Hoverfly colour patterns have often been labelled as mimetic, but only some species resemble their models closely, whereas others resemble their supposed models only vaguely, so are at best rather poor mimics (see, for example, the assessments in Howarth et al., 2000). There is a clear distinction in the literature between bumblebee mimics, which are usually accepted as such without question, and honeybee and wasp mimics, where a large proportion are generalized or imperfect to the human eye. Furthermore, there have been many conflicts among writers about the supposed models of particular species. For example, Criorhina asilica was labelled as a bumblebee mimic by Verrall (1901), but as a perfect or almost perfect honeybee mimic by most authors (e.g. Dlusskii, 1984; Roder, 1990), although Drees (1997) called it 'cryptic'.
Table 9.1 counts all the Holarctic hoverfly species that have been named as mimics in the literature, organized by mimicry ring (F. Gilbert, unpublished). The striking thing is their sheer number, 256 species from a total of 2334 Holarctic species (11%). The world totals cited above (279 from 5600) demonstrate that outside Europe the available information is very fragmentary and unsystematic, hence these numbers are almost certainly an underestimate. European insects have been studied much more intensively, and in Europe there are 138 mimics out of a total of about 630 species (22%).
Virtually all the model identifications made by the authors concerned were purely on the basis of visual similarity according to our own human perception, with no experimental or any other kind of evidence. Of course, in natural circumstances predators are required to deal with potential prey in a wide variety of circumstances, including as fast-moving evasive insects, and some potential prey represent a significant threat to well being. Identifications based upon our own perceptions may not correspond to the perceptual confusions between models and mimics generated by the eyes of predators, and this might distort our view of biological reality. One element that has been highlighted is the UV-component of colour patterns (Cuthill and Bennett, 1993; Church et al., 2004), invisible to mammalian predators, but possibly conspicuous to UV-sensitive birds or insects. A priori an unsuspected and different UV-component to the colour pattern is unlikely in Diptera, since their black colours are indole-based eumelanins: melanins strongly absorb in the UV, and therefore syrphids are unlikely to have UV patterns superimposed on any black part of their body. Photographs of social wasps and some of their hoverfly mimics in both visible and UV light have no UV patterns evident in either, nor in non-mimetic Sarcophaga flies (L. Gentle, personal communication). Similarly, Nickol (1994) took photographs of the hoverflies Volucella inanis, V. zonaria and their models
(social wasps and hornets), and also found both models and mimics to appear entirely black under UV light. Thus the ability of birds but not humans to see UV light does not seem to be a serious problem in assessing the model-mimic relationships of hoverflies. In principle, if they did exist, these kinds of distortions are simple to remove, providing that we have realistic predator-based assessments of the degree of model-mimic confusion (see Green et al., 1999); in practice, such assessments are difficult to obtain. The fact that we are able to classify some mimics as 'extremely accurate' probably implies that our perception is rather similar to that of at least some other predators.
Amongst the hoverfly mimics of bumblebees, most model identifications are reasonably obvious, and the lists of bumblebee models generated from the various suggestions by different authors are generally very similar in their colour patterns. Usually there is little ambiguity, since the quality of the mimicry is very high. The distribution of mimics among the various Mullerian complexes is very different between the Nearctic and the Palaearctic. No form seems to mimic the black bumblebees with thin yellow bands (complex B) in the western Nearctic, and this complex is absent from the Palaearctic. A large proportion of Palaearctic mimics are either black with red tails like B. lapidarius (complex A), or all tawny-coloured like B. pascuorum (complex I), or yellow-banded with a white tail like B. lucorum (complex E); all of these complexes and their mimics are largely absent from the Nearctic. In contrast, the Nearctic complex of syrphid mimics with a pattern of a yellow anterior and a black posterior, like B. impatiens (complex G), is absent from the Palaearctic, although the inconspicuous white tail of Criorhina berberina berberina and possibly one morph of Cheilosia illustrata are rather similar (and which therefore lack a closely corresponding model pattern in the Palaearctic). The white-patterned bumblebees of the Caucasus are paralleled by the white-patterned mimetic Diptera there. These distributional correspondences themselves constitute powerful corroborative evidence for the reality of mimetic relationships.
Only four Palaearctic hoverfly species have a quality of bumblebee mimicry that can be regarded as poor or unclear. In the Nearctic, very little work on models and their mimics has been done, except for the series of papers by Waldbauer and colleagues (see Waldbauer, 1988). In all their work, Waldbauer et al. decided to consider all bumblebees as members of a single Mullerian complex, and hence clearly regarded the differences among their colour patterns as irrelevant. Mimetic flies were labelled merely as generalized bumblebee mimics, without noting any closer resemblance to particular species. The authors were then able to assume that Mallota bautias was a general mimic of bumblebees, although in fact it resembles a particular group of eastern Nearctic bumblebee species rather closely. The context dependency of mimicry is highlighted, however, by the fact that M. bautias was for decades regarded as conspecific with the Palaearctic M. cimbiciformis, so closely do they resemble one another morphologically. However, M. cimbiciformis is uniformly interpreted as 'a particularly fine mimic of the honeybee' (Stubbs and Falk, 1983), and to my knowledge has never been identified as a mimic of any Palaearctic bumblebee. In the Nearctic, where honeybees were only introduced in the 19th century, the identical colour pattern can operate as a bumblebee mimic: there are even some good experimental data showing that this bumblebee mimicry is effective in protecting the fly from predation (Evans and Waldbauer, 1982).
The mimicry of honeybees by some hoverflies (mostly Eristalis species, commonly called droneflies) has been commented upon for a very long time (Osten-Sacken, 1894), and even experts can be fooled. Benton (1903) exhibited a photograph published in an apicultural journal of 'Bees working on Chrysanthemums' which were in fact Eristalis tenax. He also recounted his role in 'the famous Utter trial' (whatever that was!), where the prosecution could not distinguish between honeybees and droneflies, and therefore were unable to prove positively that bees were the cause of some alleged damage. Even experienced beekeepers were unable to make the same discrimination. However, other entomologists have been less impressed with the match between the honeybee model and Eristalis species, and Mostler (1935) attributed their lower protection in his experiments to their lesser resemblance to the model. Nicholson (1927) agreed that E. tenax was 'somewhat like the common hive-bee', but insisted that it was 'one of the least convincing cases of mimicry I know'. I suspect that most entomologists would agree with Mostler (1935). There is a range of different mimics that correspond to the colour variants of honeybees: for example, Eristalis tenax and female E. arbustorum are like the darker varieties, and male E. arbustorum resemble the lighter varieties. There are also a number of bee-like Eristalis species in North America, but honeybees are not native to the New World and we have little idea about which of the native bee fauna might have led to the evolution of mimetic colour patterns among Nearctic Eristalis species.
It is the apparently wasp-mimetic syrphids that cause the greatest difficulties in assessing the extent of mimicry among the Syrphidae. They are freely quoted as examples of mimicry, but are often unsatisfactory under critical consideration. The resemblance is often not particularly close, and the quality of mimicry varies from good to bad: many authors have made this point (e.g. Brown, 1951; Dlusskii, 1984; Waldbauer, 1988; Dittrich et al., 1993). Such problems led Waldbauer to define wasp mimicry to include only specialists that also mimic the long antennae and folded wings of vespoid wasps. However, this still does not mean that there is a one-to-one correspondence between species of models and these mimics, since often there is no particular exact replica of the mimic among available models. Some of the morphological adaptations for mimicry in this group are truly remarkable. For example, there are at least ten independently evolved solutions to the problem of mimicking wasp antennae (Waldbauer, 1970), three of which involve using the front legs. Species of the genera Spilomyia and Temnostoma and Volucella bombylans have only the normal short cyclorrhaphan type of antennae, but instead the anterior half of the forelegs is darkened, and the flies hold them up and wave them about in front of the head to create an amazingly good illusion of wasp-like antennae. Interestingly, not one bumblebee mimic has evolved elongated antennae, and only V. bombylans uses behaviour to mimic having them (although only females do this, in their final stealthy approach to the bumblebee nest in which they oviposit: Fincher, 1951; Rupp, 1989). This difference must tell us something about the salience of such features to predators; presumably long antennae are important features of identifying wasps, but the coat of hairs dominates when identifying bumblebees. Wasps do indeed wave their antennae about conspicuously, and bumblebees do not.
One characteristic of aposematic models and their mimics is that they often have harder, more durable bodies than other insects, toughened to withstand attack by predators so that the predators taste them but the prey still survive (Rettenmeyer, 1970: 58). The abdomen of many syrphine species is 'emarginate', i.e. each tergite is compressed just before the lateral margin, creating a narrow ridge or beading along the edge; this feature may have arisen in order to toughen the abdomen, since it occurs only in mimics. Specialized mimics have gone much further, and have the entire abdomen arched and convex, or cylindrical; the cuticle is punctate and hence greatly strengthened; and the joints between the overlapping tergites are very strong. If possession of an abdomen of this type is taken to define which of the black-and-yellow syrphids are truly wasp mimics, then rather few temperate species pass the test - in the northern hemisphere, only those of the genera Ceriana, Chrysotoxum, Sphecomyia, Spilomya and Temnostoma. There are, however, many genera with this type of abdomen in the tropics, perhaps an indication of a longer period of evolution among models and mimics there compared with the Holarctic.
The possibility that large Mullerian complexes of many wasp species together constitute a single model for Batesian mimics has hardly been addressed by anyone since Nicholson's (1927) largely uncited paper, except by Waldbauer and his colleagues in the eastern USA; this is an especially surprising omission among Palaearctic workers. Only Nickol (1994) has really identified this property clearly in his discussion of mimicry in Volucella zonaria, although it was also implicit in Dlusskii's (1984) important paper. Despite this omission, many such complexes appear to exist amongst black-and-yellow noxious insects and their mimics. The Mullerian complexes themselves are much less homogeneous than those of bumblebees, and overlap so that the boundaries are less distinct; presumably this is a consequence of their noxiousness.
Was this article helpful?