The significance of springtails in soil systems in South Africa is unlikely to be different from that elsewhere, recognising that springtails reach their highest species richness and abundance in moist habitats rather than in arid areas.^11,12 However, in contrast with many other regions of the world,¹⁰ and with other components of the soil fauna in South Africa, such as the ants,^13,14,15,16 springtails have, until recently, been the subject of little attention. Early research recognised that springtails are likely to play an important role in various South African biomes,¹⁷ and since the initial work of Womersley¹⁸, ongoing, but sporadic and typically restricted taxonomic investigations have been made.^{19,20,21,22,23,24,25,26,27,28} Few investigations have sought to understand comprehensively the diversity of the group (at present the only available, but unpublished list by P. Greenslade, places the fauna at about 43 genera and 90 species for the Western Cape), their contribution to ecosystem functioning, and their utility for bioindication,²⁹ although occasional attempts at doing so have been made.³⁰ The exception is research on the group undertaken in the sub-Antarctic Tundra and Polar Desert biomes of the Prince Edward Islands, which, at least geopolitically, form part of South Africa.³¹ For these island biomes, the limited fauna of 16 species is relatively well understood.^{32,33,34,35,36} Perhaps more significantly, it is this island work that has precipitated a recent and substantial change in current understanding of springtail ecology and systematics in continental South Africa.

On the basis of a comprehensive comparison of the likely effects of climate change on indigenous versus invasive springtails on the Prince Edward Islands, funded by a bilateral grant under the South Africa–Norway agreement (Table 1),³⁴ a second project was developed to compare life histories of species in polar systems (namely, the Prince Edward Islands and Svalbard, Norway) and more temperate systems (southern Norway and the Western Cape Province, South Africa). This work sought also to investigate the implications thereof for ecosystem functioning and drew in further expertise from Sweden and support from a South Africa–Sweden bilateral (Table 1). In both cases, it was recognised from the outset that the springtails in South Africa are poorly known systematically, and that this taxonomic impediment³⁷ would therefore constrain work substantially. Thus, additional collaborations were established with systematic experts under a third bilateral agreement (South Africa–France) and with the support of the International Barcode of Life project (iBOL - http://ibol.org) (Table 1). Here we report on research so far undertaken across these major projects, how systematic and ecological understanding of the group is progressing for South Africa, and the prospects for ongoing research and future collaborations amongst researchers with an interest in soil diversity and ecosystem functioning.

Initial assessments of the published taxonomic research indicated a faunal assortment of 90 species in 16 families, but, excluding incorrect identifications, only 57 valid and recognisable species (Table 2a). Now, based on a geographically extensive collection of more than 450 samples from the Western Cape (Figure 1), it is clear that the faunal component is much larger. The collection methods included high gradient extraction of litter bags; Berlese–Tullgren extraction of leaf litter, moss and rotten wood; vegetation beating and collection by hand (especially in caves and streams); pitfall trapping; litter sifting; and soil washing (of beach and deep soil sand). The current faunal assemblage stands at 136 morphospecies in 19 families (Table 2b), but only half of the samples have been sorted to species sofar and not all microhabitats have been surveyed to completion. Moreover, assessments of previous lists (in particular Paclt¹⁹) for the country, the examination of museum specimens and barcoding work have indicated that many species have been misidentified and that several groups have either cryptic species or have shown substantial radiations. For example, more than 11 Cryptopygus and 7 Parisotoma species were found during our survey, in contrast to only 2 species that have been previously recorded – Cryptopygus caecus and Parisotoma notabilis.¹⁹ In addition, 23 described species of Seira are found in South Africa,²² whilst only 1 species is described from Australia.³⁸ These groups that have radiated in the Cape Floristic Region are now the subject of closer investigation from the classical taxonomic^39,40 and barcoding perspectives. Thus, although research to date has more than doubled the number of known species, based on our experience from elsewhere it seems likely that at least 300 species will be characteristic of the regional (i.e. Cape Floristic Region) fauna. For South Africa as a whole the richness is certainly much higher. Even so, the Collembola likely has fewer species than are found in the highly diverse insect orders, such as the Coleoptera or Lepidoptera.⁴¹ Many sequences obtained from barcoding have been connected to traditional taxonomic species, although certain species with poor morphological descriptions have been more difficult(as in the case of several Seira species). The barcoding results have also led to careful re-examination of the individuals traditionally assigned to the widespread species Parisotoma notabilis, which actually represent a group of several similar species of endemic Parisotoma. In addition, the barcoding has proven not only to be of particular importance to detect introduced species of complex groups (e.g. Isotomurus maculatus), but also to show that some forms expected to be European actually do not match any European cluster (such as one cluster of Isotomurus and several of Ceratophysella).

FIGURE 1: Systematic (circle with dot in centre) and ad hoc (black circle) sampling sites for the current springtail diversity assessments in the Western Cape Province, South Africa. Protected areas are overlaid in shading.

FIGURE 2: Sample-based species rarefaction curves for the (a) Cape Flats Sand Fynbos and (b) Pinus radiata sites.

During the investigation of springtail diversity in renosterveld fragments, decomposition of the litter of three representative plant species was also investigated. Decomposition of these species varied between 0.00674/day for Galenia africana, a shrub favoured by disturbances such as overgrazing, to 0.00222/day for renosterbos Elytropappus rhinocerotis and 0.00029/day for the sturdy geophyte Watsonia borbonica, corresponding to litter half-lives of 0.3, 0.8 and 6 years, respectively.⁴² The rates for Galenia and renosterbos are much faster than those previously found for litter of the fynbos species Leucadendron parile and Protea repens.^51,52 However, they are not unusual compared with other fynbos species we have studied (Figure 3; personal observation).

If the mean fire return time for fynbos systems is about 11 years,⁵⁰ and if decomposition of the less hardy species proceeds such that many litters have half-lives that are less than half this time, then biological decomposition as a nutrient recycling process is much more significant than previously estimated.⁵³ Ultimately, the significance of this decomposition will depend on the relative contributions of species with more readily or less readily decomposed litter to the litter pool of any given system. To date, such estimates have not been made. Species with high decomposition rates of the kinds we investigated clearly make a significant contribution to the fynbos flora, and it is obvious that litter of many species do not accumulate on the ground in the way that some Protea species do (personal observations). Thus, the current research on ecosystem functioning has demonstrated that important as fire is in the Fynbos biome, it may be complemented and sometimes surpassed by other processes.

FIGURE 3: Decomposition rates per day (mean ± s.d.) of various renosterveld and fynbos species: Galenia africana, renosterbos (Elytropappus rhinocerotis) and Watsonia borbonica42 (black triangle); Protea exima53 (black circle); and Leucospermum parile and Protea repens51,52 (outlined circle).