Response of African savanna ants to long-term fire regimes


S.L. Chown, Department of Zoology, University of Stellenbosch, Private Bag X1, Matieland 7602, South Africa (fax + 27 21 8082405; e-mail


  • 1Despite the fact that fire is considered an important disturbance in savannas across the world and is used widely as a management tool in conservation areas, little is known about the effects of burning on their insect communities.
  • 2This study made use of a 50-year fire experiment to investigate the responses of ant assemblages to long-term burning regimes. The effects of fire frequency, season and time since fire (fuel age) were tested on epigaeic ants across three savanna vegetation types (Mopane woodland, Acacia savanna and Terminalia woodland) in Kruger National Park, South Africa.
  • 3There was no significant effect of burning on mean ant species richness and abundance between treatments, although there were significant differences in ant assemblage composition between burned (treatment) and unburned (control) plots. The effects of season, frequency of burn and plot age on assemblage structure were weak and often not significant.
  • 4Epigaeic ant assemblages in this savanna system appeared to be highly resistant and resilient to burning. The response of ants to fire was linked to changes in habitat cover and structure: the effect of fire on vegetation and ants was less pronounced in lower rainfall areas, where differences in vegetation structure between burnt and unburnt plots were less pronounced than in higher rainfall areas.
  • 5Synthesis and applications. The effect of fire on ant assemblages in the mid- to northern Kruger National Park depends on whether a patch burns or not, rather than the specifics of the burning treatment. Thus, conservation managers can focus concerns regarding the subtleties of fire regimes on other taxa or areas of particular concern.


In southern African savannas, fire is regarded as a major disturbance. Fire modifies broad patterns in the abundance and distribution of vertebrates and plants, which are determined predominantly by rainfall and edaphic factors (O’Brien, Field & Whittaker 2000; Van Rensburg, Chown & Gaston 2002), and thus plays a crucial role in determining system structure and function (Scholes & Walker 1993; Van Langevelde et al. 2003). While the effects of fire on vegetation are reasonably well documented in southern African savannas (Tainton & Mentis 1984; Bond 1997; Trollope et al. 1998b), for animals, and invertebrates in particular, there are few studies (Trollope 1984; Parr & Chown 2003) and results have been mixed (Ferrar 1982; Gandar 1982; Parr, Bond & Robertson 2002).

The lack of understanding of fire effects is of particular concern given that fire is widely used as a management tool in many of the region's protected areas, where the primary aim has either been the management of large game or, more recently, the conservation of biodiversity as a whole (Mentis & Bailey 1990; Van Wilgen, Biggs & Potgieter 1998; Brockett, Biggs & Van Wilgen 2001). Both within reserves and in surrounding extensive livestock farming areas, fire is used for a variety of purposes (Trollope 1982; Edwards 1984; Hough 1993). Managers of protected areas alter the fire regime by deciding when, where and how fires should be lit. If the effects of these alterations on a variety of taxa are not well understood, then neither is their role in conserving biodiversity as a whole (Parr & Brockett 1999; Keith, Williams & Woinarski 2002).

Unfortunately, many fire studies in southern Africa, as elsewhere, tend to be short term and small scale, while in practice fire policies in protected areas tend to be applied to larger areas for longer periods. Moreover, many fire studies use only a limited range of fire regimes to test community responses, use fire regimes that bear no resemblance to naturally occurring fires, or focus on responses to single fires (Parr & Brockett 1999; Parr & Chown 2003). Because the ecological effects of a single fire or the cumulative effects resulting from the imposition of a fire regime may be quite different (Bradstock, Keith & Auld 1995; Moretti et al. 2002), it is important that fire studies make a careful distinction between the two. For studies investigating fire regime effects, long-term fire studies are particularly valuable because they are more likely to detect long-term ecological responses (Andersen et al. 1998). Experimental limitations in short-term studies thus often compromise the robustness of findings (Orgeas & Andersen 2001). Moreover, if the studies themselves and the fire regime applied to an area differ considerably, then the utility of the experimental results for management purposes is questionable.

Given that there is a need to provide conservation managers with reliable information on the effects of burning on invertebrates, this study made use of a long-term fire experiment to investigate the responses of an ant assemblage to a long-term burning regime. The experiment, initiated in 1954 in the Kruger National Park, South Africa, was originally established to test the effect of different seasons and frequencies of burn on vegetation (Brynard 1964). The experimental treatments were established in four major vegetation types across the park (Trollope et al. 1998a; Biggs et al. 2003), and within each vegetation type different fire treatment plots (14 different season and frequency combinations) were replicated four times. These treatments are still maintained. Although the fire experiment was designed with the predominant focus on fire regime effects, in this study the effects of individual fires were also considered.

The aims of the current study were to (i) determine what effects season and frequency of burn and post-fire fuel age (time since fire) have on ant species richness and relative abundance; (ii) compare the response of ant assemblages to fire in the different vegetation types, and (iii) assess whether any changes in ant assemblage structure can be linked to differences in vegetation structure and habitat composition. We were particularly interested to determine the extent of the resistance (i.e. the extent of change following a disturbance) and resilience (i.e. the rate of return to a pre-disturbance state) of the ant assemblages to burning (Pimm 1984).


study area

The study was conducted in three savanna habitats, Mopane woodland (Mopane area), Acacia savanna (Satara area) and Terminalia woodland (Pretoriuskop area). Ant sampling was carried out on experimental burn plots that form part of the long-term burning experiment. Although there were four replicates within each vegetation type, because the fourth replicate differed in soil type from the other three it was considered unrepresentative (Venter 1999). Therefore, this study only focused on three replicates, each situated 10–20 km apart. The replicates in the Mopane area were Tsende (23°41′S 31°31′E), Mooiplaas (23°34′S 31°27′E) and Dzombo (23°26′S 31°22′E); in the Satara area N’wanetsi (24°26′S 31°51′E), Marheya (24°32′S 31°46′E) and Satara (24°24′S 31°45′E); and in the Pretoriuskop area Kambeni (25°15′S 31°26′E), Numbi (25°13′S 31°20′E) and Shambeni (25°12′S 31°23′E). Each replicate was divided into 12–14 plots laid out in a strip. Each plot measured approximately 380 × 180 m (7 ha) and represented a different burning regime (season and frequency combination; Trollope et al. 1998a). Fire-break roads separated the plots in each replicate. In the Mopane and Satara areas, sampling was carried out on the following burn plot treatments: August annual, August biennial, August triennial, April biennial, April triennial and, importantly, an unburnt control that had remained unburnt since 1954. The exception to this was one of the three Satara area control plots, which was accidentally burnt in April 2001. August burns represent late winter burns at the end of the dry season, while April burns represent autumn burns at the end of the wet season. In the Pretoriuskop area, because of sampling constraints, ants were sampled on two burn plot treatments only: the unburnt control plot and an August annual plot.

Additional sampling was also carried out adjacent to each replicate in the general landscape to assess the effect of the general burning regime (usually spring fires) that had been applied by Kruger National Park management; these were referred to as ‘variable’ plots but were not considered controls. They were neither prevented from burning nor subjected to a specific experimental burning regime. Fire records for the Mopane and Satara areas (Kruger National Park Scientific Services, unpublished data) indicated that the variable plots were last burnt at least 4 years prior to sampling and had burnt very infrequently (i.e. twice in the last 40 years). Two variable plots (adjacent to the Dzombo and Satara replicates) were older than 20 years. The variable plot for the N’wanetsi replicate burnt in 2001.

The Mopane plots were situated in the Mopane shrubveld, a mopane (Colophospermum mopane J. Kirk ex J. Léonard)-dominated habitat with few other woody species (Low & Rebelo 1996). The altitude of the area ranges from 300 to 340 m a.s.l., and mean annual rainfall from 450 to 500 mm (Gertenbach 1983). The Satara plots were situated in a mixed knobthorn Acacia nigrescens Oliv. and marula Sclerocarya birrea (A. Rich. Hochst.) savanna in the sweet lowveld bushveld (Low & Rebelo 1996). Mean annual rainfall here is 550 mm, and altitude ranges from 240 to 320 m a.s.l. (Gertenbach 1983). Satara and Mopane areas share basalt-derived clay soils (Gertenbach 1983). The vegetation of the Pretoriuskop area is sour lowveld bushveld (Low & Rebelo 1996). This is an open tree savanna dominated by silver clusterleaf Terminalia sericea Burch. ex DC but also with bushwillow Combretum collinum Fresen. Soils are sandy and granitic-derived. The area has a mean annual rainfall of 700 mm and altitude ranges from 560 to 640 m a.s.l. (Gertenbach 1983).

ant sampling

Epigaeic ants were collected by pitfall trapping during two sampling periods: November–December 2000, and January–February 2002 (hereafter sampling periods 2000 and 2002). In the Pretoriuskop area, sampling was carried out in January and February 2002 only. In the summer rainfall region of South Africa, ants are most active and abundant at this time (Swart, Richardson & Ferguson 1999). Collections could not be made in November and December 2001 because heavy rains prevented sampling. Pitfall trapping was determined to be the optimum sampling method based on a pilot study comparing winkler and pitfall sampling (Parr & Chown 2001).

On each plot, 20 pitfall traps (Ø 62 mm) were laid out in a grid (5 × 4) with 10-m spacing between traps. All grids were situated at least 50 m from the plot edge to reduce the possibility of ants from adjacent areas being collected in the traps, and to reduce edge effects. Pitfall traps contained 50 mL of a 50% solution of propylene glycol, which neither significantly attracts nor repel ants (Adis 1979). The pitfalls were left to settle to reduce the ‘digging-in’ effect (Greenslade 1973) and then all traps were opened for a period of 5 days. A pilot study indicated that this period was sufficient for reasonably complete sampling, without collecting excessive numbers of ants. There was no rain during the periods that the pitfall traps were open, and the weather during both sampling periods was hot and humid. Pitfall samples that had been disturbed by animals were excluded from the analyses.

Samples were washed and sorted in the laboratory. Whenever possible ants were identified to species, otherwise they were assigned to morphospecies. Voucher specimens are held at the Iziko Museum of Cape Town, Cape Town, South Africa.

vegetation sampling

To assess the overall effect of fire regime on vegetation, vegetation foliage height profiles (for the 2002 sampling period) were determined based on the methods discussed by Rotenberry & Wiens (1980) and Bestelmeyer & Wiens (1996). Vegetation height was measured at four points located 90° apart on a 1·5-m radius centred on each pitfall trap. At each point, a 1·5-m long pole was placed vertically, and the number of times vegetation came into contact with the pole in each height class (1 = 0–0·25 m, 2 = 0·26–0·50 m, 3 = 0·51–1·00 m, 4 = 1·00–1·50 m) was recorded. Ground cover was estimated on each plot by placing a 1-m2 quadrat next to each pitfall and estimating the percentage cover of grass, bare ground, litter and dead grass and forbs.

analyses: ant assemblages

Total species richness and abundance were compared between and within habitat types for both sampling periods using analysis of variance (anova), and for each habitat type species richness and abundance were also compared between sampling periods. anovas were used to determine if there were any significant differences in species richness and abundance between replicates and burn plot treatments for each sampling period separately, and combined. This was done for the Mopane and Satara areas separately. Because the use of anova requires that data are normally distributed, data were log transformed where necessary. For data that could not easily be transformed, non-parametric tests, Mann–Whitney U and Kruskal–Wallis anova, were applied to the data. Sequential Bonferroni corrections were applied to adjust the statistical significance for multiple tests (Rice 1989).

In addition to the richness values obtained from the sampled ants, the non-parametric incidence-based coverage estimator (ICE) provided in EstimateS (Version 5; Colwell 1997) was used to improve the estimate of species richness per plot. ICE is a promising, and reliable, estimator of species richness (Chazdon et al. 1998; Longino, Coddington & Colwell 2002) because it stabilizes fairly well, and provides an estimate independent of sample size. ICE is based on the number of species found in 10 or fewer sampling units (Lee & Chao 1994). While estimators are valuable tools they should not be viewed uncritically, and yield minimum estimates of species richness (Longino, Coddington & Colwell 2002).

Multivariate community analyses were undertaken using primer v.5.0 (Clarke & Gorley 2001) to assess overall changes in ant assemblage composition. Cluster analyses using group averaging and Bray–Curtis similarity measures were used to determine whether ant assemblage structure varied between years, and within- and between- habitats. Data were fourth-root transformed prior to analyses to reduce the weight of common species. Analyses of similarity (anosim) were used to establish if there were significant differences in the ant assemblages on plots that differed in burn season (August, April, control and variable), frequency (annual, biennial, triennial, control and variable) and age (i.e. time since fire). The anosim procedure of primer is a non-parametric permutation procedure applied to rank similarity matrices underlying sample ordinations (Clarke & Warwick 2001). anosim produces a global R-statistic, which is an absolute measure of distance between groups. An R-value approaching one indicates strongly distinct assemblages, whereas an R-value close to zero indicates that the assemblages are barely separable. These R-values were used to compare ant assemblages between habitat types, and burn plot treatments within and between sampling periods. R-values may occasionally be very low, indicating that assemblages are barely separable, but these values may also be significantly different from zero. This reflects a high number of replicates or samples, and the fact that R is inconsequentially small is of greater importance (Clarke & Warwick 1994). The converse may also be found, where R-values may be very high (indicating that assemblages are almost completely different) but these values are not significant. This situation occurs when the sample size is small, and in such instances the R-value is of greater importance (Clarke & Gorley 2001). The relationships between habitat types and burn plot treatments for both sampling periods combined were displayed using non-metric multidimensional scaling (nMDS) ordinations. These were iterated several times from at least 10 different starting values to ensure that a global optimum was achieved (indicated by no decline in the stress value) (Clarke & Gorley 2001). Although the experiment was designed originally to test season and frequency aspects of the fire regime, the effect of individual fires cannot be ignored as different post-fire fuel ages may result in successional effects (York 1994). For age analyses, plots were classified as follows: young = 4–5 months since fire; intermediate = 8–16 months since fire; old = 20–28 months since fire; unburnt = control and variable.

The effects of frequency, season and age were initially assessed using a series of pair-wise anosim tests. First, the age of the plots was varied (young, intermediate and old) while controlling for season and frequency. Ant assemblage composition did not differ with age for these pair-wise tests. Additional pair-wise anosim tests were necessary to determine whether it was possible to combine all frequencies or both seasons when doing subsequent analyses. Thus, frequency of burn was controlled for while season was altered (pair-wise tests: August biennial vs. April biennial and August triennial vs. April triennial), and season of burn was controlled for while frequency of burn was altered (pair-wise tests: August biennial vs. August triennial, August annual vs. August biennial, August annual vs. August triennial and April biennial vs. August triennial). There were no significant differences in ant composition between any of the burn treatment pairs. Therefore, for all further analyses, both seasons were combined when testing for frequency, and frequencies were combined when testing for effects of season.

Finally, ant species characteristic of the three habitat types, and of burnt and unburnt plots in each area, were identified using the Indicator Value method (Dufrêne & Legendre 1997; This method assesses the degree to which a species fulfils the criteria of specificity (uniqueness to a particular habitat) and fidelity (frequency of occurrence). A high indicator value (IndVal, expressed as percentage) indicates that a species can be considered characteristic of a particular habitat or site. This method can derive indicators for hierarchical and non-hierarchical site classifications, and is robust to differences in the numbers of sites between site groups (McGeoch & Chown 1998). Indicator values for each species were calculated based on a species abundance matrix, and Dufrêne & Legendre's (1997) random reallocation procedure of sites among site groups was used to test for the significance of IndVal measures for each species. Species with significant IndVals > 70% were considered as species characteristic of the site or habitat in question (subjective benchmark; McGeoch, Van Rensburg & Botes 2002).

analyses: vegetation

Differences in vegetation structure were assessed for each area by comparing the mean number of hits per plot in each foliage height category using anova. Vegetation structure was compared between replicates within an area, and between different burn plot treatments. Percentage cover for each habitat cover component was summed across all quadrats on each burn plot treatment for each replicate. Kruskal–Wallis tests were used to determine whether total habitat cover differed significantly between burn plot treatments. The bioenv procedure in primer was used to examine the relationship between habitat cover on the plots and the ant assemblages (Clarke & Gorley 2001). A single, among-site species similarity matrix was constructed using Bray–Curtis similarity measures, while several similarity matrices were constructed for each of the possible combinations of the specified habitat cover variables also using Bray–Curtis similarity measures because the measurement scale for all variables was percentage cover. Spearman's rank correlation coefficients (ρ) were then calculated for the species matrix and each of the possible habitat matrices. The variable or set of variables that have the highest ρ-value are those that best explain the species data (Clarke & Gorley 2001). Unfortunately this ρ-value does not produce an associated significance value.

In addition, anosim analyses were carried out for the Mopane and Satara areas to determine the effect of plot age (single fire effect) on habitat cover composition. The effect of season and frequency of burn (i.e. for fire regime) on habitat cover could not be investigated because anosim tests revealed that there were significant differences in vegetation cover between burn treatments of different ages. Vegetation analyses did not include Pretoriuskop because data from that area were only available for one sampling period.


species richness, relative abundance and fire effects

A total of 54 736 ants, comprising 160 ant species in 37 genera, was collected (Table 1). There was a significant difference in species richness across the three habitats (anova, F2,45 = 34·69, P < 0·0001). Mean species richness per plot was greatest at Pretoriuskop (mean 39·0 species plot−1, SE = 3·38), intermediate at Satara (28·19 species plot−1, SE = 1·78) and lowest at Mopane (mean 15·67 species plot−1, SE = 0·97). There was no significant difference in mean species richness between sampling periods (2000 and 2002) for Satara, although in Mopane mean species richness was higher in 2000 (anova, F1,40 = 11·04; mean species richness per plot ± SE = 20·09 ± 0·87 and 15·67 ± 0·97, for 2000 and 2002 sampling periods, respectively). At Mopane and Satara, when all burn plot treatments were compared there was no significant difference in mean species richness between burn plot treatments for either sampling period or when both sampling periods were combined. At Pretoriuskop, mean species richness was marginally significantly higher on annual than on control plots (Mann–Whitney U-test, P= 0·049, d.f. = 2). There was also no significant difference in the estimated total number of species using ICE between burn plot treatments in both the Mopane and Satara areas (anova, F6,14 = 1·34, P= 0·30 and F6,14 = 1·14, P= 0·39 for Mopane and Satara, respectively) (Fig. 1).

Table 1.  Species richness and abundance for each area in Kruger National Park per sampling period, and for both sampling periods combined. n= total number of pitfall traps summed across all plot types, and sampling years
 Mopane (n= 791)Satara (n = 815)Pretoriuskop (n = 120)
Species richness
2000    81    95  –
2002    69   111  89
2000 & 2002    98   121  –
2000 8 30616 169  –
2002 2 87824 5672816
2000 & 200211 18440 736  –
Figure 1.

Estimated species richness (ICE) for Mopane and Satara, for each burn plot treatment. Aug 1 = August annual burn; Aug 2 = August biennial burn; Aug 3 = August triennial burn; Apr 2 = April biennial burn; Apr 3 = April triennial burn.

The effect of single fires on species richness was also assessed for the Mopane and Satara areas. No clear trends could be detected in terms of responses to individual fires when species richness per plot for each sampling year was plotted for each replicate (Fig. 2a–f). Single fire events therefore did not result in consistent predictable trends, and local site differences (i.e. differences between replicates) appeared to be more important in determining species richness.

Figure 2.

Species richness for burn plots on each replicate for sampling periods 2000 (▪) and 2002 (□). (a–c) Mopane: Dzombo, Mooiplaas and Tsende replicates, respectively; (d–f) Satara: N’wanetsi, Marheya and Satara replicates, respectively. A star indicates plots that burnt in the year prior to the sampling period. Overall for each replicate there is little difference in species richness patterns between sampling periods, which indicates that individual fires do not have a large and consistent effect of species richness. Aug 1 = August annual burn; Aug 2 = August biennial burn; Aug 3 = August triennial burn; Apr 2 = April biennial burn; Apr 3 = April triennial burn.

There was a significant difference in ant abundance across the three habitats (anova, F2,45 = 40·01, P < 0·0001). Mean abundance (log) was greatest at Satara (2·92 ± 0·07, mean ± SE), intermediate at Pretoriuskop (2·60 ± 0·09) and lowest at Mopane (2·04 ± 0·07). There was no significant difference in abundance (log) between burn plot treatments for both Mopane and Satara. This was consistent for both sampling periods: 2000 (F(6,14) = 2·00, P= 0·13 and F(6,14) = 1·75, P= 0·18 for Mopane and Satara, respectively) and 2002 (F(6,14) = 1·62, P= 0·21 and F(6,14)= 1·96, P= 0·14 for Mopane and Satara, respectively).

community compositional changes

Within each habitat, ant assemblages were barely separable between sampling periods (R = 0·178, P= 0·001 and R= 0·085, P= 0·005 for Mopane and Satara, respectively). Thus, for all subsequent analyses for each area, data for 2000 and 2002 were combined. There were significant differences between the ant assemblages occupying the different habitat types (Global R= 0·448, P < 0·001; Fig. 3).

Figure 3.

Non-metric multi-dimensional scaling ordination of abundance of ant species in three different savanna habitats in Kruger National Park (stress = 0·23). Satara vs. Mopane, R= 0·349, P= 0·001; Mopane vs. Pretoriuskop, R= 0·788, P = 0·001; Satara vs. Pretoriuskop, R= 0·831, P= 0·001.


A two-way crossed anosim revealed that there was a significant difference in ant community composition between replicates, with the Tsende replicate being significantly different from Dzombo (R = 0·538, P= 0·001) and from Mooiplaas (R = 0·344, P= 0·001). To remove the effect of landscape from our analyses, data for the Tsende replicate were excluded for all other analyses.

All anosim tests revealed no effect of fire frequency on ant assemblage composition, and showed only a small effect of season of fire, with ant assemblages on August plots being significantly different from those on unburnt control and variable plots (Table 2). In terms of age, or time since fire, control plots, which were never burnt, were significantly different from young and intermediate-aged plots, irrespective of season and frequency. Although young plots were significantly different to old plots, the R-value was low, indicating substantial overlap between assemblages (Table 2).

Table 2.  Analysis of similarity for ant assemblages between plots types for Mopane and Satara areas (data for 2000 and 2002 sampling periods combined). The R-statistic is a measure of the similarity of assemblages, and reflects the degree of separation of assemblages: the closer the value to 1, the greater the difference in assemblage composition (Clarke & Warwick 2001). P is significant at the α-level of 0·05 using sequential Bonferroni tests for each fire parameter (Rice 1989). Significance level is denoted by *P < 0·05, **P < 0·01, ***P < 0·001. Age: Y = 4–5 months since fire; I = 8–16 months since fire; O = 20–28 months since fire; U = unburnt
Fire parameterMopaneSatara
  • Variable plots vary in age (time since fire) from 4 years to > 20 years old. They are generally older than the experimental burnt plots, and several are much closer in age to control plots.

Control vs. August 0·444NS0·654**
Control vs. April−0·009NS0·383NS
August vs. April 0·180NS0·102NS
August vs. variable 0·472NS0·357**
April vs. variable 0·053NS0·121NS
Control vs. annual 0·219NS0·909**
Control vs. biennial 0·143NS0·317NS
Control vs. triennial 0·085NS0·709***
Annual vs. biennial−0·125NS0·180NS
Annual vs. triennial−0·057NS0·115NS
Biennial vs. triennial 0·098NS0·060NS
Annual vs. variable 0·250NS0·533**
Biennial vs. variable 0·171NS0·098NS
Triennial vs. variable 0·138NS0·440**
U vs. Y 0·458**0·589***
U vs. I 0·274NS0·569**
U vs. O−0·167NS0·417**
Y vs. I 0·098NS0·101*
Y vs. O 0·250***0·103*
I vs. O 0·035NS0·192**
Control vs. variable−0·250NS0·726**


Ant assemblages did not differ significantly between the three replicates. Cluster analyses revealed that ant assemblage composition on the Satara replicate's control plot (both sampling periods) was more similar to annually burnt plots than to control plots on the other replicates. Thus, the Satara replicate's control plot was excluded from further analysis because, having burnt previously, it no longer provided the necessary baseline control comparison.

Pair-wise anosim tests revealed that differences in ant assemblage composition lay between burnt and unburnt plots, rather than between burning treatments (Table 2). Thus, ant assemblages on unburnt control plots differed significantly from those on variable plots, from those on plots burnt in August and April, and from annual and triennial burn plots. Moreover, differences in ant assemblage composition were much more pronounced between August and control plots than between April and control plots (higher R- and lower P-values). Ant assemblages on plots with annual and triennial burn frequencies were more dissimilar to control and variable plots than assemblages on plots with biennial burn frequencies (Table 2 and Fig. 4b).

Figure 4.

Non-metric MDS ordination of abundance of ant species in Satara, Kruger National Park, based on: (a) season of burn, (b) frequency of burn and (c) age (stress = 0·21). Frequencies: 1 = annually burnt; 2 = biennially burnt; 3 = triennially burnt. Age: young = 4–5 months since fire; intermediate = 8–16 months since fire; old = 20–28 months since fire; unburnt = control; variable plots = > 4 years old. For anosim results see Table 2.

There was no significant difference in assemblage composition between burnt plots of different ages (not including unburnt plots). Of the burnt plots, young and intermediate-aged plots were most different from unburnt plots (R = 0·589, P= 0·001 and R= 0·568, P= < 0·01, for young vs. control and intermediate vs. control, respectively).


Although the ant assemblages on control and August annually burnt plots differed considerably (anosim, R= 0·630, P= 0·1), the small number of replicates (n = 3) meant the difference was not significant (Clarke & Gorley 2001).

fire regime effect on individual species

Given that the responses of ant assemblages to fire lay between burnt and unburnt control plots only, and not between burning treatments, IndVal analyses were carried out to identify characteristic species, first with control and August annually burnt plots only (based on hierarchical clustering), and secondly using non-hierarchical clustering with plots classified according to area and plot type (control and burnt plots) prior to analysis (McGeoch & Chown 1998). This second IndVal analysis, using all burn plots classified either as burnt (including variable plots) or control plots, allowed the robustness and applicability of potential indicators identified from the first analysis to be confirmed.

Using abundance data from all plots, IndVal analyses revealed that there were three species that could be considered as generalists and characteristic of all three areas (Tetramorium frigidum Arnold, Monomorium albopilosum Emery and Tetramorium setigerum Mayr). Potentially reliable indicators for habitat type could only be found for the Satara and Pretoriuskop areas (Table 3). There were no species that could be regarded as indicators of plot type (annually burnt, burnt or control) in the Mopane area. At Satara, there were only three species characteristic of control plots (when annually burnt and control plots only were used), and no species characteristic of either burnt or unburnt areas when all burn plot types were used. Satara had fewer characteristic species with high abundance and high site fidelity than Pretoriuskop.

Table 3.  Percentage indicator values (IndVal > 70%) of ant species for each area and burn plot type (hierarchical clustering: annually burnt and control plots; non-hierarchically clustered: burnt plots and control plots). *Maximum indicator values
ClassificationSatara% IndValPretoriuskop% IndVal
AreasPheidole sp. 2 87·3*Pheidole sp. X 94·3
Monomorium zulu 82·5*Crematogaster sp. 2 82·9*
  Monomorium sp. 7 81·5*
  Lepisiota sp. 8 80·0
  Pyramica sp. 1 79·6*
  Tetramorium sp. 2 78·0*
  Oligomyrmex sp. 2 71·7*
Annual and control plots
AnnualMonomorium notulum100·0*
  Lepisiota sp. 6 98·7*
  Monomorium sp. 2 86·6*
  Tapinoma sp. 1 76·8
  Pheidole sp. X 70·1*
ControlTetramorium sepositum100·0*Plagiolepis sp. 4100·0*
Pheidole sp. 7 97·4*Lepisiota sp. 1 82·0*
Tetramorium gladstonei 93·8*  
Burnt and control plots
BurntMonomorium notulum 98·0*
  Lepisiota sp. 696·7*
  Lepisiota sp. 8 85·9*
  Monomorium sp. 2 80·4*
  Pheidole sp. X 80·0*
  Tapinoma sp. 1 79·4*
  Crematogaster sp. 2 70·0
ControlSolenopsis sp. 1 82·1

For Pretoriuskop, all five species identified as indicators of annually burnt plots were indicators of burnt plots when plots were classified as either burnt or control plots, i.e. with a coarser classification (Table 3). This served to confirm the robustness of species such as Monomorium notulum Forel and Lepisiota sp. 6 as indicators of burnt areas.

vegetation vertical complexity and fire effects

In both the Mopane and Satara areas, replicates did not differ significantly in the vertical complexity of vegetation, but for each area when all burn plot treatments were compared there were several significant differences in the vertical complexity of vegetation between plots (Table 4). At Mopane there was no significant difference in vegetation complexity between burn plot treatments for the two lowest height strata (i.e. up to 50 cm). However, for the upper foliage height classes (51–100 cm and 101–150 cm stratum) vegetation on the control plots was significantly denser and taller than that on all other burn plot treatments (Table 4).

Table 4.  One-way anova tests results for differences in vegetation structural complexity between burn plot treatments for each height class (sampling period 2002). Significance level is denoted by *P < 0·05, **P < 0·01, ***P < 0·001
Height (m)MopaneSatara
  • Significant difference between annually burnt plots and other plot types.

  • Significant difference between control plots and other plot types.

0–0·25 2·20NS 3·82*
0·26–0·50 2·07NS 7·29***
0·51–1·0 5·83B.** 7·12**

At Satara, for each height stratum there was a significant difference in vegetation complexity between burn plot treatments (Table 4). Foliage height profiles for the two lowest strata indicated that annually burnt plots had a lower height and density of vegetation compared with all other plot treatments, while differences in structural complexity for the upper strata indicated the presence of taller vegetation (grasses and shrubs) on control and variable plots. Differences in vertical complexity at Pretoriuskop were most pronounced for the upper strata, with much greater density of vegetation on control plots.

habitat cover and fire effects

bioenv analysis revealed that at Mopane the overall pattern in ant assemblage composition was best explained by grass cover (ρ = 0·214). At Satara, the habitat cover variables that best explained the overall pattern in ant assemblages were bare ground and litter (ρ = 0·203) and bare ground (ρ = 0·272) for sampling periods 2000 and 2002, respectively.

There was a significant difference in habitat cover overall between sampling periods at Satara (R = 0·583, P= 0·001) but not at Mopane. Thus, for subsequent anosim analyses, the two sampling periods for Satara were analysed separately. The difference in cover between sampling periods at Satara indicated that the effect of individual fires on habitat cover was greater at Satara than at Mopane.

In both areas there was a significant effect of plot age (effect of single fires) on habitat cover, with young plots (recently burnt) differing significantly from unburnt plots (Table 5). The effect of plot age on habitat cover was more pronounced at Satara than at Mopane; not only were the R-values higher at Satara, but also there were differences in habitat cover between intermediate aged and unburnt plots, and young and old plots which were not apparent at Mopane. The response of ant assemblages to plot age was similar to that found for habitat cover: not only were R-values higher at Satara than Mopane, but also ant assemblages at Satara responded to a wider range of age combinations. The vegetation analyses also illustrated that fire had a greater effect on vegetation than on ant assemblages because, for the habitat cover analyses, the R-values were often higher (indicating a stronger effect) and there were also differences between burnt plots rather than just between burnt and unburnt plots (Tables 2 and 5).

Table 5. anosim results for effect of post-fire fuel age on habitat cover for Mopane and Satara. The R-statistic is a measure of the similarity, and reflects the degree of separation of assemblages: the closer the value to 1, the greater the difference in assemblage composition (Clarke & Warwick 2001). P is significant at the α-level of 0·05 using sequential Bonferroni tests for each column (Rice 1989). Significance level is denoted by *P < 0·05, **P < 0·01. Age: Y = 4–5 months since fire; I = 8–16 months since fire; O = 20–28 months since fire; U = unburnt
U vs. Y 0·429* 0·958** 0·982NS
U vs. I 0·175NS 0·667NS 0·486*
U vs. O−0·102NS−0·012NS−0·073NS
Y vs. I 0·127NS−0·020NS−0·026NS
Y vs. O 0·260** 0·789** 0·524**
I vs. O 0·027NS 0·333NS−0·019NS

The immediate effect of burning is the consumption and removal of vegetation by fire, which increases the amount of bare ground. In both Mopane and Satara there was a significant difference in bare ground cover between burn treatment plots in both sampling periods: Kruskal–Wallis, H= 15·73, P= 0·015 and H= 13·70, P= 0·018 for Mopane and Satara in 2000, respectively, and Kruskal–Wallis, H= 14·78, P= 0·02 and H= 10·46, P= 0·005 for Mopane and Satara in 2002, respectively. Control plots had very little bare ground but tended to have higher litter loads: Kruskal–Wallis, H= 16·71, P= 0·01 and H= 7·93, P= 0·02 for Mopane and Satara 2002 sampling, respectively.


ant assemblages and response to fire

Ant species richness and abundance differed significantly between the three areas in Kruger National Park. These differences between areas correspond to broad underlying differences in habitat type and soil type (clay and sand). Net primary productivity (NPP) has been closely correlated with rainfall in Southern Africa (O’Brien, Field & Whittaker 2000). It therefore seems reasonable to assume that differences in rainfall between the three areas also correspond to differences in NPP, and therefore that differences in ant species richness may be linked to variation in NPP: ant richness was lowest in Mopane, which has the lowest rainfall (and presumably NPP), intermediate at Satara and highest at Pretoriuskop, where rainfall is highest (Gertenbach 1983). These findings are in keeping with those for ants elsewhere (Kaspari, O'Donnell & Kercher 2000).

However, within each area when all burn plot treatments were compared, fire had no effect on either mean species richness or mean abundance. This contrasts with other studies carried out across a range of habitat types, which have found that ant species richness is generally lower in areas that remained unburnt for long periods (Donnelly & Giliomee 1985; Andersen 1991; York 1994) and higher in areas that have recently burnt (Andersen & Yen 1985). Furthermore, contrary to other studies, including several conducted in savanna systems, ant species richness in this savanna system did not decline with time since fire (Donnelly & Giliomee 1985; Andersen 1991; York 1994; Parr, Bond & Robertson 2002; Hoffmann 2003). Although some studies have documented an increase in the abundance of ants following a fire (Andersen 1991; York 1999) and others have recorded a decrease (Andersen & Yen 1985), in this system no consistent trends were found.

Despite the absence of a significant effect of burning on mean ant species richness and abundance between burn plot treatments, there were clear differences in ant assemblage structure, in keeping with many other studies across a wide range of habitats (e.g. tropical savanna woodland, Andersen 1991; dry open forest, Vanderwoude, Andersen & House 1997; temperate forests, York 1999; low-latitude steppe, Farji-Brener, Corley & Bettinelli 2002; desert grassland, Zimmer & Parmenter 1998). Across all areas the only pronounced and consistent response that could be detected was between burnt plots and the unburnt control plots, and not between different burning treatments. The ant assemblages in Kruger National Park thus exhibited a high degree of resistance (Pimm 1984) to burning, at least for the burning regimes used in this study. Indeed, ant assemblages at Mopane exhibited quite remarkable resistance to burning; even after > 40 years of annual burning, ant assemblages on this plot did not differ significantly from the fire exclusion, control plot. In contrast, at Satara, ant assemblages displayed a wider range of responses to fire than those in Mopane. These effects were also indicated by the complete absence (Mopane area) or paucity (Satara area) of indicator species for distinguishing burnt from unburnt plots. These findings are in keeping with those from other savanna systems where ant assemblages were also found to be highly resistant to fire, with relatively few species exhibiting clear and significant responses to fire (Parr, Bond & Robertson 2002; Hoffmann 2003).

Ant assemblages also displayed little change between sampling periods (i.e. with single fires): trends in species richness for each replicate were quite consistent across sampling periods, particularly at Satara, despite variation between sampling periods in the number and identity of plots that burnt (Fig. 2). In fact, patterns in species richness appeared to be determined more by the replicate site than burn plot treatment, with patterns in species richness varying substantially between replicates within the same area (Fig. 2). A variety of factors have been shown to cause assemblages to differ between replicates (Drake 1991; Lockwood et al. 1997; Fukami 2001). The application of the same long-term treatment to replicates therefore does not guarantee the same deterministic path (Drake 1991; Chambers & Samways 1998).

The ant assemblages investigated in this study were also highly resilient to fire. This was particularly evident at Mopane where, although there was a significant difference between young and unburnt plots, there was no significant difference between intermediate aged and unburnt plots. Thus, only 8 months after a fire, the assemblage has returned to its pre-fire state.

vegetation and ants: fire regime and single-fire effects

The results from the Satara area confirmed the importance of bare ground and litter cover as factors responsible for overall structuring of ant assemblages (Andersen 1991; Crist & Wiens 1994; Bestelmeyer & Wiens 1996). Reduced litter cover on burnt plots is likely to result in the loss of some cryptic species from these areas (Robertson 1999). In most savanna systems, however, fire has little direct impact on epigaeic ants because many ant species nest in the soil, and they are therefore largely protected from fire (Andersen & Yen 1985). More importantly for ants, the effects of repeated fires are primarily considered to be indirect through changes in vegetation structure and habitat composition (vegetation change: Trapnell 1959; Trollope 1984; Trollope & Tainton 1986; ant responses: Andersen & Yen 1985; Chambers & Samways 1998). Here, the effects of fire on vegetation structure were more pronounced at Satara than at Mopane (Table 4). At Mopane, low biomass accumulation at ground level due to low rainfall probably accounted for the lack of difference in vegetation complexity between plot burn treatments for the two lower height strata. The effects of individual fires were also much more pronounced on vegetation cover than on ant assemblages (Tables 2 and 5). The high resistance and resilience of Mopane ants to burning, and the responses of ant assemblages to burning at Satara, can thus also be interpreted as a response by ants to changes in habitat structure as well as habitat cover. Similar responses by ants to changes in vegetation have been shown elsewhere (Greenslade & Greenslade 1977; York 1999).

why such resistance?

Although other studies have also found that ant assemblages can be relatively resistant to fire (Friend & Williams 1996; Andersen & Müller 2000; Parr, Bond & Robertson 2002; Hoffmann 2003), the ant assemblages in this savanna system exhibited a remarkable degree of both resistance and resilience to burning. The degree of response of ant assemblages is likely to be related to two main contributory factors: mean annual rainfall and changes in vegetation structure with burning, and the assemblage's history of association with fire. Our results have clearly demonstrated that differences in vegetation structure contribute to differences in the extent to which burning affects ant assemblages. Where biomass loads were generally lower (such as at Mopane), fire typically had less of an effect on vegetation structure, and thus the effect on ant assemblages was reduced (Farji-Brener, Corley & Bettinelli 2002; Parr, Bond & Robertson 2002; Hoffmann 2003). In contrast, in Pretoriuskop, high rainfall and the relatively wooded habitat resulted in very pronounced differences between burnt and unburnt plots, and thus very different habitats for ants.

Furthermore, where taxa have had a long association with fire, they are likely to display considerable resistance and resilience to burning (Andersen & Müller 2000; Orgeas & Andersen 2001). In southern and eastern African savannas, the biota has had a long history of frequent, extensive fires (i.e. human-ignited fires, and not just lightning fires), with the controlled use of fire dating back about 1·0–1·5 million years (Gowlett et al. 1981; Brain & Sillen 1988) and the start of more extensive use estimated at 250 000 years ago (Pennisi 1999). In contrast, in Australia, where ant assemblages have been shown to exhibit a greater response to burning, extensive, frequent use of fires by humans is much more recent, dating from about 50 000–40 000 years ago (Kershaw 1986).

synthesis and conservation implications

In this savanna system, fire has a small, although significant and probably indirect, effect on ant assemblages. The response of ants to fire was less affected by the subtleties of when and how often an area burnt, than the simple distinction of whether an area burnt or not. Although there are some effects of individual fires, these do not persist for long (less than a year). A possible exception to this overall trend is in higher rainfall areas where fire can play a greater role, and more detailed investigation in these areas is required. Nonetheless, it seems clear that for the central and northern areas of Kruger National Park, and indeed for similar savanna areas in southern Africa (Parr, Bond & Robertson 2002), at least for ants the type of fire does not have a marked effect on assemblages. Rather the effects have to do with the presence or absence of fire, and even then ant assemblages appear to be remarkably resistant and resilient to fire. This finding contrasts strongly with the idea that fire is an important disturbance in savanna systems (Scholes & Walker 1993; Van Langevelde et al. 2003).

This resistance of southern African savanna ant assemblages to burning has considerable implications for conservation and fire management of African dry savanna areas. It means that, as far as ant diversity is concerned, managers can afford to give more attention to other taxa when considering burning regimes. For ants, the most important consideration is whether or not an area burns. Nonetheless, because burning does affect ants, and is likely to affect other invertebrate and vertebrate taxa in ways different to that found for ants (Swengel 2001), caution should be exercised in applying burning regimes in protected areas.


C. L. Parr gratefully acknowledges fieldwork assistance from members of the University of Pretoria Zoology Department. The referees are thanked for their comments. This work was supported by a National Research Foundation (NRF) grant GUN 2053570 to S. L. Chown, and a University of Pretoria doctoral grant to C. L. Parr. South African National Parks provided logistic support.