@phdthesis{Menzel2009,
  author    = {Menzel, Florian},
  title     = {Mechanisms and adaptive significance of interspecific associations between tropical ant species},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-37251},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2009},
  abstract  = {Aggression between ants from different colonies or species is ubiquitous. Exceptions to this rule exist in the form of supercolonies (within a species) and interspecific associations (between species). Probably the most intimate interspecific association is the parabiosis, where two ant species live together in a common nest. They keep their brood separate but jointly use trails and often share food resources. Parabioses are restricted to few species pairings and occur in South American and Southeast Asian rainforests. While the South American parabioses have been studied, albeit poorly, almost nothing is known about their Southeast Asian counterparts. My PhD project focuses on Southeast Asian parabioses between the myrmicine Crematogaster modiglianii Emery 1900 and the considerably larger formicine Camponotus rufifemur Emery 1900. The two species frequently nest together in hollow trees in the tropical lowland rainforest of Borneo. The basic question of my PhD project is why these two species live together. I investigated both proximate and ultimate aspects of this question. For comparative purposes, I included studies on a trail-sharing association in the same habitat. On the proximate level, I investigated which mechanisms facilitate tolerance towards hetero-spe¬ci¬fic nestmates. Ants generally discriminate nestmates from non-nestmates via cuticular hydro¬carbons that function as colony recognition cues. I studied the specificity of nestmate recognition within and between the two parabiotic species. Using gas chromatography-mass spectrometry (GC-MS), I analyzed the cuticular substances in both ant species to find potential differences to non-parabiotic species, and to estimate the substance overlap among the two species. A high substance overlap would e.g. suggest that interspecific tolerance is caused by chemical mimicry. Finally, bioassays were conducted to evaluate the function of different cuticular compounds. Interspecific tolerance in the two parabiotic species was species-specific but not colony-specific. Ca. rufifemur tolerated all Cr. modiglianii individuals, even those from foreign colonies, but strongly attacked workers of other Crematogaster species. Cr. modiglianii, in turn, tolerated Ca. rufifemur workers of certain foreign colonies but attacked those of others. Chemical analyses revealed two sympatric, chemically distinct Ca. rufifemur varieties ('red' and 'black') with almost no hydrocarbon overlap. Cr. modiglianii only tolerated foreign Ca. rufifemur workers if they belonged to the same chemical variety as their own Ca. rufifemur partner. It also attacked other, non-parabiotic Camponotus species. Thus, reciprocal interspecific tolerance was restricted to the species Cr. modiglianii and Ca. rufifemur. Ca. rufifemur frequently tolerated conspecific non-nestmates of the same chemical variety. Minor workers were more often tolerated than majors, possibly because they possess two to three times lower hydrocarbon quantities per body surface than majors. In contrast, Cr. modiglianii nearly always attacked conspecific non-nestmates. Both species possessed hydrocarbons with considerably higher chain lengths than congeneric, non-parabiotic ant species. Long-chain hydrocarbons are less volatile than shorter ones and thus harder to perceive. They may thus considerably facilitate interspecific tolerance. Moreover, up to 98\% of the cuticular hydrocarbons in Ca. rufifemur were methylbranched alkenes, which are highly unusual among insect cuticular hydrocarbons. Cr. modiglianii and Ca. rufifemur had almost no hydrocarbons in common, refuting chemical mimicry as a possible cause of interspecific tolerance. The only hydrocarbons common to both species were two methylbranched alkenes, which constituted 89\% of the 'red' Ca. rufifemur hydrocarbon profile and also occurred in those Cr. modiglianii colonies that lived together with this Ca. rufifemur variety. Cr. modiglianii presumably acquired these two compounds from its red Ca. rufifemur partner. Cr. modiglianii was significantly less aggressive towards foreign Cr. modiglianii workers that were associated with the same Ca. rufifemur variety than to those associated with the respective other one. Hence, this species seemed to use recognition cues acquired from its parabiotic partner. Apart from hydrocarbons, both species possessed a set of hitherto unknown substances on their cuticle. The quantitative composition of the unknown compounds varied between parabiotic nests but was similar among the two species of a nest. They are probably produced in the Dufour glanf of Cr. modiglianii and transferred to their Ca. rufifemur partner. Possible transfer mechanisms include interspecific trophallaxis and 'mounting behaviour', where Cr. modiglianii climbed onto Ca. rufifemur workers without being displaced. Although the composition of the unknown compounds greatly varied between nests, they did not function as nestmate recognition cues since both species used hydrocarbons for nestmate recognition. However, the unknown compounds significantly reduced aggression in Ca. rufifemur. The ultimate, i.e. ecological and evolutionary aspects of my PhD research deal with potential costs and benefits that Cr. modiglianii and Ca. rufifemur may derive from the parabiotic association, their interactions with other species, and population genetic analyses. Additional studies on a trail-sharing association between three other ant species deal with two possible mechanisms that may cause or facilitate trail-sharing. Whether parabioses are parasitic, commensalistic, or mutualistic, is largely unknown and depends on the costs and benefits each party derives from the association. I therefore investigated food competition (as one of the most probable costs), differentiation of foraging niches (which can reduce competition), and several potential benefits of the parabiotic way of life. Besides, I studied interactions between the ant species and the hemiepiphyte Poikilospermum cordifolium. The foraging niches of the two species differed regarding foraging range, daily activity pattern, and food preferences. None of the two species aggressively displaced its partner species from baits. Thus, interference competition for food seemed to be low or absent. For both ant species, a number of benefits from the parabiotic lifestyle seem possible. They include interspecific trail-following, joint nest defence, provision of nest space by the partner species, food exchange via trophallaxis, and mutual brood care. If an ant species follows another species' pheromone trails, it can reach food resources found by the other species. As shown by artificial extract trails, Ca. rufifemur workers indeed followed trails of Cr. modiglianii but not vice versa. Thus, Ca. rufifemur benefited from Cr. modiglianii's knowledge on food sources (informational parasitism). In turn, Cr. modiglianii seemed to profit from nest defence by Ca. rufifemur. Ca. rufifemur majors are substantially larger than Cr. modiglianii workers. Although Cr. modiglianii often effectively defended the nest as well, it seemed likely that this species derived a benefit from its partner's defensive abilities. In neotropical parabioses (ant-gardens), mutualistic epiphytes play an important role in providing nest space. The neotropical Camponotus benefits its Crematogaster partner by planting epiphyte seeds, for which Crematogaster is too small. Similarly, the Bornean parabioses often were inhabited by the hemiepiphyte Poikilospermum cordifolium (Barg.-Petr.) Merr (Cecropiaceae). P. cordifolium seedlings, saplings and sometimes larger indivi¬duals abundantly grew at the entrances of parabiotic nests. However, P. cordifolium provides no additional nest space and, apart from nutritive elaiosomes, perianths, and extrafloral nectar probably plays a less important role for the ants than the neotropical epiphytes. In conclusion, the parabiosis is probably beneficial to both species. The main benefits seem to be nest defence (for Cr. modiglianii) and interspecific trail-following (for Ca. rufifemur). However, Ca. rufifemur seems to be more dependent on its partner than vice versa. For both parabiotic species, I analyzed mitochondrial DNA of ants from different regions in Borneo. My data suggest that there are four genetically and chemically distinct, but closely related varieties of Camponotus rufifemur. In contrast, Crematogaster modiglianii showed high genetic differentiation between distant populations but was not differentiated into genetic or chemical varieties. This argues against variety-specific cocladogenesis between Cr. modiglianii and Ca. rufifemur, although a less specific coevolution of the two species is highly likely. In Bornean rainforests, trail-sharing associations of Polyrhachis (Polyrhachis) ypsilon Emery 1887 and Camponotus (Colobopsis) saundersi Emery 1889 are common and often include further species such as Dolichoderus cuspidatus Smith 1857. I investigated a trail-sharing association between these three species and studied two mechanisms that may cause or facilitate these associations: interspecific trail-following, i.e. workers following another species' pheromone trail, and differential inter¬specific aggression. In trail-following assays, D. cuspidatus regularly followed extract trails of the other two species, thus probably parasitizing on their information on food sources. In contrast, only few P. ypsilon and Ca. saundersi workers followed hetero¬speci¬fic extract trails. Hence, the association between P. ypsilon and Ca. saundersi cannot be ex¬plained by foragers following heterospecific trails. In this case, trail-sharing may originate from few scout ants that do follow heterospecific pheromone trails and then lay their own trails. Interspecific aggression among P. ypsilon, Ca. saundersi and D. cuspidatus was strongly asymmetric, Ca. saundersi being submissive to the other two species. All three species discriminated between heterospecific workers from the same and a distant trail-sharing site. Thus, it seems likely that the species of a given trail-sharing site habituate to one another. Differential tolerance by dominant ant species may be mediated by selective habituation towards submissive species, and thereby influence the assembly of trail-sharing associations.},
  subject      = {Ameisen},
  language  = {en}
}