Is increased energy utilization linked to greater cultural complexity? Energy utilization by Australian Aboriginals and traditional swidden agriculturalists

Abstract

Theories have been proposed that link increases in energy utilization to increases in cultural complexity. Indeed, available estimates of per capita non-food energy utilization by hunter – gatherers and by people practising swidden agriculture in wooded areas, focusing on fuel wood use, are roughly 1 – 2 orders of magnitude lower than for industrial societies. The latter are in the range of 0.8 – 3.4 × 10⁵ MJ year⁻¹. However, apart from the use of fuel wood, the former estimates have not included work performed by burning vegetation. Here quantitative estimates are given of recent energy utilization linked to burning biomass by Australian Aboriginals and people practising traditional swidden agriculture. Per capita energy utilization linked to biomass burning by Australian Aboriginals is estimated at 1.6 × 10⁶ to 4.0 × 10⁷ MJ year⁻¹. Estimated per capita energy utilization associated with burning biomass in traditional swidden agriculture in the tropical rainforests of Kalimantan and Venezuela, the dry forest of north-eastern Brazil and the miombo woodland of Zambia is in the range of 1.0 × 10⁵ to 6.3 × 10⁵ MJ year⁻¹. The values for non-food energy utilization reported here are at variance with theories that link increases in energy utilization to increases in cultural complexity.

Keywords:

• Hunter – gatherer
• swidden agriculture
• energy utilization
• cultural complexity
• Australian Aboriginals

1. Introduction

Hypotheses have been proposed linking the utilization of energy with the evolution of cultures, suggesting that cultures evolve and can attain greater social or cultural complexity as the per capita utilization of energy increases (White 1959; Orlove 1980; Smil 1991). Bodley (2001) feels that the utilization of fossil fuels and nuclear energy makes a crucial difference in this respect and distinguishes between high-energy cultures, such as the western industrialized societies and low-energy cultures that include the hunter – gatherer and traditional swidden cultures. In swidden agriculture forest biomass is slashed and then burnt, creating a plot for annual crops. After a short period annual of cropping there is regrowth of forest biomass (Conklin 1961; De Jong 1997). Recently, intensity of energy utilization has also been linked to increasing cultural complexity by Chaisson (2001) and Spier (2005). Cultural complexity is not uniformly defined, but definitions of cultural complexity often include population-related aspects such as population density, (maximum) settlement size and permanence of settlement, aspects relating to division of labour and aspects relating to societal integration (Minegal & Dwyer 1998; Denton 2004). If such aspects are considered, there is evidence that other factors than energy utilization may be important for cultural complexity. Evidence from New Guinea has for instance shown that cultural complexity among hunter gatherers (measured in terms of population related aspects) may be dependent on the utilization of protein rich aquatic food resources (Roscoe 2006). It has also been pointed out that specific technologies and trade can be conducive to cultural complexity (e.g. Algaze 2001). Still the possibility that energy utilization is also linked to cultural complexity remains.

This paper deals with the hypothetical link between increased energy utilization and increased cultural complexity, as judged by population density, maximum settlement size, permanence of settlement, division of labour and social integration, by considering non-food energy utilization by Australian Aboriginal hunter – gatherers, traditional swidden agriculturalists and inhabitants of western industrialized countries. If the above-mentioned theories linking increased energy utilization to increased cultural complexity are correct, then per capita energy utilization by Australian Aboriginals and by traditional swidden agriculturalists should be lower than per capita energy utilization in western industrialized countries.

In current physics, energy is defined as the capacity to perform work. It is not self-evident what should be included in accounting for energy utilization by humans. Clearly humans, like all other life, are critically dependent on solar irradiation, as solar irradiation powers photosynthesis, the cycling of water, carbon and nutrients and provides for suitable ambient temperatures. The intercept of solar radiation by the Earth is about 174 260 tera (10¹²) Watt (TW) (Smil 1991). However, solar irradiation is virtually independent of human agency and is therefore in energy accounting usually considered to be ‘free’. The input of energy into the economy by human agency, here called (non-food) energy utilization, is much smaller than the intercept of solar radiation by the Earth. Worldwide this input is currently estimated to be about 12 – 14 TW (e.g. Smil 1991).

In accounting for energy inputs by humans into the economy, one category of energy inputs concerns human work and work by domesticated animals. Another category regards the use of fire. Thirdly, a variety of other physical processes and phenomena can be used to supply energy to economies. These include: nuclear fission, wind power, several varieties of water power, geothermal processes and the use of devices for the conversion of solar energy into electricity and heat.

In this paper the focus will be on energy utilization linked with burning (including burning of harvest residues and the purposeful burning of vegetation) and, as far as industrial economies are concerned, other physical processes involved in energy supply such as nuclear fission and hydropower.

The paper progresses as follows. Firstly, current estimates of energy utilization in industrialized countries by hunter – gatherers and by swidden agriculturalists will be summarized in section 1. Section 2 will consider the purposeful burning of vegetation by Aboriginal Australians. This qualifies as energy utilization because work is performed by such burning. Section 3 deals with energy utilization linked to biomass burning in traditional swidden agriculture. The question of whether per capita energy utilization by Australian Aboriginals and traditional swidden agriculturalists is indeed lower than in western industrialized countries is dealt with in section 4.

1.1 Estimates of energy utilization by industrialized economies

Current (non-food) energy accounting of industrialized economies usually takes into account the purposeful burning of fossil fuels, the burning of biomass (such as fuel wood) to provide heat and other physical processes involved in energy supply such as nuclear fission and the generation of hydro-electricity. Table I gives estimates for per capita energy utilization in some western industrialized countries (Sheffield 1998; EEA 2000). These estimates range between 0.8 × 10⁵ and 3.4 × 10⁵ MJ (= 10⁶ Joules) year⁻¹ per capita.

Table I. Estimates of yearly per capita energy utilization in western industrialized countries (excluding inputs of human energy and burning of harvest residues in the field).

Country	Estimated yearly per capita energy utilization in 1997 (×10^5 MJ)
Austria	1.4
Belgium	2.2
Canada	3.3
Finland	2.5
France	1.8
Germany	1.8
Italy	1.2
Portugal	0.8
Spain	1.1
UK	1.7
USA	3.4
Sources: Sheffield (1998); EEA (2000).

In these estimates of energy utilization in industrial economies the inputs of human and animal work are not included. As to industrial economies, this leads to an error in estimated energy input of about 1% (Smil 1991). The burning of harvest residues ‘in the field’ should in principle be included in energy accounting, as it serves the purpose of generating easily available nutrients and/or disposal of waste, but is also not included in the estimates given in Table I. An estimate of burning of biomass in the field for 1985 suggests that in the industrialized countries of North America, the former Soviet Union and Western Europe the amount of biomass burnt in the field was less than 10% of the amount of biomass burnt in these countries for the generation of heat (Yevich & Logan 2003). The latter refers mainly to the burning of fuel wood and is included in the estimates of Table I. The share of biomass in overall energy supply in the European Union was about 3.7% by the end of the 20th century (EEA 2000). In a few countries (Austria, Finland and Sweden) the share was over 10% (12.4 – 17%). This would mean that the error caused by neglect of burning biomass in the field would usually be less than 1% of the overall estimate for energy input in industrial economies. It may well be that in current estimates of energy inputs into industrial economies the error is smaller than in 1985 because in European Union countries such burning of harvest residues is increasingly prohibited by law (Yevich & Logan 2003).

1.2 Available estimates of energy utilization by hunter – gatherers

Limited empirical studies are available concerning energy utilization by hunter – gatherers. Estimates of the ratio between energy in the food output of hunting and gathering and the energy input in the form of human labour range roughly from 3 to 10 (Sorensen & Leonard 2001). In an empirical study of Fish Creek hunter – gatherers in Arnhem Land Australia, Whitehead (1987) estimated daily fuel wood use for cooking and heating. Such wood use corresponds with a per capita energy utilization of about 8 × 10³ MJ year⁻¹. Weisz et al. (2001) and Haberl (2002) suggest a utilization of firewood by hunter – gatherers in the order of 3.5 × 10³ MJ person⁻¹ year⁻¹. Such estimates for energy utilization are much lower than those for per capita non-food energy utilization in western industrialized countries ( Table I). The estimates of Whitehead (1987), Weisz et al. (2001) and Haberl (2002) for per capita energy utilization linked to fuel wood use by hunter – gatherers are 1 – 2 orders of magnitude lower than (non-food) energy utilization in western industrialized countries.

1.3 Available estimates of energy utilization by swidden agriculturalists

Swidden agriculture has been practised in temperate, subtropical and tropical forest environments in Eurasia, Africa, the Americas and Polynesia (Conklin 1961; Kabo 1985; Carcaillet et al. 2002; Tinner et al. 2005). It has a long tradition, often emerging early in the history of agricultural production in originally forested areas. Swidden agriculture was often combined with selective retention of valuable tree species (Wiersum 1997). Studies of energy utilization in traditional swidden agriculture have focused on the input of human labour and occasionally on the use of firewood (Rappaport 1971; Norman 1978; Uhl & Murphy 1981; Mishra & Ramakrishnan 1982; Sillitoe et al. 2002). The ratio between energy in the food output of swidden agriculture and energy in the labour input may be ∼5 – 54. Mishra and Ramakrishnan (1982) studied fuel wood use in a Khasi village with swidden agriculture. According to these authors per capita firewood use was equivalent to 1.6 × 10³ MJ year⁻¹, which is roughly two orders of magnitude lower than current (non-food) energy utilization in western industrialized countries (see Table I).

2 Estimate of energy utilization linked with purposeful burning of vegetation by Aboriginal Australians

There is evidence that hunter – gatherers in a range of environments did not restrict the use of fire to burning fuel wood for cooking and heating (Bush et al. 1992; Thomas & Kirkpatrick 1996; Moore 2000; Van der Kaars et al. 2000; Boyd 2002; Brockway et al. 2002; Carcaillet et al. 2002; Kirch 2005). Relatively well studied in this context are Aboriginal hunter – gatherers in Australia. Current practices of Australian Aboriginals, which seem largely similar to practices reported in recent centuries (Bowman et al. 2004), suggest that Aboriginals used fire systematically in a purposeful way in north and south Australia (Russel-Smith et al. 1997; Bowman 1998; Yibarbuk et al. 2001; Preece 2002; Verran 2002; Bowman & Prior 2004; Butzer & Helgren 2005; Gott 2005). According to these studies, purposeful burning of vegetation (also called ‘fire regime’) performed work such as cleaning campsites, controlling insects and vermin, stimulating and conserving preferred plants, making preferred tuberous plants more accessible and creating places to attract animal species for the purpose of hunting. Understoreys in bushes were ‘cleaned up’ by fire, presumably for the facilitation of travel and making game visible (Vigilante & Bowman 2004). Vegetation fires have also been used by Aboriginals for signalling, rites and waging war. Additionally, fires were used to hunt down kangaroos, to flush out or entrap small game such as bandicoots, rats and quails, to smoke out animals from burrows and to asphyxiate bats in caves. In flood plains fire set to vegetation was used to expose snakes, lizards, rodents and turtles.

So far no quantitative estimates have been made concerning the energy utilization associated with the purposeful burning of vegetation in hunter – gatherer societies.

However, in Australia several studies have been performed that allow for estimates of such energy utilization. Yibarbuk et al. (2001), Bowman et al. (2004), Vigilante et al. (2004) and Gott (2005) considered contemporary practices. Ward et al. (2001) studied practices pre-dating contact with Europeans.

On the basis of these studies, while excluding fuel wood utilization, high and low estimates for non-food energy utilization (E: in MJ person⁻¹ year⁻¹) can be calculated, according to equation 1:

where E is the energy utilization in MJ person⁻¹ year⁻¹, f is the frequency/year that the per capita area is burnt, F is the fuel load in kg km⁻², e is the energy content of fuel (in MJ kg⁻¹), and p ⁻¹ is the inverse of population density (in km²/person).

The values for the parameters applied when using Equation 1 and the high and low estimates of energy utilization are given in Table II.

Table II. Estimate of yearly per capita energy utilization linked to purposeful burning of vegetation by Australian Aboriginals.

Parameter: abbreviation	Parameter	Low estimate	High estimate	Source
F	Frequency of burning (year^−1)	0.2	0.5	Yibarbuk et al. (2001); Ward et al. (2001); Bowman et al. (2004); Vigilante et al. (2004)
f	Fuel load (kg km^−2)	4×10^5	5×10^5	Yibarbuk et al. (2001); Saarnak (2001)
e	Energy content of fuel: grass, twigs, leaves (MJ kg^−1)	15	20	Smil (1991); Woodard and Pine (1993); Srivastava (1995); Neitzke (2002)
p ^−1	Inverse of population density (km^2/person)	1.3	8	Lourandos (1980); Russel-Smith et al. (1997)
E	Energy utilization (MJ person^−1 year^−1)	1.6×10^6	4×10^7

Overall non-food energy utilization can be obtained by adding fuel wood utilization: 8 × 10³ MJ person⁻¹ year⁻¹ as estimated by Whitehead (1987) to the estimates of energy utilization in Table II. Energy utilization associated with purposeful burning of biomass by Australian Aboriginals can thus be estimated at 1.6 × 10⁶ to 4 × 10⁷ MJ person⁻¹ year⁻¹. Roughly, this is 1 – 2 orders of magnitude larger than the values for per capita (non-food) energy utilization in western industrialized countries given in Table I.

3 Estimates of energy utilization linked to biomass burning in traditional swidden agriculture

Mishra and Ramakrishnan (1982) stated that burning slash in the field, characteristic for swidden agriculture, is ‘free’ to the farmers and thus should not be included in calculations of energy utilization. However burning slash in the field performs work. One of its purposes is adding readily available nutrients to the soil. Levels of readily available nutrients such as nitrate, potassium, phosphate, magnesium and calcium may increase in soils by burning (Uhl & Jordan 1984; Montagnini & Buschbacher 1989; Kauffman et al. 1993; Hölscher et al. 1997; Giardina et al. 2000), although it should be noted that there are large overall losses of nutrients to air and by leaching from ecosystems due to swidden practices (Kauffman et al. 1993; Hölscher et al. 1997; Giardina et al. 2000). Nitrogen fixation that adds to nutrient stocks may also be stimulated by burning vegetation (Gonzalez-Perez et al. 2004). Additionally, burning slash performs work such as reduction of weeds and pests (e.g. insects), disease control, improved water infiltration, increase in soil pH and improved access for sowing (Unruh 1988; Sillitoe et al. 2002).

Swidden agriculture is currently practised by hundreds of millions of humans mainly in the tropics. It covers a wide variety of practices and is often combined with other economic pursuits (Dove 1983; Porro 2005). The contribution of swidden agriculture to the overall supply of food varies widely, and so do cultivation and fallow practices. Here ‘traditional’ forms of swidden agriculture will be considered that were practised at low densities of human population. Such traditional forms of swidden agriculture were characterized by one or a few years of annual cropping followed by long fallows (at least 10 years) or abandonment (Conklin 1961; Toky & Ramakrishnan 1982; Chidumayo 1987; Inoue & Lahjie 1990; Roder et al. 1995; Cairns & Garrity 1999; Lambin et al. 2003).

Several studies have appeared that can be a basis for estimates of energy utilization linked to vegetation burning by traditional swidden agriculturalists, as they give information about per capita plot size, length of cultivation and the amount of burnt biomass. These are the studies of swidden agriculture in north-eastern India by Mishra and Ramakrishnan (1982) and Toky and Ramakrishnan (1983), on Kalimantan (Indonesia) by Dove (1983) and Sulistyawati et al. (2005), in Venezuela (Amazonia) by Uhl (1987) and Uhl et al. (1988), in Zambia by Chidumayo (1987) and in a dry forest area in north eastern Brazil by Kauffman et al. (1993).

Based on these data, per capita non-food energy utilization linked to burning slashed vegetation in the field can be calculated as indicated in Equation 2

where E is the energy utilization in MJ person⁻¹ year⁻¹; F′ is the fuel load in kg ha⁻¹; h is the hectare cropping area in ha per person; e is the energy content of fuel (in MJ kg⁻¹) c is the conversion factor: part of fuel actually burned (0 < c < 1); and y ⁻¹ is the inverse of the number of years that a plot is under annual cropping before abandonment or forest fallow.

To obtain total energy utilization linked to burning biomass, the use of firewood should be added. The latter is estimated to be about 1.6 × 10³ MJ person⁻¹ year⁻¹, in line with the study of Mishra and Ramakrishnan (1982).

Table III gives estimates for energy utilization linked to burning biomass by a number of swidden agriculturalist groups that have been studied in sufficient detail.

Table III. Estimates of yearly per capita energy utilization linked to burning biomass (vegetation, fuel wood) for specified groups of swidden agriculturalists.

Group	Estimated energy utilization (MJ person^−1 year^−1)	Source	Assumptions
Swidden agriculturalists in tropical rainforest of Venezuela	3.9×10^5	Uhl (1987); Uhl et al. (1988)	Per capita area 0.25 ha, burnt biomass 260 tons ha^−1 of 20×10^3 MJ ton^−1 with 90% efficiency; years under cultivation: 3
Swidden agriculturalists in dry forest of Northern Brazil	1.0 – 3.3×10^5	Kauffman et al. (1993)	Per capita area 0.5 ha, burnt biomass; 74 tons ha^−1 of 20×10^3 MJ ton^−1 with 90% efficiency; years under cultivation: 2 – 7
Kantu' swidden agriculturalists in the tropical rainforest of South Kalimantan	3.1 – 6.3×10^5	Dove (1983); Sulistyawati et al. (2005)	Per capita area: 0.44 ha; burnt biomass 80 tons ha^−1 of 20×10^3 MJ ton^−1 with 90% efficiency; years under cultivation: 1 – 2
Khasi swidden agriculturalists in north-eastern Indian montane forests	1.8×10^4	Mishra and Ramakrishnan (1982)
Chitemene agriculturalists in miombo woodland Zambia	2.1×10^5	Chidumayo (1987)	Per capita area: 1.1 ha, burned biomass 21.14 tons ha^−1 of 20×10^3 MJ ton^−1 with 90% efficiency; years under cultivation: 2

Estimates for utilization of energy from biomass by swidden agriculturalists in the tropical rainforests of Kalimantan and Venezuela, the dry forest of north-eastern Brazil and the miombo woodland of Zambia are in the range of 1.0 × 10⁵ to 6.3 × 10⁶ MJ. The estimate of yearly energy utilization by Khasi swidden agriculturalists is about an order of magnitude lower.

4 Discussion

Tables I –  III show that, if compared with western industrialized countries, the levels of per capita energy utilization estimated for Australian Aboriginals and for traditional swidden agriculturalists in tropical rainforests of Kalimantan and Venezuela, the dry forest of north eastern Brazil and the Miombo woodland of Zambia are at variance with the hypothetical link between increased energy utilization and increased cultural complexity discussed in the Introduction. As judged by (maximum) settlement size, permanence of settlement, population density, division of labour and societal integration, western industrialized societies show greater cultural complexity than traditional swidden agriculturalists and Australian Aboriginal hunter – gatherers, but their per capita energy utilization is lower than the per capita energy utilization of Australian Aboriginals and of the same order of magnitude as per capita energy utilization by the traditional swidden agriculturalists in Kalimantan, Venezuela, Brazil and Zambia discussed here. Also, if increased energy utilization and increased cultural complexity are linked, one would not expect that energy utilization by Australian Aboriginals is much higher than energy utilization by swidden agriculturalists.

This conclusion is not affected by the failure of current estimates of energy utilization in industrialized counties to include the burning of harvest residues. As pointed out in the Introduction this leads to an error that will be usually less than 1%. However, there are two other possible objections to the estimates of energy utilization presented here. The first is that the estimates of energy utilization by Australian Aboriginals and swidden agriculturalists are underestimates because only burning of above-ground biomass is considered. The second objection is that the practices of Australian Aboriginals and swidden agriculturalists that underlie the estimates of sections 2 and 3 may be very different from practices in the more distant past, just as the potlachs (festivals) of the Kwakiutl (American Northwest Coast Indians) that were described in the 19th and 20th centuries were very different from potlachs pre-dating European arrival in terms of the amount and nature of goods distributed and the use of deliberate destruction (Jonaitis 1992). Section 4.1 will discuss the matter of whether the neglect of soil carbon in sections 2 and 3 leads to an underestimate of energy utilization. Section 4.2 raises the question of how far in the past fire regimes can be inferred.

Section 4.3 confronts energy utilization by Australian Aboriginals and traditional swidden agriculturalists as estimated here with objections against using the term ‘energy utilization’ for purposeful burning of vegetation by hunter – gatherers (McDaniel & Borton 2002) and the claims that traditional swidden agriculture is ecologically sustainable (Dufour 1990; De Jong 1997) or environmentally safe (Ramakrishnan 1993). Section 4 ends with concluding remarks.

4.1 Soil organic carbon

In sections 2 and 3 the focus was exclusively on burning above-ground biomass. The fate of soil carbon has been neglected. At high temperatures, soil organic carbon may be burned and as this contributed to easily available nutrients this may be considered work, and thus contribute to energy utilization. There is only limited empirical evidence to address this matter. In a study of the impact of slash-and-burn practices on the carbon stocks of savannah Vertisoils in Cameroon Obale-Ebanga et al. (2003) found an increase in soil carbon. As for slash-and-burn practices in forests, it is known that dependent on soil type, slope and temperature of the fire there may or may not be losses of soil organic carbon (Garcia-Oliva et al. 1999; Gonzalez-Perez 2004). It would seem that burning of soil carbon would be likely when fires are intense, leading to high soil temperatures. Reported fires used by hunter – gatherers and swidden agriculturalists are usually of relatively low intensity. Thus, it seems unlikely that the data presented in Tables II and III are underestimates due to neglect of the fate of soil carbon, except in cases that high intensity fires are used.

4.2 How far in the past can fire regimes be inferred?

The purposeful practices of Australian Aboriginals considered here are contemporary or date from recent centuries. Thus, the question arises how far into the past such a fire regime can be inferred. It has been argued that when the population density was lower, fire use may have been different (Russel-Smith et al. 1997). One may also consider the possibility that Australian Aboriginals may have been confined to more marginal (energy restricted) environments and that this would have influenced energy utilization. Changes in climate may too have led to changes in fire regimes (McKenzie & Kershaw 2000). Burning practices may have been dissimilar too when there was more megafauna (Roberts et al. 2001; Miller et al. 2005). Then, hunting practices may well have differed from recent practices and the presence of large herbivores may have led to differences in the presence of flammable biomass (Russel-Smith et al. 1997; Bond & Keeley 2005). All in all, Butzer and Helgren (2005) feel that the systematic burning noted in the nineteenth century and shown by contemporary practices, that underlies the estimate of this paper, can only be inferred for the late Holocene. However, studies by Genever et al. (2003) and Bickford and Gell (2005) suggest that recent fire regimes may have a longer history, because during the mid and late Holocene there are roughly constant levels of vegetation burning in the area of Whitehaven Swamp (Queensland) and in the upland wetlands on the Fleurieu Peninsula in South Australia, as evidenced by charcoal records.

Purposeful burning of vegetation by hunter – gatherers was and is not restricted to Australia. This practice was probably current amongst hunter – gatherers in a range of environments (Kabo 1985; Thomas & Kirkpatrick 1996; Moore 2000; Van der Kaars et al. 2000; Boyd 2002; Brockway et al. 2002; Kirch 2005). There is, for instance, evidence suggesting purposeful vegetation burning by Mesolithic hunter – gatherers in Europe, presumably to facilitate hunting and favouring preferred plant species (Mason 2000; Innes & Blackford 2003). In the past, use of fire has probably also been made by hunter – gatherers in the Americas and Africa, presumably to increase the abundance of preferred plants, as a means to attract game to, or away from, certain locales and for purposes of pest reduction, clearing areas for home sites, signalling and warfare (Bush et al. 1992; Boyd 2002; Brockway et al. 2002; Kirch 2005). This practice has probably led to the expansion of grasslands (Brockway et al. 2002; Kirch 2005; Bond & Keeley 2005). However, the author found no data that allow for quantitative estimates of energy utilization linked to (present or) past vegetation burning by other hunter – gatherers than Australian Aboriginals.

Whether traditional swidden agriculture associated with high levels of energy utilization as described here, is similar to swidden agriculture in forested areas during a more distant past is uncertain. One study concerning swidden agriculture in the north-eastern Cambodian monsoon forests, suggests on the basis of a sediment charcoal record that practices may have been roughly the same over the last 2500 years (Maxwell 2004). There are also other arguments for the point of view that swidden practices with levels of energy utilization in the same order of magnitude as current energy utilization by swidden agriculturalists considered to be traditional may well have occurred in the more distant past. It is likely that a relatively short period of annual cropping, abandonment and long fallows, practices that underlie the estimates of Table III, were common when population pressure was low (Roder et al. 1995; Coomes et al. 2000; De Clerck & Negreros-Castillo 2000; Prasad et al. 2001). Also, in view of sustainable use, long fallows are often needed in tropical rainforests to restore nutrient stocks that are depleted due to burning and harvesting (Hölscher et al. 1997). In temperate areas with relatively poor soils and in dry (sub)tropical areas, where carbon sequestration during forest fallows is lower than in tropical rainforests, forests fallows much exceeding 30 years may well have been necessary to maintain soil fertility indefinitely (Otto & Anderson 1982; Kauffman et al. 1993; Giardina et al. 2000).

All in all, there is no evidence that the practices of Australian Aboriginals and traditional swidden agriculturalists that underlie the estimates in Tables II and III are very different from practices in the more distant past.

4.3 Energy utilization, impact on living nature and sustainability

There has been an objection against using the term ‘energy utilization’ to denote purposeful burning of vegetation by hunter – gatherers because hunter – gatherers are different from humans participating in western industrialized societies in not substantially co-opting net primary production from other organisms or reducing the productivity of ecosystems, thereby evading biological diversity loss that is linked to current human energy utilization (McDaniel & Borton 2002). And regarding swidden agriculture, it has been stated that the traditional practices that are considered here are ecologically sustainable (Dufour 1990; De Jong 1997) or environmentally safe (Ramakrishnan 1993).

Purposeful burning of vegetation performs work and thus, as pointed out in the Introduction, qualifies as energy utilization. Burning of vegetation does also have environmental impacts. Purposeful vegetation burning by hunter – gatherers has changed Australian ecosystems greatly and may have contributed to the extinction of some plant and animal species (Bowman 1998; Lunt 1998; Woinarski et al. 2004; Bickford & Gell 2005). The agro-ecosystems that originate in swidden practices have lower levels of carbon sequestration than the forests that they replace (Houghton 1991; Prasad et al. 2001; Palm et al. 2004). There is speculation that this difference may have contributed to the increase in atmospheric carbon levels that occurred between 7000 and 1000 years before present (Carcaillet et al. 2002) and to an increase in atmospheric methane levels from 1500 to 1700 AD (Ferretti et al. 2005). Past swidden practices in Polynesia were presumably an important factor in the extinction of a number of bird species (Steadman 1989) and swidden practices in central Amazonia may reduce biodiversity in spite of long fallows (Gehring et al. 2005).

It should also be noted that similarity in energy inputs in the economy does not equal similarity in sustainability, when sustainability refers to ‘a steady state’ economy that allows for indefinite continuation of current practices regarding the use of biomass (Reijnders 2006).

Industrial countries currently depend overwhelmingly on fuels that have been formed in slow geological processes, such as fossil fuels and uranium. Current depletion of such resources greatly exceeds addition to stock. For instance: conventional oil resources have been formed over a period of about 400 million years, and current depletion is at a yearly rate that equals about one million years of resource formation (Patzek & Pimentel 2005). Such inputs cannot be continued indefinitely and are thus not sustainable.

On the other hand, swidden agriculture in tropical forests may well be sustainable. As pointed out in section 3, burning practices lead to loss of nutrients from the combined above- and below-ground ecosystem due to volatilization and leaching. Nutrients may be lost too when nutrients present in harvests are not returned to cultivated soils (Unruh 1988). Also, as pointed out in section 4.1, intense burning may reduce organic carbon levels in soils. Losses of nutrients and soil organic carbon may negatively affect future productivity (Corbeels et al. 2005). However, both nutrient and soil organic carbon levels can be restored under fallow. When a plot of land is left fallow in such a way that nutrient and organic carbon levels are restored, indefinite continuation of swidden practices at the same level of productivity is possible. The actual length of the fallow that is necessary to restore nutrient and soil organic carbon levels is dependent on a variety of factors, including burning practices, nature of the soil, erosion, the absence or presence of flooding by a river, deposition of nutrients from air, climate and nature of crops and fallow (Fujisaka 1991; Zarin et al. 1998; Sillitoe & Shiel 1999).

Loss of nutrients due to volatilization also occurs when, for example, savannah vegetation is burned by hunter gatherers. In this case again restoration of nutrient levels is possible when vegetation regrows. A study in an open cerrado (Brazilian savannah) has, for instance, suggested that after burning nutrient levels will be restored after 3 years (Pivello & Coutinho 1992).

4.4 Concluding remarks

The estimates presented here for per capita energy utilization by traditional swidden agriculturalists, Australian Aboriginals and western industrialized countries are at variance with current hypotheses that link increased energy utilization to increased cultural complexity. On the other hand, it is hard to imagine that energy utilization would not have any impact on cultural complexity at all. It is beyond the scope of this paper to come up with a new hypothesis. But it should be noted that the efficiency with which primary energy (e.g. fuel) is converted into work may be a determinant of cultural complexity. In this respect, there are large differences between practices of Australian Aboriginals and the swidden agriculturalists considered here and practices in western industrialized countries. For instance, burning vegetation provides readily available nutrients in swidden agriculture, and so does the use of manure in current organic agriculture. But the latter is much more energy efficient in providing agriculture with nutrients than traditional swidden agricultural practices (Pimentel et al. 2005). Also, current compact fluorescent lamps are about three orders of magnitude more fuel efficient in providing lighting than burning biomass (Fouquet & Pearson 2006).