towns including the complexity of social networks, kinship and family ties. Later Hinderink and Titus1 also challenged optimistic assumptions about the developmental role of small towns. They suggested that the divergent character of different contexts and differential impacts of regional conditions made generalisations about the role of small towns difficult.
Two core themes in modern literature are relevant in small town development dynamics in South Africa: locality-based development and post-productivism.3 The former has its roots in locality theory and endogenous development theory and indicates that local resources, human assets and partnerships are important. This theme contributed to an emphasis on local economic development strategies.3 This study raises questions about certain assumptions that seem to permeate thinking about local economic development, namely that training of entrepreneurs and promotion of entrepreneurship could solve many developmental problems. Our results emphasise the systemic nature of business development. The degree to which economic value is added through products and services that have a distinct external market is the foremost limiting factor. Entrepreneurial spaces have limits, and unless these limits are expanded, little success will be achieved. To paraphrase a comment made25: you cannot expect champagne outcomes from beer systems. The right leverage points must be sought to achieve success and the quantified insights developed here might be helpful tools in this regard.
The declining role of agriculture in South Africa as shown by a decline in the number of farmers and agricultural workers has contributed to an erosion of traditional livelihoods and the displacement of thousands of people.3 A rise in post-productivist activities in rural South Africa such as the development of farm tourism, game farming and production of products for niche markets (e.g. olive products) have helped to bridge the decline of traditional agriculture3 - a conclusion which is supported by the results presented in Table 5. Nevertheless, the dynamics of enterprise richness of South African towns remain overall the same irrespective of whether towns are productivist or post-productivist (Table 7). The usefulness of the application of the post-productivist label to South African towns needs further elucidation.
A question was posed about enterprise richness in and the stability of South African towns. The fact that enterprise richness is a function of all town clusters identified and that small and large towns clustered together (Table 7), suggests that fewer enterprise types do not necessarily indicate the potential for greater instability. The stability question needs further investigation, perhaps by a greater focus on the functional diversity of enterprises in South African towns.
The financial support of the Centre for Environmental Management, University of the Free State, the help of Marie Watson with the use of PRIMER, help by Falko Buschke, the library support of Annamarie du Preez and Estie Pretorius, and the analytical support of Marie Toerien are gratefully acknowledged.
D.F.T. was responsible for the conceptualisation of the study, the analyses of the enterprise diversity, the interpretation of the results and the writeup. M.T.S. refined some concepts and provided advice on ecological aspects of diversity.
1. Hinderink J, Titus M. Small towns and regional development: Major findings and policy implications from comparative research. Urban Stud. 2002;39(3):379-391. http://dx.doi.org/10.1080/00420980220112748 [ Links ]
2. Donaldson R, Marais L. Preface: Small town geographies. In: Donaldson R, Marais L, editors. Small town geographies in Africa: Experiences from South Africa and elsewhere. New York: Nova Science Publishers; 2012. p. ix-xviii. [ Links ]
3. Hoogedoorn G, Nel E. Exploring small town development dynamics in rural South Africa's post-productivist landscapes. In: Donaldson R, Marais L editors. Small town geographies in Africa: Experiences from South Africa and elsewhere. New York: Nova Science Publishers; 2012. p. 21-34. [ Links ]
4. Beinhocker ED. The origin of wealth: Evolution, complexity, and the radical remaking of economics. Boston, MA: Harvard Business School Press; 2006. [ Links ]
6. MacArthur RRH, Wilson EO. The theory of island biogeography. Princeton: Princeton University Press; 1967. [ Links ]
7. Toerien DF, Seaman MT. Evidence of island effects in South African enterprise ecosystems. In: Mahamane A, editor. The functioning of ecosystems. Rijeka: Intech; 2012. p. 229-248. http://dx.doi.org/10.5772/36641 [ Links ]
10. Toerien DF. Enterprise proportionalities in the tourism sector of South African towns. In: Kasimoglu M, editor. Visions of global tourism industry: Creating and sustaining competitive strategies. Rijeka: Intech; 2012. p. 113-138. [ Links ]
11. Tilman D. The ecological consequences of changes in biodiversity: A search for general principles. Ecology. 1999;80(5):1455-1474. [ Links ]
13. Colwell RK. Biodiversity: Concepts, patterns and measurement. In: Levin SA, editor. The Princeton guide to ecology. Princeton: Princeton University Press; 2009. p. 257-263. [ Links ]
14. Spellberg IF, Fedor PJ. A tribute to Claude Shannon (1916-2001) and a plea for more rigorous use of species richness, species diversity and the 'Shannon-Wiener' Index. Global Ecol Biogeogr. 2003;12:177-179. http://dx.doi.org/10.1046/j.1466-822X.2003.00015.x [ Links ]
15. Petchey OL, Gaston KJ. Functional diversity (FD), species richness and community composition. Ecol Lett. 2002;5(3):402-411. http://dx.doi.org/10.1046/j.1461-0248.2002.00339.x [ Links ]
17. Elton CS. The ecology of invasions by animals and plants. London: Methuen; 1958. [ Links ]
20. Clarke KR, Gorley RN. Primer v6: User manual/tutorial. Plymouth: Primer-E Ltd; 2006. [ Links ]
21. Marais L. From small town to tourism Mecca: The Clarens fairy tale. In: Rogerson C, Visser G, editors. Tourism and development issues in contemporary South Africa. Pretoria: Africa Institute of Southern Africa; 2004. p. 420-435. [ Links ]
22. Toerien D. Prince Albert: A fourth economic bubble or sustainable development? In: Donaldson R, Marais L, editors. Small town geographies in Africa: Experiences from South Africa and elsewhere. New York: Nova Science Publishers 2012. p. 143-162. [ Links ]
23. Van der Merwe IJ, Nel A. Die stad en sy omgewing: 'n Studie in nedersettingsgeografie [The city and its surroundings: A study in settlement geography]. [ Links ] Stellenbosch/Grahamstown: Universiteits-uitgewers en Boekhandelaars; 1975. Afrikaans.
25. Toerien DF, Seaman MT. Paradoxes, the tyranny of structures and enterprise development in South African towns. Presented at: Strategies to Overcome Poverty and Inequality: Towards Carnegie3; 2012 Sep 3-7; Cape Town, South Africa. [ Links ]
29. Rondinelli D, Ruddle K. Urbanization and rural development: A spatial policy for equitable growth. New York: Praeger; 1978. [ Links ]
30. Hardoy JE, Satterthwaite D. Small and intermediate urban centres: Their role in regional and national development in the Third World. London: Hodder and Stoughton; 1986. [ Links ]
31. Southall A. Small towns in Africa revisited. Afr Stud Rev. 1988;31(3):379-391. [ Links ]
Centre for Environmental Management
PO Box 339
Received: 16 Jan. 2014
Revised: 05 Mar. 2014
Accepted: 24 Mar. 2014