The South African citrus thrips (Scirtothrips aurantii) is a serious economic pest of citrus and mango in its native range, which encompasses a large part of Africa. Consequently, the adventitious establishment of this species in Queensland, Australia in 2002 was a major cause for concern in horticultural circles. In Australia, however, this thrips species has remained on plants of a single host plant genus Bryophyllum (Crassulaceae), but primarily on B. delagoense (mother-of-millions), a plant toxic to livestock and the third most invasive weed species in Queensland. No reports of S. aurantii attacking horticultural hosts in Australia have surfaced, despite several targeted surveys. I therefore tackled the central question of why South African citrus thrips (S. aurantii), a reportedly polyphagous horticultural pest in its native range (Africa), should be restricted to a single genus of plants (Bryophyllum) in Australia?
Three mutually exclusive hypotheses have been generated about the host specificity so far shown by S. aurantii in Australia, but none have yet been tested to any extent. (1) The invading population may have been so small as to have comprised only a small subset of the species’ ecological potential. (2) The S. aurantii population may be truly polyphagous and only temporarily restricted to Bryophyllum in Australia. (3) The species name S. aurantii may mistakenly include more than one cryptic species, a possibility that is rarely considered when exploring generalist host relationships.
This thesis investigated the S. aurantii host use paradox through several avenues related to understanding species and their ecology from the fundamental basis provided by the recognition concept of species. These avenues included fundamental host range testing, characterisation of the mating behaviour (or specific mate-recognition system) of the population on Bryophyllum in Australia, reciprocal cross mating experiments between insects from Australia and South African citrus, and population genetics studies across host species and the two countries. In addition, the effectiveness of S. aurantii as a potential biological control agent for Bryophyllum was investigated, because livestock producers in Queensland were starting to move insects substantial distances to control mother-of-millions in pasture areas.
The central finding of this thesis is that the species name S. aurantii mistakenly includes at least three host-associated cryptic species. Host testing revealed that the fundamental host range of the Bryophyllum population of S. aurantii includes Macadamia integrifolia, Mangifera indica and Kalanchoe blossfeldianna. However, the choice tests (involving B. delagoense) and a field survey of Man. indica demonstrated conclusively that the realised host range of S. aurantii in Australia is restricted to Crassulaceae. Reciprocal cross-mating experiments between the Australian Bryophyllum insects and South African S. aurantii from horticultural host plants showed that mating frequencies were significantly lower in test crosses (Bryophyllum x horticultural) than in controls (Bryophyllum x Bryophyllum and horticultural x horticultural), which indicates there are at least two distinct species within S. aurantii and suggested further tests of this interpretation. I therefore examined the genetic relationships between the Bryophyllum associated population in Australia and populations sampled from a range of host plant species (including Bryophyllum) in South Africa. Individual thrips collected from B. delagoense in South Africa and Australia form a single monophyletic clade, for the COI marker, with the average pairwise difference between the citrus clade and the B. delagoense clade being 2.96%. A microsatellite analysis of gene flow confirmed that the populations associated with B. delagoense do not hybridise with the thrips on horticultural species under natural conditions and are therefore host-associated cryptic species.
The existence of host-specific cryptic species within the taxon S. aurantii explains the paradox of the host plant specificity of S. aurantii in Australia, for the insects on Bryophyllum belong to a species that is cryptic in the sense of having no obvious morphological differences from the populations on other host plants (including citrus). All the evidence available demonstrates that insects of this species feed exclusively on plants from the family Crassulaceae in the field, both in Australia and South Africa, whereas the species associated with citrus does indeed use multiple host species (including Caesalpinia pulcherrima, Macadamia integrifolia and Punica granatum). The thrips on mango still need to be tested in this regard.