, Volume 97, Issue 12, pp 1107–1112

Sticky snack for sengis: The Cape rock elephant-shrew, Elephantulus edwardii (Macroscelidea), as a pollinator of the Pagoda lily, Whiteheadia bifolia (Hyacinthaceae)


    • Department of Botany and ZoologyUniversity of Stellenbosch
Short Communication

DOI: 10.1007/s00114-010-0723-6

Cite this article as:
Wester, P. Naturwissenschaften (2010) 97: 1107. doi:10.1007/s00114-010-0723-6


Following the recent discovery of rodent pollination in the Pagoda lily, Whiteheadia bifolia (Hyacinthaceae) in South Africa, now the Cape rock elephant-shrew, Elephantulus edwardii (Macroscelidea, Afrotheria) is reported as an additional pollinator. Elephant-shrews, live-trapped near W. bifolia plants, were released in two terraria, containing the plants. The animals licked nectar with their long and slender tongues while being dusted with pollen and touching the stigmas of the flowers with their long and flexible noses. The captured elephant-shrews had W. bifolia pollen in their faeces, likely as a result of grooming their fur as they visited the flowers without eating or destroying them. The animals mostly preferred nectar over other food. This is the first record of pollination and nectar consumption in the primarily insectivorous E. edwardii, contributing to the very sparse knowledge about the behaviour of this unique clade of African mammals, as well as pollination by small mammals.


Cape rock elephant-shrewElephantulus edwardiiMacroscelideaPollinationTherophilyWhiteheadia bifolia


The Pagoda Lily, sometimes called Bobbejaanskoen (baboon shoe), Whiteheadia bifolia (Hyacinthaceae), is restricted to the arid winter-rainfall region of South Africa and southern Namibia, growing in sheltered shady rock sites (Fig. 1a). This bulbous geophyte is a bizarre looking plant with its two large fleshy leaves lying flat on the ground and the multi-flowered inflorescence with large green bracts (Fig. 1a). The inconspicuous green flowers offer extremely viscous, sticky nectar. Recently, the species was discovered to be pollinated by nocturnal rodents (Wester 2009; Wester et al. 2009). Rodent pollination is a rare phenomenon and is known only for a few plant species in South Africa, tropical Africa, China, India, Malaysia, Australia and the Neotropics (Carpenter 1978; Wiens and Rourke 1978; Lumer 1980; Grünmeier 1990; Cocucci and Sérsic 1998; Yumoto et al. 1999; Tandon et al. 2003; Wang et al. 2008; Wiens et al. 2008; Wester et al. 2009).
Fig. 1

aWhiteheadia bifolia plants growing in rock crevices. bElephantulus edwardii in rock crevice behind old W. bifolia inflorescence. E. edwardii visiting W. bifolia flowers (ce). cE. edwardii licking nectar with its pink tongue; note the pollen on the long nose. dE. edwardii leaning on the inflorescence, licking nectar with its tongue and touching the pollen-sacs. eE. edwardii touching the pollen-sacs and getting dusted with pollen on its nose. fE. edwardii yawning, bending its flexible nose upwards, showing the long and slender tongue (here not fully protruded). Scale bar = 1 cm (cf taken in a terrarium)

In the above-mentioned Whiteheadia study, Namaqua rock mice (Aethomys namaquensis; family Muridae, order Rodentia) were observed in the field and in a laboratory experiment pollinating the flowers of the species, whereas no other flower visitors (insects, birds) were recorded. W. bifolia shares characters of the rodent pollination floral syndrome such as visually inconspicuous, robust, bowl-shaped flowers close to the ground, with easily accessible nectar, exserted reproductive organs, as well as a characteristic scent (Wester 2009; Wester et al. 2009).

Subsequent live-trapping retrieved two individuals of the Cape rock elephant-shrew, Elephantulus edwardii, an endemic small mammal of South Africa. The common English name refers to the resemblance between their long slender mobile noses and the trunk of an elephant, as well as an assumed relationship with the true shrews (family Soricidae, order Insectivora) because of superficial similarities. As they are not closely related to the true shrews many authors use the Bantu term ‘sengi’ (Skinner and Chimimba 2005; Rathbun and Kingdon 2006). They belong to the order Macroscelidea (family Macroscelididae) currently including 17 species in four genera distributed in Africa (Rathbun 2009). Based largely on morphological traits, it was considered that they are related to lagomorphs and rodents (including mice), but more evidence from fossils and molecular data strongly suggest that they are a part of the clade Afrotheria (including aardvarks, tenrecs, golden moles, hyracoids, sirenians and elephants) (Skinner and Chimimba 2005).

As elephant-shrews are similar to mice regarding morphology and size, one could imagine that they might also act as pollinators; however, their primarily insectivorous diet (van Deventer and Nel 2006; Rathbun 2009, see also Skinner and Chimimba 2005) would make them unexpected flower visitors. As almost nothing is known about their behaviour at flowers (Wiens and Rourke 1978; Wiens et al. 1983; Fleming and Nicolson 2002; Johnson et al. 2008; Smithies and Burgoyne 2010), observations were carried out in order to investigate whether these animals are also possible pollinators of W. bifolia.

Material and methods

Plants and study site

Live-trapping and observations of captive animals were carried out from 25 to 29 July 2009 at a ca. 0.25 ha plot along the Sevilla rock art trail on Travellers’ Rest Farm, Agter-Pakhuis in the northern Cederberg area of South Africa (S 32° 04′ 31.52″, E 19° 04′ 48.42″; elevation 320 m). Here, approximately 100 individual plants of W. bifolia occurred in shady rock crevices and cavities.

Trapping of mammals and pollen loads in faeces

Over three nights, mammal traps, baited with a mixture of peanut butter and rolled oats, were laid out near W. bifolia plants in the late afternoon and were checked the next morning (41/38/31 traps). Captured animals were removed and identified. Faeces from the elephant-shrews were collected from the traps, stored in a fridge and later crushed in molten fuchsin glycerine gelatine on a slide and examined microscopically for the presence of pollen.

Observations of captive animals

Two captured elephant-shrews were released in two different glass terraria. Each terrarium was equipped with a ca. 5 cm deep layer of sand, some stones providing hiding places, and one to three flowering W. bifolia plants. Foraging behaviour of the elephant-shrews was observed for 14 h in total over four nights and two afternoons between noon and midnight.


Trapping of mammals and pollen loads

Twenty-two Namaqua rock mice and two Cape rock elephant-shrews (Fig. 1b) were caught. The faecal samples of the elephant-shrews contained pollen of W. bifolia (2,256, 82 grains) and other species (different species of Asteraceae: 128, 461; three other unidentified species, 28, 1).

Foraging behaviour of the elephant-shrews

After release in the terrarium, the two elephant-shrews hid themselves between the rocks, but after approximately half an hour, they emerged and visited the flowers. More than 50 flower visits were observed on 4 days between 1400 and 2300 hours. The animals went to the different inflorescences one after another, visiting several flowers per inflorescence (Fig. 1 c–e; Supplementary online material: slides 1–42). They spent several seconds at each flower. With their long and flexible noses, they went between the stamens and licked the viscous nectar between the ovary and the stamens with their long, slender tongues (Fig. 1c, d, f). In doing so, they always touched the pollen-sacs and got visibly dusted with pollen on the long nose (Fig. 1c–e); they also touched the stigmas. The elephant-shrews lapped the nectar without feeding on pollen directly or eating, damaging or destroying the flowers. Foraging animals sometimes lent on the inflorescences with their paws (Fig. 1d, e). Sometimes the elephant-shrews visited the flowers very eagerly and protruded their snouts very forcefully into the flowers, especially after presenting fresh inflorescences with more nectar and a stronger scent. The elephant-shrews sometimes groomed their fur between the flower visits. Mostly, the animals preferred the nectar over other food (e.g., peanut butter with rolled oats and apples) or water, and while eating the food or drinking water, they returned to the flowers again and again.


Evidence of pollination by elephant-shrews

The present data confirm the original hypothesis that elephant-shrews pollinate W. bifolia. Captured elephant-shrews went purposefully to different inflorescences one after another and visited the flowers keenly and non-destructively. While licking up the nectar, they touched the pollen-sacs and stigmas and got dusted with pollen on their long noses.

In comparison to mice that touch the pollen-sacs and stigmas with their fur around the nose and snout, the elephant-shrews only touch the flowers with their long noses as they are able to protrude their tongues far beyond their snout. However, they also changed position with their nose and tongue in order to lap the nectar around the ovary. Thus, although their nose is slender, it is more than obvious that they transfer pollen between the plants.

The faeces of the elephant-shrews contained not only large amounts of W. bifolia pollen, but also pollen of other plant species. This can be explained by the fact that E. edwardii is omnivorous and includes plant material in its diet (van Deventer and Nel 2006). In principle, E. edwardii might be able to extract protein from pollen (van Tets 1997). However, the occurrence of W. bifolia pollen in the elephant-shrews’ faeces is likely a result of grooming the fur after flower visits as it was observed that the animals never fed directly on pollen (see also Fleming and Nicolson 2002).

Elephant-shrews: understudied pollinators

Data on elephant-shrews as pollinators are extremely scarce. Flower visits were indirectly proven by Wiens et al. (1983) who found Protea pollen on the noses and in faeces of E. edwardii captured near rodent-pollinated Protea species. In their study, the captured animals only occasionally licked the surface of the inflorescences and never lapped nectar from the nectar reservoirs of the flowers presented to them. Fleming and Nicolson (2002) also found pollen of the rodent-pollinated Protea humiflora on the noses and in the faeces of captured E. edwardii. However, they presumed that the animals fed on insects that were abundant in the inflorescences (see also Fleming and Nicolson 2003). This assumption probably stems from the fact that at that time elephant-shrews were thought to be strictly insectivorous (Perrin 1997).

The present study is the first showing clearly that E. edwardii visits flowers for nectar. A related species, Elephantulus brachyrhynchus, was recently shown to visit and possibly pollinate the parasite Cytinus visseri (Cytinaceae; Johnson et al. 2008). It is very likely that further species of elephant-shrews might be pollinators.

Elephant-shrews: part of the non-flying mammal pollination syndrome

Like rodents, elephant-shrews are an additional group of pollinating animals that can be categorised as ‘non-flying mammals’ to differentiate them from bats. This group also includes marsupials and primates (Carthew and Goldingay 1997). The plants have some floral characters in common that indicate adaptation to these animals, forming the pollination syndrome called therophily (Rebelo and Breytenbach 1987; Proctor et al. 1996). However, the morphology and behaviour of those mammals varies greatly, resulting in a wide range of floral diversity within the syndrome (Carthew and Goldingay 1997). For example, flowers pollinated by primates are in general very large; those pollinated by marsupials are normally found in the canopy, whereas flowers pollinated by rodents seem to be rather small and near the ground (Turner 1982; Kress et al. 1994; Wester et al. 2009). Elephant-shrews are similar to rodents concerning size and foraging behaviour at flowers, and thus, despite their phylogenetically unique position, seem to be unlikely to act as a unique selective force driving floral traits. However, little is currently known about the role of elephant-shrews as pollinators.


Thanks go to Haffie Strauss for friendly permission to work on her property, Connie Krug for providing traps, Anton Pauw for support (funding from the National Research Foundation of South Africa), him and two anonymous reviewers for their helpful comments on the manuscript, the Claude Leon foundation for awarding a research fellowship as well as the Ethics Committee of the University of Stellenbosch and CapeNature (Western Cape Nature Conservation Board, South Africa) for the necessary permits.

Supplementary material

114_2010_723_MOESM1_ESM.pdf (6 mb)
Slides 1–12(PDF 6,097 kb)
114_2010_723_MOESM2_ESM.pdf (7.1 mb)
Slides 13–26(PDF 7,238 kb)
114_2010_723_MOESM3_ESM.pdf (8 mb)
Slides 27–42(PDF 8,177 kb)

Copyright information

© Springer-Verlag 2010