Arthropod fauna of mammal-pollinated Protea humiflora: ants as an attractant for insectivore pollinators?
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Arthropod fauna of mammal-pollinated Protea humiflora: ants as an attractant for
insectivore pollinators?
P.A. Fleming* & S.W. Nicolson
Department of Zoology and Entomology, University of Pretoria, Pretoria, 0002 South Africa
Protea humiflora Andrews inflorescences are cryptic, but strongly scented and borne close to
the ground (geoflorous) for ready access by small, non-flying mammals. During a study of
P. humiflora pollination, we found that insectivorous elephant shrews (Macroscelididae:
Elephantulus edwardii (A. Smith)) carried higher pollen loads on their snouts than simulta-
neously-trapped rodent species. Elephant shrews seem to be acquiring pollen while forag-
ing for insects in the inflorescences. Compared with the larger bird-pollinated inflorescences
of P. repens (L.) L., P. humiflora inflorescences have a substantially lower mass of arthropods,
relatively fewer beetles (12 % of arthropod dry mass) and more ants (13 %). The large
numbers of ants in these inflorescences may attract insectivore pollinators, suggesting an
indirect, mutualistic relationship between plant, insect and insectivore.
Key words: non-flying mammal pollination, fynbos, satellite fauna, ants, nectar, mutualism.
INTRODUCTION
Protea species of the section Hypocephalae are Acomys subspinosus (Waterhouse) and Aethomys
known as the ‘rodent sugarbushes’, and are namaquensis (A. Smith)). It is more likely that pollen
characterized by cryptic, strong-smelling inflores- is acquired during foraging for insects in the inflo-
cences, borne close to the ground under dense rescences. The aim of the present study was to
foliage (Rebelo 1995). This allows ready access by quantify and characterize the arthropod fauna in
small, non-flying mammals (‘therophily’), which P. humiflora inflorescences to gain insight into why
account for half of the pollination events that lead elephant shrews might be visiting these inflores-
to seed set (Wiens et al. 1983; Fleming & Nicolson cences, and compare this with published data for
2002). Protea species are self-incompatible (Horn other, typically bird-pollinated, Protea species. In
1962), and it seems likely that the remaining polli- addition, we estimated the nectar standing crop
nation is carried out by insects (Coetzee & and compared the energy available from nectar
Giliomee 1985; Wright et al. 1991). and arthropods.
During a recent study of the pollination system
of Protea humiflora Andrews (Fleming and METHODS
Nicolson 2002), we were surprised to find that
Cape rock elephant shrews (Macroscelididae: We examined P. humiflora inflorescences collected
Elephantulus edwardii (A. Smith)) carried higher from two sites near Villiersdorp in the Rivier-
pollen loads on their snouts than simultaneously- sonderend Mountains, in the southern Cape,
trapped rodent species. The possibility that these South Africa, between July and October 2000. We
medium-sized insectivores (mean ± 1 S.D.: 49 ± analysed inflorescences at all stages of opening;
5 g, n = 11 adults) could be significant pollinators from early, recently opened, through to near-
of Protea species has not been seriously considered spent inflorescences, provided that pollen was still
previously (Wiens et al. 1983; Rebelo & Breyten- present.
bach 1987). The preponderance of pollen on the Arthropods were collected from 48 inflorescences,
snouts of E. edwardii compared with its scarcity in which were cut and placed immediately in plastic
their faeces (pollen grain exines pass through the bags and analysed on return to the laboratory.
digestive system intact; van Tets 1997), suggests Nectar standing crop was quantified for an addi-
that these shrews are unlikely to feed directly on tional 36 inflorescences: arthropods were re-
Protea pollen, as do rodent visitors (Muridae: moved with forceps on site and stored in alcohol.
*To whom correspondence should be addressed.
For nectar analysis, the styles of 12 about-to-
E-mail: tfleming@zoology.up.ac.za open florets were marked with permanent marker
African Entomology 11(1): 9–14 (2003)10 African Entomology Vol. 11, No. 1, 2003
pen and nectar was collected from their bases 37.3 mg sugar or 656 J per inflorescence over its
using 10 µl micropipettes and slight suction. complete development. However, not all florets
(Nectar was found in florets that were about to produce nectar simultaneously and only about a
open, while older peripheral florets and younger tenth of the nectar (there being around 12 whorls
central florets were largely dry.) Nectar volume of flowers, opening over 10–14 days) is estimated
was recorded (proportion of micropipette length) to be available at any one time (i.e. about 66 J).
and sugar concentration (%weight/weight) was Arthropod yields from P. humiflora inflorescen-
measured with 0–50 % and 45–80 % refractom- ces were also highly variable, ranging from 0.1 to
eters (Bellingham & Stanley, U.K.) for each floret 45.9 mg dry mass per inflorescence. The mean dry
separately. After conversion to %weight/volume mass of arthropods (7.6 ± 10.8 mg per inflores-
(Bolten et al. 1979), the quantity of sugar obtained cence) represents an energy content of 175 J
from each floret was calculated. The average value (energy values from Mostert et al. 1980).
was then extrapolated to estimate total nectar A total of 756 macro-arthropods (646 mg) repre-
sugar for the 265 ± 36 florets per inflorescence senting seven orders was recovered. The dry mass
(floret number counted in an additional 42 inflo- of macro-arthropods was significantly higher
rescences). Dry florets were recorded as a zero in during mid-winter (Table 1, K-S: P < 0.025). There
the data. were no differences in arthropod mass between
A total of 1190 thrips (Thysanoptera) was recor- the two sites sampled.
ded from 60 inflorescences that were exhaustively Over half the total dry mass of arthropods recov-
searched; however their minute size would rou- ered (58 %) was lepidopteran larvae which were
tinely yield insignificant amounts of biomass only found in inflorescences sampled during
and thrips were therefore not considered in the winter, at the beginning of the flowering season
study. In order to assess the dry mass of arthro- (Table 1). The larvae bore into inflorescence recep-
pods from each inflorescence, collections of tacles and sever the water supply to the florets,
macro-arthropods (excluding thrips; no mites causing their abortion and death (A.G. Rebelo,
(Acarina) were recovered) were categorized into pers. comm.). The dry mass of Lepidoptera larvae
‘morphospecies’. These collections were dried to from inflorescences that appeared to be aborted
constant mass at 60 °C and weighed, and the mean (i.e. were dry, brownish, or had failed to open, n =
individual mass for each morphospecies was used 19) was around three times that of inflorescences
to calculate dry masses for each inflorescence. Dry that appeared fresh (n = 65, K-S: P < 0.05).
masses for arthropod orders have been considered Lepidoptera larvae were not recovered from inflo-
in preference to numbers of individuals for statisti- rescences sampled from August onwards. A
cal comparison, given the substantial size differ- greater mass of Coleoptera was also recovered
ences between orders. Analysis of differences in during winter (KS: P < 0.001).
total dry mass and that of each arthropod order Eight species of ants were identified from
between site and season were carried out by P. humiflora flowers: Anoplolepis custodiens
Kolmogorov-Smirnov test. Data are presented as (F. Smith), (31 % of individuals, 20 % of dry mass),
the means ± 1 S.D. throughout. Crematogaster sp. 1 (31 %, 45 %), Camponotus
niveosetosus Mayr (18 %, 24 %), Crematogaster sp. 3
RESULTS (6 %, 9 %), Lepisiota capensis (Mayr)(9 %, 2 %), and
small numbers of individuals of Crematogaster sp. 2
Nectar yields were extremely variable between (2 % of total ant individuals), Camponotus sp. 2 (nr
inflorescences. Nectar yields did not vary signifi- angusticeps Emery) (2 %), and Tetramorium erectum
cantly with the degree of inflorescence opening Emery (1 %). Ant specimens have been lodged
(Fig. 1A), but are generally low probably due to the with the South African Museum. No Argentine
low rainfall during the study season as well as ants (Linepithema (Iridomyrmex) humilis (Mayr))
difficulty in sampling the concentrated and vis- were present. Ants were present throughout
cous nectar. The average volume per floret was flowering, even during the middle of winter. An
0.39 ± 0.56 µl (n = 36 inflorescences, 12 florets from average of 6.5 ants (1.55 ± 3.07 mg) was recovered
each), while nectar concentration averaged 49.0 ± from each inflorescence sampled later in the flow-
22.1 %w/w (n = 24 inflorescences, 12 florets from ering season (61 % of the macro-arthropod dry
each). Our figures extrapolate to approximately mass), compared with only 1.6 individuals (0.78 ±Fleming & Nicolson: Arthropod fauna of mammal-pollinated Protea humiflora 11
Fig. 1. Nectar standing crop (A) and the dry mass of ants (B) in Protea humiflora inflorescences compared with the
stage of flower opening, and the dry mass of ants compared with nectar volume (C).
1.34 mg) in winter (8.5 %; K-S: P < 0.01). We found considerably higher than the mean of 37.8 %
ants present at all stages of opening, from buds to measured for freshly secreted P. humiflora nectar
near-spent inflorescences (Fig. 1B). Ant abundance by Wiens et al. (1983). Our energy estimates of
was associated with a decrease in nectar volume about 66 J of energy available per day from nectar
(Fig. 1C), presumably as a result of their feeding on may therefore be an underestimate of the energy
nectar (Fig. 2). available to a small flower visitor that would be
able to feed on quantities not measurable by
DISCUSSION ourselves. Estimates of energy available from
macro-arthropods (175 J) were therefore higher
In this study, we found that P. humiflora nectar than that for the nectar standing crop. Calf (2000)
was highly concentrated and extremely viscous. similarly found much greater arthropod energy
Our average concentration of 49 % (w/w) was available for sugarbirds in bird-pollinated Protea12 African Entomology Vol. 11, No. 1, 2003
Table 1. Arthropod dry mass composition for Protea humiflora inflorescences sampled in mid-winter (July, n = 62) and
late winter/spring (August to October, n = 22). Figures are the mean mass per inflorescence (mg ± 1 S.D.) and
percentages of total macro-arthropod dry mass. P-values are for results of analysis by Kolmogorov-Smirnov test
comparing seasons.
Mid-winter Late winter & spring P
mg ± 1 S.D % mg ± 1 S.D. %
Arachnida 0.32 ± 0.94 3.45 0.11 ± 0.52 4.35 >0.10
Blattodea 0.96 ± 3.84 10.49 0.72 ± 2.34 28.26 >0.10
Coleoptera1 1.17 ± 3.34 12.80 0.07 ± 0.17 2.80 0.10
Hemiptera 0.18 ± 0.43 1.97 0.05 ± 0.17 2.09 >0.10
Hymenoptera (Formicidae) 0.78 ± 1.34 8.51 1.55 ± 3.07 60.76 0.10
Total biomass 9.41 ± 11.99 2.59 ± 4.00Fleming & Nicolson: Arthropod fauna of mammal-pollinated Protea humiflora 13 plant material (Scholtz & Holm 1985) and were inflorescences as an attractant and energy source more abundant in inflorescences which were for bird pollinators was emphasized by Mostert heavily infested with Lepidoptera larvae. While et al. (1980), who analysed stomach contents of Chirodica were not recovered from P. humiflora, Cape sugarbirds. We suggest a similar role of ar- conversely Corylophidae have not previously thropod fauna in pollination of P. humiflora by the been recorded in inflorescences of other Protea elephant shrew E. edwardii. Elephant shrews species. trapped during flowering of P. humiflora (Fleming Indigenous ants were a significant component of & Nicolson, in press) produced faeces containing the arthropod fauna of P. humiflora (35 % of pieces of ant exoskeleton (three out of eight scats numbers and 13 % of mass), but in bird-pollinated samples from E. edwardii individuals trapped inflorescences they are scarce. For example, in during winter contained positively-identifiable P. repens, Coetzee & Giliomee (1985) recorded that ant exoskeletons; however, the majority of the the alien Argentine ant was the most abundant of exoskeleton material is not identifiable). Ants (and eight ant species, while Mostert et al. (1980) recov- termites) are recorded as the main diet of ered only Argentine ants. Similarly, indigenous Macroscelididae (Perrin 1997). When foraging on ants are extremely rare in P. nitida inflorescences ants in P. humiflora inflorescences, elephant shrews (Visser et al. 1996). would be additionally rewarded by the nectar al- Ants are efficient nectar collectors and are easily ready consumed by these ants (e.g. Fig. 2). able to handle sugar solutions as concentrated as Having an insectivore as pollen vector has the nectar of P. humiflora (Josens et al. 1998). How- advantages for the plant, since these animals are ever, ants are generally described as nectar thieves less likely to consume flower parts than rodent (Hickman 1974; Inouye 1980; Galen 1983). They visitors (e.g. Vlok 1995) and also have greater home are considered to be ineffective pollinators due to ranges (Withers 1979; Fleming & Nicolson, their small size (limiting contact with anthers and unpubl. data). Insectivores may also pollinate stigmas), smooth integument, frequent grooming, Australian Proteaceae: a number of dasyurid antibiotic metapleural gland secretions, and, marsupial species recorded as regular visitors to finally, their rather limited mobility (Schubart & Banksia inflorescences (Turner 1982; Carthew & Anderson 1978; Beattie 1985; Peakall et al. 1991). Goldingay 1997; Goldingay 2000). These observa- Many plants therefore have developed protective tions may change the perceived role of arthro- devices designed to keep ants out: hiding nectar in pods, particularly ants, in pollination of Protea- tubes and closed blossoms, closed throats or sticky ceae. If arthropods contribute in some part to hairs (Geurrant & Fiedler 1981; Peakall et al. 1991). attract insectivores, then Proteaceae would derive Other plants may distract ants by luring them an indirect benefit from their arthropod visitors. away from flowers to extra-floral nectaries (e.g. The relationship between ant and plant under Zachariades & Midgley 1999). these circumstances may therefore be more Therophilous Protea species, including P. humiflora, mutualistic (Maloof & Inouye 2000) than illicit, and seem to have little defense against nectar-thieving the blanket description of ants as ‘nectar robbers’ ants. Their inflorescences are geoflorous, a short may be doing them an extreme injustice. distance separating them from ants on the ground, and there is little protection of nectar ACKNOWLEDGEMENTS within the flowers. Protea species do not possess We thank H. Robertson and K. Parr for identify- extra-floral nectaries (Rebelo 1995). Nectar may in ing ants, C. Scholtz for identifying Corylophi- fact be more accessible to insects than it is to verte- dae, T. Rebelo for advice and comments, and the brate would-be pollinators, particularly mamma- National Research Foundation for financial sup- lian visitors that lack the finesse of an avian beak. port. P.A.F. is supported by a University of Pretoria The role of ‘satellite’ arthropods in Protea repens Postdoctoral Fellowship. REFERENCES BEATTIE, A.J. 1985. The Evolutionary Ecology of Ant-plant tration in flower nectar. Oecologia 41: 302–304. Mutualism. Cambridge University Press, Cambridge. CALF, K.M. 2000. The role of male territory size and BOLTEN, A.B., FEINSINGER, P., BAKER, H.G. & quality in mating and reproductive success of Cape BAKER, I. 1979. On the calculation of sugar concen- and Gurney ’s sugarbirds, Promerops cafer and
14 African Entomology Vol. 11, No. 1, 2003
Promerops gurneyi. M.Sc. thesis, University of PERRIN, M. 1997. Cape rock elephant shrew, Elephan-
Stellenbosch, Stellenbosch. tulus edwardii. In: Mills, G. & Hes, L. (Eds) The
CARTHEW, S.M. & GOLDINGAY, R.L. 1997. Non-flying Complete Book of Southern African Mammals. 66. Struik
mammals as pollinators. Trends in Ecology and Winchester, Cape Town.
Evolution 12: 104–108. REBELO, A.G. & BREYTENBACH, G.J. 1987. Mammal
COETZEE, J.H. & GILIOMEE, J.H. 1985. Insects in pollination in the Cape flora. In: Rebelo, A.G. (Ed.) A
association with the inflorescence of Protea repens (L.) Preliminary Synthesis of Pollination Biology in the
L. (Proteaceae) and their role in pollination. Journal Cape Flora. 109–125. South African National Scientific
of the Entomological Society of southern Africa 48: 303– Programmes Report 141, C.S.I.R., Pretoria.
314. REBELO, T. 1995. Sasol Proteas. A Field Guide to the Proteas
COLLINS, B.G. & REBELO, T. 1987. Pollination biology of of Southern Africa. Fernwood Press in association with
the Proteaceae in Australia and southern Africa. the National Botanical Institute, Vlaeberg.
Australian Journal of Ecology 12: 387–421. SCHOLTZ, C.H. & HOLM, E. 1985. Insects of Southern
FLEMING, P.A. & NICOLSON, S.W. 2002. How impor- Africa. Butterworths, Durban.
tant is the relationship between Protea humiflora SCHUBART, H.D. & ANDERSON, A.B. 1978. Why don’t
(Proteaceae) and its non-flying mammal pollinators? ants visit flowers? A reply to D.H. Janzen. Biotropica
Oecologia 132: 361–368. 10: 310–311.
GALEN, C. 1983. The effects of nectar thieving ants on TURNER, V. 1982. Marsupials as pollinators in Australia.
seedset in floral scent morphs of Polemonium viscosum. In: Armstrong, J.A., Powell, J.M. & Richards, A.J. (Eds)
Oikos 41: 245–249. Pollination and Evolution. 55–66. Royal Botanic Gar-
GEURRANT, E.O.J. & FIEDLER, P.L. 1981. Flower dens, Sydney.
defenses against nectar-pilferage by ants. Biotropica VAN TETS, I.G. 1997. Extraction of nutrients from Protea
(Reproductive Botany Suppl.) 13: 211–214. pollen by African rodents. Belgium Journal of Zoology
GOLDINGAY, R.L. 2000. Small dasyurid marsupials – are 127: 59–65.
they effective pollinators? Australian Journal of VISSER, D., WRIGHT, M.G. & GILIOMEE, J.H. 1996. The
Zoology 48: 597–606. effect of the Argentine ant, Linepithema humile (Mayr)
HICKMAN, J.C. 1974. Pollination by ants: a low-energy (Hymenoptera: Formicidae), on flower-visiting
system. Science 184: 1290–1292. insects of Protea nitida Mill. (Proteaceae). African
HORN, W. 1962. Breeding research on South African Entomology 4: 285–287.
plants: II. Fertility of Proteaceae. Journal of South VLOK, J. 1995. Fynbos and rodents. Veld & Flora 81:
African Botany 73: 259–268. 105–107.
INOUYE, D.W. 1980. The terminology of floral larceny. WIENS, D., ROURKE, J.P., CASPER, B.B., RICKART, E.A.,
Ecology 61: 1251–1253. LAPINE, T.R., PETERSON, C.J. & CHANNING, A.
JOSENS, R.B., FARINA, W.M. & ROCES, F. 1998. Nectar 1983. Nonflying mammal pollination of southern
feeding by the ant Camponotus mus: intake rate and African proteas: a non-coevolved system. Annals of
crop filling as a function of sucrose concentration. the Missouri Botanical Gardens 70: 1–31.
Journal of Insect Physiology 44: 579–585. WITHERS, P.C. 1979. Ecology of a small mammal
MALOOF, J.E. & INOUYE, D.W. 2000. Are nectar robbers community on a rocky outcrop in the Namib Desert.
cheaters or mutualists? Ecology 81: 2651–2661. Madoqua 11: 229–246.
MOSTERT, D.P., SIEGFRIED, W.R. & LOUW, G.N. 1980. WRIGHT, M.G., VISSER, D., COETZEE, J.H. &
Protea nectar and satellite fauna in relation to the food GILIOMEE, J.H. 1991. Insect and bird pollination of
requirements and pollinating role of the Cape sugar- Protea species in the western Cape: further data.
bird. South African Journal of Science 76: 409–412. South African Journal of Science 87: 214–215.
PEAKALL, R., HANDEL, S.N. & BEATTIE, A.J. 1991. The ZACHARIADES, C. & MIDGLEY, J.J. 1999. Extrafloral
evidence for, and importance of, ant pollination. In: nectaries of South African Proteaceae attract insects
Huxley, C.R. & Cutler, D.F. (Eds) Ant–plant Interac- but do not reduce herbivory. African Entomology 7:
tions. 421–429. Oxford University Press, Oxford. 67–76.
Accepted 14 June 2002You can also read
