Abstract

Background and Aims

Although pollination of plants that attract flies by resembling their carrion brood and food sites has been reported in several angiosperm families, there has been very little work done on the level of specificity in carrion mimicry systems and the importance of plant cues in mediating such specialization. Specificity may be expected, as carrion-frequenting flies often exploit different niches, which has been interpreted as avoidance of interspecific competition. Interactions between the orchid Satyrium pumilum and a local assemblage of carrion flies were investigated, and the functional significance of floral traits, especially scent, tested. Pollination success and the incidence of pollinator-mediated self-pollination were measured and these were compared with values for orchids with sexual- and food-deceptive pollination systems.

Methods and Key Results

Observations of insect visitation to animal carcasses and to flowers showed that the local assemblage of carrion flies was dominated by blow flies (Calliphoridae), house flies (Muscidae) and flesh flies (Sarcophagidae), but flowers of the orchid were pollinated exclusively by flesh flies, with a strong bias towards females that sometimes deposited live larvae on flowers. A trend towards similar partitioning of fly taxa was found in an experiment that tested the effect of large versus small carrion quantities on fly attraction. GC-MS analysis showed that floral scent is dominated by oligosulfides, 2-heptanone, p-cresol and indole, compounds that also dominate carrion scent. Flesh flies did not distinguish between floral and carrion scent in a choice experiment using olfactory cues only, which also showed that scent alone is responsible for fly attraction. Pollination success was relatively high (31·5 % of flowers), but tracking of stained pollinia also revealed that a relatively high percentage (46 %) of pollen deposited on stigmas originates from the same plant.

Conclusions

Satyrium pumilum selectively attracts flesh flies, probably because its relatively weak scent resembles that of the small carrion on which these flies predominate. In this way, the plants exploit a specific subset of the insect assemblage associated with carrion. Pollination rates and levels of self-pollination were high compared with those in other deceptive orchids and it is therefore unlikely that this mimicry system evolved to promote outcrossing.

INTRODUCTION

Pollination by carrion insects has been documented in a wide range of plant families (Vogel, 1954; Fægri and van der Pijl, 1979; Meve and Liede, 1994; Proctor et al., 1996; Stensmyr et al., 2002; Johnson and Jürgens, 2010; Shuttleworth and Johnson, 2010). Plants attract these pollinators by mimicking the substrate on which the insects rely as a brood or food site (Proctor et al., 1996) through a diverse array of traits such as dark brown or purple flower coloration (Beaman et al., 1988), the frequent presence of bizarre, filliform appendages (Vogel, 1990) and proportionally large flowers and inflorescences (Davis et al., 2008). Perhaps the most characteristic trait is the putrid floral scent emitted by the flowers, which closely resembles that of real carrion (Kite et al., 1998; Stensmyr et al., 2002; Jürgens et al., 2006; Johnson and Jürgens, 2010).

Globally, the most important carrion insects include members of the Dipteran families Calliphoridae and Sarcophagidae (Payne, 1965, Blackith and Blackith, 1990; Hewadikaram and Goff, 1991; Carvalho et al., 2000; Satoshi, 2001; Archer and Elgar, 2003; Shi et al., 2009). The universally observed co-existence of calliphorids and sarcophagids on an ephemeral resource such as carrion, has led to the suggestion that they occupy slightly different niches (Denno and Cothran, 1975). Niche differentiation between these flies occurs along temporal axes such as seasonal variation in carrion abundance (Johnson, 1975; Hanski, 1987; Grassberger and Frank, 2004; Shi et al., 2009), as well as due to specialization according to carrion size (Denno and Cothran, 1975; Kneidel, 1984a), carrion type (Fuller, 1934; Cornaby, 1974; Kneidel, 1984a) and carrion successional stage (Payne, 1965; Cornaby, 1974; Grassberger and Frank, 2004).

It is unclear whether carrion-mimicking flowers attract a subset of the local assemblage of carrion flies. If so, this could be mediated by differentiation in specific plant traits that are similar to those that determine niche specialization of flies on real carrion. A comparison among closely related plant species indeed shows variation in traits such as flower size for Rafflesia and stapeliads (Beaman et al., 1988; Meve and Liede, 1994) and scent composition among arums and stapeliads (Kite et al., 1998; Jürgens et al., 2006; Johnson and Jürgens, 2010), which may reflect mimicry of different sizes and types of carrion. Additionally, carrion mimics are often visited by only one or a few species of carrion fly (e.g. Stensmyr et al., 2002), which further suggests that they only attract a subset of the local fly assemblage. To ultimately establish the level of specialization of the plant–insect interaction for carrion mimics, however, it would be necessary to take the entire assemblage of carrion flies that co-occur with plants into consideration. Apart from one study on carrion-mimicking stinkhorn fungi (Sleeman et al., 1997), no one, as far as is known, has studied this for flowering plants.

Pollination systems that rely on carrion mimicry are often characterized by the absence of a reward and thereby constitute a form of pollination through deception (Dafni, 1984; Schiestl, 2005). Floral deception is particularly widespread among Orchidaceae (Jersakova et al., 2006). The most common types of deception in orchids are food deception, where flowers exploit innate preferences of pollinators by a general resemblance to food plants in a community, or conditioned preferences of pollinators by resembling specific food plants (Batesian mimicry), and sexual deception where flowers attract male pollinators by mimicking female insects (Jersakova et al., 2006). The evolution and widespread occurrence of deception in orchids is most commonly explained as a mechanism that reduces geitonogamous self-pollination, and thereby promotes outcrossing (Cozzolino and Widmer, 2005; Jersakova et al., 2006), as floral visitors have a tendency to visit fewer flowers on rewardless plants (Johnson and Nilsson, 1999; Jersakova and Johnson, 2006). Apart from elevated levels of outcrossing, deceptive pollination is typically also associated with extremely low levels of pollination success (Tremblay et al., 2005; Scopece et al., 2010). The evolution of mimicry, as a special case of floral deception, may have been driven by selection for increased pollination success to escape high levels of pollen limitation. Johnson (1994), for example, found that fruit set in a Batesian mimic is comparable to that of a rewarding orchid with which it shares habitat and pollinator. Although carrion mimicry is known to occur in orchids (van der Pijl and Dodson, 1966), as far as is known, no studies on carrion mimicry in orchids have documented pollination success and outcrossing rates. It remains to be tested, therefore, whether this form of mimicry also promotes outcrossing and is characterized by relatively high pollination success, or whether it may have evolved for different reasons.

These questions were addressed using Satyrium pumilum as the model system. This peculiar member of the large, mainly African orchid genus Satyrium (van der Niet et al., 2005), is characterized by the virtual absence of a flowering stalk resulting in the flowers being borne at ground level (Fig. 1), which clearly sets it apart from all other satyriums. Johnson (1997) showed that its sister species S. bracteatum (van der Niet et al., 2005) is pollinated by small muscid and sarcophagid flies and suggested, in accordance with earlier speculations by Vogel (1954), that S. pumilum is most likely carrion fly-pollinated. In this study this prediction is confirmed. Furthermore, this study specifically aimed to (a) understand the association between the pollinators and the assemblage of co-occurring carrion flies, (b) test the significance of floral scent in this mimicry system, particularly with reference to potential pollinator specialization, and (c) quantify levels of pollination success and outcrossing, and compare these with other deceptive systems in orchids.

Fig. 1.

Satyrium pumilum and its sarcophagid fly pollinators. (A) The habitat of S. pumilum at Leliefontein in the Namaqualand Kamiesberg. (B) Satyrium pumilum in situ (scale bar = 1 cm). (C–E) Pollination sequence of a S. pumilum flower by a sarcophagid fly in an arena (scale bar for all three photos = 0·5 cm); (C) the fly carrying five pollinaria from other S. pumilum flowers enters an unpollinated flower – note the empty stigma (1), and that both its pollinaria are present (2); (D) as the fly moves deeper into the flower towards the right-hand spur, it presses an attached pollinium against the stigma, and its thorax against the right-hand viscidium; (E) as it leaves the flower, the fly has deposited two massulae on the stigma (1), and removed a pollinarium (2) – it now carries six pollinaria.

Fig. 1.

Satyrium pumilum and its sarcophagid fly pollinators. (A) The habitat of S. pumilum at Leliefontein in the Namaqualand Kamiesberg. (B) Satyrium pumilum in situ (scale bar = 1 cm). (C–E) Pollination sequence of a S. pumilum flower by a sarcophagid fly in an arena (scale bar for all three photos = 0·5 cm); (C) the fly carrying five pollinaria from other S. pumilum flowers enters an unpollinated flower – note the empty stigma (1), and that both its pollinaria are present (2); (D) as the fly moves deeper into the flower towards the right-hand spur, it presses an attached pollinium against the stigma, and its thorax against the right-hand viscidium; (E) as it leaves the flower, the fly has deposited two massulae on the stigma (1), and removed a pollinarium (2) – it now carries six pollinaria.

MATERIALS AND METHODS

Study species and site

Satyrium pumilum Thunb. is distributed throughout the western half of the greater Cape Floristic Region (Born et al., 2007), from Namaqualand in the north, to Riversdale in the east. The study was carried out at a site near Leliefontein, in the Kamiesberg region in Namaqualand, in a large population with several thousand plants scattered over approx. 1 ha. Here, S. pumilum grows in sandy, moist conditions near small streams, in open or shaded habitats under renosterbos bushes (Elytropappus rhinocerotis Less), often forming dense clumps of several dozen plants. The site is part of communal farm land and is grazed by sheep for several weeks in spring. Large numbers of flowering plants in this population were observed in four consecutive years (2006–2009), all of which were characterized by relatively high winter and spring rainfall. In 2008, when most of the fieldwork was carried out, the population flowered for about 8 weeks between early September and late October.

Plants have 3·13 ± 1·42 open flowers (mean ± s.d., n = 1327) with new flowers opening every 3·02 ± 1·25 d (mean ± s.d., n = 17). Individual flowers are open for 12·5 ± 1·5 d (mean ± s.d., n = 17). Stigmas become receptive on the day of anthesis, or latest the day thereafter. In contrast to Vogel (1954), no nectar was observed in any of the spurs examined (T. van der Niet, unpublished data). Moreover, the spurs are flattened and thus inaccessible to flower visitors.

Pollinator observations

Insect visitors to plants in the field were observed throughout the duration of the fieldwork during two consecutive years (5 d in September 2007 and approx. 5 weeks in September–October 2008). If insects were seen close to, or on a plant, their behaviour was observed. Initial observations revealed that visits were very rare. Therefore carrion bait was laid out to attract carrion-visiting insects during several days of the 2008 flowering season. Carcasses of the rock hyrax Procavia capensis, a common mammal from the area, which were found along the road, were used. Insects on carrion that carried S. pumilum pollinaria were caught, killed using ethyl acetate fumes, and identified and sexed by specialists (see Acknowledgements). The number of orchid pollinaria on each insect was counted.

Carrion fly assemblage

To identify the species composition of the carrion fly assemblage at the S. pumilum population, P. capensis carcasses, several days old, were laid out as a bait and samples of visiting flies caught at random using an insect net during several days in September–October 2008. Caught flies were killed using ethyl acetate fumes, and identified to family, genus or species level and sexed by specialists (see Acknowledgements).

Arena experiments

To examine whether different local carrion fly species are capable of removing pollinaria and depositing massulae in flowers of S. pumilum, 12 arenas made out of transparent empty soft drink bottles were set up. Plants with unvisited flowers were put in soil in the bottom of the bottles and flies (up to five, but generally only a single individual) that had been caught on carrion in the field released into these arenas. The behaviour of the flies was observed and pollinarium removal and massulae deposition were recorded after 24 h.

Carrion scent quantity and fly attraction

To test whether scent quantity has an effect on fly attraction, carrion was offered in small versus large quantities to flies in the field. Part of a leg of P. capensis was used to represent small carrion quantities, and entire P. capensis carcasses represented large quantities. Carrion was placed on the bottom of a bucket (approx. 30 × 30 cm) that was buried flush with the ground. To exclude visual cues, the bucket was covered with a carton lid that resembled the colour of the surrounding soil. The lid contained a square hole, approx. 10 × 10 cm in size, over which was mounted a piece of carton, approx. 1 cm above the lid level to let scent escape. An empty bucket with the same cover was also buried to control for fly attraction to the bucket. Any flies approaching the buckets were caught, and identified to family level. The experiment was repeated for 3 d during the warmest period of the day (mid-day), as this was found to correspond to the peak of fly activity. A 30-min observation period with the small quantity was always followed by a 30-min observation period with the large quantity. The reverse sequence could have potentially resulted in a bias in fly attraction during exposure of the small carrion quantity if flies, that would not have been present otherwise, were attracted to the vicinity of the large carrion quantity without being caught.

Results (counts for each fly family attracted to small versus large carrion quantities) were analysed using a χ2 contingency test. The probabilities obtained from separate tests performed on data collected on different days were also combined to create an overall test for significance (Fisher, 1954). This can be calculated as follows: –2Σ(ln P), where P is the probability for each separate test. If all the null hypotheses were true, then this quantity would be χ2-distributed, with 2k degrees of freedom, where k is the number of separate tests (Sokal and Rohlf, 1995).

Scent analysis

Floral scent was collected in 2007–2009 from plants in the field and in the laboratory. Samples were taken of entire plants, and of cut flowers, to test for spatial variation in scent production. Flower partitioning was based on the results of soaking flowers for 24 h in histological 0·01 % neutral red stain to identify osmophores (Vogel, 1990). Two small pieces of P. capensis carcass, of comparable size to that used as ‘small quantity’ in the experiment that tested for the effect of carrion scent quantity on fly attraction were also sampled. Samples obtained using dynamic headspace extraction methods, were analysed by coupled gas chromatography and mass spectrometry (GC-MS). Samples were taken by enclosing plants, flower parts and carrion in polyacetate bags prior to sampling. Air from these bags was then immediately pumped for 20 min through small cartridges filled with 1 mg of Tenax® and 1 mg of Carbotrap® at a flow rate of 50 mL min−1. For every sampling session, controls were taken from an empty polyacetate bag sampled for the same duration.

GC-MS analysis of these samples was carried out using a Varian CP-3800 GC (Varian, Palo Alto, CA. USA) equipped with an Alltech EC-WAX column (Alltech Associates, Inc., Deerfield, IL, USA) with a 30 m × 0·25 mm internal diameter and a 0·25-μm film thickness, coupled to a Varian 1200 quadrupole mass spectrometer in electron-impact ionization mode. To confirm compound identification, samples were also analysed using a Varian VF-5ms non-polar column (Varian, Palo Alto, CA, USA) with a 30 m × 0·25 mm internal diameter and a 0·25-μm film thickness. Cartridges were placed in a Varian 1079 injector equipped with a Chromatoprobe thermal desorbtion device (Gordin and Amirav, 2000; Dötterl et al., 2005). The flow of helium carrier gas was 1 mL min−1. The injector was held at 40 °C for 2 min with a 20 : 1 split ratio and then increased to 200 °C at 200 °C min−1 in splitless mode for thermal desorbtion. After a 3-min hold at 40 °C, the temperature of the GC oven was ramped up to 240 °C at 10 °C min−1 and held there for 12 min. Compounds were identified using the Varian Workstation software with the NIST05 mass spectral library and compounds were verified using retention times of authentic standards (apart from 2-heptanone and propyl isovalerate) and published Kovats indices. Compounds present at similar abundance in the controls were considered to be contaminants and excluded from the analyses. For quantification of emission rates, standards were injected into cartridges and thermally desorbed under identical conditions to the samples.

Effects of floral scent on fly behaviour

A choice experiment was set up to test whether scent, in the absence of visual cues, triggers fly attraction and whether flies discriminate between floral versus carrion scent. A small quantity of carrion (piece of a leg of a P. capensis carcass) and approx. ten S. pumilum plants were placed in separate plant pots, with an additional empty pot used as a control. The three pots were dug into the soil in a triangle with approx. 1 m distance between all pots and with the top of the pot at ground level. To exclude visual cues, the pots were covered with carton lids that resembled the colour of the surrounding soil. Each lid contained a built-in fly trap that allowed flies to enter the pot but not to escape. At the end of each day, the number of flies of each family caught in each pot was recorded. Experiments were run for 6 entire days (approx. 6 h d−1) and new plants were used for each day.

Breeding system, pollination success, pollen transfer efficiency (PTE) and self-pollination rates

To test whether S. pumilum relies on pollinators for fruit set, a breeding system experiment was set up with four treatments per plant: (1) cross-pollination (using three pollinia as pollen donors, removed from plants at least several metres away from a focal plant); (2) self-pollination (using a pollinium from the focal plant); (3) unmanipulated flowers with the pollinaria intact, to test for autogamy; (4) removal of both pollinaria and leaving a flower unpollinated to test for apomixis. All flowers used for the experiment had unpollinated stigmas. Treated plants were covered with mesh bags after the treatment to exclude all insects for the duration of the experiment. All four treatments were applied to a total of 20 plants. Fruit set was recorded after approx. 6 weeks.

Pollination rates of S. pumilum were assessed in four surveys during the flowering season in 2008 to account for intra-seasonal variation. At each survey, 50 plants at least 1 m away from each other were randomly selected. One recently wilted flower from each plant was inspected for pollination by recording removal of pollinaria and/or massulae deposition.

To assess natural fruit set, all ovaries of 21 randomly sampled plants from the field were inspected for swelling at the end of the flowering season in October 2008.

PTE, the proportion of removed massulae deposited on conspecific stigmas (e.g. Johnson et al., 2005) was calculated, using the sample of 200 flowers from the surveys of pollination success. A proxy for the total number of massulae deposited was calculated as the product of the number of flowers with massulae deposition, multiplied by the average number of massulae deposited on pollinated stigmas of 32 randomly selected flowers (eight flowers collected during each of four surveys). The total number of massulae removed was calculated as the product of the number of pollinia removed multiplied by the average number of massulae per pollinium of 27 randomly selected flowers (one pollinium counted per flower sampled from different plants, collected during the 2007 and 2008 flowering seasons).

To assess rates of pollinator-mediated self-pollination, the PTE of massulae transported to stigmas of the same plant was quantified using pollinia labelled with histological stains throughout the 2008 flowering season (Peakall, 1989). Groups of stained plants planted within a 1 m radius, only contained one plant of each stain colour (1 % fast green, 0·2 % rhodamine pink, 1 % methyl blue, 1 % gentian violet, all stock solutions mixed with a drop of tween soap) and different groups were separated by at least 10 m. This distance was deemed adequate for avoiding between-plot stain movement in this dense population. Plants were monitored in the field on a daily basis and once a plant was visited it was taken to the laboratory for counting of the removed pollinaria and stained massulae deposition. The self-pollination rate is calculated by dividing the PTE of self-pollination by the overall PTE (which includes self-pollination + outcrossing). To calculate the PTE of self-pollination, the total number of stained selfed massulae deposited was divided by the number of stained massulae removed (see above) (Johnson and Brown, 2004).

RESULTS

Pollinator observations

Although insects were rarely seen visiting plants, a large number of flies carrying S. pumilum pollinaria were observed and caught. In all cases these were Sarcophagidae (Table 1 and Figs 1 and 2). Fifteen flies with pollinaria (1 male, 14 females) were caught in association with plants and 60 flies with pollinaria (12 males, 48 females) were caught in association with carrion. The male : female ratio did not differ significantly between pollinators caught in association with plants or carrion (χ21 = 1·49, P = 0·22). Male flies carried significantly fewer pollinaria than females [males = 2·8 ± 2·2 (mean ± s.d., n = 13), females = 8·7 ± 7·8 (mean ± s.d., n = 62); Mann–Whitney, U = 176·5, P = 0·0015].

Fig. 2.

Members of the carrion fly assemblage at Leliefontein on a carcass. The majority of flies are Chrysomya chloropygia (blue), but also several Muscidae (Orthellia sp.; green) are present. Note the single sarcophagid fly in the middle, carrying the yellow orchid pollinaria (see insert).

Fig. 2.

Members of the carrion fly assemblage at Leliefontein on a carcass. The majority of flies are Chrysomya chloropygia (blue), but also several Muscidae (Orthellia sp.; green) are present. Note the single sarcophagid fly in the middle, carrying the yellow orchid pollinaria (see insert).

Table 1.

Satyrium pumilum pollinators (as identified by the presence of orchid pollinaria on the sarcophagid flies) caught in association with plants and carrion during 2007 and 2008

 Plant
 
Carrion
 
 Male Female Male Female 
Sarcophaga (Bercaea) sp. (arno and/or inaequalis)* 6 (2·7 ± 1·9) 
Sarcophaga (Liosarcophaga) redux 1 (1) 9 (10·6 ± 6·1) 12 (3·0 ± 2·2) 30 (8·1 ± 6·9) 
Sarcophaga (Nesbittia) guillarmodi 1 (11) 
Sarcophagidae indet. spp. 5 (7·0 ± 4·5) 11 (12·4 ± 12·1) 
Total 1 (1) 14 (9·3 ± 5·7) 12 (3·0 ± 2·2) 48 (8·5 ± 8·3) 
 Plant
 
Carrion
 
 Male Female Male Female 
Sarcophaga (Bercaea) sp. (arno and/or inaequalis)* 6 (2·7 ± 1·9) 
Sarcophaga (Liosarcophaga) redux 1 (1) 9 (10·6 ± 6·1) 12 (3·0 ± 2·2) 30 (8·1 ± 6·9) 
Sarcophaga (Nesbittia) guillarmodi 1 (11) 
Sarcophagidae indet. spp. 5 (7·0 ± 4·5) 11 (12·4 ± 12·1) 
Total 1 (1) 14 (9·3 ± 5·7) 12 (3·0 ± 2·2) 48 (8·5 ± 8·3) 

Fly species names are given according to Pape (1996).

The numbers represent the total number of male and female flies caught of each species, with the mean number of pollinaria/fly ± s.d. in parenthesis.

* Females of these species are currently inseparable (T. Pape, Natural History Museum, Denmark, pers. comm.)

Carrion fly assemblage

On carrion laid out as bait, a total of 314 flies were caught (Table 2). The main visitors were callophorid flies (90·8 %) of the genera Chrysomya (C. chloropyga represented 82·5 % of the flies caught), Lucilia, and Calliphora (Table 2 and Figs 2 and 3). Muscid flies made up 5·4 %, mainly of the genera Orthellia and Ophyra. Sarcophagid flies made up 3·5 % of the assemblage. A single individual belonging to the Anthomyiidae was caught. In contrast, of all flies caught in association with plants (on plants, on experimental set-ups including plants, or carrying pollinaria – which by definition means they visited plants) 98·5 % were Sarcophagidae, mostly of the genus Sarcophaga (Table 2 and Figs 2 and 3).

Fig. 3.

Fly families caught on different sources: (A) flies caught in association with plants versus carrion; (B) flies caught in association with large versus small quantities of carrion scent. Numbers of caught flies are given above the bars.

Fig. 3.

Fly families caught on different sources: (A) flies caught in association with plants versus carrion; (B) flies caught in association with large versus small quantities of carrion scent. Numbers of caught flies are given above the bars.

Table 2.

Members of the 2008 carrion fly assemblage at Leliefontein, South Africa caught in association with S. pumilum plants and carrion, respectively

  Plant (%)
 
Carrion (%)
 
  Male Female Male Female 
Anthmyiidae Anthomyia tempestatum  0  0  0 1 (0·3) 
Calliphoridae Calliphora croceipalpis  0  0 5 (1·6) 9 (2·9) 
 Chrysomyia chloropyga  0  0 41 (13·1) 218 (69·4) 
 Chrysomyia marginata  0  0 1 (0·3) 3 (1·0) 
 Lucilia cuprina  0  0  0 5 (1·6) 
 Lucilia sericata  0  0 1 (0·3) 2 (0·6) 
Muscidae Morellia sp. 1  0  0 1 (0·3) 1 (0·3) 
 Musca sp. 1  0  0  0 5 (1·6) 
 Muscina sp. 1  0  1 (1·5)  0   0 
 Ophyra capensis  0  0 1 (0·3) 3 (1·0) 
 Orthellia sp. 1  0  0  0 2 (0·6) 
 Orthellia sp. 2  0  0  0 1 (0·3) 
 Orthellia sp. 3  0  0  0 3 (1·0) 
Sarcophagidae Sarcophaga (Bercaea) africa  0  0  0 1 (0·3) 
 Sarcophaga (Nesbittia) guillarmodi  0  1 (1·5)  0 2 (0·6) 
 Sarcophaga (Liopygia) nodosa  0  0  0 1 (0·3) 
 Sarcophaga (Liosarcophaga) redux 12 (17·9) 37 (55·2) 5 (1·6) 1 (0·3) 
 Sarcophaga (Bercaea) sp. (arno and/or inequalis)*  0  4 (6·0)  0 1 (0·3) 
 Indet. spp.  0 12 (17·9)  0   0 
Total  12 55 55 259 
  Plant (%)
 
Carrion (%)
 
  Male Female Male Female 
Anthmyiidae Anthomyia tempestatum  0  0  0 1 (0·3) 
Calliphoridae Calliphora croceipalpis  0  0 5 (1·6) 9 (2·9) 
 Chrysomyia chloropyga  0  0 41 (13·1) 218 (69·4) 
 Chrysomyia marginata  0  0 1 (0·3) 3 (1·0) 
 Lucilia cuprina  0  0  0 5 (1·6) 
 Lucilia sericata  0  0 1 (0·3) 2 (0·6) 
Muscidae Morellia sp. 1  0  0 1 (0·3) 1 (0·3) 
 Musca sp. 1  0  0  0 5 (1·6) 
 Muscina sp. 1  0  1 (1·5)  0   0 
 Ophyra capensis  0  0 1 (0·3) 3 (1·0) 
 Orthellia sp. 1  0  0  0 2 (0·6) 
 Orthellia sp. 2  0  0  0 1 (0·3) 
 Orthellia sp. 3  0  0  0 3 (1·0) 
Sarcophagidae Sarcophaga (Bercaea) africa  0  0  0 1 (0·3) 
 Sarcophaga (Nesbittia) guillarmodi  0  1 (1·5)  0 2 (0·6) 
 Sarcophaga (Liopygia) nodosa  0  0  0 1 (0·3) 
 Sarcophaga (Liosarcophaga) redux 12 (17·9) 37 (55·2) 5 (1·6) 1 (0·3) 
 Sarcophaga (Bercaea) sp. (arno and/or inequalis)*  0  4 (6·0)  0 1 (0·3) 
 Indet. spp.  0 12 (17·9)  0   0 
Total  12 55 55 259 

Sarcophagid species names are given according to Pape (1996).

* Females of these species are currently inseparable (T. Pape, Natural History Museum, Denmark, pers. comm.).

Arena experiments

Sarcophagidae released in arenas with plants visited and pollinated S. pumilum flowers in 44 % of the cases (n = 16; Fig. 1; video in Supplementary Data, available online). In one case, sarcophagid larvae were found in flowers inside the arena. Several calliphorid flies [Calliphora croceipalpus 4 out of 4 (including 1 male), Chrysomya chloropygia 3 out of 6 (all females)] also successfully removed S. pumilum pollinaria.

Small versus large quantity carrion

Sarcophagids visited small carrion in significantly higher proportions than members of other fly families when compared with large carrion (Fig. 3) (test results for separate days: χ21 = 2·82, P = 0·093, χ21 = 3·99, P = 0·046, χ21 = 5·82, P = 0·016; test for overall significance: χ26 = 19·208, P < 0·01).

Scent analysis

Neutral red staining resulted in highly localized absorption of stain (Fig. 4). On the outside surface of the flower, absorption mainly occurred in the spur region and at the side of the labellum towards the landing platform. Inside the flower, stain absorption was restricted to the floor of the spur apex, which is densely carpeted with papillae.

Fig. 4.

Local variation in scent emission of S. pumilum flowers. (A–D) Patterns of neutral red staining (scale bar = 5 mm): (A) unstained flower lateral view; (B) stained flower lateral view – the grey arrow highlights the spur region with high levels of neutral red stain absorption; (C) cut unstained flower, (D) cut stained flower – the grey arrow highlights the floor of the spur region with high levels of neutral red stain absorption. (E) Scent emission rates of individual compounds in the spur region versus the rest of the flower. Total emission rates were 3·70 ng flower−1 min−1 in the spur and 0·92 ng flower−1 min−1, respectively.

Fig. 4.

Local variation in scent emission of S. pumilum flowers. (A–D) Patterns of neutral red staining (scale bar = 5 mm): (A) unstained flower lateral view; (B) stained flower lateral view – the grey arrow highlights the spur region with high levels of neutral red stain absorption; (C) cut unstained flower, (D) cut stained flower – the grey arrow highlights the floor of the spur region with high levels of neutral red stain absorption. (E) Scent emission rates of individual compounds in the spur region versus the rest of the flower. Total emission rates were 3·70 ng flower−1 min−1 in the spur and 0·92 ng flower−1 min−1, respectively.

The scent of S. pumilum flowers contains only six compounds (Table 3). This was confirmed using two different GC columns, but only results using the EC-WAX column – which gave the best separations – are reported here. The dominant compound is dimethyl disulfide (DMDS) which comprises >90 % of the blend.

Table 3.

Mean percentages (± s.e.) of individual scent compounds in the blend of whole plants and small carrion, arranged according to retention times of the EC-WAX column, and mean emission rates (± s.e.)

 Whole plant (n = 3) Small carrion (n = 2) 
DMDS 93·80 ± 2·92 17·49 
Propyl isovalerate 0·14 ± 0·14 0·0 
2-Heptanone 1·91 ± 0·57 Trace 
DMTS 1·54 ± 1·19 34·40 
p-Cresol 0·63 ± 0·63 13·82 
Indole 1·97 ± 1·47 0·82 
Other compounds 33·46* 
Emission rates (ng flower−1 min−10·80 ± 0·33 31·41 
 Whole plant (n = 3) Small carrion (n = 2) 
DMDS 93·80 ± 2·92 17·49 
Propyl isovalerate 0·14 ± 0·14 0·0 
2-Heptanone 1·91 ± 0·57 Trace 
DMTS 1·54 ± 1·19 34·40 
p-Cresol 0·63 ± 0·63 13·82 
Indole 1·97 ± 1·47 0·82 
Other compounds 33·46* 
Emission rates (ng flower−1 min−10·80 ± 0·33 31·41 

* Twenty-eight additional compounds were produced in carrion, with an average proportion of 1·19 %.

Calculated including only the compounds DMDS + DMTS + p-cresol + indole.

There is clear spatial quantitative and qualititative intrafloral variation in scent production, with the oligosulfides and 2-heptanone and propyl isovalerate produced in the spur region (Fig. 4), and not (or only in trace amounts) by the rest of the flower, whereas p-cresol and indole are produced by all flower parts. The spur region also has higher emission rates than the rest of the flower (Fig. 4). Although cut flowers have a quantitatively different scent composition compared with whole plants (e.g. elevated rates of p-cresol), this cutting effect would affect both flower parts equally, as the cut surface on each part has by definition the same surface area.

Apart from propyl isovalerate, all flower compounds [DMDS and dimethyl trisulfide (DMTS), 2-heptanone, p-cresol and indole] were also found in carrion scent, where they comprise a large proportion of the entire blend (Table 3). Emission rates of small pieces of carrion were much higher than of individual flowers.

Effects of floral scent on fly behaviour

Only Sarcophagidae were attracted in the choice experiment. These flies were equally attracted to carrion and plant scent [number of flies d−1: 3·00 ± 1·84 in plant trap versus 3·33 ± 0·49 in carrion trap (mean ± s.e., n = 6 pairs); Wilcoxon signed-rank test, Z = 0·95, P = 0·34]. Empty control buckets did not attract any flies at all. Flies therefore significantly prefer carrion scent and plant scent to no scent, respectively (carrion versus empty: Wilcoxon signed-rank test, Z = 2·21, P = 0·027; plants versus empty: Wilcoxon signed-rank test, Z = 2·03, P = 0·042).

Breeding system, pollination success, PTE and self-pollination rates

There were no fruits set in bagged flowers with the pollinaria removed. Outcrossing resulted in 94 %, and self-pollination in 100 % fruit set, respectively. Fruit set in bagged intact flowers was 11 %.

Evidence of visitation was found in 31·5 % of wilted flowers. Natural fruit set was 38·0 ± 34·8 % (mean ± s.d., n = 21 plants). Overall, 10·8 % of removed massulae were deposited on stigmas. Of removed stained massulae, 5·0 % were deposited on stigmas of the same plant (n = 22). Combining these results leads to a self-pollination rate of 45·8 %. There was no correlation between the proportion of self-pollination and the number of flowers per plant (Spearman's ρ = –0·21, P = 0·34).

DISCUSSION

Results from this study show that S. pumilum is pollinated exclusively by sarcophagid carrion flies, mainly females, despite the presence of large numbers of carrion flies from other families in the same area. The experiment testing for the effect of scent quantity on fly attraction revealed that the flies attracted to small quantities of carrion showed a resemblance to those attracted to S. pumilum. Given the key role of S. pumilum scent, which is similar to that of real carrion, and which was equally attractive to sarcophagid flies, we speculate that the basis for the selective attraction of sarcophagid flies by S. pumilum flowers and small items of carrion is that the quantity of scent emitted in both cases is much lower than that for whole animal carcasses. Finally, it was shown that pollination rates are comparatively high in this deceptive orchid, although this is associated with high rates of pollinator-mediated self-pollination.

Plant pollinators and the carrion fly assemblage

The results, which demonstrate pollination of S. pumilum by carrion flies, confirm the predictions made by Vogel (1954) and Johnson (1997) based on plant traits such as the putrid scent and dull brown coloration. Bolus (1893–1896) had already noted that S. pumilum resembles a Stapelia in both its colour and scent, a genus well-known for its carrion fly-pollinated flowers (Meve and Liede, 1994). Apocynaceae pollinaria (in one case identified as Stapelia hirsuta var. hirsuta; P. Bruyns, University of Cape Town, South Africa, pers. comm.) were found attached to several flies which also carried S. pumilum pollinaria. Interestingly, S. pumilum and many Stapelia flowers share a tendency to present their flowers at ground-level, a trait which we interpret as facilitating pollination by carrion flies which tend to alight on the ground before walking or flying further towards flowers. The fact that among all other satyriums this character state is only found in S. pumilum, along with the single occurrence of carrion mimicry, strongly suggests that it is an adaptation in the context of this pollination system (cf. Coddington, 1988). Finally, inflorescence stalks of S. pumilum have been repeatedly observed to elongate after flowering (T. van der Niet, unpubl. res.; H. Stärker, Vienna, Austria, pers. comm.). This reinforces the notion that ground-borne flowers are an adaptation that is only beneficial during the time of flowering.

Despite very rarely seeing flies visiting S. pumilum plants in the field, many flies carrying the orchid's pollinaria were caught, mainly in association with carrion. We interpret this as evidence that the flies have visited the plants and are effective pollinators. This is further supported by the arena experiments, which showed that flies pollinate S. pumilum. Finally, the pollinaria carried by the flies caught in the field match those of S. pumilum, and not of any other of the few co-flowering orchid species.

One of the most notable findings of this study was that, in comparison to the full assemblage of flies attracted to real carrion, there was a strong bias in taxonomic composition of carrion flies that pollinated the orchid. All caught flies that carried orchid pollinaria were Sarcophagidae. In contrast, members of this family represented only a small proportion of the local carrion fly assemblage, which was dominated by C. chloropyga (Calliphoridae). Specific pollination by sarcophagid flies, with a similar bias towards females, has been documented previously in Rafflesiaceae by Pape and Bänziger (2000), and Bänziger and Pape (2004). While Pape and Bänziger (2000) did note fly family-specific attraction to flowers by carrion flies, they did not quantify the abundance of local carrion flies that were not attracted to flowers, nor did they suggest potential causes for selective attraction.

The present results raise the question of why the large numbers of non-sarcophagid carrion flies observed in the vicinity of the orchid never carried pollinaria. Calliphorids are known pollinators of plants with similar traits as S. pumilum, such as the dead horse arums (Stensmyr et al., 2002). Shuttleworth and Johnson (2010) showed that members of Calliphoridae, Muscidae and Sarcophagidae were all pollinators of two species of Eucomis (Hyacinthaceae). The present arena experiments showed that two calliphorid species (including C. chloropyga) were capable of removing pollinaria. We therefore argue that calliphorids and muscids are potential pollinators of S. pumilum and that a mechanical mismatch between plant and insect cannot explain the absence of non-sarcophagids with pollinaria in the field.

A second explanation may be that S. pumilum mimics a specific brood site which only attracts sarcophagids, but not members of other carrion fly families. On carrion laid out as bait, however, sarcophagids with orchid pollinaria were found frequently co-occurring with calliphorids and muscids (Fig. 2). We therefore conclude that brood site specialization is not likely to explain the absence of non-sarcophagids with pollinaria in the field, either.

Instead, we argue that the specialized interaction between S. pumilum and sarcophagid flies, with a bias towards females, can be explained by an aspect of carrion fly niche differentiation which is different from brood site specialization. The initial observation that only Sarcophagidae carried pollinaria of S. pumilum, despite the presence of a much larger carrion fly assemblage including Calliphoridae, led us to attempt to replicate a situation in the field where only sarcophagids were attracted to a particular, real carrion source. Another preliminary observation, that only sarcophagids were attracted to trace scent of carrion, led us to hypothesize that the quantity of carrion scent may determine which flies are attracted to a particular source. Indeed, the experiment using small versus large quantities of carrion resulted in the attraction of significantly fewer Calliphoridae and Muscidae to small carrion quantities, much resembling the pollinator community (Figs 2 and 3). Even though the small quantity of carrion in the experiment still attracted some non-sarcophagid flies, as opposed to S. pumilum, we suspect that a further reduction of carrion scent quantity would have ultimately led to a complete elimination of the non-sarcophagid element from the fly assemblage attracted to the carrion. The approx. 10-fold smaller scent emission rates of typical S. pumilum plants compared with that of the small carrion used in the experiment, may thus explain the complete absence of non-sarcophagids among the orchid's pollinators.

Although the possibility cannot be excluded that the small carrion used in the experiment differed not only quantitatively (smaller absolute amount of volatiles emitted), but also qualitatively (fewer volatiles emitted) compared with the large carrion, we argue that a quantitative explanation of the results is more likely. If the small carrion lacked certain compounds that are essential to attract Calliphoridae and Muscidae, a complete absence of these families during the time that small carrion was exposed would be expected. Instead, only a reduction in the number of these flies attracted to the carrion scent was found.

The differential visitation of carrion and plants between sarcophagids and other fly families may be related to fundamental differences in life-history traits. Sarcophagidae lay few, live, larvae whereas Calliphoridae lay many eggs (Kamal, 1958; Denno and Cothran, 1976). This allows Sarcophagidae to exploit smaller pieces of carrion as it reduces the amount of intraspecific competition (e.g. Denno and Cothran, 1975; Braack, 1987). Some researchers argue that specialization along carrion size gradients is contentious (e.g. Kuusela and Hanski, 1982; Kneidel, 1984b) and that co-existence between different carrion fly families is rather a function of variation in seasonal carrion abundance (Hanski and Kuusela, 1980; Hanski 1987), carrion succession stage (Payne, 1965; Cornaby, 1974; Shi et al., 2009) or resource patchiness (e.g. Hanski, 1987). Several studies including the present one, however, support the notion that Sarcophagidae are proportionally more common on small carrion (e.g. Mönnig and Cilliers, 1944; Denno and Cothran, 1975).

A final line of indirect evidence in favour of carrion size as a major determinant of fly attraction is a morphological difference between Sarcophagidae and Calliphoridae: Sarcophagidae have more sensory pits in their antennae than Calliphoridae (e.g. Chapman, 1982; Sukonstason et al., 2004). With olfactory reception being the most likely function of these sensory pits (e.g. Sukonstason et al., 2004 and references therein), this difference could explain how Sarcophagidae are better able to detect small quantities of carrion scent (Sukontason et al., 2004). This particularly applies to the females, who have more sensory pits than males (Chapman, 1982; Sukontason et al., 2004). This, in turn, may explain the bias towards females that was found among sarcophagid flies carrying pollinaria, which was also found in other carrion mimicking plants (e.g. Pape and Bänziger, 2000; Stensmyr et al., 2002).

Ecological niche differentiation among carrion flies may then explain why only Sarcophagidae, and not Calliphoridae, visit S. pumilum. The scent quantity of S. pumilum probably mimics that of small carrion, on which Sarcophagidae proportionally outnumber Calliphoridae while the reverse is found on large carrion. This differentiation may result from the scent quantity of small carrion being too weak to be detected by Calliphoridae under field conditions. Alternatively, these flies are simply preferentially attracted to stronger carrion scent cues (Spivak et al., 1991).

The interaction between S. pumilum and sarcophagid flies can be regarded as highly specialized, especially in the light of the local carrion fly assemblage. We do not fully understand, however, why this specialized interaction has evolved. The morphology of S. pumilum requires flies of a certain size for effective contact between the fly body and the orchid viscidia (i.e. pollination). A size difference between Sarcophagidae and Calliphoridae, could have driven selective attraction of the former family. We found no evidence for this, however, among fly families at the study site. Instead of viewing the selective attraction of Sarcophagidae, using floral scent quantity as a filter, as an extreme example of specialization, with a fitness benefit over a generalized pollination system in which calliphorids and muscids would also be attracted, our observation may rather be the result of a constraint of scent emission rates. As the size of floral parts, scent emission rates and insect attraction can be correlated (Miyake and Yafuso, 2003), flowers of S. pumilum (already among the largest in the genus) would have to increase further in size to elevate scent emission, which could lead to a conflict with other floral functions such as plant–pollinator fit.

Attractant cues

The observation that a female sarcophagid deposited live larvae on a S. pumilum flower suggests successful carrion mimicry by the plant. Carrion flies use olfactory and visual cues to find carrion (Wall and Fisher, 2001), so the expectation is that mimics resemble their models in these aspects. Although visual cues were not explicitly tested in this study, S. pumilum, with its dull maroon-brown coloration (Fig. 1), has the typical appearance of a carrion mimic (e.g. Fægri and van der Pijl, 1979; Beaman et al., 1988; Proctor et al., 1996). Flowers of S. pumilum clearly mimic carrion scent, although the floral scent compounds only represent a subset of the compounds found in carrion. This result is consistent with scent analyses in some carrion-mimicking stapeliads (Jürgens et al., 2006). Stensmyr et al. (2002) found that mainly oligosulfide compounds from both carcasses and carrion-mimicking arums elicited strong antennal responses of blowflies, even though the blend of both carcass and plants contained several additional compounds. This suggests that for a plant to be a successful carrion mimic, the scent of its flowers may need only to contain a few key compounds of the model to attract the operator.

Sarcophagid flies were equally attracted to carrion and plant scent. Additionally, a comparison between sarcophagid fly species that visit carrion and plants shows strong overlap at the species level (Table 2). Although, theoretically, flesh flies may be attracted by different scent compounds in plants and carrion, a phenomenon that has been reported for calliphorid flies (Kaib, 1974), we consider this unlikely. All the compounds in the floral bouquet of S. pumilum (apart from propyl isovalerate, which comprises <0·14 % of the entire bouquet) are found among two-thirds of the volatiles emitted by the carcass of a rock hyrax, a common mammal across the distribution range of S. pumilum, and whose carcasses are visited by sarcophagid flies. In addition, the most important scent compound in S. pumilum, dimethyl disulfide, has been shown to elicit strong antennal responses in carrion-frequenting insects (Stensmyr et al., 2002; Kalinová et al., 2009) and, recently, Shuttleworth and Johnson (2010) showed that adding a mix of oligosulfides to a wasp-pollinated plant is sufficient to attract sarcophagid flies.

The co-occurrence of oligosulfides and p-cresol + indole in the scent of S. pumilum, respectively, seems at odds with comparative scent patterns found in Araceae (Kite et al., 1998) and stapeliads (Jürgens et al., 2006). In both cases two distinct scent profiles were recognized: scent of plant species dominated by oligosulfides was associated with carrion mimicry, whereas scent of plant species dominated by p-cresol, indole and 2-heptanone was associated with dung mimicry. Neither of these studies, however, included the scent of carrion and dung in their analyses. In a recent analysis, Johnson and Jürgens (2010) indeed showed that carrion samples were dominated by oligosulfides, whereas dung samples contained p-cresol and indole. Several plant and stinkhorn fungus samples in their analysis, however, contained a combination of oligosulfides as well as p-cresol and indole, similar to the present findings. Dekeirsschieter et al. (2009) showed that volatiles emitted by a pig carcass, at several decomposition stages, also included these compounds together, in agreement with the present finding of compounds emitted by rock hyrax carcasses. A partitioning of plant mimicry systems into carrion versus dung mimicry according to scent profiles is therefore not yet supported by conclusive evidence and a broader sampling of particularly various kinds of carrion and dung is required.

For rewarding plant species, the location of the nectaries often facilitates positioning of insect visitors so as to assure contact with anthers and stigma. Deceptive plant species require alternative mechanisms for insect positioning to effect pollination. Due to the proportionally large size of carrion-mimicking flowers (Davis et al., 2008), contacting anthers and stigma is further complicated. Some species have evolved complex trap flowers, where insects are guided by olfactory and visual cues towards a trap chamber where pollination takes place while insects are held captive, such as in Aristolochia and Ceropegia (Vogel, 1990; Proctor et al., 1996; Burgess et al., 2004). Carrion-mimicking orchids represent perhaps the most complicated case, as the possession of a gynostemium requires even more precise positioning and renders trap mechanisms often impossible. The present results from the neutral red staining and the scent analysis suggest that correct positioning of the insect with regard to the viscidia and stigma in S. pumilum is achieved by spatial variation in scent production. Flies are thus initially attracted to the entire inflorescence as a unit, as it perceives the whole scent bouquet from a distance. Once a fly has reached an inflorescence, it is strongly attracted to the particular scent that is emitted from the spur region, which could be considered an osmophore (Vogel, 1990). As a fly explores a flower and approaches this spur region, it is forced to enter the flower's hood, thereby touching the viscidia and/or depositing massulae from previously attached pollinaria on its abdomen, and so effecting pollination (Fig. 1).

Pollination and self-pollination rates

Pollination rates and fruit set in S. pumilum are high compared with those of other deceptive orchids (Neiland and Wilcock, 1998; Tremblay et al., 2005; Scopece et al., 2010). This is particularly so in comparison with sexually deceptive orchids (Neiland and Wilcock, 1998; Scopece et al., 2010), another system which relies mainly on scent cues for insect attraction (Schiestl, 2005). A key difference between sexually deceptive orchids and S. pumilum may be the presence of a large number of non-active scent compounds in the former group (e.g. Mant et al., 2005). In contrast, S. pumilum has a relatively simple scent bouquet which only contains compounds that largely overlap with compounds from real carrion. As these compounds are unusual for flowers that do not mimic carrion, we presume that they therefore have a functional role in attracting carrion insects. Similarly compound-poor bouquets were found for carrion-mimicking stapeliads (Jürgens et al., 2006; Johnson and Jürgens, 2010). The absence of non-active compounds may limit the opportunity for insects to learn to associate certain scent cues with the deceptive nature of the orchid, which could prevent avoidance behaviour towards the orchid by experienced individuals (e.g. Kunze and Gumbert, 2001), although this mechanism remains to be tested for flies.

Compared with the few studies available that measure levels of outcrossing for deceptive versus rewarding orchids, the proportion of massulae deposited as a result of self-pollination in S. pumilum is more typical for rewarding species than for deceptive species (Jersakova et al., 2006). This means that both pollination success and self-pollination rates of this carrion-mimicking deceptive orchid are closer to the ranges found for rewarding orchids. Johnson (1994) already showed that fruit set in a Batesian orchid mimic was comparable to that of a rewarding species that shared the same pollinator. There are unfortunately very few comparative data on fruit set and pollinator-mediated self-pollination rates available across deceptive pollination systems to arrive at broad-scale generalizations about the evolution of certain mimicry systems with regards to pollination success and outcrossing. The results seem, however, to indicate that deception through carrion mimicry may have evolved for different reasons than to promote outcrossing, which to date is considered one of the major drivers of the evolution of deceit pollination in orchids (Cozzolino and Widmer, 2005; Jersakova et al., 2006). This is further supported by the absence of protandry in S. pumilum, which also promotes outcrossing in orchids (Jersakova and Johnson, 2007). Stigmas are receptive from the first day after anthesis on the multi-flowered inflorescences which last for at least 2 weeks.

An alternative explanation for the evolution of carrion mimicry, may be Stebbins' most effective pollinator principle (Stebbins, 1970) which emphasizes the role of a geographically variable pollinator mosaic in floral evolution (Johnson, 2010). Plants adapt to pollinators which are locally most effective. The arid Karoo region of South Africa, the centre of diversity of many carrion mimics such as the Stapelieae (White and Sloane, 1933), is characterized by low diversity of flowering plants and of important pollinator groups such as bees (Kuhlmann, 2009). On the other hand, herbivorous mammals and their carcasses have likely been present historically (Milton et al., 1990), and calliphorid and sarcophagid carrion flies are recorded as being common in the area (Hepburn, 1943). The locally large proportion of flies, compared with other major pollinator groups, may have rendered them relatively effective pollinators, and may have driven the evolution of carrion mimicry in several unrelated plant groups.

Conclusions

We have conclusively shown that, of all insect species visiting carrion, only a very minor proportion pollinated the carrion-mimicking plant species S. pumilum. We argued that this most likely reflects niche differentation among carrion insects, with a critical role for scent quantity in brood- and food-site location. One implication of this study is that it may be necessary to reinterpret the sapromyiophilous pollination syndrome from one considered to be generalized for ‘carrion flies’ to one that may reflect several specialized interactions involving different groups of flies. Our study underlines the need to define specialization in terms of the proportion of animal species in local assemblages that could potentially pollinate flowers, and to understand this specificity in terms of the ecology and life history of the insects involved.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of a video showing a sarcophagid fly that carries several S. pumilum pollinaria repeatedly entering a S. pumilum flower.

ACKNOWLEDGMENTS

We thank Ruth Cozien for extensive help during field work, including unpleasant experiments involving carrion. We thank Annelise le Roux for providing excellent research facilities at the Succulent Karoo Knowledge Center, and Sarah and Joseph Grootman for facilitating fieldwork on communal land. Thomas Pape (Natural History Museum of Denmark) identified Sarcophagidae, Mervyn Mansell (University of Pretoria) identified Calliphoridae, Ray Miller (University of KwaZulu-Natal) identified Calliphoridae and Muscidae. Adam Shuttleworth and Ruth Cozien provided useful comments on an earlier version of this manuscript.

LITERATURE CITED

Archer
MS
Elgar
MA
Effects of decomposition on carcass attendance in a assemblage of carrion-breeding flies
Medical and Veterinary Entomology
 , 
2003
, vol. 
17
 (pg. 
263
-
271
)
Bänziger
H
Pape
T
Flowers, faeces and cadavers: natural feeding and laying habits of flesh flies in Thailand (Diptera: Sarcophagidae, Sarcophaga spp.)
Journal of Natural History
 , 
2004
, vol. 
38
 (pg. 
1677
-
1694
)
Beaman
RS
Decker
PJ
Beaman
JH
Pollination of Rafflesia (Rafflesiaceae)
American Journal of Botany
 , 
1988
, vol. 
75
 (pg. 
1148
-
1162
)
Blackith
RE
Blackith
RM
Insect manifestations of small corpses
Journal of Natural History
 , 
1990
, vol. 
24
 (pg. 
699
-
709
)
Bolus
H
Icones orchidearum austro-africanarum extra-tropicarum
 , 
1893
, vol. 
1
 
London
William Wesley & Son
 
(in two parts)
Born
J
Linder
HP
Desmet
P
The greater Cape Floristic Region
Journal of Biogeography
 , 
2007
, vol. 
34
 (pg. 
147
-
162
)
Braack
LEO
Community dynamics of carrion-attendant arthropods in tropical African woodland
Oecologia
 , 
1987
, vol. 
72
 (pg. 
402
-
409
)
Burgess
KS
Singfield
J
Melendez
V
Kevan
PG
Pollination biology of Aristolochia grandiflora (Aristolochiaceae) in Veracruz, Mexico
Annals of the Missouri Botanical Garden
 , 
2004
, vol. 
91
 (pg. 
346
-
356
)
Carvalho
LML
Thyssen
PJ
Linhares
AX
Palhares
FAB
A checklist of arthropods associated with pig carrion and human corpses in southeastern Brazil
Memorias do Instituto Oswaldo Cruz
 , 
2000
, vol. 
95
 (pg. 
135
-
138
)
Chapman
RF
The insects: structure and function.
 , 
1982
Cambridge, MA
Harvard University Press
Coddington
JA
Cladistic tests of adaptational hypotheses
Cladistics
 , 
1988
, vol. 
4
 (pg. 
3
-
22
)
Cornaby
BW
Carrion reduction by animals in contrasting tropical habitats
Biotropica
 , 
1974
, vol. 
6
 (pg. 
51
-
63
)
Cozzolino
S
Widmer
A
Orchid diversity: an evolutionary consequence of deception?
Trends in Ecology and Evolution
 , 
2005
, vol. 
20
 (pg. 
487
-
494
)
Dafni
A
Mimicry and deception in pollination
Annual Review of Ecology and Systematics
 , 
1984
, vol. 
15
 (pg. 
259
-
278
)
Davis
CC
Endress
PK
Baum
DA
The evolution of floral gigantism
Current Opinion in Plant Biology
 , 
2008
, vol. 
11
 (pg. 
49
-
57
)
Dekeirsschieter
J
Verheggen
FJ
Gohy
M
, et al.  . 
Cadaveric volatile organic compounds released by decaying pig carcasses (Sus domesticus L.) in different biotopes
Forensic Science International
 , 
2009
, vol. 
189
 (pg. 
46
-
53
)
Denno
RF
Cothran
WR
Niche relationships of a assemblage of necrophagous flies
Annals of the Entomological Society of America
 , 
1975
, vol. 
68
 (pg. 
741
-
754
)
Denno
FR
Cothran
WR
Competitive interactions and ecological strategies of sarcophagid and calliphord flies inhabiting rabbit carrion
Annals of the Entomological Society
 , 
1976
, vol. 
69
 (pg. 
109
-
113
)
Dötterl
S
Wolfe
LM
Jürgens
A
Qualitative and quantitative analyses of flower scent in Silene latifolia
Phytochemistry
 , 
2005
, vol. 
66
 (pg. 
203
-
213
)
Fægri
K
van der Pijl
L
The principles of pollination ecology
 , 
1979
New York, NY
Pergamon
Fisher
RA
Statistical methods for research workers
 , 
1954
Edinburgh
Oliver & Boyd
Fuller
ME
The insect inhabitants of carrion: a study in animal ecology
Bulletin of the Council for Scientific and Industrial Research
 , 
1934
, vol. 
82
 (pg. 
5
-
62
)
Gordin
A
Amirav
A
SnifProbe: new method and device for vapor and gas sampling
Journal of Chromatography
 , 
2000
, vol. 
903
 (pg. 
155
-
172
)
Grassberger
M
Frank
C
Initial study of Arthropod succession on pig carrion in a central European urban habitat
Journal of Medical Entomology
 , 
2004
, vol. 
41
 (pg. 
511
-
523
)
Hanski
I
Carrion fly community dynamics: patchiness, seasonality, and coexistence
Ecological Entomology
 , 
1987
, vol. 
12
 (pg. 
257
-
266
)
Hanski
I
Kuusela
S
The structure of carrion fly communities: differences in breeding seasons
Annales Zoologici Fennici
 , 
1980
, vol. 
17
 (pg. 
185
-
190
)
Hepburn
GA
Sheep blowfly research. V. Carcasses as sources of blowflies
Onderstepoort Journal of Veterinary Science and Animal Industry
 , 
1943
, vol. 
18
 (pg. 
59
-
72
)
Hewadikaram
KA
Goff
ML
Effect of carcass size on rate of decomposition and arthropod succession patterns
The American Journal of Forensic Medicine and Pathology
 , 
1991
, vol. 
12
 (pg. 
235
-
240
)
Jersakova
J
Johnson
SD
Lack of floral nectar reduces self-pollination in a fly-pollinated orchid
Oecologia
 , 
2006
, vol. 
147
 (pg. 
60
-
68
)
Jersakova
J
Johnson
SD
Protandry promotes male pollination success in a moth-pollinated orchid
Functional Ecology
 , 
2007
, vol. 
21
 (pg. 
496
-
504
)
Jersakova
J
Johnson
SD
Kindlmann
P
Mechanisms and evolution of deceptive pollination in orchids
Biological Reviews
 , 
2006
, vol. 
81
 (pg. 
219
-
235
)
Johnson
MD
Seasonal and microseral variations in the insect populations on carrion
American Midland Naturalist
 , 
1975
, vol. 
93
 (pg. 
79
-
90
)
Johnson
SD
Evidence for Batesian mimicry in a butterfly-pollinated orchid
Biological Journal of the Linnean Society
 , 
1994
, vol. 
53
 (pg. 
91
-
104
)
Johnson
SD
Insect pollination and floral mechanisms in south African species of Satyrium (Orchidaceae)
Plant Systematics and Evolution
 , 
1997
, vol. 
204
 (pg. 
195
-
206
)
Johnson
SD
The pollination niche and its role in the diversification and maintenance of the southern African flora
Philosophical Transactions of the Royal Society B – Biological Sciences
 , 
2010
, vol. 
365
 (pg. 
499
-
516
)
Johnson
SD
Brown
M
Transfer of pollinaria on birds' feet: a new pollination system in orchids
Plant Systematics and Evolution
 , 
2004
, vol. 
244
 (pg. 
181
-
188
)
Johnson
SD
Jürgens
A
Convergent evolution of carrion and faecal scent mimicry in fly-pollinated flowers and a stinkhorn fungus
South African Journal of Botany
 , 
2010
, vol. 
76
 (pg. 
796
-
807
)
Johnson
SD
Nilsson
LA
Pollen carryover, geitonogamy, and the evolution of deceptive pollination systems in orchids
Ecology
 , 
1999
, vol. 
80
 (pg. 
2607
-
2619
)
Johnson
SD
Neal
PR
Harder
LD
Pollen fates and the limits on male reproductive success in an orchid population
Biological Journal of the Linnean Society
 , 
2005
, vol. 
86
 (pg. 
175
-
190
)
Jürgens
A
Dötterl
S
Meve
U
The chemical nature of fetid floral scents in stapeliads (Apocynaceae-Asclepoadoideae-Ceropegieae)
New Phytologist
 , 
2006
, vol. 
172
 (pg. 
452
-
468
)
Kaib
M
Die Fleisch- und Blumenduftrezeptoren auf der Antenne der Schmeissfliege Calliphora vicina
Journal of Comparative Physiology
 , 
1974
, vol. 
95
 (pg. 
105
-
121
)
Kalinová
B
Podskalská
H
Růžička
J
Hoskovec
M
Irresistible bouquet of death – how are burying beetles (Coleoptera: Silphidae: Nicrophorus) attracted by carcasses
Naturwissenschaften
 , 
2009
, vol. 
96
 (pg. 
889
-
899
)
Kamal
AS
Comparative study of thirteen species of sarcophagous Calliphoridae and Sarcophagidae (Diptera). I. Bionomics
Annals of the Entomological Society
 , 
1958
, vol. 
51
 (pg. 
261
-
271
)
Kite
GC
Hetterscheid
WLA
Lewis
MJ
, et al.  . 
Owens
SJ
Rudall
PJ
Inflorescence scents and pollinators of Arum and Amorphophallus (Araceae)
Reproductive biology
 , 
1998
Kew, Richmond
Royal Botanic Gardens
(pg. 
295
-
315
)
Kneidel
KA
Influence of carcass taxon and size on species composition of carrion-breeding Diptera
American Midland Naturalist
 , 
1984
, vol. 
111
 (pg. 
57
-
63
)
Kneidel
KA
Competition and disturbance in communities of carrion-breeding diptera
Journal of Animal Ecology
 , 
1984
, vol. 
53
 (pg. 
849
-
865
)
Kuhlmann
M
Patterns of diversity, endemism and distribution of bees (Insecta: Hymenoptera: Anthophila) in southern Africa
South African Journal of Botany
 , 
2009
, vol. 
75
 (pg. 
726
-
738
)
Kunze
J
Gumbert
A
The combined effect of color and odor on flower choice behavior of bumble bees in flower mimicry systems
Behavioral Ecology
 , 
2001
, vol. 
12
 (pg. 
447
-
456
)
Kuusela
S
Hanski
I
The structure of carrion fly communities: the size and the type of carrion
Holarctic Ecology
 , 
1982
, vol. 
5
 (pg. 
337
-
348
)
Mant
J
Peakall
R
Schiestl
FP
Does selection on floral odor promote differentiation among populations and species of the sexually deceptive orchid genus Ophrys?
Evolution
 , 
2005
, vol. 
59
 (pg. 
1449
-
1463
)
Meve
U
Liede
S
Floral biology and pollination in stapeliads – new results and a literature review
Plant Systematics and Evolution
 , 
1994
, vol. 
192
 (pg. 
99
-
116
)
Milton
SJ
Siegfried
WR
Dean
WRJ
The distribution of Epizoochoric plant species: a clue to the prehistoric use of arid Karoo rangelands by large herbivores
Journal of Biogeography
 , 
1990
, vol. 
17
 (pg. 
24
-
34
)
Miyake
T
Yafuso
M
Floral scents affect reproductive success in fly-pollinated Alocasia odora (Araceae)
American Journal of Botany
 , 
2003
, vol. 
90
 (pg. 
370
-
376
)
Mönnig
HO
Cilliers
PA
Sheep blowfly research. VII. Investigations in the cape winter-rainfall areas
Onderstepoort Journal of Veterinary Science and Animal Industry
 , 
1944
, vol. 
19
 (pg. 
71
-
77
)
Neiland
MRM
Wilcock
CC
Fruit set, nectar reward, and rarity in the Orchidaceae
American Journal of Botany
 , 
1998
, vol. 
85
 (pg. 
1657
-
1671
)
van der Niet
T
Linder
HP
Bytebier
B
Bellstedt
DU
Molecular markers reject monophyly of the subgenera of Satyrium (Orchidaceae)
Systematic Botany
 , 
2005
, vol. 
30
 (pg. 
263
-
274
)
Pape
T
Catalogue of the Sarcophagidae of the world (Insecta: Diptera)
Memoirs of Entomology International
 , 
1996
, vol. 
8
 (pg. 
1
-
558
)
Pape
T
Bänziger
H
Two new species of Sarcophaga (Diptera: Sarcophagidae) among pollinators of newly discovered Sapria ram (Rafflesiaceae)
The Raffles Bulletin of Zoology
 , 
2000
, vol. 
48
 (pg. 
201
-
208
)
Payne
JA
A summer carrion study of the baby pig Sus scrofa Linnaeus
Ecology
 , 
1965
, vol. 
46
 (pg. 
592
-
602
)
Peakall
R
A new technique for monitoring pollen flow in orchids
Oecologia
 , 
1989
, vol. 
79
 (pg. 
361
-
365
)
van der Pijl
L
Dodson
CH
Orchid flowers: their pollination and evolution
 , 
1966
Coral Gables, FL
University of Miami Press
Proctor
M
Yeo
P
Lack
A
The natural history of pollination
 , 
1996
Portland, OR
Timber Press
Satoshi
S
Filth flies collected by the late Dr. T. Ohse in Ethiopia and thirteen African countries with record of Dr. Kano's collection in Africa. 2. Calliphoridae and Sarcophagidae
Medical Entomology and Zoology
 , 
2001
, vol. 
52
 (pg. 
209
-
217
)
Schiestl
FP
On the success of a swindle: pollination by deception in orchids
Naturwissenschaften
 , 
2005
, vol. 
92
 (pg. 
255
-
264
)
Scopece
G
Cozzolino
S
Johnson
SD
Schiestl
FP
Pollination efficiency and the evolution of specialized deceptive pollination systems
American Naturalist
 , 
2010
, vol. 
175
 (pg. 
98
-
105
)
Shi
Y-W
Liu
X-S
Wang
H-Y
Zhang
R-J
Seasonality of insect succession on exposed rabbit carrion in Guangzhou, China
Insect Science
 , 
2009
, vol. 
16
 (pg. 
425
-
439
)
Shuttleworth
A
Johnson
SD
The missing stink: sulphur compounds can mediate a shift between fly and wasp pollination systems
Proceedings of the Royal Society B – Biological Sciences
 , 
2010
, vol. 
277
 (pg. 
2811
-
2819
)
Sleeman
DP
Jones
P
Cronin
JN
Investigations of an association between the stinkhorn fungus and badger setts
Journal of Natural History
 , 
1997
, vol. 
31
 (pg. 
983
-
992
)
Sokal
RR
Rohlf
FJ
Biometry: the principles and practice of statistics in biological research.
 , 
1995
New York, NY
W.H. Freeman & Co
Spivak
M
Conlon
D
Bell
WJ
Wind-guided landing and search behavior in fleshflies and blowflies exploiting a resource patch (Diptera: Sarcophagidae, Calliphoridae)
Annals of the Entomological Society of America
 , 
1991
, vol. 
84
 (pg. 
447
-
452
)
Stebbins
GL
Adaptive radiation of reproductive characteristics in angiosperms. I. Pollination mechanisms
Annual Review of Ecology and Systematics
 , 
1970
, vol. 
1
 (pg. 
307
-
326
)
Stensmyr
MC
Urru
I
Collu
I
Celander
M
Hansson
BS
Angioy
AM
Rotting smell of dead-horse arum florets
Nature
 , 
2002
, vol. 
420
 (pg. 
625
-
626
)
Sukontason
K
Sukontason
KL
Piangjai
S
, et al.  . 
Antennal sensilla of some forensically important flies in families Calliphoridae, Sarcophagidae and Muscidae
Micron
 , 
2004
, vol. 
35
 (pg. 
671
-
679
)
Tremblay
RL
Ackerman
JD
Zimmerman
JK
Calvo
RN
Variation in sexual reproduction in orchids and its evolutionary consequences: a spasmodic journey to diversification
Biological Journal of the Linnean Society
 , 
2005
, vol. 
84
 (pg. 
1
-
54
)
Vogel
S
Blütenbiologische Typen als Elemente der Sippengliederung
 , 
1954
Jena
Gustav Fischer Verlag
Vogel
S
The role of scent glands in pollination: on the structure and function of osmophores.
 , 
1990
Washington, DC
Smithsonian Institution Libraries and National Science Foundation
Wall
R
Fisher
P
Visual and olfactory cue interaction in resource-location by the blowfly, Lucilla sericata
Physiological Entomology
 , 
2001
, vol. 
26
 (pg. 
212
-
218
)
White
A
Sloane
BL
The Stapelieae: an introduction to the study of this tribe of Asclepiadaceae
 , 
1933
Pasadena, CA
Abbey San Encino Press

Comments

0 Comments