The endemic South African velvet worm genus Peripatopsis currently contains eight recognized species described from variable morphological characters and the current taxonomy is unsatisfactory. In an attempt to investigate evolutionary relationships within Peripatopsis, we collected 137 individuals from 34 sample localities for six of the eight species. Sequence data derived from two partial mitochondrial (mt)DNA gene loci (COI and 12S rRNA), as well as partial sequence data from the ribosomal nuclear 18S rDNA locus in combination with gross morphological characters and scanning electron microscopy (SEM), was used to examine evolutionary relationships. Phylogenetic relationships were investigated using minimum evolution (ME) and Bayesian inferences (BI). Additionally, we also undertook a maximum likelihood (ML) analyses on the combined DNA sequence data set. The combined DNA evidence topologies derived from the ME, BI, and ML was highly congruent and was characterized by the presence of multiple lineages within recognized taxa. Peripatopsis clavigera, Peripatopsis moseleyi, and Peripatopsis sedgwicki each comprised two evolutionary lineages; Peripatopsis capensis comprised three; and Peripatopsis balfouri comprised six operational taxonomic units respectively. Genealogical exclusivity at both mtDNA and nuclear DNA among the geographically coherent groups coupled with pronounced sequence divergence suggested a two-fold increase in the number of species within Peripatopsis. Previously used gross morphological characters (such as the number of leg pairs and colour) were either highly variable within operational taxonomic units, or were invariant, suggesting that alternative morphological characters are necessary for species discrimination. SEM results revealed potentially useful diagnostic characters that can discriminate between at least discriminate some of the newly-identified lineages.

Examination of eroded and intact earth mounds in the Clanwilliam district, South Africa, indicates that they are well-established active termitaria of the harvester termite Microhodotermes viator. Unoccupied lower portions of the mounds contain ubiquitous trace-fossil evidence of earlier inhabitation by the same species. Previous studies indicating that fossorial molerats played a major role in the formation of the mounds are not supported by the observations presented here. Calcretization of the basal parts of the earth mounds has been caused by groundwater interaction with the more alkaline mound soil. C dating of this calcrete indicates that the earth mounds have been in existence for at least 4000 years, an order of magnitude greater than any previously recorded longevity for termitarium inhabitation.

We addressed the phylogeny of cockroaches using DNA sequence data from a broad taxon sample of Dictyoptera and other non‐endopterygotan insect orders. We paid special attention to several taxa in which relationships are controversial, or where no molecular evidence has been used previously: Nocticolidae, a family of small, often cave‐dwelling cockroaches, has been suggested to be the sister group of the predaceous Mantodea or of the cockroach family Polyphagidae; Lamproblatta, traditionally placed in Blattidae, has recently been given family status and placed as sister to Polyphagidae; and Saltoblattella montistabularis Bohn, Picker, Klass & Colville, a jumping cockroach, which has not yet been included in any phylogenetic studies. We used mitochondrial (COI + COII and 16S) and nuclear (18S and 28S) genes, and analysed the data using Bayesian inference (BI) and maximum likelihood (ML). Nocticolidae was recovered as sister to Polyphagidae. Lamproblatta was recovered as sister to Blattidae, consistent with the traditional placement (not based on phylogenetic analysis). However, because of the limited support for this relationship and conflict with earlier morphology‐based phylogenetic hypotheses, we retain Lamproblattidae. S. montistabularis was consistently placed as sister to Ectobius sylvestris Poda (Blaberoidea: Ectobinae), indicating that the saltatorial hindlegs of this genus are a relatively recent adaptation. Isoptera was placed within Blattodea as sister to Cryptocercidae. Nocticolidae + Polyphagidae was sister to Isoptera + Cryptocercidae, and Blaberoidea was sister to the remaining Blattodea.

The genus Ceropegia has largely radiated without evolutionary shifts in pollinator functional specialization, maintaining its interactions with small Diptera. Intriguing biogeographic and phylogenetic patterns may reflect processes of regional dispersal, diversification and subsequent specialization onto a narrower range of pollinators, though some of the findings may be caused by inconsistent sampling. Comparisons are made with other plant genera in the Aristolochiaceae and Araceae that have evolved flask-shaped flowers that trap female flies seeking oviposition sites.

A total of 322 records were available from the literature on fauna1 taxa endemic to the Cape Peninsula, South Africa. Excluding possible pseudoendemics, dubious records and many invertebrate groups for which little or no information exists, these records represent 112 species (1 vertebrate and the rest invertebrates) in 47 families. This number excludes many other potential endemics having distributions that extend just off the Peninsula. When mapped according to a 590 1 km x 1 km grid, these endemics were clustered in several, largely montane nodes and palaeogenic (palaeoclimatically stable) zones typically located in upper reach forest streams, riverine forest and caves (the latter supported 14 endemics). Endemics were over-represented on steep slopes. For many taxa, a very high percentage of the Peninsula representatives were endemics. There were more plant than animal endemics per 1 km' cell, although in total there were more animal than plant endemics. A significant correlation existed between the distribution of plant and animal endemics on the Peninsula, and especially on Table Mountain. The relationship, however, appears not to be causal, and is possibly related to parallel responses to historical isolation and topography. As the endemic fauna is mostly relictual, conservation of umbrella plant communities and the sandstone caves is essential. This may avert further extinction (some invertebrate endemics are likely to be extinct at this stage). Others have suffered declines in population numbers through development, invasion of alien vegetation, and possibly through predation by the introduced Argentine ant.

No abstract

1. Parts of the Namibian landscape show extensive surface perturbation in the form of long-lived, yet dynamic 'fairy circles'. While exerting profound ecological effects on 7.3% of the land surface, the origin and nature of these large bare discs embedded in an arid grassland matrix remains unresolved.2. We found no evidence to support the current hypothesis of a termite origin for fairy circles but instead observed a strong spatial association between fairy circles and large nests of the ant Black pugnacious ant Anoplolepis steingroeveri Forel, with much higher ant abundances on the circles compared with the matrix.3. Aggression trials showed that different colonies of A. steingroeveri were located on different circles, and that the species was polydomous.4. Fairy circles and Pogonomyrmex ant nests both have a bare disc surrounding the nest, are overdispersed (evenly spaced), and are associated with elevated soil moisture. Fairy circle soils exhibited a five-fold increase in soil moisture when compared with the matrix. 5. Senescent Stipagrostis obtusa (Delile) Nees seedlings were only observed on the circles and not in the matrix, and were found to have a reduction in both root length and number of roots.6. Anoplolepis steingroeveri excavated the root system of both S. obtusa seedlings on the disc and Stipagrostis ciliata (Desf.) de Winter grasses on the perimeter of the circles, where they tended honeydew-secreting Meenoplidae bugs that fed on grass roots and culms. The bugs occurred almost exclusively on grasses associated with the circles. This ant-bug interaction is a possible mechanism for the observed reduction in root length and number of senescent grass seedlings on the circles.

1. The growth pattern of Namibian fairy circles was examined in relation to environmental, termite, and plant variables to provide support for the sand termite (Psammotermes allocerus Silvestri) hypothesis of circle origin.2. New and young circles were associated with the highest number of sand termites and their foraging grass sheetings which were both considerably lower in mature and senescent circles.3. Circles increased in size with age, and apart from the senescent stage had higher soil moisture levels than the matrix. 4. In laboratory trials sand termites browsed seedling roots, negatively impacting root and leaf number.5. This provides a potential mechanism for circle formation through central-based foraging by P. allocerus which eliminates Stipagrostis grass clumps around its nest system.6. The resulting bare disc is postulated to be maintained through a combination of a depleted seed bank, termite herbivory on seedlings, and excavations by pugnacious ants (Anoplolepis steingroeveri Forel).