INTERACTIONS BETWEEN FIGS
(FICUS SPP., MORACEAE)
AND FIG WASPS
(CHALCIDOIDEA, AGAONIDAE)
THESIS
Submitted in fulfillment of the requirements for the
Degree of Doctor of Philosophy at Rhodes University
by
ANTHONY BRIAN WARE
December 1992
CONTENTS
PREFACE .......•.•....................................................... iv
LIST OF FIGURES ...••..•............•...•.•..•.•.......................••. v
LIST OF TABLES ...................•.....••................................ xii
ABSTRACT ................................................................ xvi
CHAPTER 1: Introduction ................................................... 1
CHAPTER 2:
Fig WJSp host specificity ....................•.................... . 17
Paper 1: African figs and fig wasps: the wasps eye view of Ficus species. Mitteilungen
aus dem Institut fur Allgemeine Botanik Hamburg 24, in press. (Compton,
S.G.,Ware, A.B. and van Noort, S.) ...............•........... 17
Paper 2: Host specificity in some African fig wasps. Submitted to Annals of the
Missouri Botanical Garden (Ware, A.B. ,Compton, S.G. and Phillipson, P .B) .
...................................................... 29
CHAPTER3:
Biological evidence for the presence of volatile attractants ................. 48
Paper 3: Pollinator specific volatile attractants released from the figs of Ficus burtt-
davyi. South African Journal of Science 85,323- 324 (1989). (van Noort, S.,
Ware, A.B. and Compton, S.G.) .............................. 49
Paper 4: Fig wasp responses to host plant volatiles. Submitted to Journal of Insect
Behavior. (Ware, A.B. and Compton, S.G.) .........•............ 51
CHAPTER 4:
Chemical evidence for the presence of Ficus attractants ........•.......••• 67
Paper 5: Fig volatiles: their role in attracting pollinators and maintaining pollinator
specificity. Plant Systematics and Evolution (In press). (Ware, A.B., Kaye,
P.T.,Compton, S.G. and van Noort, S) ......................... 68
CHAPTERS:
Studies of fig wasp behaviour ...................••....•............ 83
Paper 6: Dispersal of adult female fig wasps I. Arrivals and departures. Submitted
to Emomologia Experimelltalis et applicara. (Ware, A.B. and Compton,
S.G.) ................................................... 84
ii
Paper 7: Dispersal of adult female fig wasps II. Movements between trees. Submitted
to Elltomoiogia E'{perimelltalis et applicara. (Ware, A.B. and Compton,
S.G.) .................................................. 100
CHAPTER 6:
Perception of volatiles ............................................ 118
Paper 8: Preparation of small, delicate insects for scanning electron microscopy.
Proceedings of the Electron Microscopy Society of southern Africa 19,39-40
(1989). (Ware, A.B. and Cross, R.H.M.) ...................••.. 119
Paper 9: Repeated evolution of elongate multiporous plate sensilla in female fig
wasps (Hymenoptera: Agaonidae: Agaoninae). Proceedings of the Koninlijke
Nederlandse Akademie vall Wetenschappen 95,275-292 (1992). (Ware, A.B.
and Compton, S.G.) ............•..•............•.......... 121
CHAPTER 7:
Breakdown of host specificity .................•.................... 139
Paper 10: Studies of Ceratosolen gaZili, a non-pollinating agaonid fig wasp. Biotropica
23, 188-194 (1991). (Compton, S.G., Holton, K.C., Rashbrook, V.K., van
Noort, S., Vincent, S. and Ware, A.B.) .........•.•............. 140
Paper 11: Breakdown of pollinator specificity in an African fig tree. Biotropica 24, in
press. (Ware, A.B. and Compton, S.G.) .............•.......... 147
CHAPTER 8:
Fig wasp parasitoids ..................•..•..•.................... 153
Paper 12: African fig wasp parasitoid communities. In Parasitoid Community Ecology
(Eds Hawkins, B.A. and Sheenan, W.) in press. (Compton, S.G. Rasplus, J.Y. and Ware, A.B.) ........................................ 154
CHAPTER 9:
Synopsis ...................................................... 179
iii
PREFACE
Natura nusquam magis quam in minimis tota est
(Nature is nowhere more perfect than in the minutest of her works)
Pliny: Roman naturalist and philosopher 1 A.D.
Most research cannot be done in isolation and these studies are no exception.
Although the
contribution of colleagues is acknowledged in each section, I would like to make special mention of
.
of the following people: my senior supervisor, Dr Steve Compton, for providing me the opportunity
to investigate fig/fig wasp biology and for his considerable input in the investigations; Prof. Perry
Kaye provided much needed assistance in the chemical aspects of the study; Profs M. Brown and V.
Moran for having enough faith in my ability to give me a second chance at Rhodes University; the
'fig team', in particular Sally Ross, Simon van Noort and Costas Zachariades, provided many hours
of field assistance and company. Finally I would like to thank my wife, Kathy Holton, for her support
and encouragement during my studies.
iv
FIGURES
CHAPTER 1
Fig. 1.
An electon micrograph of the interior of a fig showing the ostiole (0) with
accompanying bracts (B) and the ovules (F) lining the inside of the flowers ...... 3
Fig. 2. Electron micrograph of the mandibles used by the pollinating fig wasp to force its way
through the ostiole in order to gain access to the flowers within the fig ....••••• 2
Fig.3. Fig-fig wasp development cycles. Parasitoids will arrive later than the seed predators
and oviposit from outside the fig after probing with their long ovipositors. See text
for general description (Modified from Galil and Eisikowitch, 1968) ...•..•.•. 7
CHAPTER 2
Paper 2:Fig. 1. Our distribution records of southern African F. sycomorus and associated pollinator,
C. arabicus Co); other southern African distribution records of F. sycomorus (0) are
from van Greuning (1990) and von Breitenbach (1986) and are without pollinator
records .....................•................................. 35
Fig. 2. Our distribution records of Malagasy and Comorian F. sycomorus (lot) together with
their associated pollinators; other Malagasy and Comorian F. sycomorus (0) records
are from Perrier de la Bathie (1928, 1952) and are without pollinator records •.• 56
Fig. 3.
Our distribution records of F. sakalavarum (e) together with their associated
.
pollinators; other F. sakalavarum records (0) are from Perrier de la Bathie C1928,
1952) and are without pollinator records ...••..•••••....•............•. 36
Fig. 4. Our distribution records of southern African F. cordata subsp. cordata together with
their associated pollinators (e) and F. cordata subsp. salicifolia ~);
other F. cordata
subsp. cordata (0) and F. cordata subsp. salicifolia records (+) are from van Greuning
(1990) and von Breitenbach
1986) and are without pollinator records •...•..•• 39
Fig. 5. Our distribution records of southern African Ficus "thonningii I natalensis" together
with their associated pollinators (.); other Ficus "thonningii I natalensis" records
(0)
are from van Greuning (1990) and von Breitenbach (1986) and are without
pollinator records.
F. lhollllingii (i) is only associated with one pollinator species
while F. lUuaiellsis (ii and iii) is associated with two different agaonines . .....• 41
v
CHAPTER 3.
Paper 3.Fig. 1.
Elisabethiella baijnathi trapped next to control (empty) cotton bags and bags
containing receptive F. bunt-davyi or F. thonningii.
More wasps were attracted to
the figs of F. bunt-davyithan those of F. thollllingii or controls (t[34] = 3.96,P<0.001
and
1[34]
= 3.65,P<0.01,respectfully). Equal numbers of wasps were trapped near
control bags and those containing F. thollllingii figs (t[34] = 0.41 ,P> 0.05) ....... 49
Fig. 2.
Elisabethiella baijnathi trapped next to control (empty) cotton bags and bags
containing receptive figs of F. bunt-davyi or F. sur. More wasps were trapped near
F. burtt-davyi than F. sur (t[34
=
6.89,P<0.001)or controls (t[34]
=
6.97,P<0.OOl).
There was no difference in the numbers of wasps trapped at F. sur figs and controls
(t[34] = 0.054,P>0.05) .. ........•..........•.....•.......•........ 50
Fig. 3.
Elisaberhiella baijnathi trapped next to control (empty) cotton bags containing
receptive figs of F. bunt-davyiwith beeswax sealing their ostioles or applied to their
bases. Figs with their ostioles covered did not attract wasps when compared with
controls
attractive
U[lS]
(t[18]
= 1.17, P>0.1). In contrast, figs with basal wax remained highly
= 3.37,P<O.0l) .. ...••.•......•.•.................... 50
Paper 4:Fig.1. Portion of the 1820 Settlers Botanical Garden (Grahamstown, South Africa) showing
the relative positions of F. thonningii (+), F. burtt-davyi (0) and F. sJA.{rees.
Additional exotic fig trees are represented by the open symbol (0). The numbers
indicate those trees used to monitor the arrivals and departures of the fig wasps. 55
Fig.2. The effect of bagging pre-receptive (= pre-female) figs (hatched bar) of F. burtt-davyi
on the numbers of pollinating wasps, E. baijnathi, trapped. The solid bars indicate
the number of wasps trapped on a similar tree which remained unbagged ...... 56
Fig. 3.
The numbers of wasps, together with their relative percentages, simultaneously
trapped near bags containing receptive figs of F. thollningii or F. burtt-davyi. Empty
bags acted as controls •.........•••.•....••..•.••.....••••..•...•. 58
Fig. 4. The fruiting phenologies of 10 fig trees used in the long term monitoring of fig wasp
arrivals and departures. Intercrop periods are shown with a thin solid line while the
period when trees are bearing fruit are denoted by solid blocks.............. 59
vi
Fig. 5. Identity and numbers of fig wasps trapped at a F. burrt-davyi tree. The shaded areas
represent those periods when figs were present on the tree ................. 60
Fig. 6. Identity and numbers of fig wasps trapped in a F. thOJlllingii tree. The shaded areas
indicate those periods figs were present on the tree •.......•.•.....•.... 62
CHAPTER 4
Paper 5:Fig. 1. Chromatograms of volatiles from female phase figs from six individual trees of six
Ficus species. A full amplitude response at the detector represents at least 5 ng of
material while the retention time indicates how long the volatiles remain on the
column before reaching the detector.
The smaller more volatile compounds
generally elute first while the oven temperature is still low. See text for instrumental
parameters. ",...............................................,73
Fig. 2. The volatile profiles of female phase figs from three cultivars of F. carica .. , ..•• 74
Fig. 3. The volatile profile of prefemale and female phase figs of our individual trees of F.
burrt-davyi. The closed symbol highlights the additional volatile recorded from figs
in the female phase. The open symbol indicates that volatile which was released
from female phase figs of two individual trees .••••••.••..•.••.••.•.....• 75
Fig. 4. The gas chromatograms of prefemale, female and inter-floral phase figs from F. burrtdavyi. The closed symbol indicates the additional volatile peak occurring in the
.
volatile profile of female phase figs ........................ , ..•. , ...... 76
Fig.5. Gas chromatograms of the volatiles from prefemale, female and inter-floral stage figs
of F. illgeJJs. The symbols indicate additional volatile components provided by female
phase figs ...............•..••..........•......•.•.. , ... , ...... 77
Fig. 6. The volatile profiles of the prefemale and female stage figs of F. lutea. The symbol
indicates the additional volatile produced by female phase figs .......... , ... 78
CHAPTER 5
Paper 6:Fig. 1. The number of tigs with fig wasp exit holes during the 11 day dispersal phases of two
synchronously fruiting F. burtt-davyi trees (winter 1990) ......•......•...... 88
Fig. 2. Exit holes produced at different times of the day in two synchronously fruiting F.
bllrtt-davyi trees (winter 1990) .....•....•................ , .......... 89
vii
Fig. 3.
The morning and evening ambient temperatures in the 1820 Settlers Botanical
Garden during the 11 day dispersal phase from two F. bum-davyi trees •..•...• 89
Fig. 4. The average hourly windspeeds (+ /- standard error) experienced in Grahamstown
during March 1989 ••••••...••.......•..•••••.••••..••••....•••••• 90
Fig. 5.
The flight take-off frequencies of E. baijnathi females at varying laboratory
temperatures. • ...•••.••.•..•.•.••...•••....••.•....••...•....•. 91
Fig. 6. The numbers of E. baijllathi foundresses that entered figs of F. bum-davyi. The
circles represent a Poisson (random) distribution •••.•..•..••••........•. 94
Paper 7:Fig. 1. The wind profile at various heights above ground at a site in the 1820 Settlers
Botanical Gardens, Grahamstown,
South Africa averaged from 10 occasions
throughout the year (1 km/hr = 27.8 em/sec).
The large standard deviations are
indicative of the large variation in wind speeds experienced in the area •.....• 105
Fig.2. Densities of total fig wasps at different heights in the 1820 Settlers Botanical Gardens .
. ................................••..... .................. . 105
Fig. 3. Densities of three fig wasp species at different heights in the 1820 Settlers Botanical
Garden in Grahamstown. • •..•••••••••••.••••••••.•..•••••.•••••• 106
Fig. 4. The relative percentages of emigrating E. baijnathi trapped around a wasp producing
F. burtt-davyitree relative to the prevailing wind direction. The small arrow indicates
the
me~n
preferred angle of wasp distribution and is flanked by an arc indicating the
mean angular variation (21°) •..••••••........••••..•..•••••....••• 107
Fig. 5. Numbers of E. stuckenbergi recorded at sticky traps baited with receptive F. thollllingii
figs. The mean preferred direction is indicated with a small arrow which is flanked
by an arc representing the mean angular variation •..................•.•. 108
Fig.6. The percentages of E. baijllathi trapped around a receptive F. burtt-davyi tree relative
to the prevailing wind direction ......••.........•...........•.....•• 108
Fig. 7. The percentages of E. baijnathi trapped on all sticky traps positioned at 0.5,1 and 2
m above ground level surrounding producer and receptive F. bum-davyitrees relative
to the prevailing wind directions •.•...•.••...•...••.....••...•....•. 110
viii
CHAPTER 6
Paper 8: Fig 1. Fig wasp - cryo (treatment 1) showing loss of appendages (arrowed). x 60 ...... 120
Fig. 2. Fig wasp - critical-point dried (treatment 3) showing some collapse of eye (E),
antennae (A) and abdomen CAB). x 66 .............•................. 120
Fig. 3. Fig wasp - abbreviated acetone treatment of fresh material (treatment 4) showing
extensive collapse of eye and antennae. x 125 ........................•.. 120
Paper 9.Figs 1-4. Scanning electron micrographs of agaonid antennal segments illustrating the four
types of MPS arrangements. 1. Elisabethiella stuckenbergi (Type I), 2. Allotriozoon
heterandromorphum (Type II), 3. Courtella armata (Type III), and 4. Elisabethiella
baijnathi (Type IV) .•.•••..••••.•...•...••....•.•....•••.....••.. 124
Figs 4-11. Agaonine antennal MPS arrangements 5. Fifth antennal segment of Elisabethiella
pectinata (redrawn from Joseph, 1959).6. Eighth segment of Platyscapa quadriceps
(redrawn from Grandi, 1923).7. Sixth segment of Blastophaga silvestriana (redrawn
from Hill, 1967).8. Tenth antennal segment of Blastophaga clavigera(redrawn from
Grandi, 1928). 9. Elongation of antennal segments as seen in the second funicle
segment of Ceratosolen tentacularis (redrawn from Grandi, 1928) 10. thickening of
antennal segments as in Deilagaon chrysolepidis (redrawn from Boucek, 1988). 11.
branching of the seventh antenna! segment of Dolichoris j1abellata (redrawn from
Wiebes, 1978)................................................. 126
Fig. 12. The phylogeny of Agaonine genera (modified from Wiebes, 1982) and evolution of
elongate MPS •....•.••.••..•.•..••............................. 128
CHAPTER 7
Paper 10: Fig. 1. Records of Ceratosolell species in collections of F. sycomorus figs. The dotted line
indicates the approximate southern limit of the distribution of the fig species.
Squares are F. s. sycomorus, circles F. s. gnaphalocarpa. Open squares / circles
indicate the presence of C. arabicus, closed squares C. galili. Mixed squares indicate
that both wasp species were present................................. 141
ix
CHAPTER 8
Paper 12: Fig. 1. A comparison of the sizes of F. sur figs probed by gall-making fig wasps ...•.•. 159
Fig.2. A comparison of the sizes of F. sur figs probed by parasitoid fig wasps •...•... 159
Fig. 3.
Variation in the ovipositor lengths of parasitoids associated with F. sur. Mean
ovipositor lengths (+ S.D.) were: Sycoscapter sp. 1,9.2 + 0.4, n = 13; Sycoscapter
sp. 2, 10.3,n
= 12; Apocrypta guilleensis7.1
l.I,n
= 20 •••.•.•••....•..•. 160
Fig. 4. Variation in the mean style lengths of flowers occupied by Elisabethiella baijllathi
progeny in relation to the number of foundress females entering the figs. The
distribution of female progeny changes with increasing density, with more wasps
closer to the periphery of the figs, where they are potentially easier to reach by the
parasitoids ovipositing from the outside of the figs •..........•........... 161
Fig. 5. Frequency histogram indicating the lengths of Sycoscapter (top) and Philotrypesis
(middle) in relation to the distance they must travel from the outside of the figs to
reach the ovules of F. bunt-davyi. The distance from the ovules was measured using
figs at the 'inter-floral' phase. Philotrypesisoviposits slightly earlier than Sycoscapter
during this period .....................•..•....................• 162
Fig. 6. Frequency histograms showing the distribution of flowers of varying style lengths
within the figs of F. bunt-davyi (top), the number of flowers containing fig wasps
(middle) and the relative number of parasitoids (bottom).
The fig wasps present
were Elisabethiella baijnathi, Otitesella sesqialleilata and Otitesella uluzi (gallers),
together with Sycoryctes sp. and Philotrypesis sp. (parasitoids) ............... 163
Fig. 7. Changes in the distance that parasitoids must probe with increasing diameter of F.
sur figs. Also indicated are the potential distances that the parasitoids can probe, in
relation to the size of the figs at the times when they oviposit. Each block defines
one standard deviation from the mean ovipositor length and the mean diameter of
tigs that were probed by each species ............................... 163
Fig. 8. The fruiting patterns of 52 F. bunt-davyi trees growing in Grahamstown. The small
crops of very short duration (for example on trees 26 and 49) aborted at an early
stage of development. . .......................•............•... 164
x
Fig. 9. The numbers of F. bunt-davyi trees bearing figs in Grahamstown over a two year
period. Some figs are present in the area throughout the year, but the abundance of
fruiting trees tends to decline during mid-summer and mid-winter• . . . . . . • . . 165
Fig. 10. The fruiting patterns of 18 F. sur trees growing around Grahamstown. Vertical lines
within the bars indicate periods when wasps were emerging while unpollinated figs
were present on the same tree • . • • • . • • • . . . . • . • . . . . . • . . . . . . . . . . • 165
Fig. 11. The numbers of F. sur trees bearing figs around Grahamstown over a two year period.
Figs are present in the area throughout the year, with no clear seasonal trends in
abundance. Count numbers 8 and 9 are slight under-estimates, resulting from certain
trees being inaccessible due to flooding . • • . • • . . . . . • . • . . . . . . . . . • . . • 166
Fig. 12. Sample recruitment curves for collection of wasps from figs of F. sur in four subregions of southern Africa.
'North' includes Zimbabwe, Zambia, and Malawi.
Transvaal, Natal and Cape are provinces of South Africa. The flattening of the curves
suggests that all the species associated with F. sur in the Cape have been collected..
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • 170
CHAPTER 9
Fig. 1. Gas chromatogram of 4-hexene-l-ol acetate and its coelution with the volatiles from
receptive F. bunt-davyi figs • . . . . . . . . . . . • . . . . . • . . . • . . • . . • . . . . . . 182
xi
TABLES
CHAPTER 2
Paper 1: Table 1. African Ficus species with different agaonines associated with their subspecies or
synonyms .......•............•..................•............. 21
Table 2. African Ficus species with more than one associated agaonine (excluding cases
listed in Table 1) •••••••••••••••••••••••••••••••••••••••••••••••• 22
Table 3. African Ficus species which share agaonines with congeners ............... 23
Table 4. Host relationships of selected sycoecine fig wasps .•..................... 24
Table 5. Possible causes of exceptions to the one agaonine: one Ficus species relationship.26
Paper 2:Table 1. A list of southern African Ficus species (numbers from Berg (1989» together with
their associated agaonines (numbers from Wiebes and Compton (1990» together
with the number of trees sampled .................................... 32
Table 2. A list of Ficus species from Madagascar and The Comores (numbers from Berg
(1989» and their associated agaonines (numbers from Wiebes and Compton (1990»
together with the number of trees sampled for their pollinators...........•.. 33
Table 3. Distinguishing characters between F. sycomorus and F. sakalavarum • .•..•..•. 38
CHAPTER 3
Paper 3:Table 1. Mean (+ /- s.e.) numbers of Elisabethiella baijnathi collected on sticky traps placed
next to cotton bags containing 'receptive' figs or control (empty) bags ......... 49
Paper 4.Table 1. Indigenous Ficus spp. present in the Grahamstown Botanical Garden, together with
the wasps normally found associated with the trees in Grahamstown. . ........ S3
Table 2. The fig wasps trapped near cotton bags containing either pollinated or unpollinated
(receptive) F. thonningii figs. The control bags were empty. Combined results from
two trials ...................................................... S7
Table 3. The fig wasps trapped near cotton bags containing either unpollinated (receptive)
figs of F. thOllllillgii or F. bunt-davyi. Control bags were empty .............. 57
Table 4. The wasps caught in sticky traps on F. thonningii and F. bunt-dmyi trees and their
normal host Ficus • ...••...•..•...•..•..•...•...•.......•.•...... S9
xii
Table 5. The fig wasps (all species) trapped at F. thollllingii and F. burtt-davyl trees during
their intercrop, receptive (first half of crop period) and producer (latter half of crop
period) stages ••..•...••••.........••.•...•.••.•.••...•.•.•..... 61
Table 6. Comparisons of the numbers of wasps trapped during intercrop periods with the
numbers trapped during the first half of each crop period (which includes the
receptive female phase of fig development) ..........•...•..........•..• 63
CHAPTER 5
Paper 6.Table 1. The timing of fig wasp emergence as indicated by the number of F. burtt-davyi figs
with exit holes ...•.........•••••......•..••..•.•..•••••....•••.. 87
Table 2. The number of fig wasps trapped on sticky traps positioned in an area where fig
trees were growing. The traps were replaced every morning at 06hOO and again in
the evenings at 18hOO. Monitoring was over three, one week periods in both winter
and summer...•............••.•••...•••....•••..•.•..••....... 92
Table 3. The number of pollinating fig wasps (E. baijllathi) on bagged receptive figs of a
single F. burtt-davyi tree. The wasps were removed after being counted ......... 92
Table 4. The arrival of the pollinating fig wasp E. baijnathi at branches of two receptive F.
burtt-davyi trees •..............•........•.•............•......•.. 93
Table 5. Pollinator fig wasp (E. baijnathi) searching behaviour for suitable figs of F. burttdavyi i!1 which to oviposit.......................................... 94
Paper 7:Table 1. Fig wasp flight speeds measured at 25°C over a distance of 1 m•....•....••• 106
Table 2. Comparisons between the numbers of wasps caught at different heights upwind and
downwind on F. burtt-davyi trees producing fig wasps and those receptive trees
attracting fig wasps. Producer trees have wasps emerging from the figs. Receptive
trees have figs that are ready to be pollinated ........•...•.••....••.... 109
Table 3. Pollinating fig wasps (E. stuckenbergi) trapped at sticky traps near cotton bags
containing receptive figs of F. thOllllillgii. The direction from where the wind was
blowing was used as the reference point (0°) for the circular statistics •......•. 111
xii i
Table 4. Comparison of numbers of two species of wasp trapped upwind and downwind of two
receptive trees of F. bunt-daryL The respective numbers of wasps trapped at 2 m,l m
and 0.5 m are given in parenthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
CHAPTER 6
Paper 9: Table 1. The distribution of the major antennal sensilla arrangements within the genera of
Agaoninae. See text for description of the types of MPS arrangements . . . . . . . 125
Table 2. A comparison of the exposed surface areas of the MPS on the antennae of fig wasps
with sensilla tinearia and sensilla chaetica. Numbers of sensilla and their total surface
area refer to pairs of antennae . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Appendix 1. The structure and position of the multi porous plate sensil1a (MPS) of female
agaonines. Data were derived from the literature and/or from examining dry mounted
specimens (*). A
+ indicates those species which possess elongated sensi1la. See text
for a description of the types of MPS arrangements . . . . . . . . . . . . . . . . . . . . 130
CHi\PTER 7
Paper 10: Table 1. Descriptions of F. sycomorus collections in Natal. . . . . . . . . . . . . . • . . . . . 142
Table 2.
The contents of figs colonized by C. arabicus and\or C. gaZili.
Flowers which
produced wasps are not included . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Table 3. Comparisons of the dry weights of C. arabicus and C. galili at three locations in
southern Africa. . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Table 4. The combinations of wasps entering the figs of F. sycomorus • .•...•.••••• 143
Table 5. The frequencies of C. arabicus, C. galili and S. sycomori females reared from figs
of F. sycomorus . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . • . . . . . . • . • . 144
Table 6. The numbers of wasps successfully entering the figs of F. sycomorus. Sample sizes
were 50 figs per tree. The ranges are given in parenthesis. . . . . . . . . . . . . . . 144
Table 7. The numbers of wasps that failed to gain entry into figs and were trapped in the
ostiolar bracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Table 8. The comparison of fig wasp assemblages reared from figs of F. sycomorus. Counts
are of females only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
xiv
Paper 11: Table 1. Fig wasps collected over a 35 week period on sticky traps placed in Liqustrum
lucidum and Ficus lutea trees. The receptive period when the figs were potentially
attractive to pollinators was approximately 7 weeks •...•••.•••.•••..••• 149
Table 2. Fig wasps found in the figs of a F. lutea tree growing out of its natural range in
Grahamstown, South Africa •......•••.•.•.......•....••.•••.• 150
CHAPTER 8
Paper 12: Table 1. The longevities of adult female fig wasps maintained at 20"C and 75-80% relative
humidity on diets of either distilled water or a 10 % sucrose solution. Mean longevities
for Philotrypesis sp. and Apocrypta guineensis with sugar are underestimates as the
trials were terminated after 40 and 60 days respectfully .•.•..•..••.•.•.. 167
Table 2. Egg loads of galler and parasitoid fig wasps associated with F. bunt-davyi and F. sur.
"Internal" ovipositing species lay their eggs after entering the figs, while "external"
species lay their eggs from outside the figs. Females of syn-ovigenic species contained
both mature and developing eggs, whereas pro-ovigenic females contained only mature
eggs• • . . • . • • . . . . . . . . . . . • . • . . . . • • . . . • • . . . . . . . . • . . . . . • 168
Table 3. Species richness of local and sub-regional fig wasp communities associated with F. sur
in southern Africa. 'North' includes Zimbabwe, Zambia and Malawi. Local parasitoid
richness is significantly lower in the Cape than in Natal (Mann-Whitney, Z = 2.88, P
< 0.01) and the Transvaal (Mann-Whitney, Z
=
1.98, P < 0.05), but not elsewhere.
Local galler richness in the Cape was not significantly different from the three other
subregions. . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . • . . • . • . 169
Table 4. A comparison of the species richness of local fig wasp communities associated with F.
bunt-davyi in forest and inland areas of the eastern Cape (South Africa) •.•.... 171
Table 5. The distribution of parasitoid genera within the southern African subgenera and
subsections of Ficus . . . . . . . . . . . . . • . • . . . . . . . • . . . • • . . . . • • . . • . 171
xv
ABSTRACT
Fig trees (Ficus spp., Moraceae) and fig wasps (Chalcidoidea, Agaonidae) are uniquely associated. In
one fig wasp group, the pollinators (Agaoninae), each species is generally host species-specific. The
relationship is one of obligate mutualism where the wasps provide pollination services and in return
utilises some of the ovules for larval development. Non-pollinating fig wasps (generally belonging to
subfamilies other than the Agaoninae) may be gallers or parasitoids, and can also be host species-specific.
In the accompanying studies we examined the factors governing the interactions between fig wasps and
their host trees.
Surveys of fig trees and their associated pollinating fig wasps conducted in southern Africa, Madagascar
and The Comores generally confirmed their specific relationships. An examination of F. sycomorlls in
Madagascar resulted in the reclassification of F. sakalavarum as a distinct species with its own specific
pollinator species. Biological and chemical evidence is presented demonstrating that the pollinators were
able to distinguish their hosts through volatiles which emanated from the figs when they were ready to
be pollinated. Environmental factors were found to influence wasp behaviour. Ambient temperature
governed the timing of wasp emergence from their natal figs. When dispersing from their natal figs, the
.
fig wasps flew upwards and then were blown downwind. Once nearing trees bearing figs ready to be
pollinated, the wasps lost height and flew upwind towards the trees. E. baijnathi females apparently
avoided figs which already contained a conspecific foundress. Scanning electron microscope studies of
pollinating female fig wasp antennae showed that while all the species possessed multiporous plate
sensilla, in only a few species were these sensilla elongated. Multiporous plate sensilla elongation is rare
or absent among other female chalcids and may have evolved within the Agaoninae in order to facilitate
their location on receptive host figs. Pollinator choice specificity appears to break down in a number of
cases. In the first case examined, two pollinator species were recorded from the figs of African F.
sycomorus. One. C. arabicus, pollinates the figs while the other, C. galili, acts as a 'cuckoo' by utilising
some of ovules for oviposition without providing pollen. In the second case three pollinating fig wasp
species were recorded from the rigs of F. lutea. Two were found to be incidental visitors and were not
xvi
specifically attracted to the tree. The hybn ; seeds from these crosses were successfully germinated but
the seedlings did not grow passed the cotyledon stage of their development. In the concluding study the
consequences of Ficus phenology and the structure of the fig's unusual inflorescence on the nonpollinating fig wasp community were examined. Various factors affecting the popUlation levels and
species richness were also examined. Future possible research directions were discussed.
xvi i
CHAPTER 1
GENERAL INTRODUCTION
The impact fig trees (Ficus spp. Moraceae) have made on man is reflected in the numerous references
made to them in folklore, religion, agriculture and health. They feature in both Greek and Roman
mythology (Condit, 1947 and references therein) but it is in religion where their significance is most
noticeable. Ficus religiosa L. (the pipal or bo tree) was accredited with its scientific name because
of its religious significance in India and was the sacred tree under which Buddha was reputed to have
meditated in order to obtain perfect knowledge and enlightenment (Corner, 1985a). Furthermore, it
is also the tree of fertility and propagation not only to the Indians, but also to the Hellenes and the
Italians (Condit, 1947). A further species, F. sycomorus L., besides being sacred to the Egyptians,
was also prized for its wood and fruit (Galil, 1967). Although not of particular religious significance
to the Jews and the Christians, the Bible makes 58 references to fig trees and their fruit (Cruden,
1955), thereby demonstrating their importance to those communities. The Moslems, on the other hand,
had a high regard for Ficus calling it the Tree of Heaven as it was considered the most intelligent
plant, being only one step removed from animals. Even today in some Central African tribes the trees
are held in sacred respect as their ancestors are believed to dwell in them (Abbiw, 1990). Numerous
references have been made to their healing properties where they are reputed to cure anything from
epilepsy to infertility (Abbiw, 1990; Ake Assi, 1990).
It is believed that figs were first cultivated in southern Arabia ca. 2900 B.C. and were later grown in
Asia Minor and along the Mediterranean (Storey, 1975). Archimedes (700 B.C.) wrote of figs being
cultivated on the Greek Island of Paros (Condit, 1947) although trees were grown in Crete as early
as 1600 B.C. (Storey, 1975). Only F. sycomorus and F. carica L. have been cultivated for food.
Fig trees
Ficus is one of 50 genera of Moraceae (Berg, 1989a) and dates from at least the Cretaceous (> 100
million years) (Galil, 1977; Murray, 1985). Figs are assumed to have evolved from a discoid or a cupshaped inflorescence similar to that seen in the other genera of Moraceae.
The closing of the
inflorescence has been considered a 'self-defense' adaptation against generalist seed predators (Berg,
1989a).
2
The classification of FicLls is based on the work of Comer (1965) as modified by Berg (1986). Tnere
are some 750 described Ficus worldwide of which about 500 species occur in Asia and Australasia, some
150 in the Neotropics and 105 in Africa (includes Madagascar and the Mascerene Islands)(Berg, 1989a).
Although approximately 50% of Ficus are gynodiecious having so called both male and female plants
(Berg, 1989b) only 10 species occur in Africa, all of which are in the subgenus Ficus (Berg, 1989b).
Two of these occur in the southern African subregion (van Greuning, 1990) and four in Madagascar
(Berg, 1986). On mainland Africa the subgenus Sycomorus is represented by five species, two of which
are in southern Africa while seven are found in Madagascar and the Comores (Berg, 1986, 1989b). The
subgenus Urostigma has 79 described species of which 72 are placed within the section Galoglychia
which is limited to Africa. The subgenus Pharmacosycea is poorly represented in Africa with four
described species, two of which occur in Madagascar (Berg, 1986) while none are found in southern
Africa.
Figure 1. An electron micrograph of the interior of a tig showing the ostiole (0) with accompanying protective bracts (B) and the
ovules (F) lining th~
inside of th~
syconillm (S.G. Compton and L. Vincent are acknowledged for the lise of the photograph).
3
Ficus is characterised by its specialized inflorescences (Figure 1). The flowers of the fig or syconium
(= sykon (fig) Greek) line the inside wall of an urn-shaped receptacle and are only accessible through
a bract-lined entrance or ostiole (Boucek, 1988). Fig trees may be either monoecious (figs having both
male and female flowers) or gynodiecious (some figs produce both pollen and gall flowers while others
seed flowers but no staminite flowers). They are predominantly tropical or sub-tropical, growing in
a diversity of habitats that range from desert to rain forest. They may grow as trees, shrubs or lianas
and be terrestrial or hemi-epiphytic.
Many of the latter growth form kill their hosts through
strangulation or by tree splitting.
Fig Wasps
Far less well known are the small Hymenoptera (Chalcidoidea, Agaonidae) which are always found
in association with the figs. Although Aristotle and his pupil, Theophrastus, (ca. 340 B.C.) appeared
to appreciate that these small 'psen' played a role in caprification (pollination) of the cultivated fig (F.
carica) the mechanism remained a mystery. Two thousand years later Ramirez (1969) and Galil and
Eisikowitch (1969) independently and simultaneously established the mechanism fig wasps (Agaoninae
sensu Boucek, 1988) used to pollinate the figs.
Pre-agaonid wasps are thought to have been
associated with the early Ficus forms as seed predators, gall makers or parasitoids (Ramirez, 1976).
The females of many pollinating wasp species possess pollen baskets (corbiculae) which are filled
before they leave their natal fig. Arriving their new host figs, the females deliberately unload the
pollen with their front legs and place it on the flower stigma (ethodynamic pollination) (Galil, 1973;
Ramirez, 1969). Where the pollinating wasps do not possess corbicula the pollen is incidently carried
on their bodies from the natal to the host tree (topocentric pollination) (Galil, 1973; Okamato and
Tashito, 1981) although a genus of South American pollinating wasp is said to eat pollen in the natal
fig and later regurgitate once finding fig flowers ready to pollinate (Ramirez, 1969).
4
The fig wasps have evolved anatomically in order to overcome the barriers presented by the syconium
in order to gain access to the flowers within the fig lumen. Their flattened heads, mandibles modified
with lamellae or teeth, and strong fore legs assisting them in their journey through the ostiole (Figure
2).
Figure 2. Electron micrograph of the mandibles used by the pollinating fig wasp to force its way through the ostiole in order to
gain access to the flowers within the fig.
However, the agaonines are not the only chalcid wasps associated ',Nith figs. The non-pollinating fig
wasps belong to the Torymidae, Orymidae, Pteromalidae and Eurytomidae (Joseph, 1954,1955,1956,
1958,1959,1964,1965; Abdurahiman and Joseph, 1978a, 1978b, 1978c, 1979; Boucek, 1988). There
have been referred to as secondary sycophiZes (Galil and Eisikowitch, 1974), mess mates (Wiebes, 1977)
or illterlopers(Bronstein, 1988) and may be phytophages, inquilines or parasitoids. Most of these nonpollinating wasps oviposit from outside the fig (Ansiri, 1967; Joseph, 1954; Ulenberg, 1985). The
exceptions are the sycoecines (Galil et ai., 1970; Newton and Lomo, 1979; Baijnath and Ramcharun.
1983; van Noor!, 1992) which, like the pollinators,
have to penetrate the fig lumen in order to
oviposit. A single species of fig tree may support more than 20 species of fig wasp (Boucek er ai..
1981; Hawkins and Compton, 1992; Hill. 1992).
5
Among fig wasps there is a marked sexual dimorphism.
Tne males of all the pollinating fig wasps
are wingless with large mandibles and, while most non-pollinating fig wasps males are flightless, some
species have fully developed wings. Fighting and non-fighting flightless male morphs have been
reported in some non-pollinating wasp species (Vincent 1991).
Fig - Fig Wasp Developmental Cycles
Fig crop development on anyone tree is usually synchronized. However, trees tend to develop out
of phase with each other at all times of the year (Janzen, 1979a; Wharton et ai., 1980; Milton et ai.,
1982; Baijnath and Ramcharun, 1983; Newton and Lomo, 1983; Corlett, 1984; Windsor et ai., 1989;
Bronstein, 1990). The development cycle of the fig has been conveniently divided into five phases
(Galil and Eisikowitch, 1968a) (Figure 3). Sequentially they are:
Phase A (Pre-female stage): Both male and female flowers are undeveloped and the ostiole opening
is closed.
Phase B (Female stage): The female flowers have matured and the ostiole opens allowing the
pollinating female wasps to penetrate the fig lumen. Making their way through the bracts many fig
wasps lose their wings and parts of their antennae and cannot leave the tig. The female pollinates the
flowers while ovipositing down some of the ovules. The female wasps are thought to secrete from the
acid gland while laying. This secretion is thought to stimulate the pathenogenetic development and
consequent galling of the endosperm which in turns provides food for the developing wasp larvae (Hill,
1976b; Joseph and Abdurahiman,
1981; Joseph, 1984; Verkerke, 1986, 1989).
Gall forming
Eurytomidae wasps are thought to adopt a similar strategy (Copland and King, 1972).
Pollination is not a prerequisite for fruit development although fig that have not been serviced usually
abscise and abort. The act of oviposition andlor the action of the secretions of the wasp gallers' acid
glands probably prevent the abortion of the figs (Berg, 1983; Verkerke, 1988a, 1988b, 1988c, 1989).
6
The experimental introduction of female wasps without pollen results in a high wasp progeny mortality
indicating that pollination is beneficial to both the fig and the wasp progeny (Galil and Eisikowitch,
1971).
Figure 3. Fig-fig wasp development cycles. Parasitoids will arrive later than the seed predators and oviposit from outside the fig
after probing with their long ovipositors. See text for general description. (Modified from Galil and Eisikowitch, 1968).
Phase C (Interfloral Stage): Agaonine larvae and seeds develop ·simultaneously. Only the sclerotised
pericarp encasing the pupae is left at pupation. Usually a single larva develops in each seed gall (GaIil
and Eisikowitch, 1971). Bladders or empty galls are thought to be galled ovules where the larvae have
died.
Phase D (Male stage): Immediately prior to the males emerging from the fig the internal atmosphere
of F. religiosa figs are rich in carbon dioxide. This is thought to inhibit both the ripening of the fig
7
and the emergence of the female wasps from their galls (Galil et ai., 1973). After a male has located
a gall containing a female, it makes a small incision through which it inserts its telescopic
(solenogastric) abdomens and copulates with the inhabitant.
Some figs, especially those of the
subgenus Sycomorus, have their lumen filled with liquid. The emerging males are probably able cope
with this environment because of their large water-repellant spiracle peritrema (Compton and
McClaren, 1989).
The agaonine males use their well-developed mandibles to make an exit hole through the fig wall; the
carbon dioxide escapes and the females are stimulated to emerge from their galls. The ethodynamic
females then seek out the male flower anthers and load pollen before leaving their natal fig in order
to fmd another host fig with receptive figs (Phase B).
Phase E (Postfloral stage): The figs ripen and become attractive to various birds (Breitwisch, 1983;
Jordano, 1983;Wheelwright, 1985;Bronstein and Hoffman, 1987; Lambert, 1989a, 1989b;Lambert and
Marshall, 1991;Midya and Brahrnachary, 1991;Waters, pers. comm.), bats (August, 1981; Morrison,
1978; Phua and Corlett, 1989; Ulzurrum and Heideman, 1991) and mammals (Lambert, 1990;
Hemingway, pers. comm.) which act as the primary dispersers of the seeds (Janzen, 1979b; Bronstein,
1988). Ants may act as secondary dispersers (Roberts and Heithaus, 1986; Kauffmann et ai., 1991).
Early researchers suggested that figs had both 'long' and 'short' styled flowers and because of their
limited ovipositor length, fig wasps were only able to deposit their eggs in the' short' styled flowers.
This was seen to be the main factor controlling the proportion of flowers producing seed and that
producing wasps (Galil and Eisikowitch, 1968a, 1968b, 1974;Ramirez, 1970,1976; Wiebes, 1977, 1979a,
1982,1984, 1986;Faegri and van der Pijl, 1979;Janzen, 1979a, 1979b; Berg, 1983, 1989a; Murray, 1985;
Kjellberg et ai., 1987a) and was thought to be critical to the evolutionary stability of the fig-pollinating
fig wasp mutualism (Kjellberg et ai., 1987a, 1987b).
However, more recent work on monoecious figs (Newton and Lomo (1979) on F. lutea Vahl, Galil and
8
Eisikowitch (1968b) on F. sycomorus, Bronstein (1988,1992) on F. penusa,Nefdt (1989) on F. cordata
subspecies salicifolia (Warb.) C.C. Berg, F. bunt-davyi Hutch., F. verrucuiosa Warb, F. iutea, F.
thonningii Bl.,F. sycomorus, F. abulilifolia(Miq.) Miq.,F. ottolliifolia (Miq.) Miq.,F. surForssk.,F.
sansibarica Warb. and F. capreifolia Delile and Baijnath and Ramcharun (1983) on F. sur) indicates
that fig flower style length is unimodal.
These observations have placed some doubt on the
evolutionary significance of monoecious 'short' and 'long'styled flowers. Bronstein (1992) discusses
the evolutionary aspects of the consequences of the 'conflict' between the fig and its fig wasps in
maximising their individual 'fitness'.
Objectives
The objective of these studies was to investigate the interactions between figs and their fig wasps.
1. Host specificity. This section examined the host specificity between African fig wasps and
their host trees. Reason for the breakdown in host specificity are discussed and one case was resolved
through the resurrection of a fig tree taxon to species level.
2. Biological evidence for volatile attractants. Evidence is presented for the presence of Ficus
volatiles. These volatiles were shown to be species specific and emanate from the fig only when the
fruit was ready to be pollinated.
3. Chemical evidence for volatile attractants. Gas chromatograms of fig volatiles showed that
not only was the composition of the volatile profile different for each species but that it changed when
the figs were ready to be pollinated.
4. Fig wasp behaviour. The emergence of pollinators from their natal figs, their subsequent
dispersal and finally their arrival at their new hosts was examined.
5. Perception of volatiles. The antennal sensilla were examined and related to their role in
perceiving the species-specific volatiles.
6. Breakdown of host specificity. Two case studies were undertaken in an attempt to explain
the presence of more than one species of pollinating fig wasp penetrating the figs of a particular Ficus
host.
9
7. Fig wasp parasitoids. The consequences of the phenologies of fig trees as well as the
structure of their inflorescences on the biology of the non-pollinating fig wasps are discussed. The
effect of homopterans and their accompanying ants on the wasp communities was described. Factors
influencing species composition of African fig wasp communities are discussed.
References
Abbiw, D.K. (1990). The traditional significance of Ficus in West Africa. Mitt. Ins!. Allg. Bot. Hamburg
23; 395-400.
Abdurahiman, V.C. and Joseph, K.J. (1978a). Biology and behaviour of Apocrypta bakeri Joseph
(Torymidae), cleptoparasite of Ceratosolen marchali Mayr (Agaonidae). Entomon 3; 31-36.
Abdurahiman, V.C. and Joseph, K.J. (1978b).
Cleptoparasitism of the fig wasps (Torymidae:
Cha1cidoidea) in Ficus hispida L. Entomon 3; 181-186.
Abdurahiman, V.C. and Joseph, K.J. (1978c).
Bionomics of the fig cha1cidoidae (Agaonidae:
Hymenoptera. J. Anim. Morph. Physiol., Silver Jubilee Volume; 18-26.
Abdurahiman, V. C. and Joseph, K.J. (1979). Observations on the oviposition behavior in Apocrypta
bakeri (Torymidae: Hymenoptera). J. Bombay Nat. Hist. Soc. 76; 219-223.
Ake Assi, L. (1990). Utilisation de diverses especes de Ficus (Moraceae) dans la pharmocopee
traditionnelle africaine en Cote d'Ivoire. Mitt. Insl. AUg. Bot. Hamburg 23; 1039-1046.
Ansiri, M.H. (1967). The process of egg laying in Idaminae (Cha1cidoidea: Hymenoptera). Indian J.
Em. 29; 380-384.
August, P. V. (1981). Fig (Ficus trigonata) fruit consumption and seed dispersal by Artibeusjamaicensis
in the lianas of Venezuela. Biotropica 13; 79-82.
Baijnath, H. and Ramcharun, S. (1983). Aspects of pollination and floral development in Ficus
capensis Thunb. (Moraceae). Bothalia 24; 883-888.
Berg, C.C. (1986). Subdivision of Ficus subg. Urostigma sect. Galoglychia (Moraceae). Bot. Proc. C
89; 121-127.
10
Berg, C. C. (1989a). Systematics and phylogeny of the Urticales. In: Evolution, Systematics, and Fossil
History of the Hamamelidae Vol 2: Higher Hamamelidae. (Eds Crane, P.R. and Blackmore,
S). Systematics Association Special Volume No. 48B, Claraden Press, Oxford. pp. 193-220.
Berg, C.C. (l989b). Annotated check-list of the Ficus species of the African floristic region, with
special reference and key to the taxa of southern Africa. KirJda 13; 253-291.
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, U.K.
Boucek, Z., Watsham, A. and Wiebes, J.T. (1981). The fig wasp fauna of the receptacles of Ficus
thOlllliligii (Hymenoptera: Cha1cidoidae). Tijd. Ellt. 124; 149-233.
Breitwisch, R. (1983). Frugivores at a fruiting Ficus vine in a southern Cameroon tropical wet forest.
Biotropica 15; 125-128.
Bronstein, J.L. (1988). Mutualism, antagonism and the fig-pollinator interactions. Ecology 69, 12981302.
Bronstein, J.L. (1992). Seed predators as mutualists: ecology and evolution of the fig / pollinator
interaction. In Insect - Plant Interactions Vol. 4 (Ed. Bernays, E.). CRC Press, Baco Raton.
pp. 1-44.
Bronstein, J.L. and Hoffman, K. (1987). Spatial and temporal variation in frugivoryat a neotropical
fig, Ficus penusa. Gikos 49; 261-268.
Compton, S.G.,and McClaren, F.A.C.(1989). Respiratory adaptation in some male fig wasps. Proc.
K. Ned. Akad. Wet. C 92; 52-7l.
Condit, 1.J. (1947). The Fig. Chronica Botanica Co. Wattham, Massachussetts. pp 222.
Copland, M.J.W .. and King, P.E. (1972). The structure of the female reproductive system in the
Torymidae. Trans. R. em. Soc. Londoll 124; 191-212.
Corlett, R. T. (1984). The phenology of Ficus belljamilla and Ficus microcarpa in Singapore. 1. Sing.
Nat. A cad. Sci. 13; 30-31.
Corner, E.J.H. (1965). Check-list of Ficus in Asia and Australasia with a key to identification. Grd.
Bull. Singapore 21; 1-186.
Corner, E.J.H. (1985a). Essays on Ficus. Allenollia 4; 125-168.
11
Corner, E.J.H. (1985b). Ficus (Moraceae) and hymenoptera (Chalcidoidae); Figs and their pollinators.
BioI. 1. Linn. Soc. 25; 187-195.
Cruden, A. (1955). Cruden's Complete Concordance to the New and Old Testament (Eds Adams,
AD.,Irwin, C.H. and Waters, S.A).Zondevan Publishing House, Grand Rapids, Michigan.
p.217.
Faegri, K. and van der Pijl, L. (1979). The Principles of Pollination Ecology. Pergamen Press, Oxford.
Galil, J. (1967). Sycomore wasps from ancient Egyptian tombs. IsraelJ. Ent. 2; 1-10.
Galil, J. (1973). Topocentric and ethodynamic pollination. In: Pollination and Dispersal (Eds Bronjes,
N.B.M. and Linskens, H.F.). Department
of Botany, University Nijmegen, Nijmegen,
Netherlands. pp. 85-100.
Galil, J. (1977). Fig Biology. Endeavour 1; 52-56.
Galil, J. and Eisikowitch, D. (1968a). On the pollination ecology of Ficus sycomorus in East Africa.
Ecology 49; 259-269.
Galil, J. and Eisikowitch, D. (1968b). Flowering cycles and fruit types of Ficus sycomorus in Israel.
New Phytol. 67; 745-758.
Galil, J. and Eisikowitch, D. (1969). Further studies on the pollination ecology of Ficus sycomorus L.
(Hymenoptera, Chalcidoidea, Agaonidae). Tijdschr. Em. 112; 1-13.
Galil, J. and Eisikowitch, D. (1971).
Studies on the mutualistic symbiosis between syconia and
sycophilous wasps in monoecious figs. New Phytol. 70; 773-787.
Galil, J. and Eisikowitch, D. (1974). Further studies on the pollination ecology in Ficus sycomorus.
II Pocket filling and emptying by Ceratosolell arabicus Mayr. New Phytol. 73; 515-528.
Galil, J.,Dulberger, R. and Rosen, D. (1970). The effects of Sycophaga sycomori L. on the structure
and development of the syconia of Ficus sycomorus L. New Phytol. 69; 103-111.
Galil, J., Zeroni, M. and Bar-Shalom, D. (1973). Carbon dioxide and ethylene effects in coordination
between the pollinator Blastophaga quadraticeps in the syconium in Ficus religiosa. New
Phytol. 72; 113-127.
Hawkins, B.A and Compton, S.G. (1992). African fig wasp communities: undersaturation and
longitudinal gradients in species richness. 1. Allim. Ecol. 61; 361-372.
12
Herrera, C.M. (1985). Determinants of plant-animal coevolution: the case of mutualistic dispersal of
seeds by vertebrates. Gikos 44; 132-141.
Hill, D.S. (1967a). Figs (Ficus spp.) and fig wasps (Cha1cidoidae). 1. Nat. Hist. 1; 413-434.
Hill, D.S. (1967b). Figs (Ficus spp.) of Hong Kong. Hong Kong University Press, Hong Kong. pp 30.
Hill, D.S. (1992). Record fig-wasp faunas. Fig News 4; 3.
Janzen, D.H. (1979a). How to be a fig. Ann. Rev. Ecol. Syst. 10; 13-51.
Janzen, D.H. (1979b). How many babies do figs pay for babies. Biotropica 11; 48-50.
Jordano, P. (1983). Fig-seed predation and dispersal by birds. Biotropica 15; 38-4l.
Joseph, K.J. (1954). Contributions to our knowledge
of fig insects (Cha1cidoidea: parasitic
Hymenoptera) from India. V - On seven species of the genus Philorrypesis Frsst., with a note
on unisexual variations and polymorphism. Agra Univ. 1. Res. 3; 43-94.
Joseph, K.J. (1955).
Observations sur la biologie de Philotrypesis caricae L. (Hym., Cha1cidiens,
Callimomidae). C.R. Acad. Sci. Paris 241; 1626-1635.
Joseph, K.J. (1956). De la presence de chimiorecepteurs sur la tariere de PhiiotrypesiscaricaeL. C.R.
A cad. Sci. Paris 143; 1163-1164.
Joseph, K.J. (1958). Recherches sur les cha1cidiens Blastophaga psenes L. et Philotrypesis caricae L.
du figuier Ficus carica L. Ann. Sci. nat. Zoo. 20; 197-260.
Joseph, K.J. (1959). The biology of Philotrypesis caricae L., a parasite of Blastophaga psenes L.
(Chalcidoidea: parasitic Hymenoptera). Proc. XVth Internat. Can. Zoo. London 1958,662-664.
Joseph, K.J. (1964).
A proposed revision of the classification of the fig insects of the families
Agaonidae and Torymidae (Hymenoptera). Proc. R. ell!. Soc. London B 33; 63-66.
Joseph, K.J. (1965). Newer trends in the taxonomy of the fig inhabiting Chalcidoidea. Bull. Nat. Illst.
India 34; 257-262.
Joseph, K.J. (1984). The reproductive strategies in fig wasps (Cha1cidoidae: Hymenoptera) - a review.
Proc. Indian nat. Sci. Acad. B SO; 449-460.
Joseph, K.J. and Abdurahiman, V.C. (1981). Oviposition behaviour of Ceratosolenfusciceps Mayr
(Agaonidae: Hymenoptera and the mechanism of pollination in Ficus racemosa L. 1. Bombay
Nat. Hist. Soc. 78; 287-291.
13
Kaufmann, S., McKey, D.B., Hossaert-McKey, M. and Horvitz, C.C. (1991). Adaptations for a twophase seed dispersal system involving vertebrates and ants in a hemiepiphytic fig (Ficus
microcarpa: Moraceae). Am. 1. Bot. 78; 971-977.
Kjellberg, F., Michaloud, G and Valderyon, G. (1987a). The Ficus - Ficus pollinator mutualism: how
can it be evolutionarly stable? In: Insects - Plams (Eds Labergie, V., Fabres, V. and Lachaise,
D.). Dr W. Junk, Dordrecht, The Netherlands. pp. 335-340.
Kjellberg, F.,Gouyon, P.H., Ibrahim, M.,Raymond, M. and Valderyon, G. (1987b). The stability of
the symbiosis between dioecious figs and their pollinators: a study of Ficus carica L. and
Blastophaga psenes L. Evolution 41; 693-704.
Kjellberg, F., Doumeshe, D. and Bronstein, J.L. (1988). Longevity of a fig wasp (Blastophaga psenes).
Proc. K. Ned. Akad. Wet.
en; 177-122.
Lambert, F.R. (1989a). Pigeons as seed predators and dispersers of figs in a Malaysian lowland forest.
Ibis 131; 512-527.
Lambert, F.R. (1989b). Fig eating by birds in a Malaysian lowland rain forest. 1. Trap. Ecol. 5; 401412.
Lambert, F.R. (1990). Some notes on fig-eating by arboreal mammals in Malaysia. Primates 31; 453458.
Lambert, F.R. and Marshall, A.G. (1991). Keystone characteristics of bird-dispersed Ficus
1Il
a
.
Malaysian lowland forest. 1. Ecol. 79; 793-809.
Midya, S. and Brahmachary, R.L. (1991). The effect of birds upon the germination of banyan (Ficus
bengalensis) seeds. 1. Trap. Ecol. 7; 537-538.
Milton, K., Windsor, D.M., Morrison, D.W. and Estribi, M.A. (1982). Fruiting phenologies of two
neotropical Ficus species. Ecology 63; 752-762.
Morrison, D. W. (1978). Foraging ecology and energetics of the frugivorous bat Artibeus jamaicellsis.
Ecology 59; 716-723.
Murray, M.G. (1985). Figs (Ficus spp.) and fig wasps (Cha1cidoidea: Agaonidae): Hypothesis for an
ancient symbiosis. BioI. 1. Lillll. Soc. 26; 69-81.
14
Nefdt, R.J.C. (1989). Illteractiolls between Fig Wasps atui Their Host Figs. Unpublished MSc thesis,
Rhodes University, Grahamstown, South Africa.
Newton, L.E. and Lomo, A. (1979). The pollination of Ficus voge/i in Ghana. Bot. 1. Linll. Soc. 78;
21-30.
Okamota, M. and Tashito, M. (1981). Mechanism of pollen transfer and pollination in Ficus erecta by
Blastophaga nippollica. Bull. Osaka Mus. Nat. Hist. 34; 7-16.
Phua, P.B. and Corlett, R.T. (1989). Seed dispersal by the lesser short-nosed fruit bat (Cynopterus
brachyotis,Pteropodidae, Megachiroptera). Malay. Nat. 1.42; 251-256.
Ramirez -B. W. (1969). Fig wasps mechanism of pollen transfer. Science 163; 580-581.
Ramirez -B., W. (1970). Host specificity of fig wasps (Agaonidae). Evolution 24; 680-691.
Ramirez -B., W. (1976). Evolution ofblastophagy. Brenesia 9; 1-13.
Ramirez -B., W. (1978).
Evolution of mechanisms to carry pollen in Agaonidae (Hymenoptera
Cha1cidoidea). Tijd. Em. 121; 279-293.
Roberts, J.T. and Heithaus, E.R. (1986). Ants rearrange the vertebrate-generated seed shadow of a
neotropical fig tree. Ecology 67; 1046-1051.
Storey, W.B., 1975. Figs. In: Advances in Fruit Breeding (Eds Janide, J. and Moore, J.N.). Purdue
University Press. pp. 568-589.
Ulenberg, S.A. (1985). The systematics of the fig wasp parasites of the genus Apocrypta Coqueral.
Proc. K. Ned. Akad. Wet. C 83; 1-7.
Ulzurrum, R.C.B. and Heideman, P.D. (1991). Differential ingestion of viable vs nonviable Ficus seeds
by fruit bats. Biotropica 23; 311-312.
van Greuning, J. V. (1990). A synopsis of the genus Ficus (Moraceae) in southern Africa. Sth. Afr. 1.
Bot. 56; 599-630.
van Noort, S. (1992). The Systematics and Phylogelletics of the Sycoecillae (Agaollidae, Chalcidoidea,
Hymenoptera). Unpublished PhD thesis, Rhodes University, Grahamstown, South Africa.
Verkerke, W. (1986). Anatomy of Ficus otfoniiJolia (Moraceae) syconia and its role in fig-fig wasp
symbiosis. Proc. K. Ned. A cad. Wet. C 89; 443-469.
15
Verkerke, W. (1988a). Syconial anatomy of Ficus asperifolia (Moraceae), a gynodiecious tropical fig.
Proc. K. Ned. Akad. Wet. C 90; 461-492.
Verkerke, W. (1988b). Flower development in F. sur Forsskal (Moraceae).
Proc. K. Ned. Akad. WeI.
C 91; 175-195.
Verkerke, W. (1988c).
(Moraceae).
Sycone morphology and its influence on the flower structure of F. sur
Proc. K. Ned. Akad. WeI. C 91; 319-344.
Verkerke, W. (1989). Structure and function of the tig. Experiemia 45; 612-622.
Vincent, S.L. (1991).
Polymorphism and Fighting ill Male Fig Wasps.
Unpublished PhD thesis,
Rhodes University, Grahamstown, South Africa.
Wharton, R.A., Tilson, J.W. and Tilson, R.L. (1980).
Asynchrony in a wild population of Ficus
sycomorus L. Sth. Afr. J. Sci. 76; 478-480.
Wheelwright, N.T. (1985). Fruit size, gape width, and the diets of fruit eating birds. Ecology 66; 808814.
Wiebes, J.T. (1977). A short history of fig wasp research. Grdn's Bull. Singapore 29; 207-233.
Windsor, D.M., Morrison, D.W., Estribi, M.A. and de Leon, B. (1989). Phenology of fruit and leaf
production by 'strangler' figs on Barro Colorado Island, Panama. Ex:periemia 45; 247-253.
16
CHAPTER 2
HOST SPECIFICITY
Paper 1: African figs and fig wasps: The wasp's eye view of Ficus species. Mitteilungen aus dem Jllstitutfur
Allgemeine Botanik Hamburg 24 (S.G. Compton, A. B. Ware and S. van Noort - 1991).
of Ficus species break down in southern Africa, Madagascar and The
Paper 2: Does pollinator sp~ciJty
Comores. Submitted to Annals of the Missouri Botanical Garden. (A.B. Ware, S.G. Compton and
P.B. Phillipson)
17
AFRICAN FIGS AND FIG WASPS:
THE WASP'S EYE VIE'W OF FICUS SPECIES
s. G.
Compton, A.B. Ware and S. van Noort
ABSTRACT
Fig trees (Ficus species, Moraceae) are pollinated by agaonine fig wasps (Hymenoptera, Agaonidae,
Agaoninae). We describe the plant characteristics that determine the host specificity of the wasps
and assess the role of fig wasps in the reproductive isolation of Ficus species. The practicalities of
using the pollinators to identify
and delimit Ficus species are examined and cases where the
classifications of the trees and wasps do not correspond are reviewed. We conclude that fig wasp host
relationships provide useful pointers to where future taxonomic studies should be directed.
INTRODUCTION
Our understanding of the systematics of African fig trees (Ficus spp., Moraceae) has improved greatly
over the last few years, thanks mainly to the work of C. C. Berg (for. example Berg, 1986,1988; Berg
and Hijman, 1989). Through the production of identification keys and adequate descriptions, Berg's
revisions have made African Ficus accessible to biologists interested in this taxonomically 'difficult'
genus. They have also resulted in the detection of large numbers of synonymies, and only 105 African
and Malagasy species are currently recognised.
Fig trees are of particular interest to ecologists and evolutionary biologists because of their unique
pollination system. All fig trees depend entirely on fig wasps (Hymenoptera: Agaonidae, subfamily
Agaoninae) for pollination (Boucek, 1988). There are numerous other groups of fig wasps, for
example the Sycoecinae, but these do not act as pollinators. The structure of the figs, together with
the trees' unusual asynchronous flowering phenology, are adaptations that facilitate pollination by the
wasps, but exclude other potential pollinators (Verkerke, 1989; Berg, 1990; Janzen, 1979). The tiny
17
flowers of Ficus are positioned on the inside of the fig, where they can only be reached by crawling
through the narrow bract-lined ostiole. Once inside, the wasps pollinate the flowers and gall a
proportion of the ovules, inside which the wasp larvae develop.
Agaonines are only found in
association with fig trees, and can breed nowhere else. As the trees provide sites for the development
of wasp larvae, while the wasps transfer pollen for the trees, the interaction is mutualistic.
The relationship between Ficus species and Agaonine species is believed to be usually highly specific,
with each tree species pollinated by only one wasp species, which does not breed in the figs of any
other Ficus. The host relationships of the fig wasp genera broadly correspond with the subdivisions
of Ficus
recognised by Berg.
Thus, trees belonging to subgenus Sycomorus, are pollinated by
Ceratosolen species, trees in subgenus Urostigma, section Urostigma are pollinated by Platyscapa
species, and so on.
An exception to this correspondence between trees and wasps is found in
Urostigma, section Galoglychia, the most species rich Section in Africa. Even here, however, the
disparity in classifications is only present in three of the six subsections (Wiebes, 1990).
The generally parallel phylogenies of the fig trees and their wasps has led to the suggestion that
speciation in the trees and the wasps may be linked and Thompson (1989) has concluded that figs and
fig wasps represent one of the strongest cases for such co-speciation having taken place. This is
because gene flow in both groups is intimately linked with that of their partners.
In this paper we
examine those physical and chemical features of the plants which influence host specificity in fig wasps
and examine the role of the wasps in the reproductive isolation of Ficus species. We then review cases
where the classifications of the trees and wasps do not correspond and examine possible reasons for
the disparities.
Reproductive isolation in African Ficus
Agaonines are effectively the sole pollinators of fig trees, although rare instances of pollen grains being
transported by other fig wasps have been reported (Newton and Lomo, 1979; Compton, Holton,
18
Rashbrook, van Noort, Vincent and Ware, 1991). Agaonine host choice therefore controls the limits
of gene flow in Ficus species.
The pollination syndrome in Ficlls has some similarities with that of the bee orchids (Ophrys spp.,
Orchidaceae). In bee orchids it is host-specific aculeate bees and wasps, fooled into pseudocopulating
with the flowers, which typically act as prepollination isolating factors. Paulus and Gack (1990a, 1990b)
argue that speciation in Ophrys has resulted from a change in pollinators and that many of the
morphologically distinct variants and subspecies of Ophrys species should be regarded as good species
because they each have their own specific pollinators. However, species need not have detectably
different morphologies, as it is only necessary that their pollinators should be able to distinguish
between them.
Figs are only attractive to their specific pollinators during a short period of their development, when
large numbers of agaonines can be collected at the trees (Bronstein, 1987). The wasps are attracted
to the trees by volatile compounds released from the figs (van Noort, Ware and Compton, 1989). T'ne
blend of these chemicals does not remain constant and attractiveness corresponds with a short period
when the ostiole opens and there is a detectable change in the smell of the figs (Ware, Kaye, Compton
and van Noort, in prep.). The Ficus species we have tested have elements of their volatile profile that
are consistent and differ 'from those of other species (Ware, Kaye, Compton and van Noort, in prep.).
These differences appear to form the basis of the specificity of their attraction.
Most plant speCIes are not isolated by single barriers, but by combinations of different factors
(Stebbins, 1950; Levin, 1978). In Ficus species the ostiole provides a physical filter that limits entry
to the fig (Janzen, 1979). This supplements the isolation generated by the specificity of the volatile
attractants. Fig wasps have anatomical modifications that facilitate entry through the ostiole. These
include a flattened head, the presence of teeth or ridges on the mandibles and short, heavy fore-legs
with strong tibial spines. Ostiole shape and size varies greatly between Ficus species (Ramirez, 1974)
and fig wasp head shapes appear to be adapted to the ostiole characteristics of their associated tree.
This is reflected in the parallel development of head shape in agaonine and sycoecine fig wasps that
19
share the same hosts, such that when the agaonid has a long thin head, so does the sycoecine (van
Noort and Compton, in prep.). Despite these adaptations, successful entry into the figs is not assured
and, for example, around one per cent of the Elisabethiella baijnathi females entering the figs of F.
burtt-davyibecome trapped in the ostiole (Compton and Robertson, in prep.). Failure rates are likely
to be much higher when wasps attempt to enter figs for which they are not adapted.
Despite the physical barrier posed by the ostiole, fig wasps do occasionally succeed in entering the
'wrong' figs, and may even succeed in reproducing (Compton, 1990). The colonisation of non-host
trees seems to result from the accidental arrival of a few wasps at trees with unpollinated figs, rather
than from a breakdown in the specificity of the volatile attractants (Ware and Compton, in prep.).
Once on a tree bearing figs, a proportion of the fig wasps appear to be drawn inside them, irrespective
of the Ficus species.
The occasional 'mistakes' made byagaonines result in the transfer of pollen between fig species, with
the possibility of hybrids being produced. In the case of a F. lutea tree growing in Grahamstown that
was pollinated by wasps from F. thonningii and F. sur, viable hybrid seeds were produced from both
crosses (Compton, 1990). This was despite F. sur and F. lutea being in separate subgenera. Hybrids
have also been produced from crosses involving the edible fig F. carica (Condit, 1950), suggesting that
cross-incompatability may be poorly developed throughout the genus. However, we have not been
able to grow successfully any F. lutea hybrids, and such hybrid weakness/inviability may also result in
reproductive isolation in Ficus .
Exceptions to the one fig: one agaonine relationship
Agaonines have been collected from about 70 % of the African Ficus species (Wiebes and Compton,
1990; Compton, unpublished), a higher proportion than that known from other Ficus-rich continents.
Africa is therefore particularly suitable for using fig wasps to assess the status of Ficus species.
20
Table 1. African Ficus species with different agaonines associated
with their subspecies or synonyms'
#*¥
"
Berg code
n
5
11
28
60
82
95
•
t
Ficus taxa
Agaonines
F. asperiJolia
Kradibia geslroi afrom
F. ureo/aris'
Kradibia hilli
F. sycomoms
CeralOsolen arabicus &
Ceralosoien galiIi
F. sakalavanlln'
CeralOsolen namorakensis
F. c. cordata
Plaryscapa desertomm
F. c. salicifolia
P/aryscapa awekei
F. n. natalensis
Elisabelhiella socOlrensis &
Aifonsiella longiscapa
F. n. lepieurii
Aifonsiella jimbriata
F. c. cyathislipula
Agaon Jasciatum
F. c. pringsheimiana
Agaon kiellandi
F. o. ononiifolia
COl/rtella camemnensis &
COl/rtella gabonensis
F. o. lucanda
COl/rtella scobinifera
i\W!IiY'ee.
Tables 1-3 are based on host records summarised by Wiebes in Wiebes and Compton (1990),
supplemented by a small number of more recent records. Table 1 lists four examples where pairs of
Ficus subspecies are pollinated by different agaonids and two where different pollinators are associated
with previously recognised tree species that, while morphologically distinct, are now regarded as
synonyms. These taxa appear to be candidates for recognition as separate species. However, the
examples in Table 1 represent only a few of the 27 African Ficus that do not display a one:one
relationship with the agaonines. These additional cases where trees have two or more associated
agaonines may indicate the presence of cryptic Ficus species (Table 2). Conversely, there are also
numerous examples of the same agaonine being collected from more than one Ficus (Table 3). This
brings into question the status of these Ficus species, although several are so different in appearance
that their specific status seems beyond doubt.
21
Table 2. African Ficus species with more than one associated agaonine
(excluding cases listed in Table 1).
-
§Mk
Berg code
-MW.
~;.
Ficus species
h38MW'
tim
Agaonines
A'A.'hWIW@
tr1HWM
F. palmata
Blastophaga psenes &
Blaszophaga vaidi
11
F. sycomonls
Ceratosolen arabicus &
Ceratosolen galili
12
F. mllCOSO
Ceratoso/en arabicllS &
Ceratosolen galiJi
13
F. sur
Ceratoso/en capensis &
Ceratosolen flabellatus & Ceratosolen
? silvestrianllS
15
F. vallis-choudae
Ceratosolen megacephalus &
Ceratosolen ? silveslrianlls
36
F.lurea
Allotriozoon heterandromorphum &
Elisabethiella stuckenbergi &
Ceratosolen capensis
47
F. ablltilifolia
Mgeriella jUsciceps &
Elisabethiella comptoni
58
F. craterostoma
Alfonsiella michloudi &
Alfonsiella sp. indesc.
60
F. n. naralensis
Elisabethiella socolrensis &
Alfons/elia longiscapa
66
F. Ihonningii
Elisabethiella sluckenbergi &
Alfonsiella brongersmai & Alfonsiella
longiscapa
95
F. olloniifolia
COl4nella camenlnensis &
Counella gabonensis
97
F. ancarpoides
Counella penicula &
COllnella hladikae
-=
iIliI
iiiIII:!iI
Numerous non-pollinating fig wasps share the figs with the agaonines. The host relationships of these
species can provide additional evidence on the status of their hosts, although these wasps have no
influence on gene flow in the plants. Sycoecine fig wasps are ovule-gallers associated with Ficus
section Galoglychia, and like agaonines must enter the figs to oviposit. Their larvae do not need the
figs to be pollinated and therefore can develop independently of the agaonines.
22
Table 3. African Ficlls species which share agaonines with congeners.
ill!
Berg code
=r
~
Ficlls species
~t"
*M
AM
R&
F. palmata
F. carica
BlaslOphaga psenes
Blaslophaga psenes
2
5
6
F. exasperara
F. asperifolia
F. capreijoJia
Kradibia geslroi afrum
Kradibia gestroi afrum
Kradibia geslroi afrum
11
F. sycomorns
12
F. mucosa
CeralOsoien
Ceratosolen
CeralOsolen
Ceratosolen
13
36
F. sur
F.lutea
Ceralosolen capensis
Ceratosolen capensis
13
15
F. sur
F. vallis-choudae
Cera lOS olen ? silvestrianus
Ceratosolen ? silvestrianus
23
24
F. variijolia
F. dicranostyla
Dolichoris jlabellata
Dolicholis jlabellata
36
66
F. Lutea
F. thonningii
Elisabethiella stuckenbergi
Elisabethiella stuckenbergi
40
41
60
F. vasra
F. wakejieldii
F. n. natalensis
Elisabethiella socotrensis
ElisabethieUa socolrensis
ElisabethieUa socotrensis
58
59
F. crateroslOma
F. lingua
Aijonsiella michaloudi
Alfonsiella michaloudi
60
66
F. n. nalaiensis
F. lhonningii
AijonsieUa longiscapa
Aijonsiella Longiscapa
60
F. n. leprieurii
F. kamenmensis
Alfonsiella jimbriala
Aijonsieila fimbriaia
Agaon kiellandi
86
F. conraui
F. cyalhistipula
pringsheimiana
F. densistipulata
Agaon kiellandi
Agaon kiellandi
90
91
F. sagiltifolia
F. subsagiltifolia
Agaon c. cicalriferens
Agaon c. multum
67
76
82
..
Shared agaonines
23
arabicus &
galiJi
arabicus &
galiJi
Table 4. Host relationships of selected sycoecine iig wasps.
Q
Berg code
.....
'*isaMM
®
Ficus species
Agaonines
Associated sycoecines
40
F. vasta
Elisabelhiella socOfrensis
Crossogaster rrifonnis
60
F. n. nalaiensis
Eliseabelhiella socorrensis
PhagoblaslUs barbams
60
F. n. naralensis
Aljonsieila longiscapa
Crossogaster A
Philocaenus A
60
F. ieprieurii
Aljonsiella jimbriara
PhagoblaslUS liodonllls
67
F. kamenmensis
r.lljonsiella jimbriata
Phagoblastus D
58
F. crateros!Oma
Aljonsiella michaloudi
Phagoblastus A
Phagoblasllls B
Phagoblastus liodonws
59
F. lingua
Aljonsiella michaloudi
Phagoblastus B
58
F. cralerostoma
Aljonsiella sp. indescr.
Phagoblastus C
Crossogasler oderans
66
F. thonningii
Elisabethiella stuckenbergi
PhagoblaslUS barbarus
Crossogaster oderans
66
F. Ihonningii
Aljonsiella brongersmai
Phagoblastlls barbarus
Phagoblastus E
Philocaenlls A
Crossogasler oderans
82
F. c. cyalhistipula
Agaon fascialum
Sycoeclls
thaumos!Ocnema
82
F. c. pringsheimiana
Agaon kiellandi
Sycoecus A
90
F. sagittijolia
Agaon c. cicalrijerens
Sycoecus B
91
F. subsagittifolia
Agaon c. multum
Sycoecus C
95
F. o. ot!Oniljolia
COllrtella camenmensis &
COllnella gabonensis
Seres A
95
F. o. ulugllrensis
Counella camerunensis
Seres B
95
F. o. lucanda
COl/nella scobinifera
Seres levis
97
F. anocarpoides
COllnella pendicula &
Courtella hladikae
Seres C
aw
•
7 iA44AS=
W·
ifiHii ,pM'
Sycoecines collected from trees where there is not a one agaonine: one Ficus relationship are listed
in Table 4. There are often several sycoecines associated with a particular Ficus, but host specificity
can be weIl developed. Although host records of sycoecines are not as numerous as those of the
pollinators, some interesting patterns do emerge. Phagoblasrus D and Philocaenus A occur in F.
thollllingii pollinated by Alfollsiella brongersmai,but have never been found in numerous samples from
F. thoJlllingii pollinated by Elisaberhiella stuckellbergi. Similarly, Phagoblastus barbarusoccurs in F.
24
II.
natalensis when pollinated by Elisabethiella socolrensis, but not when pollinated by AlfollSielia
longiscapa. In both examples the sycoecines suggest that the trees with two pollinators may actually
be closely related species.
DISCUSSION
The large number of differences between the currently recognised Ficus species and the host records
of the agaonines may have a variety of causes. Table 5 summarises various factors which could result
in apparent or real breakdowns in the one:one relationship between FicLls species and agaorune
species. Some of the explanations are well documented, while others are hypothetical.
Incorrect assignments of species-pairs can result from misidentifications of either the trees or the
wasps, contamination of collections and mislabelling of specimens. Such errors should be detected
eventually, as subsequent samples indicate anomalies, but at present a large proportion of the data is
still based on single collections. Delimitation of species also remains a problem in certain groups of
both trees and wasps. The F. thonningii species group is especially problematic, while the separation
of closely related wasps such as Elisabethiella stuckenbergi and E. socotrensis also leads to
uncertainties. Closely related species may be indistinguishable using classical taxonomic methods and
alternative approaches may be required to differentiate them.
Well documented case studies have shown that there are genuine exceptions to the one agaonine: one
Ficus species pattern. In West Africa, the nominate subspecies of F. ottolliifolia is pollinated by two
different agaonines. These show habitat preferences, with trees in the forest pollinated mainly by one
species, those in savannah by another (Michaloud, Michaloud-Pelletier, Wiebes and Berg, 1985). F.
sycomorus also has two associated agaonines, but here only one wasp species pollinates the tree (Galil
and Eisikowitch, 1968). The same two wasps are associated with F. mucoso, where it is again the
same wasp species that pollinates the tree (Wiebes, 1989). Mistakes by the wasps may also be
responsible for two or more agaonines entering figs on the same tree, as was described above with
F. lurea. Another example may include the two Alfollsiella species recorded in small numbers from
25
F. thollllingii (Boucek, Watsham and Wiebes, 1981) that appear to be the legitimate pollinators of
other tree species. Another explanation for different Ficus species sharing pollinators may be that they
have alternative fonns of reproductive isolation which have superseded the wasps. This would be
analogous to the situation with some atypical bee orchids, which remain distinct species despite sharing
the same pollinators (Paulus and Gade, 1990b).
Table 5. Possible causes of exceptions to the one agaonine:
one Ficus species relationship.
I_'g;wq
3f t ewSh?M"'W'M
PUG
444i'
e
TAXONOMIC
1.
Misidentifications of the trees or wasps will lead to incorrect
assignments of species-pairs.
2.
Natural variation in trees or wasps can result in uncertainties
about the delimitation of species.
3.
Agaonine species from different hosts may be anatomically
the same, but have different host preferences.
4.
Ficus species with different pollinators may be morphologically similar, but have cryptic differences that allow them to
be distinguished by the wasps.
BIOLOGICAL
5.
Trees may have two or more sympatric pollinators or may
have different pollinators in different habitats or in different
parts of their range.
6.
7.
8.
--
One or more of the associated agaonids may no longer act as
a pollinator.
Wasps can make 'mistakes', occasionally pollinating and even
reproducing inside the 'wrong' figs.
Some Ficus species may rely on post-fertilization isolating
mechanisms, rather than pollinator specificity.
The numerous examples of mis-matches between the wasps and the trees suggest that we have much
to learn about both the biology and taxonomy of African figs and fig wasps. In particular, with our
present state of knowledge it is often impossible to distinguish the factors that are responsible for
apparent breakdowns in the one:one relationship. Data on fig wasp host relationships are nonetheless
of immediate value to both Ficus and agaonine taxonomists, because they point to areas where future
studies should be directed.
26
ACKNOWLEDGEJ\IENTS
Thanks to P.E. Hulley for his valuable comments on the manuscript.
REFERENCES
Berg, C. C. (1986). The Ficus species (Moraceae) of Madagascar and the Comoro Islands. Bull. Mus.
!wtn. Paris 8: 17-55.
Berg, C.C. (1988). New taxa and combinations in Ficus (Moraceae) of Africa. Kew Bull. 43: 77-97.
Berg, C.C. (1990). Differentiation of flowers and inflorescences of Urticales in relation to their
protection against breeding insects and to pollination. Sommerfeltia 11: 13-34.
Berg, C.C. and Hijrnan M.E.E. (1989). Flora of Tropical East Africa. Moraceae. Balkema, Rotterdam,
The Netherlands.
Boucek, Z. (1988). Australasian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford,
Oxon, U.K.
Boucek, Z., Watsham, A. and Wiebes, J.T. (1981). The fig wasp fauna of the receptacles of Ficus
thOllllillgii (Hymenoptera, Cha1cidoidea). Tijd. Em. 124: 149-233.
Bronstein, J.L. (1987).
Maintenance of species-specificity in a Neotropical fig-pollinator wasp
mutualism. Gikos 48: 39-46.
Compton, S.G. (1990). A collapse of host specificity in some African fig wasps. Sth. Afr.l. Sci. 86:
39-40.
Compton, S.G.,Holton, K.C., Rashbrook, V.K., van Noort, S., Vincent, S.L. and Ware A.B. (1991).
Studies of Ceratosofell galili,a non-pollinating agaonid fig wasp. Biotropica 23: 188-194.
Condit, U. (1950). An interspecific hybrid in Ficus. 1. Hered. 41: 165-168.
Galil, J. and Eisikowitch, D. (1968). On the pollination ecology of Ficus sycomorus in East Africa.
Ecology 49: 259-269.
Levin, D. A. (1978). The origin of isolating mechanisms in flowering plants. Evol. BioI. 11: 185-317.
Janzen, D.H. (1979). How to be a fig. Anll. Rev. Ecol. Syst. 10: 13-51.
Michaloud, G., Michaloud-Pelletier, S.,Wiebes, J.T.and Berg, C.C.(1985). The co-occurrence of two
27
pollinating species of fig wasp and one species of fig. Proc. K. Ned. Akad. Wet. Ser. C 88: 93119.
Newton, L.E. and Lomo, A. (1979). The pollination of Ficus vogelii in Ghana. Bot. J. Linn. Soc. 78:
21-30.
Paulus, H.F. and Gack, C. (1990a). Pollination of Ophrys (Orchidaceae) in Cyprus. PI. Syst. Eval. 169:
XX-YY.
Paulus, H.F. and Gack, C. (l990b). Pollinators as prepollinating isolation factors: Evolution and
speciation in Ophrys (Orchidaceae). IsraelJ. Bot. 39: 43-79.
Ramirez, W.B. (1974). Coevolution of Ficus and Agaonidae. Ann. Missouri Bot. Gard. 61: 770-780.
Stebbins G.L. (1950). Variation and Evolution in Plants. Columbia University Press, New York.
Thompson, J.N. (1989). Concepts of coevolution. Trends Ecol. Evol. 4: 179-183.
van Noort, S., Ware, A.B. and Compton, S.G. (1989). Pollinator-specific volatile attractants released
from the figs of Ficus bunt-davyi. Sth. Afr. J. Sci. 85: 323-324.
Verkerke, W. (1989). Structure and function of the fig. Experientia 45: 612-62l.
Wiebes, J. T. (1989). Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, IV (African Ceratosolell). Proc. K. Ned. Akad. Wet. Ser. C 92: 251-266.
Wiebes, J. T. (1990). African figs and their pollinators- a brief overview. Mitt. Illst. AUg. Bot. Hamburg
23a: 425-426.
Wiebes, J.T. and Compton, S.G. (1990).
Agaonidae (Hymenoptera
Chalcidoidea) and Ficus
(Moraceae): fig wasps and their figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet. 93: 203222.
28
DOES POLLINATOR SPECIFICITY OF FICUS SPECIES BREAKDOvVN
IN SOUTHERN AFRICA,MADAGASCARAND THE COMORES?
A. B. Ware, S. G. Compton and P. B. Phillipson
ABSTRACT
Fig trees (Ficus spp.) are only pollinated by fig wasps (Hymenoptera, Agaonidae, Agaoninae) and each
Ficus species is usually pollinated by its own specific species of fig wasp. This one-to-one relationship
has led biologists to view figs and fig wasps as one of the classic examples of coevolution between
plants and animals. In this paper we summarise the host relationships of the pollinating fig wasps
recorded from South Africa, Namibia, Madagascar and The Comores and examine those cases where
the one-to-one relationship appears to break down. We discuss possible reasons for such apparent
breakdowns in specificity and how these anomolies relate to the hypothesis of coevolution between
the trees and their pollinators. A consideration of one such case leads us to propose that Ficus
sakalavarum Baker from Madagascar is a distinct species from the related F. sycomorus L.
INTRODUCTION
The Ficus-fig wasp pollinator relationship is one of obligate mutualism. Fig trees (Ficus spp., Moraceae)
are dependent upon female fig wasps (Hymenoptera, Chalcidoidea, Agaonidae, Agaoninae sensu Boucek,
1988) for pollination and in return the wasps use some of the ovules for oviposition and subsequent larval
development (Galil, 1977). Baker (1961), Hill (1967), Ramirez (1970), Galil (1977), Wiebes (1979),
Janzen (1979) and Michaloud et aI. (1983), among others, have remarked on the specificity of the
relationship between each species of fig tree and their pollinating wasps and the fig-fig wasp relationship
has been viewed as one of the best documented examples of plant-insect co-evolution (Janzen, 1979;
Thompson 1982, 1989; Bronstein and McKey, 1989). Furthermore, because Agaoninae are host specific
and usually the trees' sole pollinators (for exceptions see Newton and Lomo, 1979; Compton et aI.,
29
1991) the wasps control the limits of gene flow in Ficus species. This, together with the similarities
between the phylogenies of the Agaoninae and Ficus (Wiebes, 1982), has led to the suggestion that
speciation of the two groups is linked (Thompson, 1989).
The present classification of Ficus is based on the 'rather weak differentiating morphological and
anatomical "key" characters' of Corner (1965)(Berg, 1990). Even so there is broad agreement between
these subdivisions and the phylogeny of the Agaoninae as proposed by Wiebes (1982).
For example,
trees belonging to the subgenus Sycomorus are pollinated only by wasps of the genus Ceratosolen, and
those of the subgenus Urostigma section Urostigma by Platyscapa wasps and so on. However, within the
subgenus Urostigma section Galoglychia this correspondence breaks down (Berg, 1989; Wiebes, 1990).
The host specificity of certain fig wasp species is not always absolute and several cases of two or more
Agaoninae species found in association with one species of African Ficus have been documented (Galil
and Eisikowitch, 1968; Boucek et ai., 1981; Michaloud et ai., 1985; Compton, 1990; Compton et ai.,
1991; Wiebes and Compton, 1990; Ware and Compton, in press).
In this study we record the host specificity of fig tree pollinators from South Africa, Namibia, The
Comores and Madagascar. We discuss possible reasons for exceptions to the one Ficus species - one
pollinator species pattern, and the implications of this on the taxonomy of the fig tree species.
NATURAL HISTORY AND THE BASIS OF FIG WASP SPECIFICITY
The Ficus inflorescence (the fig or syconium) is unusual in that the flowers are contained within a
globular receptacle and access to them is through a narrow bract-lined entrance - the ostiole. Fig structure
prevents incidental pollination (Verkerke, 1989; Berg, 1990; Janzen, 1979) and the anatomy and
behaviour of the pollinating fig wasps have evolved to overcome these barriers (Ramirez, 1974). For
example, the head shapes of pollinating females are related to the ostiole structure of their host figs (van
Noort, 1992). During the passage through the ostiole to the lumen of the fig the female wasps typically
lose their wings and part of their antennae, and cannot leave. Their decision to attempt entry into a fig
is therefore essential to their future reproductive success.
30
Fig pollination and development has been divided into five phases (Galil and Eisikowitch, 1968):
1. the prefemale phase - when the female flowers are undeveloped and the ostiole is closed;
2. the female phase - when the ostiole opens allowing the pollinating wasps access to the mature female
flowers;
3. the interfloral phase - once pollination has taken place the seeds and the fig wasp larvae develop
simultaneously;
4. the male phase - when the male flowers, wasps and the seeds have reached maturity, the flightless
male wasps chew their way out of their galls and seek flower galls containing conspecific females, they
chew through the galls and copulate with the trapped females; the females then, actively or passively
collect pollen and leave their natal fig through exit holes chewed through the wall of the fig by the males;
5. the postfloral phase - the figs ripen and are eaten by frugivores which disperse the seeds.
Fig crop development is usually synchronous on each tree (for an exception see Baijnath and Ramcharun,
1983), but not between trees (Bronstein, 1987; Wharton et aI., 1980; Windsor et aI., 1989). This means
that female fig wasps must leave their natal trees in order to fmd suitable figs in which to oviposit. Fig
wasps are only attracted to trees bearing figs that are ready to be pollinated (Bronstein, 1987; Ware and
Compton, in prep A.). Volatiles emanating from the figs were shown to be the source of the attraction
(van Noort et ai., 1989; Ware et ai. in press, Ware and Compton, in prep. B). Therefore, host plant
specificity, at least in part, seems to result from the flying wasps being attracted to specific volatile
components released by their host figs.
Pollinator fig wasps are not the only wasps which are uniquely associated with figs (Boucek, 1988).
Some wasp species belonging to other subfamilies of Agaonidae also feed on the developing figs, while
others parasitise the wasp larvae.
Many of these non-pollinating wasps are also apparently host specific
(Ulenberg, 1985; van Noort, 1992) and can provide additional evidence on the species status of their
hosts, but as none of these species pollinate the figs they have no effect on fig gene flow.
Table 1. A list of southern African Ficus species (numbers from Berg (1989)) together with their associated agaonines (numbers
from Wiebes and Compton (1990)) together with the number of trees sampled.
H
*#!I.JPft
Ficus species
Nt
....
EM
Agaonine species
Wf·@
w
lBiNfBi¥f#WifMiiM#
Number
of trees
sampled
,*ib¥¥&k
- F. carica L.
1 Biastophaga psenes L.
3
6 F. capreifolia Dellie
2 Kradibia gestroi (Wiebes)
2
?
7 F. pygmaea Hiern
11 F. sycomorus L.
7 Ceratosoien arabicus Mayr
o
27
14 Ceratosoien gaiili Wiebes
13 F. sur Forssk.
11 Ceratosolen capensis Grandi
30
27 F. ingens (Miq.) Miq.
24 Platyscapa soraria Wiebes
16
28a F. cordata subsp. cordata Thunb.
25 Platyscapa desertorum Compton
17
28b F. cordata subsp. salicifoIia (Yahl) Berg
20 Platyscapa awekei Wiebes
11
29 F. verruculosa Warb.
21 Platyscapa binghami Wiebes
8
36 F. lutea Vah!
28 Allotriozoon heterandromorphum Grandi
4
38 Elisabethiella stuckenbergi Grandi
11 Ceratosolen capensis Grandi
42 F. glumosa Delile
39 Elisabethiella glumosae Wiebes
23
43 F. stuhlmanniiWarb.
48 AIfonsiella binghami Wiebes
10
45 F. tettensis Hutch.
44 Nigeriella excavata Compton
6
47 F. abutiIifolia (Miq.) Miq.
40 ElisabethielJa comptoni Wiebes
17
50 F. trichopoda Baker
32a Elisabethiella bergi bergi Wiebes
13
58 F. craterostoma Mildbr. & Burr.
60a F. natalensis subsp. natalensis Hochst.
• AlfonsielJa sp. indet.
38 ElisabethieIJa stuckenbergi Grandi
2
17
42 Elisabethiella socotrensis Mayr
53 AIfonsiella longiscapa Joseph
62 F. burtt-davyi Hutch.
36 ElisabethieJla baijnathi Wiebes
18
63 F. iIicina (Sonder) Miq.
37 EIisabethiella enriquesi (Grandi)
11
66 F. thonningii BI.
38 EJisabethiella stuckenbergi Grandi
31
96a F. tremuia subsp. tremula Warb.
75 Courtella wartii Compton
5
98a F. poJita subsp. poJita Vahl
70 CourtelJa bekiliensis (Risbec)
3
99 F. bizanae Hutch. & Burtt-Davy
- Courtella sp. indet.
3
101 F. sansibarica subsp. sansibarica Warb.
72 CourtelJa annata (Wiebes)
5
104 F. bubu Warb.
80 CourtelJa michaloudi (Wiebes)
2
@
32
•
it
Table 2. A list of Ficus species from Madagascar and The Comores (numbers from Berg (1989» and their associated agaonines
(numbers from Wiebes and Compton (1990» together with the number of trees sample for their pollinators.
*'w
4wa;;w
•
Number
of trees
sampled
Agaonine species
Ficus species
. . mee
+g
&A
i¥f¥i¥¥&
64MB
w
3 F. pachycJada subsp. arborea (Perrier) C.C. Berg
w
- Kradibia sp. inde!.
8 F. bojeri Baker
2
o
?
9 F. brachycJada Baker
4 Kradibia cowani Saunders
1
10 F. politoria Lam.
5 Kradibia saundersi Wiebes
1
11 F. sycomorus L.
7 Ceratosoien ambicus Mayr
11
14 Ceratosoien gaJili Wiebes
- F. sakaJavarum Baker
17 F. tiliifolia Baker
9 Ceratosolen namorakensis (rusbec)
13
8 Ceratosoien stupefactus Wiebes
2
18 F. torrentium Perrier
?
0
19 F. polyphlebia Baker
17 Ceratosoien longimuc:ro Wiebes
3
20 F. botryoides Baker
16 Ceratosolen blommersi Wiebes
5
21 F. trichociada Baker
?
0
22 F. karthaiensis C.C.Berg
?
0
25 F. assimilis Baker
?
0
26 F. ampana C.C.Berg
?
0
30 F. madagascariensis C.C.Berg
?
0
32 F. menabeensis Perrier
23 Piatyscapa bergi Wiebes
33 F. humbertii C.C.Berg
2
- sp. indet.
36 F. Iutea Vah!
6
28 A11otriozoon heterandromorphum
3
Grandi
50 F. trichopoda Baker
32 Elisabethiella bergi Wiebes
51 F. grevei Bail!.
11
- sp. indet.
52 F. rubra Vahl
46 Nigeriella avicola Wiebes
53 F. marmorata Baker
?
64 F. antandronarum subsp. bemardii C.C.Berg
0
4
- Nigeriella sp. indet.
54 F. bivalvata Perrier
3
0
- EJisabethieJia sp. indet.
3
65 F. refJexa subsp. refJexa Thunb.
41 Elisabethiella refJexa Wiebes
2
98 F. poJita Vahl
70 CouItella bekiliensis (Risbee)
0
•
g __ i.
i
*
33
•
MATERIALS AND .METHODS
Male phase figs were collected from trees belonging to 22 of the 23 Ficus species recorded in the
southern Africa floristic region (van Greuning, 1990), and 15 of the 26 species recorded for Madagascar
and The Comores (Berg, 1986). The figs were placed in plastic containers closed with fine netting and,
once they had emerged, the wasps were either stored dry with silica gel or in alcohol.
Voucher
specimens of representative trees are lodged at the following herbaria: BG, GRA, K, MO, P, RUH,
TAN. In general, the taxonomy of the figs follows Berg (1989) and that of the wasps Wiebes and
Compton (1990) and Boucek (1988).
RESULTS AND DISCUSSION
Agaoninae were collected from all of the Ficus species sampled (Table 1 and Table 2). When these
results are added to previously published records, pollinators have now been collected from more than
70% of 105 Ficus species known from Africa, Madagascar and The Comores (Berg, 1989).
Typically, each Ficus species is associated with a single species of Agaoninae. However, of the Ficus
collections we have examined, four species from southern Africa (F. sycomorus L., F. cordata Thunb.,
F. lutea Vah!, F. nataiensis Hochst.), and F. sycomorus from Madagascar and The Comores were found
to host more than one pollinator species (Table 1 and Table 2).
These examples of Ficus species where host/pollinator specificity appears to break down are discussed
individually below.
F. sycomorus
The distribution of F. sycomorus and its pollinator Ceratasalen arabicus Mayr within southern Africa
are shown in Figure 1. Except in the western part of the sub-continent, trees of southern African F.
sycomorus, in common with their East African counterparts (Galil and Eisikowitch, 1968), were also
34
host to C. galili Wiebes. This species acts as a 'cuckoo', which, although having fully-formed pollen
carrying apparatus and utilising the ovules for larval development, does not pollinate the figs and hence
has no influence on its host's gene flow (Galil and Eisikowitch, 1968; Compton et al., 1991).
Ceratosolell galili is not a sister species to C. arabicus (Wiebes, 1989), suggesting that it or its ancestors
colonised F. sycomorus from another Ficus species, rather than having evolved ill situ from C. arabicus.
14°
20""
32'"
".
20'"
00
o
fJ .,
00
o
0
00 0
o
o
..o
go
o
•
.'
.'
~".'
...... .:r>~·6
o
0 0
'0
o
80
o
0
.....
.....
........
o
:
..
,.. -."
1................... ;-.
.. :
o •
o
.
"
'.
:
o
•
o. :.
.. 0
o
.....,
)" '1.
20'"
?
.:: ......................)
0
0,
: ••••••••
.....
32'"
o F. sycomorus
.. C. arabicus
26'"
32°
Figure 1. Our distribution records of southern African F. syconnus and associated pollinator, C. arabicus (.); other southern
African distribution records of F. s),comoms (0) are from van Greuning (1990) and von Breitenbach (1986) and are without
pollinator records.
Ficus sycomorus also occurs in Madagascar and The Comores. Two forms of F. sycomorus have been
recognised in Madagascar, both originally described as distinct species; namely the small-fruited F.
cocculifolia Baker (1886) and the large-fruited F. sakalavarum Baker (1886).
Ficus sakalavarum was
later reduced to a variety (Perrier de la Bathie, 1928) and then to a subspecies (Perrier de la Bathie,
1952) of F.' cocculifolia. It was only much later that Berg (1986) equated both taxa with the African F.
sycomorus, and included them both in his concept of this species. Berg (1986), however, suggests that
35
the "sakalavarum" form may represent a distinct subspecies within F. sycomorus. The taxonomy of F.
sycomorus in Madagascar therefore remains problematic.
Three species of Agaoninae, namely C. arabicus, C. gaZili and C. namorakensis (Risbec), were found
associated with the figs of F. sycomorus sensu Berg (1986) in Madagascar (Figure 2 & 3). However,
the latter wasp species was only found in the "sakalavarum" form and only C. arabicus and/or C. galili
were found in the "cocculifolia" form. The distribution of the two forms overlap in Madagascar (Figure
2 & 3) indicating that geographic factors alone are not responsible for the restriction of C. namorakensis
to the figs of "sakalavarum". The non-pollinating fig wasp faunas of the two forms are also distinct
(Ulenberg, 1985; Compton, unpublished), suggesting that fig wasps as a whole distinguish between the
two forms of F. sycomorus. The differing preferences shown by the two pollinators, C. Ilamorakellsis
and C. arabicus, are even more significant, because they indicate that the two forms are reproductively
isolated. even in areas where they are sympatric.
50=
of
('l
c
/~
10=
)
0
•
•
•
1
•
o
\
IJ
l~}
••o
OF. saka/cvarum
o F. sycomorus
C. arabicus
~
• C. namorakensis
•
Figure 2. Our distribution r~cods
SYCOll1011lS
o
of Malagasy and Cornoran F.
together with their associated pollinators; other
Malagasy and Comoran F. s),comoJ71s records {o) are from Perrier
Figure 3. Our distribution records of F. sakalaval1ll1l (IJ) together
with their associated pollinators; other F. sakalaval'lllll records ~O)
are from Perrier de la Bathie (1928, 1952) and are without
pollinator records.
de la Bathie (l928. 1952) and are without pollinator records.
36
In the field we had no difficulty in differentiating between the "cocculifolia" and "sakalavarum" forms
of F. sycomorus, provided they were bearing male phase or postfloral phase figs. The male phase figs
of F. sycomorus measured up to about 20 nun in diameter, while the "sakalavarum" form
were
considerably larger (100-150 rom in diameter) with a much thicker syconium wall. A single
"sakalavarum" fig measuring only 40 nun was sampled, which produced a solitary female pollinator
(whereas hundreds are usually present). This fig was clearly abnormal. A more distinctive difference
between the two forms concerned the postfloral phase figs (when they are ready for dispersal). During
this development phase the figs of "sakalavarum" change colour only slightly, changing from green to
a somewhat yellowish green, they are glabrous and slightly soft, but they never become juicy.
In
contrast, like African F. sycomorus, the figs of the "cocculifolia" form change from green to yellow or
red, they usually remain somewhat pubescent and they become soft and juicy. There also appears to be
differences in the fruiting phenologies of the two forms. On individual trees of the "sakalavarum" form
few figs matured at anyone time, while figs of the "cocculifolia" form developed synchronously, like
those of African F. sycomorus.
The morphological and developmental differences between the two forms may reflect different dispersal
systems. Several putative avian dispersers were recorded eating postfloral phase "cocculifolia" figs in
Madagascar, while no birds were observed eating "sakalavarum" figs (Ross pers. corom.), nor did we
record any avian-associated fruit damage. The ripe figs of F. sycomorus in Africa and the "cocculifolia"
form in Madagascar are reported by Perrier de la Bathie (1952) and Palmer & Pitman (1972) as favourite
food of the closely related African and Madagascar Green Pigeons (Treron calva (Temminck) and T.
australis (L.». In mainland Africa and in Madagascar the fruit of these plants has also been recorded as
being eaten by humans and other mammals (Perrier de la Bathie, 1952; Palmer & Pitman, 1972), while
the "sakalavarum" form is reported to be inedible or even poisonous to humans (Perrier de la Bathie,
1952).
Zebu cattle, which readily feed on fallen figs, appear to be the main potential dispersers of
"sakalavarum" figs at the present time. Since zebu cattle are not indigenous in Madagascar, the figs may
originally have been dispersed by the giant lemurs or the ostrich-like Aepyornis that occurred on the
island, all of which are now extinct.
37
Unfortunately the two forms are not always easy to distinquishin the herbarium (Berg, 1986). Vegetative
differences have not been detected and collections of "sakalavarum" may have immature fruits
approximately the same size as more mature fruits of F. sycomorus. Unless the developmental phase of
these specimens has been determined or other relevant information is known, positive identification may
not be possible.
On The Comores, only the "cocculifolia" form of F. sycomorus has been recorded, and this occurs on
the islands of Anjouan, Mayotte and Grande Comore (Table 2; Perrier de la Bathie, 1952; Compton,
1992). Only C. arabicus and C. galili were found associated with these plants, and these plants appear
to be morphologically indistinguishable from African F. sycomorus.
It seems clear that the small-fruited plants in Madagascar and The Comores are conspecific with African
F. sycomorus. The large-fruited plants represent a related, but distinct, species, endemic to Madagascar,
to which the name F. sakalavarum should be applied. This species is pollinated by the Malagasy-endemic
wasp C. namorakensis. Differences between the two species are summarised in Table 3.
Table 3. Distinguishing characters between F. sycomorus and F. saJrolavarum .
'ieee
. . .*H
F. saJrolavarum
F. sycomorus
f444
AM
fie
3M
&
1. Ripe fig diameter (mm)
15-20
(40)-100-150
2. Ripe fig wall thickness (mm)
1-2
>5
3. Ripe fig colour
yellow or red
yellow-green
4. RiPe fig texture
juicy
dry
5. Fig maturation
synchronous
asynchronous
6. Pollinator wasp
C. arabicus
C. namorakensis
7. "Cuckoo" wasp
C. galili
none recorded
8. Possible seed dispersers
birds, man and various
other mammals
cattle
9. Natural distribution
Africa (widespread),
The Comores, western
Madagascar
western and
southern
Madagascar
¥
38
,.
rm
F. cordata
Two subspecies of F. cordata (subsp. cordata and subsp. salicifolia (Yahl) C.C. Berg) are recorded from
the South African/Namibian region (Berg, 1989;van Greuning, 1990), and a third (subsp. lecardii CWarb.)
C.C. Berg) is known from West Africa (Berg, 1989). From the distribution maps produced by van
Greuning (1990) and von Breitenbach (1986) and our own records it is evident that the two southern
African subspecies are allopatric (Figure 4). The two subspecies are morphologically distinguishable (Berg
& Wiebes, 1992) and they are consistently pollinated by different species of fig wasps (Platyscapa
desertorum Compton and P. awekei Wiebes, see Table 1). They also have distinct non-pollinator fig wasp
faunas (Compton, unpublished).
26'"
14°
.............
0
";.
0
00
... _.... ::.:.. + .................
0
00
o 0
0
00
,.
.....
0
0
0
+
.'
.' +
0
00
\.........,
( ......
00
26'"
".
...... \ ....
0
0
000 0 0 00 0
00
20°
:'+~
+
+
+++
+
+
.~
+
:+ ++ +
;++++++
..................
+r-t~
+
0
~
0
00
o
.~
~'\",
... ...... -.;..
i
••
0
•
•
00
00
oo
o
0
o
o
0
0ba. .....-o....... ~ ...' •.·· - 0
000
o
0
o
o
o
0
o
+
......
0
o
'.
0
'.
......
'
.. :
o
subsp. cordata
• P. desertorum
+
subsp. salicifolia
"" P. awekei
o
20°
26'"
Figure 4. Our distribution records of southern African F. cordaw subsp. cordaw together with their associated pollinators (.) and
F. cO/'daw subsp. salicifolia (*); other F. cordala subsp. cordaw (0) and F. cordara subsp. salicifolia records
Greuning (J 990) and von Breitenbach (1986) and are without pollinator records.
39
~)
are from van
Since they have different pollinators and are geographically separated, F. cordata subsp. cordata and F.
cordata subsp. salicifolia are lik,-iy to be genetically isolated. A strong case could therefore be made for
the reinstatement of F. salicifolia at the species level (see Paulus and Gack, 1990 for the treatment of a
similar situation in the Orchidaceae). However, there is theoretically no reason why a Ficus species should
not attract different pollinators in different parts of its range (Compton et aI., in press). Therefore, in
order to assess their taxonomic status, it would be important to know whether the attractant chemicals
produced by the figs of two forms differ. The third subspecies creates a further complication. Ficus
cordata subsp. lecardii is somewhat intermediate between the two southern African forms of F. cordata,
and its pollinator is unknown (C.C. Berg, pers. comm.). Clearly more work is required on this species
before the significance of its two pollinators can be assessed fully.
F. natalensis
Species within the widespread African Ficus "thonningii I natalensis" complex are taxonomically
problematic throughout their range (Berg, 1989; Dowsett-Lemaire and White, 1990). Ficus thonningii is
highly variable in appearance, and Berg and Wiebes (1992) have informally recognised 10 different forms.
Ficus natalensis is more homogeneous, the typical subspecies occurs in southern Africa and further north,
while a second (subsp. leprieurii(Miq.) C.C. Berg) occurs only in tropical Africa. Ficus thollliingii and F.
Jlaralensis are closely related and frequently confused (Berg & Wiebes, 1992).
On the basis of their pollinators three partially sympatric southern African forms within the "thonningii
I natalensis" complex can be recognised (Figure 5). In South Africa Ficus thonllingii is consistantly
pollinated by a single pollinating species, Elisabethiella stuckellbergi (Grandi), although further north (i.e.
in Zimbabwe) it has been recorded as host to Alfollsiella longiscapa Joseph and A. broJlgersmai Wiebes
(Boucek et ai., 1981). Southern African F. natalellsis subsp. natalensis has been recorded as the host of
three pollinating species, namely E. sruckenbergi, E. socotrellsis Mayr and A. longiscapa.
The species of non-pollinating fig wasps reared from F. llatalellSis subsp. llalaiellSis trees pollinated by
E. socotrellSis appear to be the same as those from figs pollinated by E. sruckenbergi, while those
40
associated with F. nalaiensis subsp. nataiellsis figs pollinated by A. longiscapa are distinct (S. Compton,
unpublished).
For example, Phagoblastus barbarus (Grandi) (Agaonidae, Sycoecinae) is found in figs
pollinated by both Elisabethiella species, but not those pollinated by A. longiscapa (van Noort, 1992). The
host preferences of the wasps therefore suggest there may be a 'cryptic' form of F. natale1lSis pollinated
only by A. longiscapa and distinct from both F. rholllliJlgii and the F. natalensis subsp. nataiensis pollinated
by species of Elisabethiella. Although we have only recorded it from a small number of trees in Natal
(Figure 5), A. IOJlgiscapa appears to be the normal pollinator of F. llataiensis subsp. natalellsis elsewhere
in Africa (Wiebes, 1988; Compton, unpublished). Yet another pollinator, Alfonsiellafimbriata Waterston
appears to be associated with F. natalellsis subsp. Zeprieurii{Berg & Wiebes, 1992).
,
.-:
\
-.....
.
.....•
";
..
-
......... ~: . ...........-.
._.................
",
-..... _...
:
(._... _:...
.....
. ..... 0°00 0
"0
0
o
o
..... /
o
II
o
• o.
o
0
....... ~.;
<> F. thonningii
10 F. nataiensis
• E. stuckenbergi
•
E. socotrensis
III
aF. nata/en;is
• A. longiscapa
Figure S. Our distribution records of southern African Ficus "thonningii I natalensis" together with their associated pollinators
(a); other Ficus "thonningii / natalensis" records (0) are from van Greuning (1990) and von Breitenbach (1986) and are without
pollinator records. F. Ihonningii (i) is only associated with one pollinator species while F. natalensis (ii and iii) is associated with
two different agaonines.
An additional complicating factor surrounds the taxonomic status of the Elisabethiella pollinators.
Differentiation between E. socotrensis and E. sruckenbergi is difficult and some South African specimens
are morphologically intermediate between the two species (Wiebes, pers. comm.). The problem is further
aggravated in that E. socotrellsis is associated with two completely distinct Ficus species, F. wakefieldii
Hutch., in Zambia and North-east African F. vasta Foissk. (Wiebes & Compton, 1990; Compton,
unpublished).
41
The pollination of some F. natalellsis subsp. natalensis by A. longiscapa in southern Africa suggests that
these trees may be closely allied to components of the "thonningii / natalensis" complex from tropical
Africa which share the same pollinator. Although these trees are sympatric with E. socotrellsis pollinated
F. llatalensis subsp. natalensis, if pollinator choice is consistent they must be reproductively isolated.
Given the close relationship or possibly conspecificity of southern African E. socotrensis and E.
stuckellbergi, together with identical non-pollinating fig wasp faunas, it seems reasonable to conclude that
the Elisabethiella pollinated trees of the "thonningii / natalensis" complex in southern Africa represent
components of a single variable species. Alternatively, if E. socotrensis is a good species, distinct from
E. stuckenbergi, this would suggest that in southern Africa E. socotrensis pollinated trees of F. natalensis
subsp. nalaiensis are reproductively isolated from F. thonningii. Both these hypotheses are consistent with
a species specific pollinator / host relationship.
Hybridization is a possible source of some of the observed morphological variability. Chromosome counts
of Ficlls spp. are mostly diploid (2n = 26) (Condit, 1933,1964; Ohri and Khoshoo, 1987), including counts
of some F. thonningii (Condit, 1964). However F. burkei (Miq.) Miq. and F. hochsetteri (Miq.) A. Rich.
which are now regarded as varieties of F. thonllingii (Berg, 1989), have been recorded as being tetraploid
(2n
=
56) (Condit, 1964). These tetraploids may be a result of interspecific hybridization and this
hypothesis may account for the diversity of "thonningii / natalensis" forms and the resultant species
delimitation difficulties experienced by taxonomists (Berg, 1990; Ramcharun et ai., 1990).
Further
progress in delimitating species within the "thonningii / natalensis" complex may require a combination
of karyological and modem molecular approaches such as DNA restriction techniques.
F. lutea
F. lurea is widely distributed in Africa, Madagascar and The Comores (Berg, 1990; Compton, 1992). In
South Africa its natural distribution is restricted to the more humid forests of Natal (van Greuning, 1990),
but it is planted as an ornamental tree elsewhere.
Within its natural range, F. iutea appears to be
42
pollinated exclusively by Allotriozooll heterandromorphum Grandi (Compton, unpublished). A tree planted
in Grahamstown, some 500 km outside the tree's normal distribution range was pollinated by A.
heterandromorphum and by small numbers of both Elisabethiella stuckenbergi and Ceratosolen capensis
Grandi (species that normally pollinate F. thonllingii and F. sur respectfully). As reported elsewhere
(Ramirez, 1988), this suggests that the normal host tree specificity exhibited by Agaoninae can breakdown
under conditions where a tree's pollinator is rare or absent.
Hybrid plants involving the edible fig F. carica have been artificially produced (Condit, 1950) and naturally
occurring hybridization has been reported (Ramirez, 1988). We have successfully germinated seed from
F. lutea which was naturally pollinated by E. stuckenbergi and C. capellsis, but have not been able to coax
these hybrids past the cotyledon stage of their development (Ware and Compton, 1992). The hybrid
weakness/inviability shown by these crosses may be widespread and effectively act as post germination
isolating mechanisms within Ficus.
CONCLUSIONS
An examination of cases of Ficus species for which more than one pollinating wasp species has been
recorded has served to highlight numerous gaps in our understanding and knowledge of fig and fig wasp
biology. In the case of F. iycomorus, work on the wasps combined with field work on the trees has helped
to redefine species limits. In the other cases further work needs to be done. The situation in the Ficus
"thonningii I natalensis" complex is of particular interest with respect to the role that hybridisation and
polyploidy may play in Ficus evolution.
ACKNOWLEDGEl\1ENTS
The Director of Tsimbazaza
Botanical
and Zoological Park (Antananarivo,
Madagascar),
Dr
Randrianzafy, Albert, is thanked for his assistance and in allowing one of his staff members to accompany
us on our travels in Madagascar.
Dr Randrianasolo, Evariste, provided us with much needed local
knowledge and language skills, without which the expedition would not have been nearly as successful.
43
The Missouri Botanical Garden staff generously provided us with facilities, equipment and advice. The
South African National Parks Board, the staff of Kruger National Park, Hans Merensky Nature Reserve,
Natal Parks Board, Kwa Zulu Bureau of Natural Resources and the Malagasy Directorate of Water and
Forests are acknowledged for their roles in allowing us to collect specimens in areas under their
jurisdiction. Finally we need to thank Kathy Holton, Simon Ma1comber, Claire Hemingway, Simon van
Noort, Costas Zachariades, Sally Ross amongst others for their assistance in the field, often under trying
conditions. The financial support by Rhodes University, Foundation for Research Development and the
Anglo American Chainnan's Fund was greatly appreciated.
REFERENCES
Baijnath, H. and Ramcharun, S. (1983). Aspects of pollination and floral development in Ficus capensis
Thunb. (Moraceae).
Bothalia 14: 883-888.
Baker, H.G. (1961). Ficus and Blastophaga. Evolution 15: 378-379.
Baker, J. (1886). Further contributions to the flora of Madagascar. 1. Linn. Soc., Bot. 22: 441-537.
Berg, C.C. (1986). The Ficus species (Moraceae) of Madagascar and the Cornaro Islands. Bull. Mus.
Natn. Hist. Paris,Ser. 4 (sect. B Adansonia 1) 8: 17-55.
Berg, C.C. (1989). Classification and distribution of Ficus. Experiemia 45: 605-611.
Berg, C.C. (1990). Annotated check-list of the Ficus species of the African Floristic region, with special
reference and a key to the taxa of southern Africa. Kirkia 13: 253-291.
Berg, C.C. and Wiebes, J.T. (1992). African Fig Trees and Fig Wasps. North-Holland.
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, England.
Boucek, Z., Watsham, A. and Wiebes, J.T. (1981). The fig wasp fauna of the receptacles of Ficus
thOll11ingii (Hymenoptera, Chalcidoidea). Tijd. Em. 124: 149-233.
Bronstein, J.L. (1989). A mutualism at the edge of its range. Experiemia 45: 622-637.
Bronstein, J.L. and McKey, D. (1989). The fig/pollinator mutualism: a model system for comparative
biology.
Experientia 45: 601-604.
Compton, S.G. (1990). A collapse of host specificity in some African fig wasps. Sth Afr. 1. Sci. 86: 39-40.
44
Compton, S.G. (1992). Moraceae. New records of Ficus and their pollinators on Grand Comores.
Bothalia 22: 46-47.
Compton, S.G., Holton, K.C., Rashbrook, V.K., van Noort, S., Vincent, S.L. and Ware, A.B. (1991).
Studies of Ceratosolen galili,a non-pollinating agaonid fig wasp. Biotropica 23: 188-194.
Condit, LJ. (1933). Chromosome number and morphology in thirty-one species. Univ. Calf. Publ. Bot. 17:
61-74.
Condit, LJ. (1950). An interspecific hybrid in Ficus. l. Hered. 41: 165-168.
Condit, LJ. (1964). Cytological studies in the genus Ficus. III. Chromosome numbers in sixty-two species.
A4adrono 17: 153-155.
Comer, E.J.H. (1965). Check-list of Ficus in Asia and Australasia with keys to identification. Gdn's Bull.
Singapore 21: 1-186.
Dowsett-Lamaire, F. and White, F. (1990). New and noteworthy plants from the evergreen forests of
Malawi. Bull. lard. Bot. Belg. 60: 73-110.
Galil, J. (1977). Fig Biology. Endeavour 1: 52-56.
Galil, J. and Esikowitch, D. (1968). On the pollination ecology of Ficus sycomorus in East Africa. Ecology
49: 259-269.
Hill, D.S. (1967). Figs of Hong Kong. Hong Kong University Press, Hong Kong.
Janzen, D.H. (1979). How to be a fig. Ann. Rev. Ecol. Syst. 10: 13-51.
Michaloud, G., Michaloud:'Pelletier, S., Wiebes, J.T. and Berg, C.C. (1985). The co-occurrence of two
pollinating species of fig wasp and one species of fig. Proc. K. Ned. Akad. Wet. Ser. C. 88: 93-119.
Newton, L.E. and Lorna, A. (1979). The pollination of Ficus vogelii in Ghana. Bot. l. Linn. Soc.
78: 21-30.
Ohri, D. and Khoshoo, T.N. (1987). Nuclear DNA contents in the genus Ficus (Moraceae). Pl. Syst. Evol.
156: 1-4.
Palmer, E. and Pitman, N. (1972). Trees of Southern Africa. A.A.,Balkema, Cape Town.
Paulus, H.F. and Gack, C. (1990). Pollinators as prepollinating isolation: Evolution and speciation
Ophrys (Orchidaceae). Israel l. Bot. 39: 43-79.
Perrier de la Bathie, H. (1928). Les Ficus de Madagascar. Arch. Bot. Bull. Mens. 2(8/9): 137-180.
45
111
Perner de la BiHhie, H. (1952). Ficus. In Perner de la Bathie, H. and Leandri, J. Flore de Madagascar,
55' Famille. Moracees.
Ramcharun, S., Baijnath, H. and van Greuning, J. V. (1990). Some aspects of the reproductive biology of
the Ficus llatalellsis complex in southern Africa. Mitt. inst. AUg. Bot. Hamburg 23: 451-455.
Ramirez B., W. (1970). Host specificity of fig wasps (Agaonidae). Evolution 24: 680-691.
Ramirez B., W. (1974). Coevolution of Ficus and Agaonidae. Ann. Missouri Bot. Gard. 61: 770-780.
Ramirez B., W. (1988). Ficus microcarpa L.,F. benjamilla L. and other species introduced in the New
World, their pollinators (Agaonidae) and other fig wasps. Rev. BioI. Trop. 36: 441-446.
Thompson, J.N. (1982). interaction alld Coevolutioll. John Wiley and Sons, New York.
Thompson, J.N. (1989). Concepts of coevolution. Trends Ecol. Evo!. 4: 179-183.
Ulenberg, S. (1985). The phylogeny of the genus Apocrypta Coqueral in relation to its hosts, Ceratosolell
Mayr (Agaonidae) and Ficus L. Verhand. K. Akad. Wet. 83: 149-176.
van Greuning, J. V. (1990). A synopsis of the genus Ficus (Moraceae) in southern Africa. S. Afr. 1. Bot.
56: 599-630.
van Noort, S. (1992). The Systematics and Phylogenetics of the Sycoecinae (Agaonidae, Chalcidoidea,
Hymenoptera). Unpublished PhD thesis. Rhodes University, Grahamstown.
van Noort, S., Ware, A.B. and Compton, S.G. (1989). Pollinator-specific volatile attractants released from
figs of Ficus burtt-da,yi. Sth. Afr. 1. Sci. 85: 323-324.
Verkerke, W. (1989). St~cure
and function of the fig. Experientia 45: 612-621.
von Breitenbach, F. (1986). National List of indigenous Trees. Dendrological Foundation, Promedia
Pub., Pretoria.
Ware, A.B. and Compton, S.G. (1992). Breakdown of pollinator specificity
III
an African fig tree.
Biotropica 24: 544-549.
Ware, A.B.,Kaye, P.T.,Compton, S.G. and van Noort, S. in press. Fig volatiles: their role in attracting
pollinators and maintaining pollinator specificity. PI. Syst. Evol.
Wharton, R.A., Tilson, J.W. and Tilson, R.L. (1980). Asynchrony in a wild population of Ficus sycomorus
L. Sth. Afr. 1. Sci. 76: 478-480.
Wiebes, J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev. Ecol. Syst. 10: 1-12.
46
Wiebes, J.T. (1982). The phylogeny of the agaonidae (Hymenoptera, ChaIcidoidea). Neth. J. Zoo. 32:
395-411.
Wiebes, J.T. (1988). Agaonidae (Hymenoptera ChaIcidoidea) and Ficus (Moraceae): fig wasps and their
figs, II (Alfonsiella). Proc. K. Ned. Akad. Wet. C 91: 429-436.
Wiebes, J.T. (1989). Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and their
figs, IV (African Ceratosolen). Proc. K. Ned. Akad. Wet. C 92: 251-266.
Wiebes, J.T. (1990). African figs and their pollinators - a brief overview. Mitt. Inst. AUg. Bot. Hamburg
23: 425-426.
Wiebes, J.T. and Compton, S.G. (1990). Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae):
fig wasps and their
figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet. 93: 203-222.
Windsor, D.M., Morrison, D.W., Estribi, M.A. and de Leon, B. (1989). Phenology of fruit and leaf
production by 'strangler' figs on Barro Colorado Island, Panama. Experientia 45: 247-253.
47
CHAPTER 3
BIOLOGICAL EVIDENCE FOR VOLATILE ATTRACTANTS
Paper 3: Pollinator-specific volatile attractants released from the figs of Ficus burtt-davyl. South African
Journal of Science 85; 323-324. (S. van Noort, A.B. Ware and S.G. Compton - 1989).
Paper 4: Fig wasp responses to host plant volatiles. Submitted to Journal of Insect Behaviour. (A.B. Ware
and S.G. Compton).
48
REPRINTED FROM
Mei 1989
Vol. 85
Suid-A/rikaanse Tydskrij vir Welenskap
323
Research Letters/N avorsingsberigte
Pollinator-specific volatile attractants
released from the figs of Ficus burtt-davyi ~
0.4
n.
«
a:
There are about 750 species of fig trees (Ficus spp., Moraceae),
all of which are pollinated by tiny fig wasps of the family
Agaonidae. 1 With rare exceptions, each species of fig tree is
pollinated by a single species of fig wasp, which is only found
in association with that one kind of tree. 2 After completing their
development inside mature figs, adult female wasps fly off in
search of 'receptive' immature figs. These are normally on other
trees because the figs on anyone tree are typically all at the same
stage of development.3.4 Once the female finds a receptive fig
she enters it via the ostiole, pollinates the flowers, lays her eggs
and dies. It has generally been assumed,5-7 supported by some
circumstantial evidence,8 that trees bearing receptive figs release
chemicals which attract fig wasps to them. Here we provide
experimental confirmation of the release of pollinator-specific
attractant volatiles from the figs of F. burtt-davyi Hutch: and
show that the volatiles emanate from the ostioles of the figs.
I-
n.
<Jl
«
3:
u..
o
o
i=
a:
on.
F. rhonningii
1
83.5 :!: 17.7
50.2 :!: 15.3
600.5 :!: 178.5
53.0 ± 11.8
92.7 :!: 21.0
80\5.9 :!: 258.5
46.7 ± 15.0
29.0 ± 3.9
71.8 :!: 37.8
2
3
4
5
6
F. sur
c::
n.
UNPOLLINATED
F. BURTT-DAVY1
IS.7:t
Controls
2.8
17.7:t 2.4
56.8 :t 28.7
29.8
31.0
60.3
14.2
19.5
56.7
0.1
o
Table 1. Mean (:!: s.e.) numbers of Elisaberhiella baijnarhi collected on
sticky traps placed next to cotton bags containing 'receptive' figs or control
(empty) bags.
F. burl1-davyi
0.2
z
Materials and methods
Unpollinated figs were collected from F. burtt-davyi, F. thanningii BI. and F. sur Forsk. trees growing in Grahamstown. To
ensure that the figs were receptive, they had been sealed inside
cotton bags until the time when the other figs on the trees had
been pollinated. The attractiveness of the figs was tested in July
1988 using Elisabethiella baijnathi Wiebes wasps emerging from
a single F. burtt-davyi growing in the 1820 Settlers Gardens in
Grahamstown. During the experiments the figs were kept in white
cotton bags which prevented visual attraction but allowed the diffusion of volatiles. The bags were suspended as 1.2 m above the
ground on 18 black wooden poles placed in a circle about 5 m
away from trye tree. The attractiveness of each bag was monitored
using an adjacent sticky trap, which consisted of a clear plastic
cylinder (diameter 5 cm, surface area 200 CQ2) sprayed with
pruning sealant. Empty bags and their associated sticky traps
acted as controls.
In the first experiment (days 1 - 3), the bags contained either
10 unpollinated F. burtt-davyi figs (A), 10 un pollinated F. thanningii figs (B), or were empty (C). The bags and associated traps
were alternated around the tree (ABCABC etc.) and replaced
every 24 hours. The experiment was then repeated on days 4 - 6,
using unpollinated F. sur figs in place of those of F. thonningii.
A third experiment determined the site where attractants were
released from the figs of F. burtt-davyi. Thirty bags and their
associated sticky traps were placed in a 10-m radius around the
tree and left for four hours. Ten bags were empty, a further 10
bags each contained 50 unpollinated figs with their ostioles sealed
by painting beeswax over the opening, and the remaining bags
contained 50 unpollinated figs that had been painted basally with
Day
0.3
<Jl
5.5
± 5.0
:t 15.1
2.7
:t 3.4
:: 18.0
:!:
=
49
UNPOLLINATED
F. THONNINGII
CONTROL
Fig. I. Elisabethiella baijnalhi trapped next to control (empty) cotton bags and bags containing receptive figs of F. burtt-davyi or
F. Ihonningii. More wasps were attracted to the figs of F. burttdavyi than to those of F. Ihonningii or controls (/{341 = 3.96,
P<O.OOI and 1{341 = 3.65, P<O.OI, respectively). Equal numbers
of wasps were trapped near control bags ancj those containing F.
.
rhonningii figs (tP'1 = 0.41, P>O.5).
beeswax. The last acted as controls for any possible attractive
properties of the beeswax.
Results
There was considerable variation in the quantity of fig wasps
trapped on different days, reflecting difer~ncs
in the numbers
of wasps emerging from the tree (Table 1). Because of this, the
data from experiments 1 and 2 were standardised by converting
the number of wasps on each trap to a proportion of the total
wasps collected that day and arc sine transformed for statistical
analysis. Significantly more fig wasps were attracted to the figs
of F. burtt-davyi than those of F. thonningii or to control bags,
but there was no difference in the numbers of wasps at the control
and F. thanningii bags (Fig. 1). E. baijnathi females were therefore attracted to the figs of the tree they pollinate, but not to
the figs of F. thanningii. Similar results were obtained in experiment 2, where significantly more wasps were attracted to the figs
of F. burtt-davyi than to those of F. sur or controls (Fig. 2). This
showed again that E. baijnathi was only attracted to the figs of
its host tree.
F. burtt-davyi figs with their ostioles covered with wax were
no more attractive than control bags (mean wasps per trap =
27.3 and 12.3, Fig. 3). In contrast, figs with basal wax remained
highly attractive (mean wasps per trap = 148.4).
Discussion
J ermy et al. 9 have emphasized the advantages of field studies
of olfaction over those which are carried out in the laboratory.
Here we have shown that under natural conditions the pollinator
of F. burtt-davyi is attracted to the smell of its receptive unpollinated figs, but is not attracted to the un pollinated figs of
two other species. The wasps were not attracted if the ostioles
of the figs were covered, showing that the source of attraction
came from within the figs.
Further experiments have shown that polina~ed
F. bUrll-davyi
cease to be attractive to the pollinator, and pre'liminary GC-MS
analysis has revealed at least one volatile compound which is
released prior to pollination, but not subsequently.1O Identifica-
324
c
South African Journal of Science
S.
0.4
w
6.
7.
8.
Co
"-<
a:
....
t/)
0.3
"-<
9.
t/)
~
u.
0
Z
10.
0.2
Q
....
a:
0
"0
0.1
g;
UNPOLLINATED
F. BURTT-DAVYI
UNPOLLINATED
F. SUR
CONTROL
Fig. 2. Elisabelhiella baijnalhi trapped next to controi (empty) cotton bags and bags containing receptive figs of F. bum-davyi or
F. sur. More wasps were trapped near F. bum-davyi than F. sur
(tIl'l = 6.89, P<O.OOI) or controls (tIl'l = 6.97, P<O.OOI). There
was no difference in the numbers of wasps trapped at F. sur figs
and controls (l1141 = 0.054, P>O.5).
400
~
"-
....
300
0:
UI
Q.
t/)
Co
f/)
200
<
~
100
OSTIOLE
PPEN
OSTIOLE
SEALED
CONTROL
Fig. 3. Elisabethiella baijnalhi trapped next to control (empty) cotton bags and bags containing receptive figs of F. burtt-davyi with
beeswax sealing their ostioles or applied at their bases. Figs with
their ostioles covered did not attract wasps and when compared
with controls (1/"1 = 1.17, P>O.I). In contrast, figs with basal
wax remained highly attractive (1/181 = 3.37, P<O.OI).
tion and synthesis of this compound is proceeding, in preparation for its bioassay.
We thank Mr J. Cameron for permission to work in the Settlers
Botanical Gardens, and Professor H.R. Hepburn for comments
on the manuscript.
S. VAN NOORT, A.B. WARE and S.G. COMPTON
Department of Zoology and Entomology,
Rhodes University,
Grahamstown, 6140 South Africa.
Received 22 February; accepted II April 1989.
I. Wiebes J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev.
£col. Sy5l. 10, 1-12.
2. ),lichaloud G., ),fichaloud·Pelietier 5., Wiebes J.T. and Berg C.C. (1985). The
CO·OCCUrT(nCe of IwO pollinating species of. fig wasp and one species of fig.
Proc. K. Ned. Akad. Wet. (C) 38. 93 - 119.
3. Baijnath H. and Ramcharun S. (1988). Reproductive biology and chalcid symbiosis in FicJts burll-davyi (Moraceae). Monogr. SYSI. BOl. Misssouri BOl. Gard.
25,227-235.
4. Wharlon R.A .• Tilson J.W. and Tilson R.L. (1980). Asynchrony in a wild
population of Ficus ,yeoman,s L. S. Afr. J. Sci. 76. 478 - 480.
50
Vol. 85
May 1989
I
Hill D.S. (1967). Fig! (Ficus sPP.) and fig wasps (Chalcidoidea). J. not. Hist.
1,413-434.
Galil J. (1977). Fig biology. Endeavour 1, 52 - 56.
Janzen D.H. (1979). How 10 be a fig. Ann. Rev. Ecol. 5yst. 10, 13 - 51.
in a neotropical figBronstein J.L. (1987). Maintenance of speci-~;nty
pollinator wasp mutualism. Oikos 48, 39 - 46.
Jermy T., Szentes; A. and Horvath J. (1988). Host plant finding in
phytophagous insects: the case of the Colorado potato beetle. Entomol. expo
appl. 49, 83-98.
Ware A.B., van Noon S., Kaye P. and Compton S.G. (in preparation).
FIG WASP RESPONSES TO HOST PLANT VOLATILES
A. B. Ware and S. G. Compton
ABSTRACT
Fig trees (Ficus spp. Moraceae) are pollinated by fig wasps belonging to the family Agaonidae. Each tree
species is usually pollinated by a single species of wasp. Previous experiments have shown that the wasps
are attracted to the trees by volatiles emanating from the figs. Using fig-bearing trees and arrays of sticky .
traps baited with figs, we investigated the specificity of wasp attraction and its timing. The pollinatol's of .
two closely related Ficus species are specifically attracted to figs of their host at the time when figs are
ready to be pollinated. Some non-pollinating fig wasps appear to use the same cues.
INTRODUCTION
Fig wasps (Chalcidoidea, Agaonidae) are intimately associated with fig trees (Ficus spp.,
Moraceae)(Boucek, 1988). Each of the 750 or so Ficus species (Berg, 1988) is generally pollinated by
a specific species of pollinating wasp belonging to the subfamily Agaoninae (Wiebes, 1979; Wiebes and
Compton, 1990). The fig trees are totally dependent on the wasps for pollination and in return provide
sites for their larval development inside the fruits - the figs.
In addition to the pollinators there are also many species of non-pollinating fig wasps with larvae that
also develop inside the figs. These belong mainly to subfamilies of the Agaonidae other than Agaoninae,
but include representatives of other chalcid fanlllies (Boucek, 1988). Some of the species gall the fig
ovules while others parasitise the gall formers. A few non-pollinating wasp species are like the pollinators
and enter the lumen of the fig prior to oviposition (van Noort, 1992), but the majority reach the ovules
from the outside, penetrating the wall of the figs with their long ovipositors.
Although the host
relationships of most non-pollinating species are unknown, some of them are like the pollinating wasps
and are exclusively associated with a single Ficus species (Ulenberg, 1985; van Noort, 1992).
51
In most Ficus species the development of fig crops tends to be synchronised within anyone tree, but is
not synchronised between trees (Wharton et al., 1980; Bronstein and Patel, 1992; Bronstein, 1988, 1992;
Bronstein et ai., 1990). Adult females of pollinating fig wasps are short-lived, surviving at most a few days
(Kjellberg et ai., 1988), while the longevity of some female non-pollinating wasps can extend to one or
two months (Joseph, 1958; Compton et ai., in prep). The gaps between fig crops on each tree may be
months or even years (Bronstein, 1987; Windsor et a/., 1989). The combination of the within-tree fruiting
synchrony and the short life-spans of the wasps means that both the pollinating and the non-pollinating
female wasps must usually leave their natal trees in order to find figs that are suitable for oviposition
(Bronstein, 1987, 1992).
Van Noort et ai. (1989) showed that the pollinating wasp Elisabethiella baijnathi Wiebes located the figs
of its host tree, Ficus burtt-davyi Hutch., using volatiles released by the figs when they were ready to be
pollinated ('receptive' or 'female phase' figs: Galll, 1977). Figs at other stages of development were not
attractive to the pollinators, nor were figs which had their ostioles covered, suggesting that the attractants
emanated from within the figs during this short period of their development (van Noort et ai., 1989). The
responses of non-pollinators were not investigated, but those species which oviposit at the same stage of
fig development as the pollinators could potentially make use of the same volatiles, whereas wasps which
oviposit into figs at a later stage of development might be expected to utilise alternate cues.
Here we examine aspects of the specificity of the volatiles used by fig wasps to find their host trees.
Using arrays of sticky traps baited with figs of different developmental stages, we determined when wasps
are attracted to the figs of F. thonningii Bl. and compared the specificity of the volatile attractants
produced by this tree and F. burtt-davyi. We also experimentally prolonged the period when figs
remained attractive to their pollinators, in order to determine the length of time figs would 'wait' for their
pollinators.
52
MATERIALS AND METHODS
The study was conducted in the 1820 Settlers Botanical Gardens situated at Grahamstown, in the eastern
Cape Province of South Africa. Three local eastern Cape Ficus species grow in the gardens. Two, F.
burtt-davyi (some 110 individuals), F. thonningii (57 trees, some of which have been planted), are closely
related and are placed in the section Galoglychia of the subgenus Urostigma, while the third, F. sur
Forssk. (10 trees), belongs to the subgenus Sycomorus. The locations of most of the trees in the gardens
are indicated in Figure 1 and the fig wasps associated with these species locally are listed in Table 1.
Table 1. Indigenous Ficus spp. present in the Grahamstown Botanical Garden, together with the wasps normally found associated
with the trees in Grahamstown.
tt44F*4
Pollinator
Ficusspp.
G¥&W44i9S
F. thoIl1lingii BI.
$A iA1¥MiR#
ttt
**
&M
Mew H
*'tn-¥
EJisabethieJla stuckenbergi Grandi
EM'E'
eWMWW,
Non-pollinators
6
M
¥*¥¥#£i$af1AMYJ'ltf'¥H#4§!i@M§.'#h
OtiteseJla tsamvi Wiebes
PhagobJastus barbarus Grandi
Syro1}'Ctes sp. *
PhiJotrypesis sp:
F. burtt·davyi Hutch.
Elisabethiella baijnathi Wiebes
OtiteseJJa uJuzi Compton
OtiteseJla sesquianeIlata van Noort
SYC01}'Ctes sp."
PhiJotrypesis sp:
F. sur Forssk.
CeratosoJen capensis Grandi
Sycophaga cyclostigma Waterston
ApoCl}'pta guineensis Grandi
Apocxytophagus spp.
1M!
"The Philotrypesis and SYC01}'Ctes species recorded from F. thonningii and F. burtt-davyi cannot be distinguished at present, and
may not be host tree specific.
Sticky traps, each consisting of a cylinder (10 em radius; 30 em length) covered with cellulose and
sprayed with pruning sealant (Frank Fehr, Durban), were used to investigate the attraction of fig wasps
to figs at different stages of fig development. Poles, bearing the sticky traps placed at a height of 1.2 m,
were placed in a 3 X 3 array about 40 m from the nearest fig tree. Each pole was positioned 5 m from
53
its nearest neighbour. Twenty-five receptive phase F. thonningii figs were placed in each of three cotton
bags (treatment A) and 25 post pollinated figs in each of a further three bags (treatment B). The final
three empty bags acted as controls (treatment C). The bags were attached to the poles immediately above
the sticky traps and placed in position (orientated ABC:BCA:CAB) at 07hrOO. The sticky traps were
removed for analysis 6 hours later. The experiment was conducted twice in December 1989.
We then investigated how long unpollinated figs could potentially remain attractive to fig wasps. F.
burtt-davyi was chosen for these experiments because it is a smaller species than F. thonnillgii and all
its figs are within reach from the ground. We selected two F. burtt-davyi trees growing about 100 m
apart that were of comparable size and had produced approximately 5000 figs at the same stage of
development. Approximately half of the figs on one of the trees were surrounded by cotton bags during
their early pre-female phase. This prevented any pollination or oviposition by fig wasps. Single sticky
traps were then placed in each tree to monitor arrivals of fig wasps and were replaced weekly.
The specificity of the volatile attractants emanating from the figs of F. thonningii and F. burtt-davyi figs
was investigated in two field choice experiments. In the first experiment a 3 X 3 array of sticky traps
was used as before, but with the cotton bags containing either 25 receptive phase figs of F. thonllillgii
(three bags) or 25 receptive phase figs of F. burtt-davyi (three bags). The last three empty bags again
acted as controls. Two replicate trials were conducted in December 1989 and January 1990.
In a long term experiment monitoring the specificity of wasp attraction, the arrivals of wasps at F.
thonningii and F. burtt-davyi trees in the Botanical Gardens were monitored over a two year period.
Single sticky traps were placed in five trees of each species. In F. burtt-davyi the traps were positioned
between 0.5 and 1.5 m above ground level, while in the taller F. thOllllillgii they were placed at a height
of approximately 2 m. The traps were replaced weekly and the numbers and identity of the trapped fig
wasps were recorded. The relative positions of the trees that contained traps are indicated in Figure 1.
54
...
.............
"
................. ,
5
",. I
.-
""
""
"\
o
\\
\
1820 SETILER \
BOTANICAL:
GARDEN
:
~
8
93
,
,
I
J
I
I
I
I
J\
I J
I I
I
1
I
J
J
I
)
\
:
I
I
I
I
'.
I
I
I.
I
I
I
I
I
I
I
,
,
I
I
; "
I
IJ
\
-""
I
""
",
..... 1
•
+
o
•
I
/ I 0. 0
I
\I
J
--- -1-'1", __ .......
I
J
I)
I
J
I
.... --__
I
I
I
.........
•
~.4
:
I
~
"•
\
,
, ~'.
G
,/
tI
,
to. o
I '
, ;'
o.
40
.
m
Figme 1. Portion of the 1820 Settlers Botanical Garden (Grahamstown, South Africa) showing the relative positions of F.
thonningii (~),
F. burtt-davyi Co) and F. sur (A) trees. Additional exotic fig trees are represented by the open symbol (0). The
numbers indicate those trees used to monitor the arrivals and departures of the fig wasps.
55
RESULTS
F. thonningii is pollinated by E. stuckenbergi, and significantly more females of this species were recorded
from sticky traps placed near receptive phase F. thonningii figs than on traps near pollinated figs or the
control bags (Table 2). There was no difference between the number of E. stuckenbergi trapped on the
control sticky traps and those near the pollinated figs (Table 2). A similar preference for unpollinated
figs of F. thonningii was shown by the non-pollinating species, Phagoblastus barbaros, Philotrypesis sp.
and Otitesella spp. although too few examples of the latter species were trapped for statistical significance
to be recorded.
In the experiment that examined the duration of fig attractiveness, figs on the control F. burtt*davyi tree
were. rapidly pollinated by their pollinating wasp (E. baijnathi) and within about two weeks the wasps
ceased to be attracted to the tree (Figure 2). In contrast, large numbers of wasps continued to arrive
at the F. burtt-davyi tree with bagged figs for a period of five weeks (Figure 2). The figs therefore
remained attractive to their pollinating wasps for an extended period when pollination was prevented. Far
fewer wasps were collected on the control tree, presumably because they avoided the traps by entering
the figs.
3100
20001
'C
\I)
0.
0.
~
«l
300
3300
I
I
2400
3
4
5
II
250
CI)
0.
CI)
«l
~
-
200
0
'\I)
..0
E
:J
z
150
100
50
0
1
2
6
7
Weeks
Figure 2 The effect of bagging pre·receptive (= pre.female) figs (hatched bar) of R burtt-davyi on the numbers of pollinating
wasps, E baijnathi, trapped. The solid bars indicate the number of wasps trapped on a similar tree which remained unbagged.
56
Table 2. The fig wasps trapped near cotton bags containing either pollinated or unpollinated (receptive) F. rhollllillgii figs. The control bags
were empty. Combined results from two trials.
~:;fu"tr!a,.I1i'lmO9D[2s&
Mann-Whitney U comparisons
Number of wasps trapped
Receptive figs
Post-pollinated
figs
Wasp species
trapped
n
Range
Meanl
trap
traps
Range
Mean!
trap
n
traps
n
traps
Mean!
trap
Control I postpollinated figs
Control I
receptive figs
Control
U
P
36.0
27.0
ns
Range
P
U
"m'4-rgfisW5§,(R3T$\JBlw!jtCQ&;eS:~7£*N1}P?9I
5-638
6
4.8
'-14
6
1.8
3.3
0-7
6
0.3
0-1
6
0
32.5
21.0
ns
6
4.3
0-8
6
0.7
0-1
6
0
33.0
21.0
ns
6
2.5
0-11
6
0
6
0
25.5
21.0
ns
E. stuckenbergi
6
162.6
P. /;aJ'iJarus
6
PhUoll}'1}('sis sp.
Otitesella spp.
0-4
ns
Wjil'q.1!LSQN;:mEr%IDg2Of9"~5?nRpw-3tGs·@*&A£uT$[§MJ
ns
P < 0.5; , ,
not significant;'
P < 0.01
U1
"'-J
Table 3. The fig wasps trapped near cotton bags containing either unpollinated (receptive) figs of F. thollllingii or F. burtt-davyi. Control bags were empty.
fjI'WMt£-289i~@:msr;lRS1$>&?gQnU%.}F)w»"<J,Y3!¥§
Number of was,?s trapped
Wasp
species
Receptive figs
Receptive figs
F. burtt-davyi
F. IhOlllziJlgii
n
traps
rn
Wi',§j)T£lt<fG3pNACm
;a
w
W';PiSfr-
Meanl
trap
l
Range
T1'i§rwpu
n
traps
2.5
0-4
36
6
1.2
1-2
23.5
2-18
6
0.3
0-1
36
0-5
6
0
31-66
6
0-3
E. stuc:kellbergi
6
53.8
E. baijl/arhi
6
10_0
4-17
6
1.5
P. barbarus
6
0.2
0-1
6
10.8
Philotlypesis sp.
6
0.5
0-3
6
1.3
P < 0.001.
u
n
traps
Meanl
trap
P
Control IF.
F. thollllillgii I F.
bul1t-davyi
bum-davyi
U
P
U
21
ns
36
P
rnfwm"aSY'?FP¥i2J!lj$N9.g~W:51,ITy)qtA%
0-6
not significant; '"
Control I F.
thoflllillgii
Range
Range
3.2
ns
Control
Meanl
trap
6
~"'il:-I.t!m3Jj1·&)fHaD2=r{4
Mann-Whitney U comparisons
27
ns
ns
36
36
25
ns
36
21
ns
23.5
ns
80% (323)
/'~",
F. thonningii
2% (8)
(65)
2% (9)
(19)
1Ii~m
4% (3)
1 % (1)
Control
62% (15)
8% (2)
29% (7)
Jill
E. stuckenbergi
P. barbarus
E. baijnathi
Phi/otrypesis sp.
Figme 3. The numbers of wasps, together with with their relative percentages, simulta neously trapped near bags containing
receptive figs of R thonningii
or R burtt-davyi.
Empty bags acted as controls.
The specificity of wasp attraction was confirmed when individuals were provided with a choice between
the receptive figs of two closely related Ficus species. The ratios of wasps on different treatments were
similar during the two trapping periods and the data have therefore been combined for analysis.
Significantly more E. stuckenbergi andP. barbarus (wasps associated withF. thonningii) were trapped near
receptive F. thonningii figs than on traps near F. burtt-davyi figs or the controls (Table 3; Figure 3).
Likewise, significantly more E. baijnathi were trapped near figs of its host species, F. burtt-davyl. No
preferences were shown by the Philotrypesis sp., which may be associated with both F. thonningii and F.
burtt-davyi.
58
The fruiting phenologies of the five F. burtt-davyi and five F. thonningii trees that were monitored for two
years are shown in Figure 4. On the four trees of each species that produced figs the crops varied in
duration from as little as 8 weeks in summer to over 20 weeks during winter. The development of the
figs on anyone tree was generally well synchronised, but the trees fruited at different ti:::1es of the year.
Most of the wasps trapped on the trees belonged to species known to be specifically associated with that
Ficus species (Table 4). F. sur is the third indigenous fig species growing in the botanical gardens. Only
very small numbers of the wasps associated with this species were recorded from traps placed in F.
thonningii and none in F. burtt-davyi trees (Table 4).
Table 4. The wasps caught in sticky traps on R thonningii and R burtt-davyi trees and their normal host Ficus.
"I bEA §+§
B.'§*?kfHfi¥?iH
Trap location
Origins of wasps on traps
F. thonningi/
444MWWM
9 ¥GNHW
2120
F.thonningii
.it
¥4R
o
1204
'''tHiS +
#i¥iiiSi#Whi$M#¥ 4i£l1W
696
22
13
F. burtt-davyl
&*E
Host tree
Indeterminate
F.sur
F. burtt·davyl
¥
120
dEaf*'
Wt?¥fi§%Vi#%#5M#!
___________________________
5~-
4~-m.,
3
-I----i:=a
g' ~·,:-m.
. M.~:
________~
________
',VI
--
__~
2~-Ii·eBl
______________________________
5~-
:~-;.'s·?
4{.
1_
Matl%1JII
_,..
_
'1"'1"'1"'1"'1""1"'1"'1'"'1"'1"'1""1'''1'''1''''1'''1'''\""\"'1"'\""1"'1""1"'1"'1'"'j'''\'''['
DJFMAMJ JASONDJ FMAMJ JASONDJFM
1990
1991
1992
Figure 4. The fruiting phenologies of 10 fig trees used in the long tenn monitoring of fig wasp arrivals and departures. Intercrop
periods are shown with thin solid lines while the period when the trees were bearing fruit are denoted by solid blocks.
59
Few wasps were present in the 10 trees when they were not bearing fruit (Table 5). During each fruit crop,
the trapped wasps initially comprised those adult female wasps which had been attracted to the tree to
oviposit. After a few weeks these were then followed during the second half of the crop period by their
progeny as they emerged from the mature figs (Table 5, Figures 5 and 6). On1y the wasps trapped during
the first half of each crop period had therefore flown to the trees from elsewhere, and only the trapping
results during this period have been included in the following analyses.
F. burtt-davyl Tree 2
Phl/otrypesls
sp.
Sycorytes sp.
Otitesella spp.
100
80
70
60
50
4a
30
20
10
o
E. baf]nathf
I
o
20
40
60
80
100
Weeks
Figure 5. Identity and numbers of fig wasps trapped at a F. bllrtt-davyi tree. The shaded areas represent those periods when figs were
present on the tree.
On F. thollllillgii trees the numbers of E. stuckenbergi and P. barbarus were significantly higher during the
first half of the crop periods than during the intercrop periods (Table 6). In contrast there were no such
differences in the trapping rates of E. baijnathi, the species that pollinates F. burtt-davyi. The reverse
situation was present on the F. burrt-davyi trees, where there were no increases in the numbers of E.
60
Table 5. The fig wasps (all species) trapped at F. thonningii and F. burtt-daryi trees during
their intercrop. receptive (first half of crop period) and producer !latter half of crop period) stages.
•
~.
Tree #
Number
of crops
m! l'iillill!
!Ii
Total
crop
period
(Wks)
Intercrop
period
(Wks)
-
Ii
!IIDlII
Number of wasps trapped
(mean I week)
Intercrop
period
Receptive
period
W&kMkk
Producer
period
¥FW 8 Mh
e
ea
F. thonningii
21
98
2.37
18.66
13.29
42
77
1.35
14.29
25.74
3
38
81
2.10
15.32
8.35
4
63
56
0.91
8.32
22.10
0
0
119
0.75
5
164
431
1.25
12.79
15.39
2
15
90
0.12
0.40
2.80
2
3
28
77
0.04
7.29
15.79
3
4
21
84
0.02
39.91
0.76
4
3
19
86
0.26
50.42
3.16
5
0
0
105
0.04
12
83
442
0.10
2
5
Total
2
F. burtt-daryi
Total
..
iii
61
24.35
-
6.75
21
£t.WIESIri:i
stuckenbergi and P. barbarns trapped on the trees during receptive periods, but numbers of E. baijnathi
did increase (Table 6). Thus, during periods of fig receptivity the three species were only preferentially
attracted to their own host trees (Table 6).
F. thonnlngiJ Tree 3
SyCOryct6S
sp.
P. barban.Js
140
130
129
110
100
90
80
70
SO
50
40
30
E. stuckenoorgl
•••••••••
20
l@!i~
10
O--rrli1,;.tl'"
o
20
I
"1' " I ' ' '
I
40
60
80
100
120
Weeks
Figure 6. Identity and numbers of fig wasps trapped in a F. thonningii tree. The shaded areas indicate those periods figs were
present on the tree.
62
Table 6. Comparisons of the numbers of wasps trapped during intercrop periods with the numbers trapped during the first half of each crop period (which includes the receptive
female phase of fig development).
5J!:IW.jT¥r",~l1'i4f2wsmnPFpN;}£taEB\·M
Tree
Number of wasps trapped (mean/week)
#
E. stuckenbergi
Receptive
period
Intercrop
period
P. barbarus
Mann-Whitney
Z
P
Receptive
p~riod
Intercrop
period
E. baijllathi
Mann-Whitney
Z
P
Receptive
period
Intercrop
period
Mann-Whitney
z
P
lW#jt~'fT!?maer;:p,wJV&01iIOMNBL%.g
F. 1/lOlIlIillgii
29.29
0.5
-2.899
2
6.85
0.42
-3.436
3
11.00
0.54
-1.019
4
9.71
0.36
-3.903
5
TOTAL
...
0.14
0.25
-2.787
0
0.01
0.227
ns
3.15
0.49
-3.676
0
0.05
0.969
ns
ns
0.60
1.05
-1.719
0.05
0.03
-0.711
ns
1.93
0.13
-3.295
0
0
1.000
ns
0.37
11.40
0.44
ns
0.20
-5.905
2.03
0.41
0.09
-6.291
0.02
0.20
0.413
ns
ns
m
w
F. burtl-dmyi
0
0
1.000
ns
0
0
1.000
ns
0.13
0.06
-1.103
2
0.07
0
0.260
ns
0
0
1.000
ns
5.29
0
-5.586
3
0
0
1.000
ns
0
0
1.000
ns
38.00
0.02
-5.548
4
0
0
1.000
ns
0
0
1.000
ns
51.00
0.17
-4.486
5
TOTAL
0
0.02
0
0
0.311
ns
ms"trfJMS!PW'iR5j=T7VZ?~18<9ENle+>;2$
0
0
...
0.03
1.000
ns
22.67
0.06
-9.473
,w:m
ns = not significant; •• = p
<
0.01; ••• = p
<
0.001
DISCUSSION
In the Grahamstown Botanical Gardens the overlap in the fruiting periods of F. burtt-davyi and F. rhonningii
and the close proximity of the trees meant that adult wasps associated with the two species could potentially
colonise the trees of either species. However, long term monitoring of wasp arrivals at F. rhonningii and
F. burtt-davyi trees showed that the trees' pollinators were only attracted to their respective host trees. The
two wasp species were thus able to distinguish their own host figs in the presence of receptive figs of the
other species. This was confirmed in the experiments using figs placed in cotton bags, which also showed
that, as in the case of F. burtt-davyi (van Noort et al., 1989), the pollinators of F. thonningii were not
attracted to their host figs unless they were at the receptive stage.
P. barbarus was the only non-pollinating wasp recorded on the traps in large numbers. This species was
also found to be attracted to receptive phase figs of only its host tree (F. thonningii). P. barbarus enters
the figs to oviposit at the same time as the pollinators, and like them may be attracted by the changes in the
volatiles that are released during the receptive period (Ware et al., in press).
These experiments have confirmed the specificity of wasp responses to the volatile attractants released by
two closely related Ficus
~pecis,
and have shown that figs normally remain attractive for a short period,
unless pollination is prevented. Gas chromatograph analysis of volatiles produced by the figs of these species
has shown that additional compounds are released during their receptive phase of development and these are
likely to be the basis for the observed specificity of attraction (Ware et al., in press). Only isolation and
bioassay of the attractant volatiles will confirm this link.
Bronstein (1992), in discussing the proximate factors that determine whether or not a fig tree will be
pollinated, suggested that localised wasp extinction could be a major factor limiting fig production. This is
because in small tree populations there may be no receptive figs for the short-lived pollinating wasps to
colonise.
However, if they remain unpollinated, the figs of F.burtt-davyi were able to maintain their
attractiveness to pollinators for extended periods.
This could potentially overcome local shortages of
64
extended receptive period in F, burtt-davyi is that a smaller number of fig trees can maintain the wasp
populations in each local area (Bronstein et ai" 1990), This is in contrast to another African species, F.
sycomOnlS, where unpollinated figs abscise only about a week after the start of the female phase (Galil
and Eisikowitch, 1969).
ACKNO~DGEMTS
We would like to thank J.D, Cameron for permission to work in the 1820 Settlers Botanical Gardens.
The financial support of the FRD to ABW is gratefully acknowledged.
REFERENCES
Berg, C.C. (1988), Classification and distribution of Ficus. Experientia 45: 605-611.
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, V.K.
Bronstein, J.L. (1987). Maintenance of species specificity in a Neotropical fig-pollinator wasp mutualism.
Gikos 48: 39-46.
Bronstein, J.L. (1989). A mutualism at the end of its range, Experientia 45: 622-636.
Bronstein, J.L. (1992). Seed predators as mutualists: ecology and evolution of the fig/pollinator
interaction. In Insect-Plant Interactions Volume IV (Ed. Bernays, E.). CRC Press Inc. Florida,
pp.1-44.
Bronstein, J,L., Gouyon, P.H., Gliddon, C., Kjellberg, F. and Michaloud, G. (1990).
Ecological
consequences of flowering asynchrony in monoecious figs: a simulation study. Ecology, 71: 21452156.
Bronstein, J.L. and Patel, A. (1992). Causes and consequences of within-tree phenological patterns in
the Florida strangling fig, Ficus aurea (Moraceae). Am. 1. Bot. 79: 41-48.
Galil, J. and Eisikowitch, D. (1969). Further studies on the pollination ecology of Ficus sycomonls L.
(Hymenoptera, Chalcidoidea, Agaouidae). Tijd. Entomol. 112: 1-13.
Galil, J. (1977). Fig biology. Endeavour 1: 52-56.
Joseph, K.1. (1958). Recherches sur les Chalcidiens Blastophaga psenes (L.) et PhiIotrypesis caricae (L.)
65
du figuier Ficus carica (L.). Am. Sci. nat. 20: 197-260.
Kjellberg, F., Doumesche, B. and Bronstein, J.L. (1988). Longevity of a fig wasp (Blastophaga psenes).
hoc. K. Ned. Akad. Wet. Series C 91: 117-122.
U1enberg, S. (1985). The phylogeny of the genus Apocrypta Coqueral in relation to its hosts, Ceratosolen
Mayr (Agaonidae) and Ficus L. Verhand. K. Akad. Wet. 83: 149-176.
van Noort, S. (1992). The systematics and phylogenetics of the Sycoecinae (Agaonidae, Chalcidoidea,
Hymenoptera). Unpublished PhD thesis. Rhodes University, Grahamstown.
van Noort, S., Ware, A.B. and Compton, S.G. (1989). Pollinator specific volatile attractants released from
the figs of Ficus burtt-daryl. Sth Afr. J. Sc. 85: 323-324.
Ware, A.B., Compton, S.G., Kaye, P.T. and van Noort, S.G., in press. Fig volatiles: Their role in
attracting pollinators and maintaining pollinator specificity. Pl. Syst. Evol.
Wharton, RA., Tilson, J.W. and Tilson, R.L. (1980).
Asynchrony in a wild population of Ficus
sycomorns. Sth Afr. 1. Sc. 76: 478-480.
Wiebes, J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev. Evol. Syst. 10: 1-12.
Wiebes, J.T. and Compton, S.G. (1990). Agaonidae (Hymenoptera, Chalcidoidea) and Ficus (Moraceae):
fig wasps and their figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet, 93: 203-222.
Windsor, D.M., Morrison, D.W.,
Estrib~
MA. and de Leon, B. (1989). Phenology of fruit and leaf
production by 'strangler' figs on Barro Colorado Island, Panama. Experientia 45: 647-653.
66
CHAPTER 4
CHEMICAL EVIDENCE FOR VOLATILE ATTRACTANTS
Paper 5: Fig volatiles: their role in attracting pollinators and maintaining pollinator specificity. In press Plant
Systematics and Evolution (A.B. Ware, P.T. Kaye, S.G. Compton and S. van Noort).
67
FIG VOLATILES:
THEIR ROLE IN ATTRACTING POLLINATORS AND
MAINTAINING POLLINATOR SPECIFICITY
A. B. Ware, P. T. Kaye, S. G. Compton and S. van Noort
ABSTRACT
Each fig tree species (Ficus) is totally dependent on a specific species of wasp for pollination and the larvae
of these wasps only develop in the ovules of their specific Ficus host. Because the fig crop on any particular
tree is generally highly synchronised, the short lived female wasps must leave their natal tree in order to find
figs which are suitable for oviposition.
Chemical volatiles produced by figs when they are ready for
pollination are thought to be the means by which the wasps detect a suitable host. Gas chromatograms of
the fig volatiles of 7 species of Ficus showed them to be species specific. Age related changes in the volatile
profiles were noted as extra volatiles are produced when the figs were ready for pollination.
lNTRODUCTION
The relationship between fig trees (Ficus spp., Moraceae) and their pollinating wasps (Chalcidoidea,
Agaonidae, Agaoninae; senSll Boucek (1988» is often considered to be the extreme example of plant-animal
coevolution (Janzen, 1979). There are some 750 Ficus species worldwide (Berg, 1988), each of which is
generally pollinated by females of its own specific species of wasp (Wiebes, 1979; Michaloud et ai., 1985;
Wiebes and Compton, 1990). The trees are totally dependent on the wasps for pollination, while the wasp
larvae develop only in the ovules of their Ficus hosts.
68
Figs (also called syconia) are hollow, roughly spherical inflorescences, lined on their inner surface with
hundreds of unisexual flowers.
The pollinating female wasps enter the fig through the bract-lined
entrance (the ostiole) and pollinate the flowers, some of which are also used for oviposition. These
foundress wasps usually lose their wings during the passage through the ostiole and are unable to leave.
Fig development can be divided into five distinct phases (Ga1i1 and Eisikowitch, 1968). In the prefemale
phase, the ostioles of the young developing figs have not yet opened. This stage is followed by the female
phase where the female flowers mature and the ostiole opens to allow the pollinators to enter the fig.
Once pollination has taken place the figs enter the inter-floral phase where both seeds and wasp larvae
are developing. The male phase commences with the maturing of the male flowers and the emergence
of the wingless males of the pollinator wasp which seek out and mate with the female wasps while they
are still in their natal galls. After emerging from their galls the pollinator females acquire a load of
pollen either actively or passively (Ga1i1 and Eisikowitch, 1973). They then leave their natal fig through
a hole chewed through the fig wall by the males. Finally the fig ripens (post-floral phase) and attracts
various avian or mammalian frugivores which disperse the seeds (Janzen, 1979).
Fig development on individual trees is normally highly synchronised, forcing the short-lived adult females
(Kjellberg et ai., 1988) to leave their natal trees and search elsewhere for figs containing flowers that are
ready to be pollinated. Factors involved in host finding and host specificity are only partially understood.
A potential attractant is chemicals released from the figs (Janzen, 1979; Ramirez, 1970). Bronstein,
(1987) provided indirect evidence for chemical attraction when she showed that large numbers of
pollinators of the neotropical F. pertusa L. arrived at their host tree only when the figs were ready to
be pollinated. Confirmation of long distance chemical attraction was provided by van Noort et ai. (1989)
who showed that the pollinators of F. burtt-davyi Hutch. were attracted only to the figs of their host Ficus
and this only occurred when the figs were at the appropriate stage of development.
There have been few previous studies of the volatiles released by fig trees. Jennings (1977) found that
the differences between the steam distillate volatiles of ripe figs from 4 cultivars (some are gynocarpic
and do not require the services of the pollinating wasps to set fruit) of F. carica L. were only quantitative.
69
Other studies have concentrated on either the leaf volatiles (Buttery et at., 1986) or the composition of
volatiles from stem exudates (Warthen and McInnes, 1989), but neither leaves nor stems playa role in
attracting fig wasps (van Noort et ai., 1989). Barker (1985) provided gas chromatograph evidence of the
existence of fig volatiles.
Host specificity of Ficus species is likely to be achieved through a combination of these long distance
volatile attractants, short range, contact stimuli provided by the fig surface and other physical
characteristics of the fig.
These may include the chemical properties of the fig surface (Ware and
Compton, 1992) and the physical characteristics of the fig ostiole (Ramirez, 1974; Janzen, 1979) through
which the wasps must crawl in order to reach the fig flowers and oviposit (Galil, 1977).
In this paper we address questions related to the chemical nature of the long distance attractants
produced by figs. Initially we determined whether the figs of each Ficus species has a characteristic
bouquet, a possible means by which the wasps could distinguish their host tree species from other Ficus.
Changes in the composition of the bouquet of the figs of several species were then examined in relation
to their developmental cycle. Changes observed in the volatile profile of the figs during the period when
the fig flowers are ready for pollination could account for the observation that wasps are attracted only
to the trees at this stage of fig development.
MATERIALS AND METHODS
The volatiles of seven Ficus species were investigated: F.sur Forssk., F. burtt-davyi, F. thonningii Bl., F.
lutea Vah!, F. ingens (Miq.) Miq., F. macrophylla Desf. from the Grahamstown area, eastern Cape
Province, South Africa, and three cultivars of F. carica (Calimyrna, Kardota and White Genoa) from the
Citrusdal area of the western Cape Province, South Africa. F. macrophylla is native to Australia while
F. carica is of Mediterranean origin. The other species are native to South Africa.
70
Cotton bags were used to enclose prefemale stage figs in order to prevent wasps from pollinating the figs.
Once the figs had reached the attractive female phase, determined by confirming that wasps had entered
other figs on the same tree, they were harvested and within 10 minutes were placed in a glass tube
(internal diameter 30mm, length 300mm). Air cleaned with activated charcoal was directed over the figs
at approximately ll/minute for 5 hours and the volatiles, chemicals in the vapour phase, that were
released trapped on activated charcoal (Orbo 32, Supelco, Bellefonte, PA). The volatiles of unpollinated
prefemale and pollinated inter-floral stage figs were processed in a similar way. With the exception of
the locally scarce F. lutea and F. macrophylla, prefemale, female and inter-floral stage figs from at least
three trees of each species were analyzed independently. The number of figs processed depended on their
size. For large figs such as F. carica as few as 8 figs were used, while for F. buTtt-davyi, the species with
the smallest figs, at least 20 figs were used during each volatile trapping experiment.
Volatiles were eluted from the charcoal traps with 1ml dichloromethane (Nlerck Cat No 6048). The eluant
was then sealed in glass ampoules and stored at 4°C. When required, the contents of each ampoule were
concentrated to approximately 10ul by evaporation with a stream of nitrogen, and lul of the resultant
concentrate was chromatographed on a fused silica capillary column (SGE; 25m with an internal diameter
of 0.22mm) on a Hewlett Packard (HP) 5890 gas chromatograph (GC) fitted with a flame ionisation
detector and using nitrogen as a carrier gas.
The instrumental parameters were: injection port
temperature 210°C, flame iollmttion detector temperature 210°C, nitrogen carrier gas 20ml/minute. The
initial oven temperature of 40°C was maintained for 1 minute and then was increased at a gradient of
SOC/min to a maximum temperature of 180°C, which was then maintained for 5 minutes. The temperature
was then increased at a rate of lOoC/min until the oven temperature reached 250°C which was maintained
for 10 minutes before the run was terminated. Purge time for the injection port was set at 0.5 minutes.
The results were analyzed on an HP 3393A integrator, the attenuation being set to zero.
71
RESULTS
Differences in volatile profiles
The volatiles released from the female phase figs of the seven Ficus species each resulted in a unique gas
chromatogram (Fig. 1; Fig. 6). All the chromatograms were complex, containing many peaks each of
which represented an individual volatile compound. Most of the volatiles were present in trace quantities
(a full scale deflection at an attenuation of zero represented approximately 5ng of material), some of
which may be caused through the degradation of the figs, insect damage or even directly from small
insects such as scale insects. The profiles from different individual trees of the same species were
generally similar (see below for an exception) showing that each tree species has its own characteristic
bouquet. For example the volatile profiles of the three cultivars of F. carica were found to be essentially
similar, differing quantitatively rather than qualitatively (Fig. 2). The general uniformity within species
was observed in the prefemale phase chromatograms of some ten F. burtt-davyi (Fig. 3). However, the
female phase figs of a further 2 trees were found to contain an additional major peak, which eluted at
ca. 12 minutes.
Age related changes in volatile profiles
The chromatograms of F. burtt-davyi figs at the female stage of development showed an additional volatile
eluting at ca. 12 minutes (Fig. 3 and Fig. 4). As mentioned above, some trees produced a further
additional peak with a slightly reduced retention time at this stage of their development (Fig. 3). The
volatile profiles of prefemale and inter-floral phase figs of F. burtt-davyi were similar (Fig. 4). Similarly,
in F. ingens the female phase figs produced extra volatiles that were not recorded before or after this
stage of development. In this case there were consistently two additional peaks, with retention times of
ca. 24 and 25 minutes (Fig. 5). One additional peak was present in the female phase chromatogram of
F. lutea with a retention time of ca. 12 minutes (Fig. 6). Unfortunately, no inter-floral fruit was available
for comparison because most of the figs of F. lutea figs were not pollinated.
72
f. carica
F. sur
F. tllonningU
F. ing#n8
F. macropllylta
o
15
30
TI ME (m 1nutes 1
Figure 1. Chromatograms of volatiles from female phase figs from six individual trees of six Ficus species. A full amplitude
response at the detector represents at least 5 ng of material while the retention time indicates how long the volatiles
remained on the column before reaching the detector. The smaller more volatile compounds generally elute first while
the oven temperature is still low. See text for instrumental parameters.
73
o
C")
'"c
!...
>,
E
'"
u
'"c0
Q)
(!)
'"
+-'
Q)
0
"0
+-'
s...
.c
::.:
~
'"
0
N
Vl
Q)
....,
:::J
C
E
o
o
3SNOd~
3NIHJVW
Figure 2. The volatile profiles of female phase figs from three cultivars of R carica.
74
FH1ALE PHASE
PREFEMALE PHASE
..,
Tree 2
w
VI
z
a
c..
VI
w
0::
0::
Tree 3
a
l-
T
u
W
lW
Cl
Tree 4
,
o
,
10
o
RETENTION TIME (min.)
20
10
20
Figure 3. The volatile profile of prefemale and female phase figs of four individual trees of F. burtt·davyi. The closed symbol
highlights the additional volatile recorded from figs in the female phase. The open symbol indicates that volatile which
was released from female phase figs of two individual trees.
75
PREFEMALE PHASE
FEMALE PHASE
INTER-FLORAL PHASE
.
o
•
10
•
20
30
40
TIME (minutes)
Figure 4. The gas chromatograms of prefemale, female and inter-floral phase figs from F. burtt-davyi. The closed symbol
indicates the additional volatile peak occurring in the volatile profile of female phase figs.
76
o
M
LU
LU
Vl
Vl
<>:
:r::
0..
<>:
:r::
0..
W
~
LU
Vl
<>:
:r::
0..
LU
~
-I
<>:
<>:
::E:
::E:
W
Ll..
~
Ll..
LU
<>:
a::
0
LU
Ll..
I
a::
a::
0..
I..LJ
!-
Z
0
N
c::
E
LU
::E:
~
I-
>
Z
0
I-
Z
UJ
IUJ
c:
0
~. 3
c::
~
Q/
L
L
3SNOd~
c
l
~OlJ31a
Figure 5. Gas chromatograms of the volatiles from prefemale, female and interfloral stage figs of F. ingens. The symbols indicate
additional volatile components produced by female phase figs.
77
0
PREFEMALE PHASE
I
w
U)
z
0
0U)
w
~
~
w
z
::r:
u
«
::;::
o
~
I
10
·20
30
FEMALE PHASE
I
I
I
40
TIME (mi nutes)
Figure 6. The volatile profiles of the prefemale and female stage figs of F. Jutes. The symbol indicates the additional volatile
produced by female phase figs.
78
DISCUSSION
Flower volatiles playa vital role as olfactory cues in attracting pollinating insects (Pellmyr and Thien,
1986). Figs are no exception in attracting their pollinators, even though their flowers are contained within
a syconium. Many insect use these olfactory cues together with visual stimuli such as colour (Tabashnik,
1985; Owens and Prokopy, 1986) and shape (Rausher, 1978; MacKay and Jones, 1989) to find their host
plant. Van Noort et ai., (1989) have shown that the wasp pollinators of F. burtt-davyi do not require these
additional visual aids to frod receptive figs of their host.
The movement of volatile molecules in the atmosphere is complex (Murlis et ai., 1992). To be effective
and reliable sources of information, volatile attractants have to be consistently emitted and easily
distinguished from the background of naturally occurring odours.
Electrophysiological studies have
shown that cues resulting from single volatile compound are probably the exception rather than the ru1e
(Visser, 1986). This implies that the fig volatiles are probably an uncommon mixture of compounds of
the immediate environment and present themselves in reasonable amounts only when the figs are ready
to be pollinated.
In pollinating systems such as those between some orchids (Ophrys) and male bees, the partnership can
also be highly specific (Hills
et ai., 1972; Borg-Karlson et ai., 1985).
Here each orchid species possesses
a unique blend of volatiles, components of which may mimic the pheromones of attractive female bees.
The plants deceive the male bees which, while attempting to copu1ate with them, pollinate the flowers.
Among such orchids, speciation potentially results from mutations which lead to changes in the plants'
attractive volatiles (Hills et ai., 1972).
Similarly, the volatiles produced by figs may facilitate the obligate relationship between fig trees and their
pollinators. The figs of each Ficus species produce a different bouquet of volatiles, which is largely
consistent within species, allowing host specific pollinators to differentiate between them. Furthermore,
additional volatile(s) are released at the time the pollinators are attracted.
Presumably it is these
additional compounds, either alone or in combination with the 'normal' volatile bouquet which form the
79
basis of attraction.
Female phase volatiles could therefore be of biological significance because this is
the period when pollinators are attracted to their respective host trees. Identification, synthesis and
bioassay of the compounds are now required in order to confirm these findings.
ACKNOWLEDGEMENTS
We would like to extend our thanks to Mr A Sonemann for preparing and maintaining the GC and to
Mr EJ. van Zyl (Fruit and Fruit Technology Research Institute) for allowing us to work on their
experimental farm at Citrusdal. Comments on the manuscript by Prof O. Pellmyr and the reviewer were
much appreciated. The bursary support of the Foundation of Research Development to ABW and SvN
is gratefully acknowledged.
REFERENCES
Barker, N.P. (1985). Evidence of a volatile attractant in F. ingens (Moraceae). Bothalia 15; 607-611.
Berg, C.C. (1988). Classification and distribution of Ficus. Experientia 45; 605-611.
Borg-Karlson, AK (1985).
Chemical basis for the relationship between Ophrys orchids and their
pollinators 1. Volatile compounds of Ophrys lutea and O. fusca as insect mimetic
attractants/excitants. Chern. Scrip. 25; 283-294.
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.AB. International, Wallingford, Oxon.
Bronstein, J.L. (1987).
Maintenance of species-specificity in a neotropical fig - pollinating wasp
mutualism. Gikos 48; 39-46.
Buttery, R.G., Flath, R.A., Mon, T.R. and Ling, L.C. (1986). Identification of germacrene D in walnut
and fig leaf volatiles. 1. Agric. Fd. Chern. 34; 820-822.
Galil, J. (1977). Fig biology. Endeavour 1; 51-56.
Galll, J. and Eisikowitch, D. (1968). Flowering cycles and fruit types of Ficus sycornorns in Israel. New
Phytol. 67; 745-758.
Galll, J. and Eisikowitch, D. (1973). Topocentric and ethnodynamic pollination. - In BRONJES, N.B.M.,
LINSKINS, H.F. (Eds): Pollination and dispersion, pp. 85-100. - Department of Botany,
University Nijmegen, Nijmegen, Netherlands.
80
Hills, H.G., Williams, N.H. and Dobson, C.H. (1972). Floral fragrances and isolating mechanisms in the
genus Cataseturn (Orchidaceae). Biotropica 4; 61-76,
Janzen, D.H. (1979). How to be a fig, Ann. Rev. Ecol. Syst. 10; 13-51.
Jennings, W,G. (1977). Volatile components of figs. Fd. Chern. 2; 185-191.
Kjellberg, F., Doumeshe, D. and Bronstein, J.L. (1988). Longevity of a fig wasp (Blastophaga psenes).
Proc. K Ned. Akad. Wet. C 91; 117-122.
MacKay, DA. and Jones, R.E. (1989). Leaf shape and host-rmding behaviour of two ovipositing
monophagous butterfly species. Ecol. Entomol. 14; 423-431.
Michaloud, G" Michaloud-Pelletier, S., Wiebes, J,T. and Berg, C.C. (1985). The cooccurence of two
species of fig wasp and one species of fig. Proc. K Ned. Akad. Wet. C 88; 93-119.
Murlis, J., Elkinton, J.S. and Carde, R.T. (1992), Odor plumes and how insects use them. Ann. Rev. Ent.
37; 505-532.
Owens, E.D. and Prokopy, RJ, (1986). Relationship between reflectance spectra of host plant surfaces
and visual detection of host fruit by Rhago/etis pomonella. Physiol. Entomol. 11; 297-307.
Pelimyr, O. and Thien, L,B. (1986). Insect reproduction and floral fragrances: keys to the evolution of
the angiosperms. Taxon 35; 76-85.
Ramirez B., W. (1970). Host specificity of fig wasps (Agaonidae), Evolution 24; 681-691.
Ramirez B., W. (1974). Coevolution of Ficus and Agaonidae. Ann. Missouri Bot. Gard. 61; 770-780.
Rausher, M.D. (1978). Search image for leaf shape in a butterfly. Science 200; 1071-1073.
Tabashnik, B.E. (1985). Deterrence of diamondback moth (Lepidoptera: Plutellidae) oviposition by plant
compounds. Environ, Ent. 14; 575-578.
van Noort, S., Ware, A.B. and Compton, S.G, (1989). Pollinator-specific volatile attractants released
from the figs of Ficus burtt-davyi. S. Afr. 1. Sci. 85; 323-324,
Visser, J.H. (1986). Host odor perception in phytophagous insects, Ann. Rev. Ent. 31; 121-144.
Ware, A.B.and Compton, S.G. (1992). Breakdown of pollinator specificity in an African fig tree.
Biotropica 24; xx-yy.
Warthen, J.D. and McInnes, D.O. (1989). Isolation and identification of male medfly attractive
components in Lichi chinensis stems and Ficus spp. stem exudates. J. Chern. Ecof. 15; 1931-1946.
Wiebes, J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev. Ecol. Syst. 10; 1-12.
81
Wiebes, J.T. and Compton, S.G. (1990). Agaonidae (Hymenoptera Chalcidoidea) andFicus (Moraceae):
fig wasps and their figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet. 93;203-222.
82
CHAPTER 5
STUDIES OF FIG WASP BEHAVIOUR
Paper 6: Dispersal of adult female fig wasps 1. Arrivals and departures. Submitted to Entomologia
expermenatalis et applicatus (A.B. Ware and S.G. Compton).
Paper 7: Dispersal of adult female fig wasps II. Movements between trees. Submitted to Entomologia
expermenatalis et applicatus (A.B. Ware and S.G. Compton).
83
DISPERSAL OF ADULT
FE~1AL
FIG WASPS.
I: ARRIVALS AND DEPARTURES
A.B. Ware and S.G. Compton
ABSTRACT
Ficus burtt-davyi, like most other fig trees (Ficus spp., Moraceae) is pollinated by its own unique species
of tig wasp, in this case Elisabethiella baijnathi (Chalcidoidea, Agaonidae).
Because fig crop
development on anyone tree is synchronised the small, short-lived female wasps have to fly to other trees
in order to find figs which are at a suitable stage of development for oviposition. This paper examines
the effects of temperature on the timing of emergence of the wasps from their natal figs, their dispersal
from the surface of the figs and their subsequent behaviour on arrival at new host trees.
INTRODUCTION
Fig trees (Ficus spp. Moraceae) and fig wasps (Chalcidoidea, Agaonidae) are intimately associated
(Boucek, 1988). Each Ficus species is usually pollinated by just one species of pollinating fig wasp
(Agaonidae, Agaoninae) (Wiebes 1979; Wiebes and Compton, 1990). The fig trees are totally dependent
on the females of their specific pollinating wasp for pollination and the fig wasps can only develop inside
the fruits of their host Ficlls. Non-pollinating fig wasps (belonging to other subfamilies of Agaonidae)
can be equally host plant specific (Ulenberg, 1985; van Noort, 1992).
Floral structure in Ficlls is unusual in that the inflorescences (the figs, also called syconia) are hollow,
roughly spherical and lined on the inside with hundreds or thousands of unisexual flowers. Entrance to
the centre of the fig (the lumen) is through a narrow bract-lined passage, the ostiole. Fig development
can be divided into five distinct phases (Galil and Eisikowitch, 1968). During the prefemale stage the
female flowers develop within the lumen of the fig and the ostiole is closed. In the next development
stage, the female stage, the female flowers are mature and are receptive to pollination. The ostiole opens
84
allowing the pollen-laden female fig wasps to penetrate the lumen of the fig in order to lay their eggs.
While passing through the ostiole the wasps lose their wings and parts of their antennae and they are
unable to leave the fig. The intert10ral phase follows during which the t10wers and wasp larvae develop
simultaneously. The male phase commences with the maturing of the pollen bearing male t1owers. The
female t10wers at that time contain fully developed seeds. The flightless male wasps have also reached
maturity and chew their way out of their ovules and seek out ovules containing conspecific females for
mating. The females leave their galls and after loading pollen leave the fig through an exit hole chewed
by the males. Finally the figs ripen (postt1oral phase), ready to be eaten by birds and mammals which
subsequently disperse the seeds (Janzen, 1979).
Fig crop development tends to be synchronised within each tree with gaps of months or even years
between crops (Bronstein, 1987; Windsor et ai., 1989). This means that the wasps cannot oviposit in
figs on their natal trees and the newly emerged wasps are forced to fly to other trees in order to fmd
suitable figs for oviposition.
Adult life spans of the pollinators are short (Kjellberg et ai., 1988;
Compton et ai., in prep.) and the wasps locate suitable figs for oviposition using volatile attractants which
emanate from the figs when they are ready to be pollinated (= female phase)(Bronstein, 1987; van Noort
et ai., 1989; Ware et ai., in press). Some non-pollinating fig wasps may utilize the same attractants as
the pollinators to fmd their hosts (Compton, submitted).
The biology of fig wasps when they are within the figs has been comparatively well documented (Galil,
1977; Janzen, 1979) but little is known about free-living adult female fig wasps (Bronstein, 1992). This
study examines factors that influence the emergence and departure of female wasps from their natal figs
and their arrival at receptive trees. Observations were also made on their behaviour when leaving their
natal trees and after fmding a suitable host tree.
MATERIALS AND lVIETHODS
Ficus burtt-dmyi Hutch. is the most common of the indigenous Ficus species occurring in the eastern
Cape Province of South Africa. At our field site in the 1820 Settlers Botanical Garden, Grahamstown,
85
the trees grow as rock-climbing shrubs against small cliffs and are pollinated by the fig wasp
Elisabethiella baijnathi Wiebes. Crop development in this species is highly synchronised (Compton et
ai., in press). Non-pollinating fig wasps associated with this species include Otitesella sesquianellata van
Noort, O. uluzi Compton, Philotrypesis sp. and Sycorycres (= Sycoscaperidea sensu Boucek) sp. Most
of these wasps breed only on F. burtt-davyi in the area, although the latter two species cannot presently
be distinguished from those that utilise F. thonningii Bl. and may not be host tree specific.
Emergence of Fig Wasps and Departure from Their Natal Figs.
In order to determine the time of day when female fig wasps emerged from their natal figs, five wasp
producing trees (male phase) were visited regularly during the daylight hours in summer while five
further crops were monitored in winter. Selected branches were marked and on each visit any figs with
wasp exit holes were counted and removed. Ambient temperatures were also recorded during visits to
two winter and two summer crops. Observations of the pollinators' preflight behaviour on the surface
of the figs were also recorded.
In a laboratory experiment, the critical take-off temperature for pollinator females was investigated.
Groups of forty wasps were initially subjected to 30 minutes pre-conditioning at each temperature in a
dark controlled-envirorllnent room before being released at the base of a box (115 cm high, 20 cm deep
and 20 cm wide, with the top and one side constructed of clear plastic sheets) placed under a fluorescent
light. A record was made of how many wasps took flight during the following 30 minute period.
In the field, the diel patterns of flight activity of E. baijJlathi were determined using five sticky traps
placed 1.5 m above the ground in the centre of the area where the fig trees were growing. Each trap
consisted of a clear cellulose acetate sheet measuring 60 cm x 20 cm and sprayed with pruning sealant
(Frank Fehr, Durban). The traps were replaced daily at 06hOO and 18hOO and all fig wasps trapped were
identified and counted. Trapping was carried out during three, one week long, periods both in summer
and winter.
86
Arrivals at Receptive Trees
The timing of wasp arrivals at a tree with receptive figs was studied by bagging pre-receptive F. bumdavyi figs. Once the figs had become receptive, the surrounding cotton bags prevented the wasps from
entering the figs and the wasps moved around on the bag surface trying to gain entry. Bags were visited
every three hours during the daylight hours over three day periods. All E. baijnathi females around the
bagged figs were counted and removed.
In a further investigation, previously bagged branches of two trees bearing pre-receptive figs were
exposed to the wasps once they had matured and were ready to be pollinated.
The behaviour of
individual pollinators as they landed and explored the branches was then recorded using a dictaphone.
Similarly, patterns of entry into individual figs was examined by exposing branches of unpollinated,
previously bagged receptive figs to the wasps for 15 minutes. The figs were then re-bagged and taken
to the laboratory where the number of foundresses in each was determined.
RESULTS
The Timing of Fig Wasp Emergence, Departure and Arrivals
Table 1. The timing of fig wasp emergence as indicated by the number of F. bZlrIt-davyi figs with exit holes.
Number of
trees
Morning
Afternoon
Night
06hOO- I2hOO
12h00-18hOO
18hOO-06hOO
Summer
5
636
154
129
Winter
5
1159
265
o
The fig wasps associated with F. burtt-davyi usually emerged from their natal figs between 06hOO and
12hOO (Table 1). No wasps emerged before 06hOO during the winter sample periods. However, in rnidsummer some wasps had emerged from their natal tig prior to the first sample of the day, possibly in
87
the period when most of the wasps emerged from the figs (Figure 2).
140
CIJ
ill
80
0
60
-I
:r:
l-
><
ill
TREE I
120
100
40
20
0
:r:
l-
240
$
220
CIJ
QJ
200
LL.
LL.
180
0
a:
Lll
CO
:2
::>
Z
TREE 2
160
140
120
100
80
60
40
20
0
2
3
4
5
6
7 . 8
9
10
11
DAYS
Figure 1. The number of figs with fig wasp exit holes during the 11 day dispersal phases of two synchronously fruiting F.
davyi trees (winter 1990).
b14rlt-
Fig crops, together with their associated fig wasps, develop more quickly during summer than winter and
subsequent wasp emergence dates are also more closely synchronised. During the summer months, the
ambient temperatures remained between 16°e and 30D e and the wasps from each of the five trees
completed their emergence in 2-3 days. Because wasp emergence was synchronised in summer and they
emerged over a short period, the effect of day to day temperature variation on the emergence rates could
not be adequately assessed. This was not the case in winter where the emergence periods were of longer
duration and ranged between 7 and 20 days (Figure 1). Figure 2 shows the variation in emergence rates
from the two winter crops where temperature data was also collected. Early in the mornings of days 5-7
a disproportionately large numbers of figs were found with exit holes (Figure 2).
Prior to these
observations (days 4-6) being made, berg wind conditions (offshore winds with accompanying high
temperatures) were prevalent (Figure 3) and this appears to have led to the timing of the wasp emergence
being brought forward.
88
07hOO
601
l-
~
J:
80i 12hOO
J
t::
$: 60j
en
"u:
40j
20jlll I
!!II
1IIIu
40, 17hOO
20
I
1i.IB.1
I
1
2
3
4
5
6
7
8
9
10 11
DAYS
Figure 2. Exit holes produced at different times of the day in two synchronously fruiting F. bum-davyi trees (winter 1990).
151
I07hOO
10
5
,---.
p
w
a:
'--"'
.~
.'-~,
~
l-
<l::
a:
ill
0..
:2!
2°l 17hOO
w
J
l- 15
10
&i
5
I
I
-,
1
2
3
4
5
6
7
l
8
9
10 11
DAYS
Figure 3. The morning and evening ambient temperatures in the 1820 Settkrs Botanical Gard"n during the II day wasp dispersal
phase from two F. bllrll-da'}'i trees.
89
12
11
10
l/~\I
9
Y
I
,,{ ,
v
8
o
7~
ill
ill
0.-
6~
5~
o
z
$:
I
i
' I
"
4 J (l
:
I'
,
I
\l
"
!
:\
\: .
' \,:
\
\Vi~Ii
I~l.'
/ 'I
1
""-t--+-J!
!
I
I
o
__~.-r
3
;
1
•
"
I
I
__
midnight
.
I · ~
'[1""-
I,
:JJ
O~i-'
:
!
J
1
1
I
:
i / :
I
C/)
I
,-.~n_oTr
6
__~'-r
9
12
__
15
~-'.
18
__
~
__
21
midnight
TIME (hours)
Figure 4. The average hourly windspeeds (+ /- standard error) experienced in Grahamstown during March 1989.
Support for temperature playing a role in the timing of emergence comes from regression analysis of
temperature and the number of wasps emerging from the figs over the day. The daily early morning
temperatures (07hOO) were significantly correlated with temperature at 17hOO the previous day (Spearman
Rank: r
= 0.724; P <
0.05). It was therefore not surprising to find that the proportion of figs with exit
holes produced early in the morning (before 07hOO) was positively correlated not only with the 07hOO
temperature on the morning of emergence (Spearman Rank: r
=
temperature experienced the previous day at 17hOO (Spearman Rank: r
0.724; P < 0.05) but also the
= 0.799; P <
0.001). Thus, the
delay in wasp emergence on day 9 (when the majority of wasps only emerged between 10hOO and 12hOO)
(Figure 2) can be related to the low temperatures experienced both that morning and the previous day
(Figure 3). Wasps therefore emerged later in the day when the mornings are cold although the previous
day's temperature may also be important in influencing the emergence pattern.
90
In the laboratory the critical takeoff temperature for E. baijnathi was found to be between 15 and 16°C.
Below this temperature no wasps were active in the air while at 20"C almost all the wasps were observed
to fly (Figure 5). The critical takeoff temperature was also the temperature at which fig wasps began
to exit from their natal figs. Thus it appears that the wasps responded to the ambient temperatures and
only emerged from their natal figs when the temperatures were likely to be high enough to allow flight
to take place.
100
80
60
40
20
O~-,
14
15
16
17
18
19
20
TEMPERATURE (oG)
Figure S. The flight take-off frequencies of E. baijnalhi females at varying laboratory temperatures.
The wasps trapped on the sticky traps indicated that all the fig wasps associated with the locally occurring
Ficus species were essentially day flying (see Table 2). In summertime a few wasps were captured on
the sticky traps before 06hOO. These wasps may have been trapped in those few hours after sunrise
before the traps had been replaced.
91
Table 2. The number of fig wasps trapped on sticky traps positioned in an area where fig trees were growing. The traps were
replaced every morning at 06hOO and again in the evenings at 18hOO. Monitoring was over three, one week periods in both winter
and summer.
flMiM
A
-riitiBMP*F¥4§fMI
AMi
'i§8tt'!¥
MS'W+"U
Number of wasps trapped
.
Summer
Species
Day
Winter
Night
Total
Day
Night
ijWjliriMM
MM
E. stuckenbergi'
20
7
49
0
P. barbarus '
3
0
37
E. baijnathr
11
8
16
0
C. capensiil
A. guieneensis
3
S. cyclostigma
2
3
-
Night
Day
&&&&*
&4AAfii¥#'
69
7
0
40
0
0
27
8
0
2
0
0
3
0
5
0
0
2
0
"3
0
Otitesella Spp.,·2
2
0
23
0
25
0
Sycoryctes Sp.'·2.3
2
3
16
0
18
3
Philotrypesis Sp.,·2
3
0
8
0
11
0
155
0
200
18
45
Total
18
N.
tiS
"f5
j%
i
-
• The fig wasps species are normally associated with the following Ficus'F. thonningii 81. 2F. burtt-davyi Hutch. and
3F. sur Forssk.
Few E. baijnathi were recorded arriving at the receptive figs after midday and the majority of the wasps
(84%) arrived at the receptive trees between 06hOO and 12hOO (Table 3). These arrivals corresponded
with the times when
th~
wasps left their natal figs (Table 1).
Table 3. The number of pollinating fig wasps (E. baijnalhi) on bagged receptive figs of a single F. bum-davyi tree. The wasps
were removed after being counted.
Percentage of wasps caught
Number
Date
06hOO
09hOO
12hOO
15hOO
I ShOO
caught
42
37
17
3
267
M
24/12/89
!=
26/12/89
6
79
12
3
0
180
28112189
6
61
20
10
4
71
Total
3
58
26
11
2
518
I
;Eli'fiBZIEI_·
•
92
Fig Wasp Behaviour
Shortly after the female E. baijnathi emerged from the figs, they positioned their wings above their body
and, after flaring their antennae, took off near vertically. They were then carried away from the trees
by the wind and were lost from sight. Fig wasps arriving at branches bearing receptive figs did not
necessarily land on the figs themselves (Table 4). Those not landing on the figs walked along the
branches, presumably searching for a suitable fig or flew away. The patrolling appeared to be more
directed than their choice of landing site as more visits were made to figs than to leaves (Table 4).
Wasps were observed to visit a total of 96 figs but only successfully entered the figs on 17 occasions
(Tables 4 and 5).
Table 4. The arrival of the pollinating fig wasp E. baijnathi at branches of two receptive F. bunt-davyi trees.
Date
Number of wasps
28/12/89
18
61 1192
15
Total
33
E. baijnathi females seloom antennated leaves and this activity was usually reserved for figs (Table 5).
Once on a fig, just over 50% of the wasps (17) eventually successfully penetrated an ostiole while some
25 % flew off without attempting entry. The remaining individuals attempted to enter a fig but aborted
their efforts and either continued searching on the same tree or flew away. For those wasps that were
observed to enter the figs the time taken from landing (n
73.0, n
= 32) to entry averaged
115.6 seconds (s.d. =
= 17; Table 5). Because of difficulties following flying insects it is not clear how many such
searching periods individual females had to make. The time taken for the wasps to disappear into a fig
once they had started entry averaged 84.12 seconds; s. d. = 36.09, n = 17; Table 5).
93
Table S. Pollinator fig wasp (E. baijnalhl) searching behaviour for suitable figs of F. bUrII-davyi in which to oviposit.
'5
Date
MM
Number of
•
Antennate
,msue
Penetrate ostiole'
wasps
None
Leaf
None
Fig
made
Failed
made
¥§Wh
*MYtiJ€
MF9b5'¥
ME
S
*9
28112/89
18
4
13
7
7
5
6/ 1/92
15
4
10
2
10
4
33
8
23
9
17
9
Total
2
•
ZUU
I
Successful
§#
bNW&MMA
·aM
iA
May abort attempt to enter the ostiole of one fig and successfully penetrate another.
Whether wasps discriminate against receptive figs that have already been entered by others was also
examined. The number of foundresses entering figs was found to be more regular than expected from
a random (Poisson) distribution with more figs having a single foundress then expected (Ch?[3]
P
= 67.67;
< 0.001) (Figure 6).
140..,
!
120
(j)
CJ
100
u:
LL
0
80
0:
ill
OJ
2
60
j
:::l
Z
40
20
o
0
0
1
2
3
4
5
NUMBER OF FOUNDRESSES
figure 6. The numbers of E. baijnazhi foundresses that entered figs of F. bum-davyi. Th" circles represent a Poisson (random)
distribution.
94
DISCUSSION
The timing of emergence of adult fig wasps from their figs is initially detennined by the males, which
chew the exit holes that allow the females to escape. In the case of E. baijnathi, exit holes were mainly
produced during the mornings, with temperature apparently influencing the precise timing of their
production.
Fig wasps are typically less than 2 mm in length, and with their slow flight have no
directional control above wind speeds of around 30 cmlsec (Ware and Compton, submitted). Wind
speeds are relatively low during the mornings and the timing of wasp emergence coincides with a time
when conditions for flight by the females are improved and may be an adaptation to avoid the
increasingly high windspeeds that develop as the day progresses.
Under laboratory conditions the females will only fly at ambient temperatures above 15°C. This compares
with recorded threshold temperatures for aphid flight of between 12.8 and 15.SOC (Robert, 1987). Fig
wasps in the field were nonetheless recorded flying below 15°C, probably due to the influence of solar
radiation, which would heat the small black bodies of the females by a few degrees during the short preflight period when they are on the surface of the figs (Lewis and Taylor, 1964).
Given that E. baijnathi generally leave their natal figs during the morning, and that arrivals at receptive
trees also occur at this'time, it appears that most of the wasps arriving at receptive trees had emerged
locally. Alternatively, conditions may inhibit the wasps from flying in the afternoons. Their dispersal
ability is, however, limited by their short life span of some 2-3 days (Compton et ai., in press) and such
deliberate prolongation of exposure to the elements seems unlikely.
Most southern African fig wasps, like E. baijllathi, are dark and are likely to be diurnally active.
Others, such as Allotriozooll heteralldromorphum Grandi, from F. lurea Yahl (Newton and Lomo, 1979;
Ware and Compton, 1992a), A/follsiella species (Compton, unpublished) and Ceratosolen arabicus Mayr,
from F. sycomorus L. (Galil and Eisikowitch, 1968; Compton et ai., 1991), fly at night and have been
collected at light traps (Ware and Compton, unpublished; H.G. Robertson, pers. comm.). These wasps
all display 'ophionid' features such as yellow coloration and large eyes, that are typical of many night-
95
flying insects (Huddleston and Gould, 1988).
The maximum extent to which E. baijllathi, or any other fig wasps, can migrate is unknown, but fig
wasps recorded on Ana.lc Krakatoa in 1984 must have flown from neighbouring islands more than 2 km
away (Compton et al., 1988). Even more impressive are the records of Allotriozoon heterandromorphum
Grandi in figs of an isolated F. lutea Vahl tree whose nearest known conspecific was 80 km away
(Compton, 1989; Ware and Compton, 1992a).
The observed flaring of the antennal sensilla when E. baijnathi females are about to takeoff from their
natal figs appears to be an ability that is limited to the few species of Agaoninae that have Type IV
sensilla arrangements (Ware and Compton, 1992b). Host fmding may be aided if the multiporous plate
sensilla on the antennae remained flared in flight because this results in a greater volume of air being
sampled (Kaissiing, 1971).
Once receptive trees had been detected, and females had landed, they generally antenna ted the figs and
not the leaves. Similarly E. baijnathi is not stimulated into antennating figs of other species (Ware and
Compton, 1992a), which suggests that recognition of the substrate occurs before antennal contact
chemoreceptors are employed.
This initial recognition of receptive figs could involve non-contact
chemoreceptors on their antennae or tarsal contact chemoreceptors. Alternatively, E. baijllathi may react
to the shape of F. burtt-davyi figs and begin antennating once the correct curvature has been detected.
Visual cues have also been shown to be important in habitat location by some hymenopteran parasitoids
(McAuslane et ai., 1991; Drost and Carde, 1992).
Ramirez (1986) found that the number of foundress pollinators in the figs of F. citrifolia did not differ
from a random (Poisson) distribution. However, the figs he used were saturated (all figs contained
foundresses) and fig wasps may be 'forced' to enter figs already containing foundresses if they cannot
find any uninhabited figs. This lack of a choice of foundress free figs was avoided in our studies, which
found that the distribution of foundress females inside the figs of E. baijnathi was overdispersed. This
could imply that the wasps are able to determine whether figs have been previously entered and avoid
96
them. Collections of naturally pollinated figs have often shown that the majority of figs receive only one
pollinator (Compton and Nefdt, 1990; Ramirez, 1986) suggesting that this result was not a consequence
of high densities of wasps vieing for entry in our experiment. Such avoidance of figs which already
contain foundresses would improve the reproductive success of E. baijnathi females, through the
avoidance of competition for oviposition sites, and would also influence their progeny sex ratios, which
become less female biased when two or more females share a fig (Nefdt, 1989).
ACKNOWLEDGEMENTS
The financial support of the Foundation for Research Development and Rhodes University to ABW is
gratefully acknowledged. Jim Cameron is thanked for granting us permission to work in the 1820 Settlers
Botanical Garden, while Mike Way is thanked for his field assistance.
REFERENCES
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, U.K.
Bronstein, J.L. (1987). Maintenance of species specificity in a Neotropical fig-pollinator wasp
mutualism. OiKOS 48: 39-46.
Bronstein, J.L. (1992).
Seed predators as mutualists: Ecology and evolution of the fig/pollinator
interaction. In Insect-Plant Interactions Vol. 4 (Ed. Bemays, E.). CRC Press, Florida.
Compton, S.G. (1990). A collapse of host specificity in some African fig wasps. Sth. Afr. 1. Sci. 86:
39-40
Compton, S.G., Holton, K.C., Rashbrook, V., van Noort, S., Vincent, S. and Ware, A.B. (1991).
Studies of Ceratosolen galili, a non-pollinating agaonid fig wasp. Biorropica 23: 188-194.
Compton, S.G. and Nefdt. R.J.C. (1990). The figs and fig wasps of Ficus bum-davyi. Mitt. Inst. AUg.
Bot. Hamburg 23: 441-450.
Compton, S. G .• Rasplus, J.-Y. and Ware, A. B. (in press). African fig wasp parasitoid communities.
In Parasitoid Commullity Ecology (Eds Hawkins. B.A. and Sheenan, W.). Oxford University
97
Press, U.K.
Compton, S.G., Thornton, 1.W.B., New, T.R. and Underhill, L. (1988). The colonization of Krakatoa
by fig wasps and other cha1cids. Phil. Trans. R. Soc. London B 322: 459-470.
Drost, Y.C. and Carde, R.T. (1992). Use of learned visual cues during habitat location by Brachymeria
imermedia. Elltomol. expo appl. 64: 217-224.
GaIil, l. (1977). Fig biology. Endeavour 1: 52-56.
Galil, l. and Eisikowitch, D. (1968a). Flowering cycles and fruit types of Ficus sycomorus in Israel.
New Phytol. 67: 745-758.
Galil, l. and Eisikowitch, D. (1968b). On the pollination ecology of Ficus sycomorus L. in east Africa.
Ecology 49: 259-269.
Galil, l. and Eisikowitch, D. (1973). Topocentric and ethodynamic pollination. In Pollination and
Dispersion (Eds Bronjes, N.S.M. and Linskins, H.F.). Department of Botany, University
Nijmegen, Nijrnegen, Netherlands.
Huddleston, T. and Gauld, 1. (1988). Parasitic wasps (Ichneurnonoidea)
1ll
British light traps.
Entomologist 107: 134-154.
Janzen, D.H. (1979). How to be a fig. Ann. Rev. EcoZ. Syst. 10: 13-5l.
Kaissling, K.- E. (1971). Insect olfaction. In Handbook of Sensory Physiology. Vol. 4. Springer, Berlin.
pp351-431.
Kjellberg, F., Dournesche, B. and Bronstein, l.L. (1988). Longevity of a fig wasp (BlaslOphaga psenes).
Proc. K. Ned. Akad. Wet. C 91: 117-122.
Lewis, T. and Taylor, L.R. (1964). Diurnal periodicity of flight by insects. Trans R. ellt. Soc. Land.
116: 393-476.
McAuslane, H.l., Vinson, S.B. and Williams, H.l. (1991). Stimuli influencing host microhabitat location
in the parasitoid Campoletis sonorensis. Entomol. expo appl. 58: 267-277.
Nefdt, R. (1989). Interactions befi.Veen Fig Wasps and Their Host Figs. Unpublished MSc thesis, Rhodes
University, Grahamstown, South Africa.
Newton, L.E. and Lomo, A. (1979). The pollination of Ficus vogelii in Ghana. Bot. 1. Linn. Soc. 78:
21-30.
Ramirez B., W. (1987). The influence of the microenvironment - the interior of the syconium - in the
98
coevolution between fig wasps (Agaonidae) and the fig (Ficus). In Insect-Plants Proceedings of
the 6th International Symposium
011
Insect - Plant Relationships (Eds Labeyrie, V., Fabres, G.
and Lachaise, D.) Dr W. Junk Publishers Dordrecht. pp 329-334.
Robert, Y. (1987). Aphids and their environment: dispersion and migration. In Aphids Their Biology,
Natural Enemies and Control-World Crop Pests Vol. 2A (Eds Minks, A.K. and Harrewijn, P).
Elsevier, Amsterdam. pp 300-310.
Ulenberg, S. (1985).
The phylogeny of the genus Apocrypta Coqueral in relation to its hosts,
Ceratosolen Mayr (Agaonidae) and Ficus L. Verhand. K. Ned. Akad. Wet. 83:
149-176.
van Noort, S. (1992). The Systematics and Phyiogelletics of the Sycoecinae (Agaonidae, Chalcidoidea,
Hymenoptera). Unpublished PhD thesis, Rhodes University, Grahamstown, South Africa.
van Noort, S., Ware, A.B. and Compton, S.G. (1989). Pollinator specific volatile attractants released
from the figs of FicLls burtt-davyi. Sth. Afr. 1. Sci. 85: 323-324.
Ware, A.B. and Compton, S.G. (1992a). Breakdown of pollinator specificity in an African fig tree.
Biotropica 24: 544-549.
Ware, A.B. and Compton, S.G. (1992b).Repeated evolution of elongate multi porous plate sensilla in
female fig wasps (Hymenoptera: Agaonidae: Agaoninae). Proc. K. Ned. Akad. Wet. 95:
275-292.
Ware, A.B., Compton, ·S.G., Kaye, P.T. and van Noort, S. (in press). Fig volatiles: their role in
attracting pollinators and maintaining pollinator specificity. PI. Syst. Evo!.
Wharton,R.A., Tilson, J.W. and Tilson, R.L. (1980). Asynchrony in a wild population of Ficus
sycomorus. Sth. Afr. 1. Sci. 76: 478-480.
Wiebes, J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev. Evol. Syst. 10: 1-12.
Wiebes, J.T. and Compton, S.G. (1990).
Agaonidae (Hymenoptera, Chalcidoidea) and Ficus
(Moraceae): fig wasps and their figs, IV (Africa concluded). Proc. K. Ned. Akad. Wet. 93:
203-222.
Windsor, D.M., Morrison, D.W., Estribi, M.A. and de Leon, B. (1989). Phenology of fruit and leaf
production by 'strangler' figs on Barro Colorado Island, Panama. Experiemia 45: 647-653.
99
DISPERSAL OF ADULT FEMALE FIG WASPS.
l.MOVENTSB~R
A.B. Ware and S.G. Compton
ABSTRACT
Fig wasps (Chalcidoidea, Agaonidae, Agaorunae) are the exclusive pollinators of fig trees (Ficus spp.,
Moraceae).
Fig development on the African fig tree, F. burtt-davyi, is normally synchronised on
individual trees, but not between trees. Consequently the females of each generation of the pollinating
species (Elisabethiella baijnathi) have to disperse to other trees to find 'receptive' figs which are suitable
for oviposition. This paper examines this aspect of fig - fig wasp biology. The flight speed of insects
is closely linked to their size and directional flight is difficult for small insects, such as fig wasps, in all
but the lightest of wind. We investigated the movements of fig wasps between trees using sticky traps
placed around fig trees or near cotton bags containing figs. Away from the trees, the densities of flying
wasps at different heights was also determined. When the wasps disperse from their natal figs they takeoff near-vertically and they are unable to exert directional control once they enter the air column and are
subsequently blown downwind. Near receptive host trees the wasps lose height and then fly upwind at
speeds of around 25 em/sec.
INTRODUCTION
Flight speed in insects is closely linked to their size with smaller insects flying more slowly than larger
insects (Lewis and Taylor, 1974). Directional flight for small insects will usually be problematic in all
but the lightest of winds, as they have no control over where they are carried.
Small species can
nonetheless achieve directional control by flying close to vegetation or to the ground, where there is a
'boundary zone' of relatively slow moving air produced by frictional drag (Taylor, 1958).
Pollinating fig wasps (Hymenoptera, Chalcidoidea, Agaonidae, Agaoninae sensu Boucek, 1988) are small
insects, usually between 1 and 3 mm in length, that are intimately associated with fig trees (Ficus spp.,
100
Moraceae). Each Ficus species is generally pollinated by one particular wasp species, which occurs on
no other Ficus species (Wiebes, 1979; Wiebes and Compton, 1990). Fig trees are unusual in that their
flowers are contained within an urn-shaped inflorescence - the fig. Pollinator access to the flowers is
limited to the 'female' phase of fig development, which is also the period when the flowers are receptive
to pollination (Galil, 1977). The female pollinating wasps penetrate the fig through a narrow bract-lined
entrance, the ostiole, and in the process usually lose their wings and part of their antennae. Once having
entered the fig, they are unable to leave. After pollination the ostiole closes (Verkerke, 1989) and the
larvae develop inside ovules galled by the females.
In most Ficus species, fig development on anyone tree is synchronised, which forces the female fig
wasps to leave their natal trees in order to find figs at the correct stage of development for oviposition
(Bronstein, 1989). As a consequence of the asynchronous fruiting, the often low densities of conspecifics
(Wharton et al., 1980; Gautier-Hion and Michaloud, 1989) and the small proportion of figs which are
suitable, the short-lived wasps (Kjellberg et aI., 1988; Compton et al., in press.) often have to fly long
distances to fmd them.
The pollinating fig wasps detect figs that are suitable for oviposition using
species-specific volatile attractants that are released from receptive figs when they are ready for
pollination (van Noort et ai., 1989; Ware et ai., in press; Ware and Compton, in prep).
The ability of the pollinators to find their hosts is impressive. Even when fig trees are isolated from their
conspecifics, such as those planted outside their natural distribution range (Compton, 1990; Ware and
Compton, 1992) or on islands previously sterilised by volcanic activity and now separated by expanses
of water (Compton et ai., 1988), at least small numbers of pollinating wasps find them.
Non-pollinating fig wasps (belonging mainly to the Agaonidae, but in subfamilies other than the
Agaoninae) may either gall the ovules like the pollinators or may parasitise the gall formers. Some of
these wasp species are also FiclIs species-specific (Ulenberg, 1985; van Noort, 1992).
Those non-
pollinating wasps that enter the figs to lay their eggs at the same time as the pollinators may utilise the
same volatile cues as the pollinators to fmd the figs while those ovipositing from the outside at a later
stage probably use other cues (Compton, in press).
101
In a previous paper (Ware and Compton, submitted) we investigated the effects of ambient temperature
on the timing of wasp emergence from their natal figs, as well as their behaviour when the pollinators
leave their natal figs and arrive at a suitable host tree. Here we describe the patterns of dispersal of
certain African fig wasps from their natal trees, within the general air column and as they approach trees
with receptive figs. Their flight speeds and the effects of wind direction on their movements is also
described.
MATERIALS AND METHODS
The Study Site
Field studies were undertaken in the 1820 Settlers Botanical Garden, Grahamstown, South Africa during
1989. A large number of shrub-like F. burtt-davyi Hutch. grow there as rock climbers on the steeper
N-E facing slopes of Gun Fire Hill. Two other indigenous Ficus species, F. thonningii Bl. and F. sur
Forssk., also grow in the gardens.
The pollinators of these three Ficus species are Elisabethiella
baijnathi Wiebes, E. stuckenbergi Grandi and Ceratosolen capensis Grandi respectively.
The
development of figs in both F. burtt-davyi and F. thonningii crops are well synchronised on individual
trees but not between the trees, and this prevents the wasp populations from cycling on the same trees
(Compton et ai., in press). This is not the case with F. sur, whose crops often contain figs at all stages
of development (Baijnathi and Ramcharun, 1983; Compton et ai., in press).
Wasp Aerial Densities
A single vertical black pole, 20 cm in diameter and 460 cm in length, was placed vertically amongst the
F. burtt-davyi trees about 20 m from the nearest fig tree. A continuous series of sticky traps, consisting
of nine cellulose transparencies sprayed with pruning sealant (Frank Fehr, Durban) were placed along
the length of the pole. The traps were 60 cm in length except for those at the bottom and top of the pole,
which were 20 cm long. The traps were exposed over six non-consecutive weekly periods in February,
March, July and August.
102
A "snap-shot" of wind speed variation with height at the site of the pole was obtained by measuring wind
speeds at nine different heights (between 0.1 and 4.5 m above ground level), using a hand-held Casella
low speed air meter.
Wind speed estimates were obtained on 10 different days between 12hOO and
15hOO. Based on these results the average wind profile for the site was produced. This information,
together with the numbers of wasps trapped at the site, was used to estimate the relative aerial densities
of fig wasps in the area.
Fig Wasp Flight Speeds
We estimated the flight speeds of three wasp species reared from F. burtt-davyi. These were E. baijnathi
and two non-pollinating species, Philotrypesis sp. and Sycoryctes (= Sycoscapter sensu Boucek, 1988)
sp. In preliminary experiments it was established that the wasps preferred to fly upwards rather than
horizontally on take-off, and flight speed estimates were for near-vertical flight.
Newly emerged
individuals were placed into a box (115 em high X 22 em long X 22 em deep) with the top and one
length composed of transparent sheets. The box was placed under an incandescent lamp in a room
maintained at 25°C.
Using a stop watch, the flight speeds of ten individuals of each species were
measured over a distance of 1 m from take off from the base of the box.
Flight Direction
The direction from which fig wasps flew to traps baited with receptive figs was examined using single
sticky traps plaee on poles at a height of 1.2 m, below to cotton bags containing receptive figs. As
insufficient pollinators of F. burtt-davyi were available we used figs of F. thollningii. The trial, with
three replicate traps, was initiated at 07hOO and terminated 5 hours later. After noting the general wind
direction, the sticky traps were removed, divided into 10 equal vertical sections and the number of wasps
trapped in each section was recorded. Wind speeds were also monitored at irregular intervals during this
period.
103
Dispersal from Natal Trees and Arrival at Trees Bearing Figs Ready for Pollination
The movements of fig wasps leaving their natal trees and arriving at receptive trees were investigated
using arrays of sticky traps. Poles (each supporting three sticky traps (30 X 21 cm) placed at 0.5, 1 and
2 m above ground level) were placed approximately 4 m from the F. bunt-davyi trees.
When the
topography allowed, eight poles were placed equidistantly about the tree, but where this was not possible
certain poles were omitted. The traps were replaced daily, when the numbers and identities of the wasps
trapped were recorded. Prevailing daily wind directions were also noted. The temperature during the
observation periods ranged between 25 and 27'C.
RESULTS
Wasp Aerial Densities
Although there was considerable variation in wind speeds between days, the pattern of increasing wind
speed with height was consistent and ranged from near zero velocity at 0.1 m from the ground up to an
average of nearly 10 km/hr at a height of 4.5 meters (Figure 1). This information, together with the
numbers of wasps trapped at each height on the vertical pole, allowed wasp density profiles to be
estimated. No fig wasps were trapped below 0.1 m. Between this level and 1.7 m, the average density
of fig wasps (all species) increased. Wasp densities then plateaued out until 4.5 m, after which they
increased markedly (Figure 2). Among the three commonly trapped wasp species, the same trend was
evident for the two species of pollinating wasp.
However, the density of the non-pollinating wasp,
Phagoblastus barbarus Grandi (which is associated with F. thonningii), remained relatively constant over
the range of heights examined (Figure 3). Based on the average wind speeds experienced in the area,
the wasp densities (all species) varied in the ratio of approximately 1: 2: 4 at heights of 0.5, 1, and 2
m respectfully.
104
15 l
14l
l
~
11
10
9~
o
W
ill
(L
(j)
o
z
S
7~
J
I
I
8~
!I
2J
1
OlL-~
1
I
__~
0:1
0.5
1.1
__~
1.7
2.3
2.9
HEIGHT (m)
__~
3.5
__~_
4.1
4.5
FIgUre 1. The wind profile at various heights above ground at a site in the 1820 Settlers Botanical Gardens, Grahamstown, South
Africa averaged from 10 occasions throughout the year (I kmlhr
27.8 em/sec). The large standard deviations are indicative of
the large variation in wind speeds experienced in the area.
=
120
......... 100
(")
b
,...
X
-
E
~
U)
Z
80
60
W
0
(L
~
(j)
40
I
20
0-,---,--0.1
0.5
1.1
1.7
2.3
2.9
TRAP HEIGHTS em)
3.5
Figure 2. Densities of total fig wasps at different heights in the 1820 Senlcrs B<;tanical Gardens.
105
4.1
4.5
80
E. stuckenbergi
1
60
40
20
..-..
JIli_.a~-"
0;
"?
~
x
:s
~
20
1
!
~
I
__.----",,,1L.
E. baijnathi
15
U5 10
Z
i
ill
5i
0..
0 ....1 _ _ __
o
i
C/)
~
2°1
i
P. barbarus
15 i
10
i
!W
j
5;
1
•
i
O~
:
i
0.1
0.5
1.1 1.7 2.3 2.9 3.5 4.1
4.5
TRAP HEIGHTS (m)
Figure 3. Densities of three fig wasp species at different heights in the 1820 Settlers Botanical Garden in Grahamstown.
Flights Speeds
In the laboratory the estimated flight speeds of the three wasp species ranged between 11 and 37 cm/sec
(equivalent to 0.4 and 1.3 kmlhr respectfully) (Table 1). This means that, based on the average wind
speeds (Figure 1), the wasps would have to fly at heights of less than 0.3 m above ground level if they
were to be able to maintain directional control (Figure 1). Gnder the. windiest conditions recorded they
would have to fly at less than 0.15 m above the ground.
Table 1. Fig wasp flight speeds measured at 2S·C over a distance of I m.
'Miii M
&
FLIGHT SPEEDS (em/sec)
SPECIES
n
MEAN
RANGE
=
Elisabethiella baljnathi
10
27.06
19.86 - 37.04
Philotrypesis sp.
10
20.67
11.00 - 34.38
Sycoryctes sp.
10
21.30
16.57-27.78
k*J4iW&mf'W
L
106
Wasp Dispersal from Natal Trees
Wasps leaving their natal trees flew mainly downwind (Table 2). Because of the rocky terrain and cliff
faces in the 1820 Settlers Botanical Garden, it was rarely possible to place a fuII complement of traps
around each tree.
However, all eight trap poles were positioned around one tree, allowing the
movements of the wasps to be assessed in detail (Figure 4). Only 2 % of the wasps recorded at this tree
were trapped upwind. Using the Rayleigh Test (Baschelet, 1981), which determines whether there is
evidence for bias in any given direction, it was found that the wasps moved in a mean preferred direction
of 11° from the recorded direction of the wind (mean angular deviation
= 21°; Z
= 1438; P
< 0.001).
0%
82%
1%
Wind Direction
Figure 4. The relative percentages of emigrating E. baijnazhi trapped around a wasp producing F. bum-davyi tree relative to the
prevailing wind direction. The small arrow indicates the mean preferred angle of wasp distribution and is flanked by an arc
indicating the mean angular variation (21").
Wasp Arrivals at Receptive Trees
E. stuckenbergi females were observed to retain directed flight when flying near bagged figs provided
wind speeds were low. No wasps were observed flying once wind speeds had increased to beyond 100
cm/sec. Unfortunately their small size did not permit behavioural observations if the wasps were further
than about 50 em from the bags. E. srucke!1bergi flying near the receptive figs displayed a casting
behaviour (a swaying flight 10 -20 em from the bags) before flying towards the bagged figs.
107
Significantly more wasps were trapped on the leeward side of the traps (Table 3) indicating the upwind
movement of the wasps towards the bagged receptive figs. In the more natural situation, wasps were
again trapped downwind of F. burtt-davyi trees bearing receptive figs (Figure 6).
Approximate
Wind Direction
Scale
50 wasps
Figure S. Numbers of E. slUckenbergi recorded at sticky traps baited with receplive F. rhonningii figs. The mean preferred
direction is indicated with a small arrow which is flanked by an arc representing the mean angular variation.
missing
26%
Wind Direction
n = 111
15%
Figure 6. The percentages of E. baijna/hi trapped around a receptive F. bllrll-davyi Iree relative to the prevailing wind direction.
108
'.
Table 2. Comparisons between the numbers of wasps caught at different heights upwind and downwind on F. burtt-davyi trees producing fig
wasps and those receptive trees attracting fig wasps. Producer trees have wasps emerging from the figs. Receptive trees have figs that are
ready to be pollinated.
~':."li!1m,f/euAtUI2Joj
...
TREE
HEIGHT
TRAP
DAYS
#
TRAPS PLACED
UPWIND
ft". .·X'SmwsswnrmfMWl:.l'wmM1)*$'
F
WASPS TRAPPED AT DIFFERENT HEIGHTS
DOWNWIND
UPWIND
DOWNWIND
me 8' nrevmemzermremsnst ,aa
0.5m
5 wsw'
1m
2m
-e'
'Iff'
CHI 2121
1m
2m
71
205
182
24.78
0
5
9
13.87
0.5
m
'W)~9i7§1rJ!
PRODUCER TREES
13
0.7
6
15
17
0
23
2.0
2
4
6
0
27
1.0
3
8
7
3
4
3
13
18
9
3.87
11
12
27
3
6
9
85
228
200
253.62
TOTAL
0
ill
5
RECEPTIVE TREES
34
2.5
7
15
15
28
55
33
58
121
50
32.93
12
0.5
4
10
12
5
3
12
29
9
26
21.01
13
0.7
3
9
6
9
10
4
6
7
4.74
99
1.8
2
6
4
5
12
16
2
32
18
31.10
36
2.3
2
5
5
2
5
2
15
8
4
0.09
17
45
43
41
84
73
108
176
105
45.03
TOTAL
trttmtrr?PF 1Rllre=tirE'?iWlrtwrmmm'r=wml3"i"""smwnnn tt"
ns = not significant; ••• =
p <
0.001
e '?JMI52F"~li4W$
P
?!W¥i~tlF.PN'fM-aA"1m3J:#S6H
·..
·..
·..
·..
ns
ns
·..
·..
ns
Wasp Densities in Relation to Trap Heights
More wasps were caught downwind than upwind, although the proportion caught downwind were much
higher around producer trees (30:1) than receptive tree (3:1) (Table 2; Figure 7). The high densities of
wasps trapped around wasp producing trees were a result of the synchrony of tree fruiting which resulted
in large numbers of wasps being trapped over short periods. Wasps leaving their natal figs were blown
downwind and most impacted on the nearby traps before gaining height (Table 2; Figure 6).
PRODUCER TREES
RECEPTIVE TREES
Trap
Heights
2%
2m
1%
1m
<1%
O.Sm
Wind Direction
Figure 7. The percentages of E. baijnarhi trapped on all sticky traps positioned at 0.5, I and 2 m above ground level surrounding
producer and receptive F. bllm-davyi trees relative to the prevailing wind directions.
The densities of wasps at different heights around the trees was expected to be similar to those recorded
away from the trees, unless the wasps had modified their behaviour. Upwind, wasp densities were
generally as predicted with most wasps collected on the more elevated traps (no significant deviations
were recorded around four of the five traps (Table 3; P
>
0.05). Therefore, when upwind of the trees,
the wasps did not modify their flying heights. In contrast, far more wasps than expected were captured
110
Tllble 3. Pollinating fig wasps (E. sllIckellbergi) trappeu at sticky traps near cotton bags containing receptive figs of F. tllOllllillgii. The direction from
where the wind was blowing was used as the reference point (a") for the circular statistical analysis.
_ _IIlII_ _"'iII
K",.*ew~r;g5s'tmWNfi!jT<lI:1=Z&S_
TRAP
NUMBER OF
WASPS TRAPPED
#
WASPS TRAPPED (%)
UPWIND
DOV,VNWIND
MEAN
PREFERRED
DIRECTION
'Zf-gI1Si;W$"~V59+!:XMp.\jJ§ltEA_¥e·*
......
RA YLEIGH TEST
z
ceg;*'1\?iM&@fItnPr¥SKWT53!/
34
66
174"37'
65'41'
17.76
2
176
30
70
165"01'
59"32'
37.07
3
675
22
78
174"38'
55"39'
182.52
P<
0.001
p
•
151
~wmrWfi'lg;"tas?!S&PezJR1:LM#.924·kcb\jp6y
-'"
MEAN
ANGULAR
DEVIATION
downwind of receptive trees on traps that were closest to the ground. The total number of pollinators
trapped at sticky traps placed at 0.5 m was equal to those trapped at the 2
ill
traps; some 4 times higher
than would have been expected. Similarly, the 1 m trap had more than 1.5 times a many wasps as the
2
ill
trap whereas the density of wasps expected at this height should have been half as many as the
number trapped at 2
ill
(Table 3; Figure 7).
These findings were significantly different from the
=
generally expected wasp densities of the area (Chj2~
149.2; P
< 0.001). Together with the increased
number of wasps trapped downwind these results suggest that wasps downwind of receptive trees alter
their general flight behaviour by losing height and t1ying upwind.
Relatively large numbers of E. stuckenbergi were trapped along with E. baijnathi at two of the receptive
F. burtt-davyi trees. This allowed us to examine whether there was some inherent component of the trees
which resulted in the large numbers of E. baijnarhi trapped low down on the leeward side of trees
bearing receptive figs independent of specific fig wasp behaviour. There were no differences in the
heights that the two species of wasps were trapped upwind. However, there were significant differences
in the heights that E. baijllathi were trapped downwind and those of E. stuckenbergi trapped either
upwind or downwind (Table 4). This was because the number of E. sTUckenbergi trapped at 0.5, 1 and
2 m did not differ significantly from the expected wasp densities for the area (ratio 1 :2:4) (Chi2[2J: upwind
=
3.46; P
> 0.5; downwind
=
0.138; P
> 0.05). More E. stuckenbergi were trapped upwind (30
wasps) than downwind '(24 wasps) of the two experimental trees.
Table 4. Comparison of numbers of two species of wasp trapped upwind and downwind of
two receptive trees of F. bllrll-davyi. The respective numbers of wasps trapped at 2m,
1m and O.Sm are given in parenthesis.
E. baijl/arhi
DOWNWIND
(33:15:33)
UPWIND
(22:12:6)
Chi
E. stuckellbergi
ns
= not significant;
•• »
2
1>1
P
Chi 2121
UPWIND
(22:18:1)
4.79
ns
21.65
DOWNWIND
(13:7:4)
0.92
ns
18.02
=P
<0.001
112
P
DISCUSSION
Wind, Wasp Densities and Flight Speeds
As expected (Taylor, 1960), wind speeds at the study site were consistently lowest near ground level and
increased with height. In unobstructed sites, Taylor (1960) showed that densities of small flying insect
decreased with height. This was Dot the case at our botanical gardens site, where the wasp densities were
relatively stable in the air column up to 4.5 m. This may reflect the vegetation and topography of the
site, as bushes in the vicinity of our trapping pole were approximately 2 m high.
Flight speeds in insects are closely related to their body size (Lewis and Taylor, 1974) with smaller
insects flying more slowly. Previous studies have shown that the vertical flight speed for the greenbug,
Schizaphis graminum (Rondani), ranges from 22-67 em/sec (Halgren, 1970) while that of another aphid,
Aphis fabrae Scopoli is between 20 and 30 em/sec (Kennedy and Booth, 1963). These species are of
comparable size to the fig wasps studied here, and their flight speeds are similar.
When the wasp's flight speeds are related to the wind speeds recorded at the study site, it is apparent
that the boundary layer for fig wasps, where controlled flight is possible, is normally less than 0.5 m
above ground level. The wasps therefore, have to fly close to vegetation or the ground if they are to
actively reach a host tree.
Dispersal from Natal Trees
Wasps leaving their natal trees initially fly upwards and are then taken downwind by the prevailing air
currents. In the laboratory, fig wasps are strongly attracted to light and, as with nitidulids (Blackmer
and Phelan, 1991) and migrating aphids (Kennedy and Booth, 1963; Kennedy and Ludlow; 1974, Robert,
1987), their initial vertical flight behaviour in the field may be phototactic. After this initial upward
movement, directional control would be lost once the insect entered the air column if the air was not too
unstable the wasps could, nonetheless to some extent, control their flight height.
113
As wind speeds
normally increase with height above ground level, the higher the wasps fly at this time the further they
are likely to disperse within any given time period. Given that the pollinating wasps are short-lived and
that receptive trees may be some distance away, this rapid dispersal from their natal trees may be
necessary in order to allow a chance for subsequent location of a suitable host plant.
Arrivals at Receptive Trees
Figs that are ready to be pollinated release volatiles that are attractive to flying wasps (van Noort et ai.,
1989; Ware et aI., in press; Ware and Compton, in prep.). Wasps utilising these volatiles will necessarily
detect them downwind of the trees and then need to fly towards the source of emission. The nature of
odour plumes and how insects use them to fmd their source has recently been reviewed by Murlis et al.
(1992). The observed casting (or zigzagging) anemotactic response of fig wasps when close to receptive
figs is similar to that of other insects tracking upwind in search of their hosts (Willis et al., 1991;
Nottingham and Croaker, 1987; Charlton and Carde, 1990). The increased numbers of E. baijl/athi
trapped close to the ground when downwind of the receptive F. burtt-davyl trees and the different heights
that E. stuckenbergi and E. baijnathi were trapped as they arrived implies that E. baijnathi alone was
responding to the volatile attractants by dropping out of the air colunm and then moving at low heights
upwind.
Compton and Robertson (in prep.) estimated that 95 % of adult female E. baijnathi produced in the 1820
Settlers Botanical Garden failed to find a receptive fig in which to oviposit. The short adult life span of
the adult fig wasps, predation and environmental effects such as dehydration, will ultimately limit the
distance that they can travel. Nevertheless, despite their small size, fig wasps are remarkedly efficient
colonisers of their host trees.
Given the mutualistic relationship between the fig trees and their
pollinating wasps, it is reasonable to speculate that evolutionary pressures have maximised the
effectiveness of the volatile attractants emanating from the figs. These studies described here suggest
how fig wasps utilise these cues to find their host figs.
These results allow us to produce an hypothesis that describes the way in which E. baijJlathi, in
114
particular, and perhaps fig wasps in general, manage to travel from fig tree to fig tree. An initial vertical
flight ensures that the wasps rapidly enter the air column where they are blown downwind.
On
perception of host tree volatile attractants the wasps lose height. Once in the boundary layer the wasps
use controlled upwind flight to search for the receptive figs releasing the volatiles.
This study complements a prevIOus investigation (Ware and Compton, submitted) where the role
temperature played on fig wasp emergence and pre-dispersal fig wasp behaviour were examined. These
studies have highlighted the roles that environmental factors play in the dispersal and host finding
behaviour of fig wasps.
ACKNOWLEDGEMENTS
We are grateful to Mr J. Cameron for permission to work in the 1820 Settlers Botanical Gardens. John
MacLaughlan and Mike Way are thanked for their field assistance. The financial support of Rhodes
University and the FRD to ABW is gratefully acknowledged.
REFERENCES
Baijnath, H. and Ramcharun, S. (1983). Aspects of pollination and floral development in Ficus capensis
Thunb. (Moraceae). Bothalia 14: 883-888.
Baschelet, E. (1981). Circular Statistics ill Biology. Academic Press Inc. London.
Blackmer, J.L. and Phelan, P.L. (1991). Behavior of Carpophilus hemipterus in a vertical flight
chamber: transition from phototactic to vegetative orientation. Emomol. expo appl. 58: 137-148.
Boucek, Z. (1988). Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, U.K.
Bronstein, J.L. (1989). A mutualism at the edge of its range. Experielltia 45: 622-636.
Charlton, R.E. and Carde, T.T. (1990). Orientation of male gypsy moths, Lymantria dispar (L.), to
pheromone sources: The role of olfactory and visual cues. 1. Insect Behav. 4: 443-469.
Compton, S.G. (1990). A collapse of host specificity in some African fig wasps. Sth. Afr. 1. Sci. 86:
39-40.
Compton, S.G., Rasplus, J.-Y. and Ware, A.B. (in press). African fig wasp parasitoid communities. In
115
Parasitoid Commullity Ecology (Eds Hawkins, B.A. and Sheenan, W.).
Compton, S.G., Thorton, LW.B., New, T.R. and Under;li1l, L. (1988). The colonization of Krakatoa
by fig wasps and other chalcids. Phil. Trails. R. Soc. Londoll (B) 322: 459-470.
Drake, V.A. and Farrow, R.A. (1989). The 'aerial plankton' and atmospheric convergence. Trends
Ecol. Evol. 4: 381-385.
Gautier-Hion, A. and Michaloud, G. (1989).
Figs: are they keystone resources for frugivorous
vertebrates throughout the tropics? A test in Gabon. Ecology 70: 1826-1833.
Galil, J. (1977). Fig Biology. Endeavour 1: 52-56.
Halgren, L.A. (1970). Flight behaviour of the greenbug, Schizaphis graminum (Homoptera: Aphididae),
in the laboratory. Anll. Emomol. Soc. Amer. 63: 938-942.
Kennedy, J.S. and Booth, C.O. (1963). Free fljght of aphids in the laboratory. 1. expo BioI. 40: 67-85.
Kennedy, J.S. and Ludlow, A.R. (1974). Co-ordination of two kinds of night activity in an aphid. J.
expo BioI. 61: 173-196.
Kjellberg, F., Doumesche, B. and Bronstein, J.L. (1988). Longevity of a fig wasp (Blastophaga psenes).
Proc. K. Ned. Akad. Wet. C 91: 117-122.
Lewis, T. and Taylor, L.R. (1974). Introduction to Experimental Ecology. Academic Press, New York.
McManus, M.L. (1988). Weather, behaviour and insect dispersal. Mem. ent. Soc. Can. 146: 71-94.
Murlis, J., Elkinton, J.S. and Carde, R.T. (1992). Odor plumes and how insects use them. Ann. Rev.
Ell!. 37: 505-532.
Nottingham, S.F. and Croaker, T.H. (1987). Changes in flight track angles of cabbage root fly, Delia
radicllm, in diffuse clouds and discrete plumes of host-plant volatile allylisothiocyanate. Ent.
expo appl. 43: 275-278.
Robert, Y. (1987). Aphids and their environment: Dispersal and migration. In Aphids: Their Biology,
Natural Enemies and COlllrol (Vol. 2A)(Eds Kinks, A.K. and Harrewijn, P.).
Taylor, L.R. (1958). Aphid dispersal and diurnal periodicity. Proc. Linn. Soc. Lond. 169: 67-73.
Taylor, L.R. (1960). The distribution of insects at low levels in the air. 1. Anim. Bioi. 29: 45-63.
Taylor, L.R. (1974). Insect migration, flight periodicity and the boundary layer. 1. Anim. Bioi. 43: 225238.
U!cnberg, S. (1985). The phylogeny of the genus Apocrypra Coqueral in relation to its hosts, CeratosoIell
116
Mayr (Agaonidae) and Ficus L. Verhand. K. Akad. Wet. 83: 149-176.
van Noort, S. (1992). The Systematics and Phylogenetics of the Sycoecinae (Agaonidae, Chalcidoidea,
Hymenoptera). Unpublished PhD thesis, Rhodes University, Grahamstown, South Africa.
van Noort, S. Ware, A.B. and Compton, S.G. (1989). Pollinator specific volatile attractants released
from the figs of Ficus burtt-davyi. Sth. Afr. 1. Sci. 85: 323-324.
Verkerke, W. (1989). Structure and function of the fig. Experientia 45: 612-622.
Ware, A.B. and Compton, S.G. (1992). Breakdown of pollinator specificity in an African fig tree.
Biotropica 24: 544-549.
Ware, A.B., Kaye, P.T. Compton, S.G. and van Noort, S. (in press). Fig volatiles: Their role in
attracting pollinators and maintaining pollinator specificity. PI. Syst. Evol.
Wharton, R.A., Tilson, J. W. and Tilson, R.L. (1980). Asynchrony in a wild population of Ficus
sycomorus. Sth. Afr. 1. Sci. 76: 478-480.
Wiebes, J.T. (1979). Co-evolution of figs and their insect pollinators. Ann. Rev. Evol. Syst. 10: 1-12.
Wiebes, J. T. and Compton, S. G. (1989). Agaonidae (Hymenoptera, Chalcidoidea) and Ficus (Moraceae):
Fig wasps and their figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet. 93: 203-222.
Willis, M.A., Murlis, J. and Carde, R.T. (1991). Pheromone-mediated upwind flight of male gypsy
moths, Lymantria dispar, in a forest. Physiol. Entomol. 16: 507-521.
117
CHAPTER 6
PERCEPTION OF VOLATILES
Paper 8: Preparation of small, delicate insects for scanning electron microscopy. Proceedings of the Electron
Microscopy Society of southern Africa 19; 39-40 (A.B. Ware and R.H.M. Cross - 1989).
Paper 9: Repeated evolution of elongate multiporous plate sensi11a in female fig wasps (Hyemoptera:
Agaonidae: Agaoninae). Proceedings of the Koninlijke Nederlandse Akademie van Wetenschappen 95;
275-292 (A.B. Ware apd S.G. Compton - 1992).
118
Elektronmikroskopieyereniging van Suideiike Afrika
ONDERSTEPOORT (1989)
PREPARATION OF SMALL, DELICATE INSECTS FOR SCANNING ELECTRON MICROSCOPY
A.B. Ware and R.H.M. Cross
Rhodes University, Grahamstown, South Africa
Problems likely to be encountered in the preparation of most
biological specimens for electron microscopy are minimised by having
access to living material at the outset, proceeding to one of a variety
of Ilapprovedll preparative procedures'. Unfortunately this ideal scenario
does not always present itself and researchers are often required to make
the best of what they have - sometimes material collected decades
previously and presented in a variety of Ilpreservatives
of dubious
nature and questionable efficiency.
ll
Fig wasps are small (length 3 mm), delicate insects which have
presented difficulties in preparation by conventional methods,
with
collapse of eyes and antennae being commonly encountered problems.
Although most conventional solvent-dependent preparative procedures have
relied upon aldehyde fixation with ethanol as the principal dehydrating
agent, the use of rapid heat-assisted air drying from acetone has been
reported 2,3 to be successful in the preparation of fresh and long-term
preserved material. Twelve different procedures were used to investigate
the effectiveness of the acetone treatments:
A. Cryo treatment of (1) fresh material quench-frozen in subcooled nitrogen, gold-coated and viewed on the SEM cryo stage.
B. Critical-point drying from liquid carbon dioxide after
glutaraldehyde fixation,
ethanol dehydration,
amyl acetate
transition and gold coating of (2) 20 year-old alcohol-preserved
material and (3) fresh material.
C. Acetone. treatment followed by accelerated hot air drying and
gold coating on: (4) fresh, (5) 48 hour frozen, (6) 48 hour
alcohol-preserved, and (7) 20 year old alcohol-preserved wasps.
D. Air-dried (4 days), gold-coated (8) 48 hour frozen and (9) 48
hour alcohol-preserved wasps.
E. Gold coating alone of (10) 48 hour alcohol-preserved, (11) 48
hour frozen and (12) freshly-collected wasps.
Although several of these treatments appear to be quite contrary to
the well-established
norms of specimen preparation for electron
microscopy, they were included for evaluation as they represent some of
the common means by which specimens are Ilpreserved in the field.
ll
CryoSEM gave good preservation of insect form; the major disadvantage
of this method being the physical damage caused during the freezing
process where appendages were easily lost or broken (fig 1). The other
conventional treatment, critical-point drying, was less successful in
preventing collapse of eyes and abdominal segments (fig 2).
All other
treatments showed some collapse of eyes, abdomen and/or antennae with the
ELECTRON MICROSCOPY SOCIETY OF SOUTHERN AFRICA - PROCEEDINGS - VOLUME 19 - 1989
119
worst case being (fig 3) when the heat-accelerated acetone vapourization
was terminated too early.
Acetone treatments appear to be as effective as critical-point drying
in attempting to preserve the natural appearance of long-term alcoholstored
specimens where artefacts arise in all cases during the
preparative process. The results of other treatments, while showing some
promise for the preparation of long-term preserved material, have been
inconsistent and are therefore inconclusive at this stage.
References
1. Cross, R.H.M. and Mansell, M.W. (1978) Proc. Electron Microsc. Soc.
South. Afr., 8, 67.
2. Truman, J.W. 11968) Ann. ent. Soc. Am. 61(3), 779.
3. Walpole, D.E., Coetzee, M. and Lalkhan,-C.M. (1988) J. ent. Soc. sth.
Afr. 51(2), 293.
Fig. 1. Fig wasp(arrowed). x 60
Fig. 2. Fig waspcollapse of eye (E),
Fig. 3. Fig wasp(treatment 4) showing
cryo
(treatment
1)
showing
loss
of
appendages
critical-point dried (treatment 3)
showing some
antennae (A) and abdomen (AS). x 66
abbreviated acetone treatment of fresh material
extensive collapse of eye and antennae. x 125
120
Proc. Kon. Ned. Akad. v. Wctensch. 95 (2), 275-292
June 22, 1992
Repeated evolution of elongate multiporous plate sensilla in female
fig wasps (Hymenoptera: Agaonidae: Agaoninae)
by A.B. Ware and S.G. Compton
Department of Zoology and Entomology. Rhodes University, Gra/wlIlslOwll. 6140.
Republic of SOUlh Africa
Communicated by Prof. J.T. Wiebes at the meeting of September 30,1991
ABSTRACT
Multiparous plate sensilla (MPS) are a characteristic feature of the antennae of chalcids
(Hymenoptera, Chalcidoidea). The elongate sensilla chaetica form of MPS occurs in the males of
many chalcid species, but is rare amongst females other than in fig wasps. Female fig wasps
(Agaonidae, Agaoninae) were classified according to the position, shape and size of their MPS. In
this group MPS elongation, with its concomitant increased surface area, has apparently evolved independently on at least nine occasions. This repeated evolution may be related to the life history
of fig wasps and their mutualism with figs.
INTRODUCTION
Multiparous plate sensilla (MPS), also known as multiparous pitted sensilla
(Zacharuk, 1980), thin walled sensilla (Slifer, 1970) or longitudinal sensilla
(Boucek, 1988), are a characteristic feature of chalcid (Hymenoptera,
Chalcidoidea) antennae (Boucek, 1988). Snodgrass (1925) distinguished two
forms of MPS: sensilla linearia (= sensilla placodea) are plate-like structures attached to the antennal segments over most of their length, while sensilla
chaetica are hair-like and detached from the antennal segments except at their
basco Sensilla linearia are almost ubiquitous among female chalcids (Miller,
1972; Weseloh, 1972; Voegele et aI., 1975; Barlin and Vinson, 1981; Dahms,
1984; Wibel et ai., 1984) and their possession can be considered as the
plesiomorphic condition. Sensilla chaetica have a more restricted distribution,
but are a feature of many male chalcids and some female fig wasps
(Agaonidae).
121
MPS are considered to have an olfactory function (Zacharuk, 1985) and
chalcids are assumed to use them to detect their hosts (Vinson, 1985). Sensilla
chaeticc: are typically more elongate than sensi/ia Iinearia. The functional
significance of sensilla elongation and its associated detachment from the
antennal surface may be related to an increase in receptor surface area, which
in turn should result in improved sensitivity. However, sensilla elongation is not
the only way in which increased receptor surface area can be achieved. An alternative is for the number of sensilla to be increased. This requires that the size
of the antennae be enlarged through the lengthening, branching or thickening
of some of the antennal segments.
The Agaonidae comprises wasps which have an intimate association with fig
trees (Ficus spp., Moraceae) (Boucek, 1988). The pollinating fig wasps belong
to the subfamily Agaoninae and are highly host specific (Michaloud et aI.,
1985; Wiebes and Compton, 1990). The relationship between trees and
agaonines is mutualistic, with the wasps both pollinating the trees and utilising
some of the ovules for egg laying (Galil, 1977; Janzen, 1979).
Fruit production on each fig tree is typically highly synchronized. This ensures cross-pollination, but means that females of each wasp generation must
seek out new trees before they can oviposit. Because the female wasps are shortlived (Kjellberg et al., 1988) they must locate a suitable tree quickly. The trees
are identified through species-specific volatile chemicals released from the figs
(Ware and Compton, in prep.). The wasps are only attracted by the volatiles
when the figs are 'receptive' and ready to be pollinated (Bronstein, 1987; van
Noort et al., 1989). Microscopic examination of fig wasp MPS has confirmed
that they are covered in the pores that are typical of olfactory receptors (Ware
and Compton, in prep.) and they are likely to be the organs by which female
fig wasps perceive their host figs.
This paper examines the MPS of female agaonine fig wasps and has two objectives: to record the presence and arrangement of the sensifla chaelica and
sensilfa /inearia, and to determine how often elongation of sensilla has evolved
within the subfamily. The functional significance of sensilla elongation is
discussed in relation to the life history of the wasps.
MATERIALS AND METHODS
Antennae from the females of 25 agaonid species were examined with a
dissecting stereomicroscope and a scanning electron microscope (JEOL JSM
84). The presence of either sensilla linearia or sensilla chaerica was noted,
together with their position on the antennal segments. A literature survey was
also undertaken to extend these observations to cover all but one of the described genera of Agaoninae.
Preliminary observations showed that, because there was a continuum of sensilla forms, the distinction between sensilla /inearia and sensilla chaetica was
not clear-cut. The following criteria were nonetheless adequate to distinguish
between them: sensi/la linearia were plate-like and were usually attached to the
antennae over all or most of their length. Where these sensilla extended beyond
122
the apical aspect of the antennal segment they were finger-like. In contrast, sensilla chaetica were attached to the antennae at their origins only, with the rest
of the structure free and ending in a distinct point. The lengths of both sensilla
linearia and sensilla chaetica were highly variable. For our analysis we defined
elongation as having occured jf the detached portions of the sensilla were at
least 1.5 times the length of the antennal segment to which they were attached.
Sensilla elongation was expected to result in an increase in the total surface
area of the MPS. To confirm this we examined the antennae of two congeneric
species, one with sensilla chaetica and the other with sensilla linearia. Based on
scanning electron micrographs, estimates of the numbers and average lengths
of their sensilla were produced. The exposed surface areas of individual sensilla
were then calculated.
Sensilla linearia approximate to cylinders, and we estimated that one third of
their surface area was attach"ed to the antennae. Their surface area was
therefore calculated as 2/3(2nrh + nr). Sensilla chaetica are cone-like with their
bases attached to the antennae. Their surface area was therefore calculated
using nrl/r 2 +h 2 •
RESULTS
Arrangements of sensilla
MPS were found on the club and funicle segments but never on the anelli,
pedicel or scape. The simplest form of MPS arrangement (designated Type I)
consisted of a single, although sometimes irregular, whorl of sensilla linearia
(Figure 1). In a modification of this arrangement, at least one antennal segment
had two or more whorls of sensilla linearia (Type II; Figure 2). Sensilla chaetica
also occured in two distinct arrangements. They either originated from the sides
of their antennal segment, sometimes from sockets (Type III, Figure 3), or from
sockets situate,d axially (Type IV; Figure 4).
Descriptions obtained from the taxonomic literature were adequate to assign
the antennae of 218 agaonine species to one of the four groups outlined above
(Table 1). Sensilfa chaetica were recorded in 22.5070 of the species while sensilla
linearia were found in 76.6%. The remaining two species possessed antennae
with both sensilla linearia and sensilla chaetica. Sensilla elongation was present
in 45 species, all of which had sensilla chaetica (Appendix 1a).
The Type I arrangement of sensilla linearia was recorded in 95 species and
Type II in 76 species. In the latter group there was considerable variation in the
number of MPS per whorl and the number of whorls per antennal segment. The
first funicle segment nonetheless consistently had only a single whorl of MPS,
even when the remaining segments had two or more.
A degree of sensilla detachment was noted among some of the species with
sensilla linearia. This was most pronounced in Elisabethiella pectinata (Joseph)
(Figure 5), Platyscapa bergi Wiebes, Pegoscapus tomentellae Wiebes and
Pegoscapus tonduzi (Grandi). In Platyscapa quadraticeps (Mayr) the MPS on
123
Figs. 1-4. Scanning electron micrographs of agaonid antennal segments illustrating the four types
of MPS arrangements. 1. Elisabethiella sl!/ckenbergi (Type 1). 2. Allotriozoon helerandromorphulIl
(Type II), 3. Courtella armata (Type III) and 4. Elisabethiella baijna/hi (Type IV).
124
Table I. The distribution of the major antenna! sensilla arrangements within the genera of
Agaoninae. See the text for description of the types of MPS arrangements.
Tribes
Agaonini
Blastophagini
Genus
Elisabethiella
Nigeriella
Courtella
Agaon
Allotriozoon
Paragaon
A fjonsiella
PleistodoJ1les
Tetrapus
Dolichoris
Blaslophaga
Wiebesia
Liporrhopalum
Platyscapa
Maniella
Deilagaon
Waterstoniella
Eupristina
Pegoscapus
Kradibia
Ceratosolen
Total Agaoninae
Number of
species examined
15
4
13
Antennal sensilla
arrangement types
+
+
II
III
IV
+
+
+
+
+
+
12
+
3
2
+
+
7
9
3
+
+
+
8
+
11
+
+
+
I
9
13
+
+
+
0
?
3
8
6
63
+
+
+
+
+
220
95
20
10
?
+
+
+
+
?
+
+
+
?
+
+
+
+
76
39
10
funicle segments 5-7 were typical sensilla linearia attached over their entire
length, whereas some of those on segment 8 arose apically and were attached
only at their or!gins (Figure 6).
The Type III arrangement of sensilla chaetica was recorded in 39 species.
They ranged from the short stocky sensilla of Blastophaga silvestriana Grandi
(Figure 7) to the long slender hair-like MPS of Blastophaga clavigera (Mayr)
(Figure 8). The Type IV sensilla arrangement was found in 11 species. The two
species with both sensilla linearia and sensilla chaetica (Nigeriella jusciceps
Wiebes and B. clavigera) had a combination of Type I and Type III MPS arrangements (Figure 8).
Distribution oj Sensilla Arrangements
The Agaoninae can be divided into two tribes, the Agaonini and the
Blastophagini (Boucek, 1988). Sensilla lineat'ia were recorded in five and sensilla chaetica in six of the nine genera of the Agaonini (Table 1). Elisabethiella
was the only genus in which all four sensilla types were found, although combinations were also recorded in two other genera.
125
Figs. 5-1 J. Agaonine antennal MPS arrangements 5. Fifth antennal segment of Efisabethiella peclinata (redrawn from Joseph, 1959). 6. Eighth segment of Platyscapa quudraticeps (redrawn from
Grandi, 1923).7. Sixth segment of Blastophaga silves/riana (redrawn from Hill, 1967).8. Tenth
antennal segment of Blastophaga clavigera (redrawn from Grandi, J 928). 9. Elongation of antennal
segments as seen in the second funicle segment of Cera/oso/en tentacularis (redrawn from Grandi,
1928). 10. thickening of antennal segments as in Deilagaon chrysolepidis (redrawn from Boucek,
1988). II. branching of the seventh antennal segment of Dolichoris flabelluta (redrawn from
Wiebes, 1978).
Sensilla linearia were recorded from almost all the Blastophagini, while sensilla chaetica were more restricted in distribution. When the two forms are compared at species level, sensilla chaetica are clearly rarer in the Blastophagini (in
12 out of the 152 species surveyed compared with 38 of 68 species).
126
Elongation of sensilla
Elisabethiella stuckenbergi (Grandi), a species with the Type 1 arrangement
of sensilla linearia (Figure 1), had an estimated total MPS surface area only half
that of Elisabethiella baijnathi Wiebes, a species with the Type IV arrangement
of sensilla chaetica (Figure 4; Table 2). This was despite E. stuckenbergi being
the larger of the two species and having more individual sensilla.
MPS elongation was recorded in 11 of the 21 genera. Seven of these also include species with sensilla that are not elongate, showing that elongation has
occured independently in each genus (Figure 12). Using the phenogram
modified from Wiebes (1982) and assuming an absence of reversals, it appears
that elongation of the MPS has arisen at least four times in the Agaonini (i.e.
in Elisabethiella, Nigeriella, AlfonsiellalParagaon and Courtellal Agaon) and
five times in the Blastophagini (in Dolichoris, Blastophaga G, Liporrhopa/um,
Eupristina and Pegoscapus). In total, sensilla elongation may therefore have
evolved on at least nine occasions within the Agaoninae. If elongation also
evolved independently within congeners then this figure will be an
underestimate.
Antennal modifications
We recorded only isolated examples of structural modifications to the antennae that would allow an increased number of sensilla to be carried. Antennal
segment elongation is present in Ceratosolen tentacularis (Grandi) (Figure 9)
and Liporrhopalum /ongicornis (Grandi); there is antennal thickening in all
Deilagaon spp. (Figure 10), and the antennae of Ceratosolenflabellatus Grandi
and Dolichoris flabellata Wiebes are branched (Figure 11).
DISCUSSION
Sensilla elongation and detachment has evolved repeatedly in the females of
agaonines, but not in the females of other chalcids, where elongation has occured mainly in males. This suggests that female fig wasps and the males of
other chalcids share common advantages in possessing elongate sensilla with
their correspondingly greater surface area. Alternate ways that surface area can
be increased include the elongation, thickening or branching of antennal
Table 2. A comparison of the exposed surface areas of the MPS on the antennae of fig wasps with
sensilla linearia and sensilla chaetica. Numbers of sensilla and their total surface area refer to pairs
of antennae.
Species
MPS
Form
Elisabethiella
stuckenbergi
Elisabethiella
baijnathi
Sensilla
linearia
Sensilla
chaetica
Type
IV
Total number
of sensilla
Area per
sensilla (mm2)
146
0.55
80
106
1.56
166
127
Total surface area
of sensilla (mm2)
Ceratoso/en
Kradibia
Blastophaga (C&D)
Wlebes/a
Manie/la
Pegoscapus
Eupristina
Blastophagini - - t - - l - Waterstoniel/a
Blastopnaga G
Deilagaon
P/atyscapa
Uporrnopa/um
Blastopnaga A
Do/ichoris
Agaoninae
Elisabeth/ella
Nigerie/la.
Courte/la
Agaon
Allotriozoon
Paragaon
Agaonini - - - I
AlfonsieJ/a
r--- Pleistodontes
-
?
m
I
m
!@
I
I
!~
I
Io
Tetrapus
100%
I
I
Fig. 12. The phylogeny of Agaonine genera (modified from Wiebes, 1982) and evolution of
elongate MPS.
,
segments, all of which are found in male chalcids in the families Pteromalidae,
Eulophidae and Encyrtidae. Such antennal modifications are comparatively
rare among female chalcids, however, including agaonids. Why the antenna!
enlargement is uncommon amongst female fig wasps is uncertain, but could be
related to the narrow confines of the figs through which the wasps must crawl
after emerging from their natal galls.
Antennae with large surface areas are likely to have developed among male
chalcid wasps to improve their efficiency at finding mates. In fig wasps, males
seek out females and mate with them before the latter leave their natal galls,
so any modifications of the female antennae are unlikely to be related to mate
detection or recognition. We therefore suggest that the repeated evolution of
elongate and detached sensilla in female agaonids has resulted from their need
to detect trace quantities of volatiles in order to find suitable oviposition sites
(van Noort et aI., 1989). As MPS elongation is evident within several different
lineages of agaonines, such selection pressures acting on host finding ability
have clearly been important during the evolution of fig wasps. In addition to
the elongation of sensifla chaetica there has also been a trend towards the place-
128
ment of the sensilla into sockets on the surface of the antennal segment. In at
least one agaonid species these allow the sensilla to be flared, which may further
increase their sensitivity (Nijhout and Sheffield, 1979; Ware and Compton, in
prep.).
ACKNOWLEDGEMENTS
We would like to thank P.E. HuIIey, S. Vincent, S. van Noort and C.
Zachariades for their constructive comments on the manuscript and the Electron Microscope Unit for their help with the photography. The post-graduate
bursary support by F.R.D. and Rhodes University to ABW is gratefully
acknowledged.
REFERENCES
Badin, M.R. & S.B. Vinson - Multiporous plate sensilla in antennae of the Chalcidoidea
(Hymenoptera). Int. J. Insect Morpho!. Einbryol. 10,29-42 (1981).
Boucek, Z. - Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, U.K.,
156-209 (1988).
Boucek. Z., A. Watsham & J.T. Wiebes - The fig fauna of the receptacles of Ficus thonningii
(Hymenoptera, Chalcidoidea). Tijdschr. Ent. 124, 149-231 (1981).
Bronstein, J.L. - Maintenance of species-specificity in a neotropical fig-pollinator wasp mutualism.
Oikos 48, 39-46 (1987).
Dahms, E.C. - An interpretation of the structure and function of the antennal sense organs of
lvfelil/obia australica (Hymenoptera, Eulophidae) with the discovery of a large dermal gland
in the male scape. Mem. Qd Mus. 21, 361-385 (1984).
Grandi, G. - Identification of some fig insects (Hymenoptera) from the British Museum (Natural
History). Bull. en!. Res. 13, 295-299 (1923).
Grandi, G. - Rivisione critica degli Agaonidi descritti da Gustavo Mayr e Catalogo ragionato delle
specie fino ad oggi descritte di tutto iI mondo. Boll. Lab. Ent. Bologna I, 107-210 (1928).
Galil, J. - Fig biology. Endeavour I, 52-56 (1977).
Hill, D.S. - Fig wasps (Chalcidoidea) of Hong Kong.!. Agaonidae. Zool. Verh. Leiden 89, 2-55
(1967).
Janzen, D.H. - How to be a fig. Ann. Rev. Ecol. Syst. 10, 13-51 (1979).
Joseph, K.J. - On a collection of fig insects (Chalcidiodea: Agaontidae) from French Guinea. Proc.
Roy. ent. Soc. London ·28,29-36 (1959).
Kjellberg, F., B. Doumesche & 1.L. Bronstein, - Longevity of a fig wasp (Blastophaga psenes).
Proc. Kon. Ned. Akad. Wet. (C) 91, 117-122 (1988).
Michaloud, G., S. Michaloud-Pelletier, J .T. Wiebes & C.C. Berg - The occurrence of two
pollinating species of fig wasp and one species of fig. Proc. Kon. Ned. Akad. Wet. (C) 88,
93-119 (1985).
Miller, M.C. - Scanning electron microscope studies of the flagellar sense receptors of Peridesmia
discus and Nasonia vilripennis (Hymenoptera: Pteromalidae). Ann. en!. Soc. Am. 65,
1119-1124 (1972).
Nijhout, H. & H. Sheffield - Antennal hair erection in male mosquitoes: A new mechanical effector
in insects. Science 206, 595-596 (1979).
Slifer, E.H. - The structure of arthropod chemoreceptors. Ann. Rev. Ent. IS, 121-142 (1970).
Snodgrass, R.E. - The morphology of insect sense organs and the sensory nervous system. Smith.
Miscel. ColI. 77, 1-80 (1925).
Van Noort, S., A.B. Ware & S.G. Compton - Pollinator-specific volatile attractants released from
the figs of Ficus burtt-davyi. Sth Afr. J. Sci. 85, 323-324 (1989).
129
Vinson, S.B. - The behavior of parasitoids. In Comprehensive Insect Physiology, Biochemistry and
Pharmacology (Volume 9) (Eds. Kerkut, G.A. and Gilbert, L.l.) Pergammon Press Ltd., Oxford, 417-470 (1985).
Voegele, J., J. Cals-Usciati, J.-P. Pihon & J. Daumal - Structure de l'antenne femelle der
Trichogrammes. Entomophaga 20, 161-169 (1975).
Weseloh, R. - Sense organs of the hyperparasite Cheiloneurus noxius (Hymenoptera: Encytridae)
important in host selection process. Ann. ent. Soc. Am. 65, 41-46 (1972).
Wibel, R.G., J.D. Cassidy, H.E. Buhse, M.R. Cummings, V.P. Bindokas, J. Charlesworth, &
D.L. Baugartner - Scanning electron microscopy of the antennal sense organ of Nasonia
vitripennis (Hymenoptera: Pteromalidae). Trans. Am. Micro. Soc. 103, 329-340 (1984).
Wiebes, J.T. - The fig wasp genus Dolichoris Hill (Hymenoptera, Chalcidoidea, Agaonidae). Proc.
Kon. Ned. Akad. Wet. (C) 82, 181-196 (1978).
Wiebes, J. T. - The phylogeny of the Agaonidae (Hymenoptera, ChaJcidoidea). Neth. J. Zool. 32,
395-411 (1982).
Wiebes, J .T. & S.G. Compton - Agaonidae (Hymenoptera Chalcidoidae) and Ficus (Moraceae): fig
wasps and their figs, IV (Africa concluded). Proc. Kon. Ned. Akad. Wet. (C) 93, 203-222 (1990).
Zacharuk, R.Y. - Ultrastructure and function of insect chemosensilla. Ann. Rev. Ent. 25, 27-47
(1980).
Zacharuk, R. Y. - Antennae and scnsilla. In Comprehensive Insect Physiology, Biochemistry and
Pharmacology (Volume 6) (Eds. Kerkut, G.A. and Gilbert, L.l.) Pergammon Press Ltd., Oxford, 1-69 (1985).
Appendix la. The structure and pOSitIOn of the multiporous plate sensilla (MPS) of female
agaonines. Data were derived from the literature and lor from examining dry-mounted specimens
(*). A + indicates those species which possess elongated sensilla. See text for a description of the
types of MPS arrangements.
Antennae
Type
References
(Appendix Ib)
AGAONINI
Grandi
allotriozoonoides (Grandi) 1916
articulata (Joseph) 1959
bergi Wiebes 1989
baijnathi Wiebes 1~89
comptoni Wiebes 1989
dyscritus (Waterston) 1920
enriquesi (Grandi) 1916
glumosae Wiebes 1989
hilli Wiebes 1989
longiscapa Wiebes 1986
pectinata (Joseph) 1959
plalyscapa Wiebes
rejlexa Wiebes 1975
socolrensis (Mayr) 1885
stuckenbergi (Grandi) 1955
ELlSABETHIELLA
III +
IV +
I
IV +
I
I
II
Wiebes
avicola Wiebes 1975
excavala Compton 1990
jusciceps Wicbes 1974
lelou::.eyi Wicbes 1974
7,61
23,61
61*
61*
61*
30
5,61 *
61*
61
59,61
23,61
61
48,61
16,61*
18,61*
NIGERIELLA
III &1 +
130
48,65
65*
47,65
47,65
Appendix la. (Contd.)
Kieffer
armata (Wiebes) 1974
bekiliensis (Risbec) 1956
bispinosa (Wiebes) 1969
camerunensis (Wiebes) 1974
gabonensis Wiebes 1985
hladikae (Wiebes) 1979
malawi Wiebes 1990
medleri (Wiebes) 1972
michaloudi (Wiebes) 1979
penicula (Wiebes) 1974
scobiniferum (Waterston) 1920
sy/viae Wiebes 1986
wardi Compton 1990
Antennae
Type
References
(Appendix lb)
III +
46,59*
43*
42
24,46
63
54
63
45
54*
24,46
27,46
56,59
65*
COURTELLA
Dalmon
acuta/um Wiebes 1989
balio/um Wiebes 1974
cicatriferens Wiebes 1989
fascia/um Waterston 1914
gabonese Wiebes 1989
kiellandi (Wiebes) 1974
megalopon Wiebes 1976
oblusum Wiebes 1989
paradoxum (Dal) 1818
spatuialUm Wiebes 1968
taiense Wiebes 1989
lridentalum Joseph 1959
lll+
1II+
1II+
III +
III +
III +
III +
III +
III
III +
1Il+
III +
AGAON
Grandi
he/erandromorphum Grandi 1916
prodigiosum Grandi 1916
nigeriense Wiebes 1974
III
III
III
III
III
+
+
+
+
+
IiI +
III +
III +
III +
IlI+
III +
1lI+
63
46,63
63
27,63
63
46,63
49,63
63
5,32,41,63
41,63
63
23
II
II
II
5*
ALLOTRIOZOON
Joseph
perp/exum Joseph 1959
josephi Wiebes 1986
5*
46
PARAGAON
Waterston
bergi Wiebes 1988
binghami Wiebes 1988
brongersmai Wiebes 1972
fimbriala Waterston 1920
longiscapa Joseph 1959
michaloudi Wiebes 1988
nalalensis Wiebes 1972
III +
III +
23,49
59
IV
IV
IV
IV
IV
IV
IV
60
60
4,44
29
4,23,44
60*
44
ALFONSIELLA
Saunders
blandus Wiebes 1963
froggatti Mayr 1906
+
+
+
+
+
+
+
PLEISTODONTES
33,40
6
131
Appendix la. (Contd.)
Antennae
Type
greenwoodi (Grandi) 1928
immaturus Wiebes 1963
imperia/is Saunders 1883
nitens (Girault) 1915
p/ebejus Wiebes 1963
rennellensis Wiebes 1968
rieki Wiebes 1963
II
I
II
Mayr
american us Mayr 1885
costaricanus Grandi 1925
mexican us Grandi 1952
References
(Appendix lb)
15
33
16
17
33
40
33
TETRAPUS
16
11
17
BLASTOPHAGINI
Hill
boschmai (Wiebes) 1964
cristala (Grandi) 1928
f/abellata Wiebes 1978
in ornata Wiebes 1979
nervosae (Hill) 1967
umbilicata Wiebes 1979
valentinae (Grandi) 1916
vasculosae Hill 1967
DOLICHORIS
II
1lI+
III +
I
II
I
II
Gravenhorst
c!avigera Mayr 1885
errata Wiebes 1966
gomberti Grandi 1928
inopinata Grandi 1926
intermedia Grandi 1926
javana Mayr 1885
psenes (Linnaeus) 1758
puncticeps Mayr 1906
pumilae Hill 1967
quadrupes Mayr 1885
silvestriana Grandi 1929
56
14,53
53
53
20,53
53
53
20
BLASTOPHAGA
II
16
39
13
16
16
16,20
II
II
II
II
16
20
16
20
Boucek
cOlllubernalis (Grandi) 1927
II
12
Waterston
cuspidatae Hill 1969
dubium (Grandi) 1926
giacomillii (Grandi) 1926
gibbosae Hill 1967
longicornis (Grandi) 1926
midotis Hill 1969
phillippinensis Hill 1969
rutherfordi Waterston 1920
II
II
II
III +
II
II
III +
III +
21
16
16
20,21
16
21
21
21,27
III & I +
II
II
II
,.
WIEBESIA
LIPORRHOPALUM
132
Appendix la. (Contd.)
subula/ae Hill 1969
uniglandulosae Hill 1969
References
(Appendix Ib)
II
III
21
21
Motschoulsky
PLATYSCAPA
arnoltiana Abdurahiman 1980
awekei Wiebes 1977
bergi Wiebes 1986
binghami Wiebes 1980
coronata (Grandi) 1928
corneri Wiebes 1980
deserlorum Compton 1990
eliennei Wiebes 1977
fisheri Wiebes 1977
ishiiana (Grandi) 1923
quadraticeps (Mayr) 1985
soraria Wiebes 1980
Ijahela (Abd & Joseph) 1975
DElLAGAON
Antennae
Type
64
51*
59
64*
II
III
14,20,51
64
65*
51
51
10,20,51
9,51
II
64*
I
1,51
II
II
II
50
3,50
7,10,50
II
II
II
II
II
57
10
57
57
10
55
57
57
Wiebes
annulalae Wiebes 1977
chrysofepidis Wiebes 1977
megarhopalum (Grandi) 1924
WATERSTONIELLA
Grandi
borneana Wiebes 1982
fiord Grandi 1924
javana Wiebes 1982
mafayana Wiebes 1982
masii (Grandi) 1921
solomonensis Wiebes 1980
suma/rana Wiebes 1982
williamsi Wiebes 1982
EUPRISTINA
Saunders
altissima Bal. & Abd. 1981
aurivillii Mayr 1906
bakeri Grandi 1927
belgaumensis Joseph 1954
masoni Saunders 1883
verticil/ala (Waterston) 1921
PEGOSCAPUS
IV +
2
16
12
3,22
26
20,31
III +
8
17
16
16
25
25
8
58
II
II
Cameron
aguilari (Grandi) 1919
baschieri (Grandi) 1952
bifossulalus Mayr 1885
brasiliensis Mayr 1928
carlosi (Ramirez) 1970
cumanensis (Ramirez) 1970
eSlherae (Grandi) 1919
flagella/us Wiebes 1983
133
Appendix 10. (Contd.)
Antennae
Type
ileanae (Ramirez) 1970
jimenezi (Grandi) 1932
kraussii (Grandi) 1952
mariae (Ramirez) 1970
oroczoi (Ramirez) 1970
silvestrii (Grandi) 1919
stand/eyi (Ramirez) 1970
lomentel/ae Wiebes 1983
/onduzi (Grandi) 1919
Irislani (Grandi) 1919
urbanae (Ramirez) 1970
williamsi (Grandi) 1923
References
(Appendix Ib)
25
8
17
25
25
8
25
58
8
8
25
9
Saunders
brownii Ashmead 1904
copiosae (Wiebes) 1980
cowani Saunders 1883
gestroi (Grandi) 1916
hilli Wiebes 1978
jacobsi (Wiebes) 1964
nigricorpus (Girault) 1915
setigera Wiebes 1978
sumatrana (Grandi) 1926
wassae (Wiebes) 1980
KRADIBlA
II
II
II
II
II
II
I
II
I
II
52
55
26,52
52*
52,65*
37
3
52
16,52
5S
II
34,56
5,62
38,56
34
20
7*
34,56
12,56
34,56
38,56
16
62
34
34
34
18,36,62"
18
62
16,20
34
34
10
16
5
CERATOSOLEN ~ayr
abnormis Wiebes 1963
acutalus Mayr 1906
adenospermae Wiebes 1965
a/bulus Wiebes 1963
appendicu/a/us (Mayr) 1885
arabicus ~ayr
1906
armipes Wiebes 1963
baked Grandi 1927.
bianchii Wiebes 1963
bimerus Wiebes 1965
bisll!catus (~ayr)
1885
blommersi Wiebes 1989
boschmai Wiebes 1963
brongersmai Wiebes 1963
ca/opilinae Wiebes 1963
capensis Grandi 1955
carayoni Grandi 1963
coeclis (Coquere1) 1855
cons/rictus (~ayr)
1882
corneri Wiebes 1963
dentifer Wiebes 1963
efisabethae Grandi 1923
emarginatus ~ayr
1906
leae Grandi 1916
II
II
II
I
II
II
II
II
II
1
II
1
134
Appendix la. (Contd.)
f/abelalus Grandi 1916
Jusciceps Mayr 1906
gaiili Wiebes 1964
grandii Wiebes 1963
gressilli Wiebes 1980
hewitti Waterston 1920
hooglandi Wiebes 1963
humalus Wiebes 1963
imbecilis Grandi 1927
immanis Wiebes 1981
indigenus Wiebes 1981
internalus Wiebes 1978
iodotrichae Wiebes 1963
josephi Wiebes 1963
jucundus Grandi 1927
julianae Grandi 1916
iongicornis Joseph 1959
longimucro Wiebes 1989
medlerianus Wiebes 1980
modera/us Wiebes 1963
namorokensis Risbec 1956
nanus Wiebes 1963
nexilis Wiebes 1979
notus (Baker) 1913
nuga/orius Grandi 1952
orientalis Wiebes 1963
pi/ipes Wiebes 1963
praestans Wiebes 1963
pygmaeus Grandi 1927
silvestrianus Grandi 1916
solitarius Wiebes 1980
solmsi complex Mayr 1885
sordidus Wiebes 1963
stupeJactus Wiebes 1989
tentacularis (Grandi) 1926
vechti Wiebes 1963
vissali Wiebes 1981
Antennae
Type
References
(Appendix Ib)
II
I
5*
62
36,62*
34
55
31
34
34
12
56
56
52
34
34
I
II
II
II
II
II
II
II
12
I
II
II
11
I
II
II
II
11
II
II
II
II
I
II
II
II
II
5
23
62
55
34
43,47
34
55,62
35
17
34
34
34
12
5
55
20
34
62
16
34
56
AppendLr: lb. References referred to in Appendix 1a.
I. Abdurahiman, U .C. & K.J. Joseph - Three new Chalcidoidea (Hymenoptera) from India.
Orient. Ins. 9, 99-109 (1975).
2. Balakrishman Nair, P., M. Joseph & U.C. Abdurahiman - New fig wasps (Hymenoptera:
Chalcidoidea) from Ficus altissima. Proc. Kon. Ned. Akad. Wet. (C) 84, 145-153 (1981).
3. Boucek, Z. - Family Agaonidae. In Australian Chalcidoidea (Hymenoptera). C.A.B. International, Wallingford, U.K., 156-209 (1988).
4. Boucek, Z., A. Watsham & J .T. Wiebes - The fig wasp fauna of the receptacles of Ficus thonningii (Hymenoptera Chalcidoidea). Tijdschr. Ent. 124 149-231 (1981).
5. Grandi, G. - Gli Agaonini (Hymenoptera Chalcididae) raccolti nell' Africa occidentale dal
Prof. F. Silvestri. Boll. Lab. Zoo I. Portici 10, 121-285 (1916).
135
6. Grandi, G. - Nota su due Agaonini (Hymenoptera Chalcididae) dell' Australia. Boll. Lab.
Zool. Portici 11, 145-159 (1916).
7.Grandi, G. - Contributo alIa conoscenza degli "Agaonini" (Hymenoptera Chalcididae)
dell'Eritrea e dell'Uganda. Boll. Soc. Ent. ltal. 48, 1-42 (1917).
8. Grandi, G. - Contributo alIa conoscenza degli Agaonini (Hymenoptera Chalcididae)
dell' America. Agaonini di Costarica. Agaonini. Boll. Lab. Zool. Portici 13, 15-56 (1919).
9. Grandi, G. - Identification of some fig-insects (Hymenoptera) from the British Museum (National History). Bull. ent. Res. 13, 295-299 (1923).
10. Grandi, G. - Agaonini e Sycophagini olartici c indo-malcsi. 13011. Lab. Zool. Portici 18, 1-31
(1924).
II. Grandi, G. - Morfologia del gen. Tetrapus Mayr e descrizione di una nuova specie della Costa
Rica. Boll. Soc. Ent. Ital. 57, 1-13 (1925).
12. Grandi, G. - J-Iymcnoplcres sycophiles rccoltcs aux lies Philippines par F.Baker. I. Agaonini.
Phil. J. Sci. 33, 309-329 (1927).
13. Grandi, G. - Hymenopteres Sycophiles recoltes dans I'Inde par Ie Frere E. Gombert. Bull. Soc.
Zool. France 53, 69-82 (1928).
14. Grandi, G. - Un nuovo genere e quattro nuove specie di Imenotteri sicofili di Sumatra. Boll.
Lab. Zool. Portici 1, 71-89 (1928).
15. Grandi, G. - Due specie di B/astophaga dell Isole Figi ed istituzione di un nuovo sottogenere.
Boll. Lab. ent. R. 1st. Bologna 1, 65-70 (1928).
16. Grandi, G. - Revisione critica degli Agaonidi descritti da Gustavo Mayr e catalogo rationato
delle specie fino ad oggi descritte di tutto iI mondo. Boll. Lab. ent. R. 1st. Bologna 1,
107-210 (1928).
17. Grandi, G. - Insetti dei Fichi messicani, malesi ed australiani. Boll. 1st. Ent. Univ. Bologna
19, 47-67 (1952).
18. Grandi, G. - Insetti dei Fichi (Hymenoptera Chalcidoidae) dell' Africa Australe. Boll. 1st. Ent.
Univ. Bologna 21, 85-106 (1955).
19. Grandi, G. - Una nuova specie di Ceratoso/en Mayr dell'Africa occidentale (Hymenoptera
Chalcidoidea). Boll. 1st. Ent. Univ. Bologna 26, 231-238 (1963).
20. Hill, D.S. - Fig-wasps (Chalcidoidea) of Hong Kong. 1. Agaonidae. Zool. Verh. 89, 2-55
(1967).
21. Hill, D.S. - Revision of the genus LiporrhapaJum Waterston, 1920 (Hymenoptera
Chalcidoidea, Agaonidae). Zool. Verh. 100, 1-36 (1969).
22. Joseph, K.J. - Contributions to our knowledge of fig insects (Chalcidoidea: Parasitic
Hymenoptera) from India. VI. - On six species of Agaonidae. Agra. Univ. J. Res. 3,
401-415 (195~).
23. Joseph, K.J. - On a collection of fig insects (Chalcidoidea: Agaonidae) from French Guinea.
Proe. Roy. ent. Soc. London 23, 417-418 (1959).
24. Michaloud, G., S. Michaloud-Pelletier, J.T. Wiebes & C.C. Berg - 1985. The oecurence of two
pollinating species of fig wasp and one species of fig. Proe. Kon. Ned. Akad. Wl;t. (C),
88, 93-119 (1985).
25. Ramirez, W.B. - Taxonomic and biological studies of neotropieai fig wasps (Hymenoptera:
Agaonidae). Univ. Kan. Sci. Bull. 49, 1-44 (1970).
26. Saunders, S.S. - Descriptions of three new genera and species of fig wasps allied to
BlaslOphaga from Calcutta, Australia, and Madagascar; with notes on their parasites and on
the affinities of the respective races. Trans. ent. Soc. London 1-27 (1883).
27. Waterston, J. - Notes on African Chalcidoidea. l. Bull. ent. Res. 5, 249-258 (1914).
28. Waterston, J. - IV. Notes on fig insects, including descriptions of three new species and a new
Blastophagine genus. Trans. ent. Soc. London, 128-136 (1920).
29. Waterston, J. - A new Blastophagine genus and species from E. Africa (HymenopteraChalcidoidae). Ent. mono Mag. 56, 197-200 (1920).
30. Waterston, J. - On a new African fig insect (BJastophaga dyscritus, sp. n.). Proe. ent. Soc.
London 417-418 (1920).
136
31. Waterston, J. - On some Bornean fig insects (Agaonidae - Hymenoptera Chalcidoidea). Bull.
enL Res. 12, 35-40 (1921).
32. Wiebes, J.T. - On the variability of Agaon paradoxum (Dalman) Grandi and Seres armipes
Waterston with remarks on other African Agaonidae (Hymenoptera Chalcidoidea). Zool.
Meded. Leiden 37, 231-240 (1961).
33. Wiebes, J.T. - Indo-Malayan and Papuan fig wasps (Hymenoptera, Chacidoidca). 2. The
genus Pleistodonles Saunders (Agaonidae). Zoo!. Meded. Leiden 38, 305-321 (1963).
34. Wiebes, J. T. - Taxonomy and host preferences of Indo-Australian fig wasps of the genus
Ceratosolen (Agaonidae). Tijdschr. Ent. 106, 1- j 12 (1963).
35. Wiebes, J.T. - Fig wasps from Ficus dzumacensis, with notes on the genus Sycobiella
Westwood. Zoo I. Meded. Leidcn 39, 19-29 (1964).
36. Wiebes, J. T. - Fig wasps from Israeli Ficus sycomorus and related East African species
(Hymenoptera Chalcidoidea). I. Agaonidae. Ent.. Ber. Arnst. 24, 187-192 (1964).
37. Wiebes, J.T. - Indo-Malayan and Papuan fig wasps (Hymenoptera Chalcidoidea). 3. Insects
from Ficus conocephalifolia, with a note on Sycophaginae. Nova Guinea, Zool. 27, 75-86
(1964).
38. Wiebes, J .T. - Indo-Malayan and Papuan fig wasps (Hymenoptera Chalcidoidea). 4.
Agaonidae from Ficus section Adenosperma. Zoo I. Meded. Leiden 40, 225-233 (1965).
39. Wiebes, J.T. - Bornean fig wasps from Ficus stupenda Miquel (Hymenoptera Chalcidoidea).
Tijdschr. Ent. 109, 163-192 (1966).
40. Wiebes, J.T. - A new Pleistodontes (Hymenoptera Chalcidoidea, Agaonidae) from Rennell
Island. Nat. Hist. RennellIsl. 5, 115-117 (1968).
41. Wiebes, J .T. - Species of Agaon from Congo (Kinshasa), with notes on synonymy
(Hymenoptera, Chalcidoidea). Proc. Kon. Ned. Akad. Wet. (C) 71, 346-355 (1968).
42. Wiebes, J.T. - Hymenopteran Agaonidae with an introductory chapter on the West African
fig wasps. Ann. Mus. Roy. Afr. centr., in 8, Zool. 175, 449-464 (1969).
43. Wiebes, J.T. - Revision of the Agaonidae described by J. Risbec and notes on their torymid
symbionts (Hymenoptera, Chalcidoidea). Zool. Meded. Leiden 45, 1-16 (1970).
44. Wiebes, J.T. - The genus Alfonsiella Waterston (Hymenoptera, Chalcidoidea, Agaonidae).
Zoo!. Meded. Leiden 47, 321-330 (1972).
45. Wiebes, J.T. - A new species of Agaon from Nigeria (Hymenoptera Chalcidoidea). Ent. Ber.
Arnst. 32, 122-124 (1972).
46. Wiebes, J .T. - Species of Agaon Dalman and Allotriozoon Grandi from Africa and Malagasy
(Hymenoptera, Chalcidoidea, Agaonidae). Zool. Meded. Leiden 48, 123-144 (1974).
47. Wiebes, J .T. - Nigeriella, a new genus of West African fig wasps allied to Elisabethiella Grandi
(Hymenoptera Chalcidoidea, Agaonidae). Zoo I. Meded. Leiden 48, 29-.42 (1974).
48. Wiebes, J.T. - Fig wasps from Aldabra (Hymenoptera Chalcidoidea). Zoo!. Meded. Leiden
49, 225-236 (1975).
49. Wiebes, J .T. - A new species of Agaon from Nigeria, and some additional records
(Hymenoptera Chalcidoidea, Agaonidae). Ent. Ber. Arnst. 36, 124-127 (1976).
50. Wiebes, J.T. - Dei/agaon, a new genus of Indo-Malayan and Papuan fig wasps (Hymenoptera,
Chalcidoidea, Agaonidae). Bijdr. Dierk. 46, 291-298 (1977).
51. Wiebes, J .T. - Agaonid fig wasps from Ficus salicifolia Vahl and some related species of the
genus Platyscapa Motschoulsky (Hym., Chal.). Neth. J. Zool. 27, 209-223 (1977).
52. Wiebes, J.T. - The genus Kradibia Saunders and an addition to Ceratosolen Mayr
(Hymenoptera Chalcidoidea, Agaonidae). Zool. Meded. Leiden 53, 165-184 (1978).
53. Wiebes, J.T. - The fig wasp genus Dolichoris Hill (Hymenoptera Chalcidoidea). Proc. Kon.
Ned. Akad. Wet. (C) 82, 181-196 (1978).
54. Wiebes, J .T. - Fig wasps from Gabon: new species of Agaon (Agaonidae) and Phagoblastus
(Torymidae) (Hymenoptera Chalcidoidae). Proc. Kon. Ned. Akad. Wet. (C) 82,391-400
(1979).
55. Wiebes, J.T. - Records and descriptions of Agaonidae from New Guinea and the Solomon
Islands. Proe. Kon. Ned. Akad. Wet. (C) 83, 89-107 (1980).
137
56. Wiebes, J.T. - The species group Ceratoso/en armipes Wiebes (Hymenoptera Chalcidoidea,
Agaonidae). Proc. Kon. Ned. Akad. Wet. (C) 84, 365-377 (1981).
57. Wiebes, J.T. - New species of Waterstoniella Grandi from the Indo-Malayan region
(Hymenoptera Chalcidoidea, Agaonidae). Proc. Kon. Ned. Akad. Wet. (C) 85, 399-411
(1982).
58. Wiebes, J. T. - Records and descriptions of Pegoscapus Cameron (Hymenoptera Chalcidoidea,
Agaonidae). Proc. Kon. Ned. Akad. Wet. (C) 86, 243-253 (1983).
59. Wiebes, J .T. - Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, I. Proc. Kon. Ned. Akad. Wet. (C) 89, 335-355 (1986).
60. Wiebes, J .T. - Agaonidae (Hymenoptera, Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, II (AlfonsielIa). Proc. Kon. Ned. Akad. Wet. (C) (91), 429-436 (1988).
61. Wiebes, J.T. - Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, III (Elisabethiella). Proc. Kon. Ned. Akad. Wet. (C) 92,117-136 (1989).
62. Wiebes, J.T. - Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, IV (African Ceratosolen). Proc. Kon. Ned. Akad. Wet. (C) 92,251-266 (1989).
63. Wiebes, J .T. - Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and
their figs, V. (Agaon). Proc. Kon.- Ned. Akad. Wet. (C) 92, 395-407 (1989).
64. Wiebes, J .T. & U.C. Abdurahiman - Additional notes on Platyscapa Motschoulsky
(Hymenoptera Chalcidoidea, Agaonidae). Proc. Kon. Ned. Akad. Wet. (C) 83, 195-207
(1980).
65. Wiebes, J.T. & S.G. Compton - Agaonidae (Hymenoptera Chalcidoidea) and Ficus
(Moraceae): fig wasps and their figs, VI (Africa concluded). Proc. Kon. Ned. Akad. Wet.
93, 203-222 (1990).
138
CHAPTER 7
BREAKDOWN OF HOST SPECIFICITY
Paper 10: Studies of Ceratosolen galili,a non-pollinating agaonid fig wasp. Biotropica 23; 188-194 (S.G.
Compton, K.C. Holton, V.K. Rashbrook, S. van Noort, S.L. Vincent and A.B. Ware - 1991).
Paper 11: Breakdown of pollinator specificity in an African fig tree. Biotropica 24; in press. (A.B. Ware and
S.G. Compton - 1992).
139
BIOTROPICA 23(2): 188-194
1991
Studies of Ceratosolen galili, a Non-Pollinating Agaonid Fig Wasp 1
S. G. Compton, K. C. Holton, V. K. Rashbrook, S. van Noort, S. L. Vincent, and A. B. Ware
Department of Zoology and Entomology, Rhodes University, Grahamstown, South Africa
ABSTRACT
The African fig tree Ficus lycomorllJ is host to two species of agaonid fig wasps, Cera/Olo/en arabicul and C. galili.
Our studies of C. ga/ili in southern Africa confirm that it does nOt actively pollinate the figs of F. J),comorus, although
some accidental pollination takes place. The absence of pollination behavior in C. galili raises questions about the
reasons why other agaonids pollinate the figs and thereby maintain the fig-fig wasp mutualism. C. ga/ili larvae did
not suffer elevated mortality rates when developing in un pollinated Bowers and the only potential "cost" of not
pollinating that we detected was that adult female C. galili were smaller than those of C. arabicuJ that developed
on the same tree.
UMCABANGO-NJE
Umkhiiwane wase Afrika, i FiCUJ s),comorUJ, ungosokhaya wohlobo olubili Iweminyouu, okuyi- Cera/OJo/en arabiC!1J
ne C. galili. Ucwaningo Iwethu lwe C. ga/ili yase Afrika yase-Ningizimu luginisekisile ukuthi ayiyiqholi neze
imikhiwane ye F. s),comorus, nakuba kwenzeka ngengozi iqholeke lemi khiwane. Ukungaqholi kwe C. gali/i kususa
imibuzo ngeziza thu ezenza ukuba eminye iminyouu eyi agaonids iziqhoie izimbali zomkhiwane ngaleyondle!a igcine
ubudlelwano phakathi kwayo iminyouu nemikhiwane. Izibungu ze C. galili azange zitshengise izinga eliphakeme
lokufa ngenkathi zikhula ezi mbalini ezingaqholiwe. Ukukhubazeka, nokho, esakubona wukuthi iminyouu yesifazane
end ala yayiyimincane ngemizimba kunaleyo yeminyouu eqhololayo, i C. ambicus eyayikhula kanye nayo esihlahlemi
esisodwa.
EACH OF THE 750 OR SO SPEClES OF FIG TREES (Ficus
is, with a few exceptions, pollinated
spp., ~lorace)
by a single species of host specific fig wasp (Hymenoptera, Agaonidae). In Africa, the exceptions'
co this general pattern include F. ottoniifolia (Miq.)
Miq. and F. sur Forssk., where twO species of agaonids are kno\vn co pollinate each of the trees (Michaloud et a/. 1985). F. s),comorus L. is also associated
with twO agaonids, but may be unique in that only
one of them pollinates the figs. WI or king in East
Africa, Galil and Eisikowitch (1968, 1969) showed
that Cera/oso/en arabiws Marc was a legitimate
pollinacor of F. s),comorus. The second -species, C.
ga/ili Wiebes colonized the figs, but had pollen
pockets that were never used. C. galiti was therefore
a "cuckoo" that exploited the murualism. Recently
\'\7iebes (1989) recorded both wasps from F. mllcoso
Ficalho, a fig tree closely related to F. J),comorus,
and again found that only the females of C. arabiclls
carried pollen.
TI1e absence of active pollination by C. galili
raises questions about how the behavior evolved in
agaonids and why they should continue ro carry out
Received 3 January 1990, revision accepted ISMay
1990.
I
this behavior, which forms the basis of the fig-fig
wasp mutualism. Kjellberg et a/. (1987) considered
that maintenance of the wasps' elaborate pollination
behavior indicated that there was consistent selection
favoring its retention. A direct advantage of pollination was shown for Blastophaga qlladraticeps
Mayr, because its larvae suffered increased mortality
rates if they developed in un pollinated flowers (Galil
& Eisikowitch 1971). Increased larval mortalities
also occur in Elisabethiella baijnathi Wiebes and
C. capensis Grandi when they develop in unpollinated flowers (Nefdt & Compton, pers. (omm.).
Pollination benefits to wasp larvae may therefore be
a general phenomenon, perhaps due to improved
larval nutrition (Verk~
1989). C. ga/ili nonetheless appears to have circumvented the problems
of developing in unpollinated flowers, and it is unclear why a similar abandonment of pollination behavior has not been observed in other species.
This paper describes studies of C. ga/ili and
some other fig wasps associated with F. sycomorus
in southern Africa. These studies aimed to answer
the following questions: Does C. galili fail to pollinate the figs of F. sycomorus also in southern Africa?
If so, then does C. galiti "pay a price" for not
pollinating the figs? Do c. ga/ili females seek out
188
140
Non-Pollinating Fig Wasp
.'
'-~
o
189
...... ....
/
..........
........
o
500 km
FIGURE 1. Records of Cera/OJo/en species in collections of F. JycomorUJ figs. The dotted line indicates the approximate
southern limit of the distribution of the fig species. Squares are F. J. JycomorllJ, circles F. J. gnaphalocarpa. Open
squares/circles indicate the presence of C. arabicuJ, closed squares C. galili. Mixed squares indicate that both wasp
species were present.
figs which contain flowers already pollinated by C.
arabicus? Are the twO agaonids equally successful
at entering the figs? Do the twO species compete for
oviposition sites? Do any of the other fig wasps
associated with F. s),comorus require fertile seeds for
their larvae and consequently fail to develop in figs
which lack C. arabicus?
F. sycomorus is distributed throughout most of
tropical ~d
subtropical Africa. Two subspecies are
generally recognized, F. s. s),comorusj which in southern Africa is found in the east, and F. s. gnapha/ocarpa in the west. The twO subspecies (or forms) are
distinguished only by the placement of the figs,
which occur on mot:lified leafless branches in F. s.
sycomorus, but are borne among the leaves by F. s.
gnaphalocarpa (Berg, in press).
Figure 1 is based on colJections of mature F.
s)'comorus figs (Compean, pers. comm.) and summarizes our distribution records for the tv.'o Ceraloso/en species in southern Africa. C. ga/ili was at
least as common as C. arabicus in the more humid
east of the subcontinent, but was not recorded from
F. s. gnapha/ocarpa growing in Namibia. In a twO
year study Wharean el a/. (1980) similarly failed
to detect C. ga/ili in Namibia.
In southern Africa F. sycomortls also suppOrtS
numerous species from the family Torymidae. One
141
of the torymids (the seed predator Sycophaga s),comori L.) behaves like an agaonid in that females enter
the figs through the ostiole to oviposit. The remaining species have long ovipositors and oviposit
into the fig flowers from the outside of the figs. The
biology of these wasps is largely unknown, but some
are seed gallers, while others may be parasitoids or
inquilines.
METHODS
Haphazardly sampled figs were obtained from thirteen F. sycomortls trees growing at various localities
in northern Natal, South Africa (Table 1). "Immature" crops were at the early inter-floral stage
(sensu Galil 1977). At this time the figs contained
remains of the female wasps which had entered to
lay their eggs. CountS of wasps which had successfulJy entered the cavities of the figs were obtained
by cutting the figs in half through the ostiole and
searching for the wasps' remains under a dissecting
microscope. Subsamples were examined for the presence of wasps which had failed in their artempts to
enter the lumens of the figs and had become trapped
in the ostiolar bracts.
"Mature" crops consisted of figs containing
wasps that had completed their larval developmenc
Compton, Holton, Rashbrook, van Noort, Vincent, and Ware
190
TABLE l.
Tree
1
2
3
4
5
6
7
8
9
10
11
12
13
DeJCriplions of F. sycomorus eollee/ions in Natal.
Collection
date
6.12.88
6.12.88
9.12.88
8.12.88
10.12.88
8.12.88
7.12.88
5.12.88
6.12.88
7.12.88
8.12.88
10.12.88
10.12.88
Locality
Degree
square
Figs sampled
Rd. south of Ndumu Game Park
Rd. south of Ndumu Game Park
Mkuzi Game Park
Mkuzi Game Park
Outside entrance ro Mkuzi Game Park
Rd. south of Ndumu Game Park
Ndumu Game Park
Tar road bridge over Pongola River
Rd. south of Ndumu Game Park
Ndumu Game Park
Ndumu Game Park
Outside entrance ro Mkuzi Game Park
Outside entrance to Mkuzi Game Park
2632CC
2632CC
2732CA
2732CA
2732CA
2632CC
2632CD
2732AB
2632CC
2632CD
2732CA
2732CA
2732CA
Immature
Immature
Immature
Immature
Immature
Immature
Immature
Mature
Mature
Mature
Mature
Matute
Mature
Sample
size
(figs)
50
50
50
50
50
10
10
25
25
25
25
25
24
and were ready to emerge. Here, figs collected hap- females. The wasps were crushed under glass cover
hazardly from the trees were placed individually in slips and examined under a compound microscope
netting-covered jars. After the wasps had emerged for the presence of pollen in their pollen baskets
they were killed and then recorded, together with and superficially on their body surfaces.
any wasps remaining inside the figs.
The dry weights of adult females of the twO
Figs from three of the mature crops were grou ped Ceratosolen species were compared. The wasps were
according to the species of Cera/oso/en which had dried at 40 c C in an oven and then weighed indiemerged from them. These were used to determine vidually on a kahn Microbalance.
whether seeds were only produced in figs comaining
C. ambicus. Flowers which had not produced wasps
RESULTS
were scored as being either seeds, "unpollinated,"
or bladders. "Unpollinated" flowers were those CAN C. GAUU POlliNATE THE FIGS OF F. SYCOMORUS?which showed no evidence of their ovules having . Comparisons of the figs colonized by C. arabicus
expanded due to pollination or galling. Bladders and C. galili confirmed that, as in East Africa, the
(sensu Galil & Eisikowitch 1971) superficially re- former species rout:inely pollinated the flowers, while
semble seeds, but: are hollow. They may represent the latter did not (Table 2). However, twO healthy
flowers where wasp larvae died at an early stage of seeds were detected in figs which only produced C.
developmem (Galil & Eisikowitch 1971).
galili, showing that occasionally this species can
Samples of 20 recently emerged C. galili fe- pollinate a few flowers. None of 140 females of C.
males were collected from each of seven figs, to- ga/ili investigated, and all of 20 C. arabicus, had
gether with one comrol sample of 20 C. arabicus pollen in their pollen baskets. Three C. ga/i/i did
TABLE 2.
The eonlenfJ 0/jigs colonized by C. arabicus and/or C. galili. Flowers whieh produced wasps are nol included.
Totals
Agaonid(s)
Tree
8
9
10
Combined
C.
C.
C.
C.
C.
C.
C.
C.
C.
arabicus
galili
arabicus
galili
arabieu!
galili
arabieus
arabicus
galili
+ C. galili
+ C. galili
+ C. galili
No. figs
Seeds
Bladders
Unpollinated
4
5
5
5
3
5
4
8
15
200
23
59
57
54
8
30
23
65
143
0
250
142
0
78
0
98
2
200
176
2
123
223
81
250
0
204
723
Non-Pollinating Fig Wasp
TABLE 3.
191
Comparison of Ihe dry weights of C. arabicus and C. galili at three focalions in sorahem Africa.
C. arabiczlJ
Localiry
Pongola River, Natal 1
Pongola River, Natal 2
Limpopo River, Botswana
PongoJa River, Natal 1
C. galili
Sex
N
Mean
(mg)
SD
N
Mean
(mg)
SD
F
P
F
F
F
20
20
16
19
0.097
0.131
0.082
0.117
0.015
0.022
0.009
0.183
20
21
16
18
0.079
0.104
0.063
0.061
0.012
0.012
0.009
0.008
18.68
23.68
33.38
6.21
<0.001
<0.001
<0.001
<0.001
M
Between localities (females): C. arabiclIs F(2&Hl
=
40.58, P < 0.001; C. galili F{2&54)
have some pollen attached superficially to their bodies (one, one, and six pollen grains, respeCtively),
which suggestS how "accidental" pollination can
take place.
=
59.26, P < 0.001.
from a honey bee pollen basket) had a volume
almost equal to that of the gaster of the wasps.
Do C.
GAUL!
FEMALES PREFER FIGS CONTAINING C.
ARABlCUS?-Adult female C. galili might be ex-
DOES C. GAUU "PAY A PRICE" FOR NOT POlliNATING
THE FIGS?-The numbers of bladders in the figs
provide a relative estimate of the larval mortalities
of C. arabicuJ and C. galiN (Table 2). More bladders were present in figs containing C. ga/ili only
(16% of the flowers, compared with 10% for C.
arabicus only), but the difference was not significant
(Z = 0.705, P > 0.05). Galil and Eisikowitch
(1971) found that differential mortalities of female
agaonid larvae occurred in figs which had not been
pollinated, resulting in a collapse of the normally
female-biased sex ratios. This was not true of C.
gaiiii. A count of 4991 individuals from 12 figs
containing only C. galiii revealed that 73.9 percent
were female, a sex ratio similar to that of C. arabicus.
The body weights of adult C. arabicus and C.
galili are compared in Table 3. Wasps from different trees varied significantly in body size; but
from anyone crop, C. galil! were consistently smaller. The difference in weights between the females
(around 0.02-0.03 mg) was not due to the pollen
load of C. arabicuJ. Pollen of this weight (extracted
TABLE 4.
pected to preferentially colonize figs that already
contain C. arabicus if their larvae gain any benefit
from developing in figs containing pollinated flowers. However, C. galili females were the most abundant wasps in the immature fig samples and were
the only occupants of about half the figs (Table 4).
Combinations of species did occur and occasionally
females of C. galiii, C. arabicuJ and S. Jycomori
were all present in a single fig. Nonetheless, figs
containing females of both agaonids were consistendy underrepresented in the samples, compared
with figs containing only one species (for combined
totals X2{IJ = 160.39, P < 0.001).
As with the immature fig saniples, C. galili was
the more numerous agaonid in the mature figs (Table 5). Figs containing combinations of the rwo
species were again underrepresented, confirming that
C. galili females do not actively seek au! figs pollinated by C. arabii:uJ.
ARE THE TWO AGAONIDS EQUALLY SUCCESSFUL AT
ENTERING THE FIGs?-In figs which contained only
The combinalions of wasps entering Ihe jigs of F. sycomorus.
Figs conraining combinations of species
Figs containing single species
Tree
No. of figs
C. arabiclIJ
C. gaIiIi
S. Jycomori
C. arabiCIIJ
+ C. ga/ili
1
2
3
50
50
50
50
50
250
20
23
12
20
14
1
0
0
1
0
2
7
10
6
1
7
31
4
5
Total
2
13
70
28
42
30
134
143
S. sycomori
+ C. ga/ili
C. arabicIIJ
+ C. ga/ili
+ S.
s),comori
1
1
1
2
0
0
0
3
4
4
0
10
192
Compton, Holton, Rashbrook, van Noort, Vincent, and Ware
TABLE 5.
The frequencies of C. arabicus, C. gal iIi and S. sycomori females reared from figs of F. sycomorus.
Numbers of figs where wasps were present as
Combinations of species
Single species
Tree
No. of figs
8
9
10
11
12
13
Total
25
25
25
25
25
24
149
C. arabicuJ
S. sycomori
19
13
21
15
15
23
106
0
2
0
0
0
1
3
5
1
2
3
1
a
12
C. ga/ili the number of females which successfully
entered the figs varied between 1 and 55, with an
overall mean of 5.66 females per fig (Table 6). C.
arabicus was never recorded at such high densities,
and neither was C. galili in figs which it was sharing
with the other species.
A proportion of the females that had attempted
to enter the figs failed to do so and became trapped
in the ostiolar bractS. C. galili had a particularly
high failure rate and on ttees 1-5, 77.2 percent of
all the females that had attempted entry were found
dead part way through the ostioles (Table 7). These
were all facing inward and were nOt females which
were attempting to exit the figs. C. arabicuJ females
were significantly more successful at gaining entry,
with only 13.6 percent failing to do so (X 2 [1] =
68.12, P < 0.001).
Do THE TWO SPECIES COMPETE FOR OVIPOSITION SITES
IN SHARED FIGs?--Competition between the agaonids was examined using data from ttee 9, where the
twO species shared a relatively high proportion of
the figs. As males of the twO species are difficult to
separate, the comparisons were based on females
only. When alone, a mean of 146.3 C. ga/ili females
TABLE 6.
C. arabicuJ
C. galili
+ C.
S. Jycomori
+ C.
gali/i
1
8
2
4
3
galiN
0
1
a
3
6
0
10
0
18
were reared from each fig, compared with a mean
of 92.9 females per fig when sharing with C. arabicuJ. Although suggestive of competition for oviposition sites, this difference was not significant (2
= 1.47, P > 0.05).
Do THE TORYMlD FIG
WASPS REQUIRE FERTILE SEEDS
FOR THEIR LAR v AE?-Nine species of torymid fig
wasps were reared from the mature figs (Table 8,
counts of the twO rare WatJhamiella spp. are combined). All of the rorymids were recorded from figs
where C. arabicus was absent, showing that none
of them are conventional seed predatOrs requiring
fertile seeds for their larval development.
DISCUSSION
These studies in the southern part of the range of
F. sycomorus confirm that C. arabicuJ is itS only
active pollinator. C. galili accidentally pollinated a
few flowers by carrying pollen on its body surface,
but the number of seeds produced in this way was
negligible. NewtOn and Lomo (1979) recorded similar accidental pollination by a sycoedne fig wasp
which enters the figs of F. 11ltea Vah!' C. ga/ili
The numbers of wasps JUfcwfuffy entering the figs of F. sycomorus. Sample JizeJ were 50 figs per tree. The
rangeJ are given in parenlheseJ.
Mean wasps per fig
Single species presenc
Tree
1
2
3
4
5
Total
C. arabicus
1.30 (1-2)
1.74 (1-4)
1.50 (1-3)
1.00 (1)
1.23 (1-2)
1.46
Both species presenc
C. galifi
2.30
2.86
3.46
i2.74
1.33
5.66
C. arabiaJJ
1.43
2.30
1.33
1.00
1.29
1.64
(1-18)
(1-6)
(1-25)
(2-55)
(1-3)
144
(1-3)
(1-10)
(1-2)
(I)
(1-2)
C. galili
1.71
2.00
2.00
1.00
1.43
1.68
(1-4)
(1-4)
(1-6)
(1)
(1-4)
Non.Poliinating Fig Wasp
TABLE 7.
Tree
The numbers 0/ wasps Ihat/ailed 10 gain entry inlo jigs and were Irapped in the 011iolar bracts.
Mean plus range per fig
No. of
figs
10
10
10
10
10
10
10
70
1
2
3
4
5
6
7
Toral
C. arabicu!
C. galili
0.5 (0-3)
9.8 (0-33)
1.4 (0-4)
3.9 (2-7)
9.80-24)
2.7(0-6)
0.3 (0-2)
0.5 (0-2)
4.06
o
o
o
o
o
0.1 (0-1)
0.09
females were less successful than those of C. ambicus
at getting through the ostioles of the figs. This may
have been due to interference resulting from the
very high densities of C. ga/ili trying to enter the
figs. A similarly high proportion of wasps become
trapped when large numbers of female E/isabethiella baijnathi Wiebes attempt to encer the figs of
F. burlt-davyi Hutch. (Nefd! & Compton, pers.
comm.).
C. gali/j larvae commonly developed in figs
lacking any pollinated flowers, and there was no
evidence that this resulted in elevated mottality rates.
However, the adult females they produced were
TABLE 8.
193
The composition
0/ fig
0.5 (0-4)
o
o
o
o
o
0.4 (0-2)
0.13
consistently smaller than those of C. arabiclIs. This
could reflect differences in the quantity or quality
of the food available to C. galili larvae developing
in unpollinated flowers. Smaller species of agaonids
contain fewer eggs, and within a species, egg loads
are correlated with body size (Nefdt & Compton,
pers. comm.). C. gaiiH females are, therefore, likely
to carry fewer eggs than those of C. arabicus emerging from the same tree. This appears to be the only
potential .. cost" to C. ga/ili of not pollinating the
flowers, although other explanations for the size
difference, such as phylogenetic constraints, are
equally plausible. If other agaonids which have for-
wasp aJScmb/ages reared /rom figs ofF. sycomorus. Count! are 0/ females only.
Tree 8
(15 figs)
Cera/oso/en
galili
Ceratoso/en
arabicu!
SycoJcapteridea
(pale) sp. indec.
Sycoscapleridea
(dark) sp. indee. Syco!capler
sp. imler.
Apocr),plophaglls
gigas Mayr
Apocrypla
iongilarsllJ
Mayr
ElIkoebelia
sycomori
Wiebes
WalJhamidla
spp. inder.
Sycophaga
sycomori
S. sycomori
Tree 10
(15 figs)
Tree 9
(15 figs)
N
(figs)
Toral
wasps
Range
per fig
N
(figs)
Total
wasps
Range
per fig
1)
13
1368
(0-234)
13
3660
(0-411)
585
(0-129)
6
368
(0-109)
3
39
(0-145)
15
321
(3-46)
15
731
(0-110)
9
118
(0-26)
13
257
(0-52)
6
72
(0-24)
5
40
(0-14)
13
227
(0-52)
0
0
0
2
2
11
93
(0-17)
3
18
(0-10)
10
98
(0-23)
5
29
(0-11 )
12
183
(0-28)
10
52
(0-12)
5
14
(0-9)
4
9
(0-4)
0
0
0
(0-1)
8
6
(0-2)
4
436
N
(figs)
Toral
wasps
10
686
(~-20
6
0
0
Range
per fig
0
145
(0-152)
(O-I)
(0-1)
0
0
0
194
Compton, Holton, Rashbrook, van Noort, Vincent, and Ware
saken pollination are detected, then it will be interesting to see if they also are smaller than their
associated pollinator.
C. arabicus and C. gaiili differ in both appearance and behavior. C. arabicus flies at night
and is often collected at light traps (Wharton et ai.
1980, Compton & RobertSon, pers. comm.). Associated with this are its "Ophionoid" features, such
as yellow coloration and enlarged eyes (HuddlestOn
& Gauld 1988). In contrast, C. gaMi is a black,
day flying species, which usually emerges from the
figs in the early afternoon (S. G. Compton, pers.
obs.). This may be the reason for the apparent rarity
of C. gaiili in dry habitats, because its diurnal flight
period should make it more prone to dehydration.
The rarity of figs containing both Cera/oso/en species
may also be related to their different flight preferences. Figs cease to be attractive to agaonids after
they have been pollinated. If the attraction wanes
within hours of pollination, then figs entered at night
may already be unsuitable by the following day.
Alternatively, females may distinguish and avoid
figs that have already been entered by the other
species.
The evolution of agaonids with the biology of
C. gaMi requires the following: there must be tv.·o
agaonids sharing the same host Ficus (otherwise the
tree will not be pollinated and will go extinct); and
a mutation for the loss of pollination behavior must
occur and be sufficiently advantageous to become
"fixed" throughout the species. The nature of these
hypothetical advantages is uncertain. Morphological
evidence suggests that C. ambicus and c. galili are
not "sister species" (Wiebes 1989) and therefore
C. gal;l; cannot be derived from C. arabiclIs. Presumably the ancestors of C. gafili originally pollinated some other fig species, and subsequently colonized F. sycomorus while it was already being
pollinated by C. arabicus. Given that there do not
seem to be major hurdles associated with forsaking
pollination, the apparent rarity of species such as
C. galili may be due to the infrequency of such
colonization events in the history of Ficus-agaonid
coevolution.
ACKNOWLEDGMENTS
We would like to thank the Natal Parks Board and the
KwaZulu Government Service for permission to carry out
research in their reserves. P. E. Hulley and J. 1. Bronstein
kindly provided comments on the manuscript. G. Dube
produced the Zulu abstract.
LITERATURE CITED
BERG, C. C. In press. Annotated check-list of the Ficus species of the African floristic region, with special reference
and a key to [he taxa of southern Africa. Kirkia.
GAUl., J. 1977. Fig biology. Endeavour 1: 52-56.
- - - , AND D. EISIKOWITCH. 1968. On the pollination ecology of Ficus sycomQrus in East Africa. Ecology 49:
259-269.
- - - , AND - - - . 1969. Further studies of the pollination ecology of Ficus sycomorus 1. Tijdschr. Eneomol.
112: 1-13.
- - - , AND - - - . 1971. Studies onmutualistic symbiosis between syconia and sycophilous wasps in monoecious
figs. New Phyrol. 70: 773-787.
HUDDLESTON, T., AND 1. GAULD. 1988'- Parasitic wasps Ochneumonoidea) in British light-traps. Entomologist 107:
134-154.
KJEllBERG, F., G. M!CHALOUD, AND G. V ALDEYRON. 1987. The Ficus-pollinator mutualism: how can it be evolutionary
stable? Proe. 6th Int. Symp. Insects-planes, pp. 335-340. W. Junk, Dordrecht.
MICHALOUD, G., S. MICHALOUo-PEllETIER, J. T. WIEBES, AND C. C. BERG. 1985. The co-occurrence of twO pollinating
species of fig wasp and one species of fig. Proc. K. Ned. Akad. \X'et. C 88: 93-119.
NEWTON, L. E., AND A. Lo!>lO. 1979. The pollination of Ficus voge/ii in Ghana. Bot. J. Linn. Soc. 78: 21-30.
VERKERKE, \X'. 1989. Structure and function of the fig. Experientia 45: 612-621.
WHARTON, R. A., J. \X'. TILSON, AND R. 1. TILSON. 1980. Asynchrony in a wild population of FiClis J)'comorus 1.
5th. Afr. J. Sci. j6: 478-480.
.
WIEBES, J. T. 1989. Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and their figs, IV
(African Ceratosolen). Proe. K. ~ed.
Akad. Wet. C 92: 251-266.
146
BIOTROPICA 24(3): OO-DO
1992
Breakdown of Pollinator Specificity in an African Fig Tree I
Anthony B. Ware and Stephen G. Compton
Department of Zoology and Entomology, Rhodes University, Grahamstown, South Africa
ABSTRACT
A single giam.leafed fig m:e (Ficus lulla) is planted on the Rhodes Unin:rsity campus in Grahamstown, South
Africa, some 500 km outside its normal distribution range. Small numbers of fig wasps (Hymenoptera, Agaonidae)
which normally pollinate tWO other Fiol! species emered and successfully pollinated the figs of chis tree. One of the
wasp species reproduced successfully. MonitOring of adult fig wasps arriving ac the tree established thac these alien
species were noe attraaed to F. fuua. However, from IaboratOty studies it appears thac once having landed on F.
lutea figs, these wasps were stimulated to search for the ostiole, through which they gained emrance to the fig cavity.
Females of a third pollinacof species were also ptesent on the cree, but they failed to initiate osriole searching behavior
when on the figs. Hybrid seeds resulting from the entry of the alien ",.. asps germinated successfully, but did not
progress pasc the coryledon stage, indicating postgerminarion deficiencies in the hybrids.
Key words:
Agaonidat; cowolu/ion; Ficus; hOIl spteijidly; Hymenopltra; mUllialism; pollination specijicil)'; Soulh Africa.
FIG TREES (Fiou spp.; Moraceae) and their polli- range scimuli on the fig surface, the physical barrier
nating wasps (Chalodoidea, Agaonidae, Agaoninae, imposed by the ostiole, and the suitable lengths of
the styles. Fig crees are also hosts to numerous species
sensu Boucek 1988) have an obligatory murualiscic
relacion.ship. Each of the 750 or so species of fig of nonpoUinating fig wasps, mainly belonging to
cree (Berg 1988) is generally pollinated by a single other subfamilies of Agaonidae (Boucek 1988).
species of fig wasp, which is uniquely associated These either have larvae that develop inside ovules
with thac cree (Wiebes 1979, Michaloud et al. they have galled, like those of the pollinators, or
1985). The maintenance of the specificity of the are parasitOids'of ocher fig wasps. Many nonpollirelacion.ship between fig cree species and their par- nating fig wasps f!!ay also be host cree specific (Bouticular agaonine pollinatOrs has long been held as cek et al. 1981, Olenberg 1985, van Noort, pers.
the excreme example of coevolution (Janzen 1979). comrn).
However, the mechanisms determining this speci- NewtOn and Lomo (1979) studied the pollination
biology of the giant-leafed fig cree, Ficus Julea vaW
fiory are not dearly understood.
Crop development on individual fig crees is (= F. vogelii ,CMiq.) Miq.), in its natural habitat in
often synchronized, forcing adult female pollinating cropical Africa. Although the southernmost limit of
wasps (foundresses) to leave their natal crees in order its distribution is Natal, South Afcica (van Greuning
1990), F. Iulea is planted furrher south as an orto find crees with figs suirable for oviposition. They
appear to recognize suitable hose crees. through Fi- namental cree. One such cree is present in GraClu-specific volarues released from the figs when they hamstOwn, some 500 k.m outside its normal range.
are ready for pollinacion (= female phase; Galil Compton (1990) recorded that females of twO spe1977) (van Noorr et al. 1989). On finding receptive cies of agaonines not normally associated with F.
figs, the pollinatOrs must then negotiate a bracr- /uJea had emered and pollinated 'the figs of this
lined pore (the ostioJe) in order to gain access to cree, and chat one of the wasp species reproduced .
the female Rowers lining the in.side of the fig (the successfully. Furthermore, he nored that individuals
lumen). The flowers are pollinated while the wasps of three non pollina ring fig wasp species normally
oviposit down some of the styles (Galil 1977, Jan- associated with other Fi(uJ spp. also reproduced
zen 1979). Ovipositor lengths of fig wasp species successfully. The objectives of this paper are to adare highly correlated with the mean sryle lengths of dress questions raised by these inirial.observation.s.
the ficus species they utilize (Nefdt 1989). Host Are the volaciles released from (he figS of F. julea
specificiry in fig wasps may therefore be determined acuarove to a range of pollinator species and thereby a comb ina cion of long range acuaction, short fore not as species specific as supposed? How imporrant are shorr-range stimuli on the fig surface in
determining
host specificity? Docs osriole scrucrure
I RecciveJ 9 July
1991, rC"ision accepted 24 January;
playa
role
in
preventing alien wasps from entering
1992.
147
B - Biotropica - 24(3) ... 5285 ... Gal. 71
the "wrong" figs? Does the oviposition and pollination behavior of alien wasps change when they
emer [he "wrong" figs? If fig wasps can pollinate
the "wrong" cree, do [he hybrid seeds grow successfully?
MATERIALS AND METHODS
minacion success of seeds from crosses wich F.III/ea,
F. sur Forssk and F. ,hanningii BI. (macernal parent
always F. IUlea). The seeds werc germinatcd on
moistened filter paper in Petri dishes at room temperatures (2D-28°C). After germination, thc seeds
were transfetred to pors containing a 25:75 mixture
of vermiculite and sterilized porting soil, and grown
indoors.
A 10 m high (220 cm DBH) F. /utea cree growing
on the Rhodes Universiry campus in Grahamsrown SHORT-RANGE RESPONSES OF FIG wASPS.-Branches
(eastern Cape Province of South Africa) was the of F. /ulea bearing female phase figs were placed
objeer of this invesrigarion. As far as could be as- in glass containers (50 x 50 x 30 em). Into each
certained, irs nearest known conspecific is another concainer a different pollinatOr species was released;
planted specimen at the Addo Elephant Nacional approximately 500 Cera/aso/en capensis Grandi from
Park some 80 km ro the west (ComptOn 1990). A F. sur were placed with 16 figs, 500 Elisabethiella
6 m high Liquslrnm lucidum Ait. (Oleaceae) plant- stuckenbergi Grandi from F. thanningii with 58 figs,
ed some 20 m away from the Grahamsrown F. and 700 Elisabelhiefia baijnathi Wiebes from F.
lulea was used as a concrol cree.
burtt-davyi Hutch. with 32 figs. To ascertain whether
The first F. lulea figs appeared in March 1990 the surface hairs on the figs of F. lutea were imand at irs peak in May the crop size was estimated portant in prevencing E. baijnathi from penerrating
to be 250,000. At (his time many of the figs were
the lumen of the fig (the figs of F. burtt-davyi are
aborting as they had nor been pollinated (no foun- . glabrous), 200 E. btl/jnathi females were released
dresses were recorded from samples of fallen figs).
OntO 10 F. /utea figs that had their surface hairs
The flowering/fruiting cycle of L. lucidum was De- removed. At the end of each observation period of
cember-January.
approximately 6 hr, the tOtal numbers of wasps that
successfully penerrated the fig lumens were recorded.
MONITORING OF FIG WASPS ARRIVING AT THE TREES.- C. capensis, which was found to readily enter F.
\VI asps visiting the crees were deteered using sricky /ulea figs in the female phase, was used as a control
craps made from cellulose sheers (21 x 30 em)
co (est that the figs provided to the othet wasps
secured to wh.ite cylinders (10 em radius). The sheers were suitable for enrry.
were made scicky by spraying with pruning sealant
(Frank Fehr Ltd., Durban). Three craps were hung
W/ ASP BEHAVIOR WITHIN THE FIGS OF F. LUTEA.in each of the F. lutea and L. lucidum crees at
Figs were rransversely bisected while wasps were
heighrs of 1.5, 2, and 4 m. The rraps were replaced
passing through the oscioles. The CUt edge of the
weekly and any fig wasps caught were counted and
half fig containing the wasp was placed onto a glass
identified. MonitOring of fig wasp arrivals began in
slide, where it became firmly attached by the exNovember 1989, three months after the previous
uding latex. The behavior of the wasp within the
F. lutea crop had finished, and at about the rime
lumen could be observed through the microscope
when the new crop W'a5 in.iciated. Trapping concinslide using a dissecting microscope. Oviposition and
ued for 35 weeks.
pollinacion by E. baijnathi, E. sluckenbergi, and C.
capensis was observed both in their usual hose figs
FOUNDRESSES AND THEIR PROGENY.-Approximately
and in those of F. IUlea.
nine weeks after the initiation of the F. /u/ea crop
most of the figs had not been pollinated and began
to abort. The remaining fruit were harvested once
RESULTS
the figs had ripened (= male phase; Galil 1977).
Any figs that already had exit holes produced by MONITORING OF FIG \\7ASP ARRIVALS.-Species com",:asp progeny were ignored. Each fig was bisecred posing the wasp fauna normally associated with F.
and, where possible, the idenriry of the foundresscs /utea in irs nacive range were never recorded from
was established. Some nonpollinaring fig wasps ovi- the sticky traps on the rree. Only small numbers of
posit through the fig wall from the ourside and the other fig wasps were coUeered (N = 51 representing
progcny of these develop withom evidence of a 0.49 wasps/trap/week; Table 1). The rwo most
frequently trapped fig wasps were C. capensis and
foundress.
\'Ve examined the germination and poscgcr- Sycaphaga cyclostigma \'Vacer5tOn, both normallr as-
148
T ABLE I.
Fig waspJ (ollerlfd o"er a 35 wuk ptriod on Jricky IrapJ placed in Liqusrrum lucidum and Ficus lurea
IrUJ. The reaplit'e period u,hm !h~ jig! wert potmlially attraClit.t 10 pollinalorJ waJ approximalely 7 wukJ.
L. /lJCidum
Species
Pollinaeors
C. capemiJ
E. baijnalhi
E. JllJCkenbergi
N onpoilinaeors
A. guineemis
S. cyclostigma
P. barbaruJ
Toeal
Mann Whimey U Statistic
(wasps on L. lucidum
and F. lu/ea)
F. luua
Total
period
Receptive
period
Taral
period
Recepeive
period
Toeal sample
period
Recepeive
period only
10
6
0
21
2
19
0
2
4
4
0.65
0.37
0.68
0.22
1.00
0.78
0
2
1
0
0
0
9
13
2
0
0.03'
0.37
0.78
0.37
1.00
0.79
16
8
51
27
0
3
3
1
• P < 0.05.
sociated with F. sur. Lower densiries of the pollinarors normally associated with F. Ihonningii and
F. burtt-davyi were also recorded.
Counts from the sock')' craps in the concrol L.
lucidum cree were equally low and there was no
indicarion thac any of the pollinating wasp species
were significantly more abundant in the F. IUlea
than the concrol cree (Table 1) (P > 0.05 for all
pollinating species; Mann \\lhitney U Statistic). Only
the parasitoid Apocrypla guineensis Grandi normally
associated with F. sur (Compton & RobertSOn 1988,
Ulenberg 1985), was collected significantly more
often from the F. lurea cree than the concrol cree
(P = 0.03; Mann Whitney U starisric). When counts
for all the pollinating wasps were combined, there
was again no significant difference in the number
of wasps crapped on F. Iliita and L. lucidum'over
the whole period (P = 0.811; Mann Whitney U
Statisric) , nor during the period when the fig cree
was potenrially arnaccive (P = 0.474; Mann Whitney U Statisric). Furthennore, if the wasps were
being arnaaed differentially co the fig cree during
the period when the figs were receptive, then the
number of wasps trapped on the F. lutea should
have increased reiarive to those on L. lucidum. 1bi.s
was not the case <X2 with Yates' correction = 1.62;
P > 0.05).
FOUNORESSES ANO niEJR PROGE....'Y.-By the end of
July the figs had matured to the male phase, 97 of
which were sampled (of these, 8 figs had wasp exit
holes and were excluded from the following counts).
Foundresses of Allorriozoon hellrandromorphum
Grandi, the pollinator normally associated with F.
149
lulea (Newton & Lomo 1979, Wiebes & Compron
1990), were recorded from 61.8 percent of the figs
(Table 2). These wasps reproduced successfully in
all the figs in which foundresses were found, as well
as in an additional 8 figs from which the wasps are
assumed to have escaped after laying their eggs
(Table 2). A single female SycorycteJ sp. was reared
from a fig containing A. helerandromorphllm. Some
other species of this genus are known to be parasicoids (e.g., Compton & Nefdt 1990) and A. heterandromorphum is likely to have been its host. Neither C. capensis nor S. cyclostigma, the cwo wasps
usually associated with F. sur, succeeded in reproducing in the figs of F. lulea, despite foundresses
being found in 29 percent of the figs (Table 2). In
conc:ra.st, E. sluckenbtrgi, the pollinator normally
associated with F. thenning;;, produced progeny in
all the figs in which foundresses were recorded (Table 2). E. baijnathi, the pollinator of the most
common Ficus in the area, F. bUrJI-davyi, were
never recorded as foundresses in the lumen of F.
/ulea figs, noe were its progeny recorded.
GERMINATION 51110IES.-1n three series of germinacion trials, the seeds from F. /ufea/F. !harming;i
and F. lurea / F. fur hybrid crosses took 8 CO 13
days to germinate while the pure F. IUlea seeds took
from 34 to 38 days. However, despite weir rapid
germination rimes, the hybrid seedlings were unsuccessful and, under our growing condirions, POStgermination survival was zero with no hybrids progressing beyond (he cotyledon stage. In concrast,
over 90 percent of the pure F. furea seedlings grew
successfully ro at least the first crue leaf seage.
8 - Biotropica - 24(3) .•. 5285 .•• Gal. 72
TABLE 2.
Fig WtOpJ found in rh: Jigs of a F. lurea /ru growing oul of iII natura! range in Graha1!IJ/oU'n, 501llh Africa.
\XI asp progen y
Number
of figs
Foundress(es)
40
A. htfUandromorphu1f1
Species
A. htterandromorphu1f1
SY(OY),cw sp.
A. heurandromorphllm
+5. e)'e/oJligma
A. heterandro1f1orphum
+ CrossogaJler JilveJlrii
C. capens;s
C. (apensis
12
3
3
3
3
A. heurandromorphuTn
C. Ji/vmrii
None present
None present
3
5
E. sfuekmbtrgi
A. htltrandromorphum
3
1
total of 303 female C.
capensi.; entered the lumens of 13 F. illtea figs,
while 29 female E. sllIekenbergi entered 21 figs.
LABORATORY STIJDIES.-A
Their behavior appeared to
be identIcal to that when they searched for the
ostiolar openings on their usual host figs. In contrast,
no E. baijnathi females entered the syconia of the
F. ill tea figs. Females of this species amennated the
surface of their host figs, but this behavior ""as not
evident when they were in contact with the figs of
F. illlea, even when the covering of surface hairs
had been removed.
Once inside the F. Itllea figs, the ovipositorprobing behavior of both E. !tllckenbergi and C.
capensi! appeared no different from that observed
in their own host figs. Once probing had commenced, the wasps removed pollen from their pollen
basker::s with their front legs and proceeded to deposit it into the nearby stigmas.
DISCUSSION
In natural siruations, receptive figs can arrract large
numbers of their associated pollinatOrs over relatively short periods (Bronstein 1987). No A. helmrndromorphum (the normal pollinaror of F. Illtea)
were recorded from sticky traps placed in the F.
iulea rree, showing that this wasp species was uncommon in OUf srudy area. Nevertheless, despite
(he trcc's isolated location, 55 of the figs were found
by these pollinators (an estimated 0.022% of the
'150
1
A. helerandromorphu1"
14
None preJenf
40
12
9
+S. cyclostigma
E. sluckmbergi
S. cydouigma
Frequency
(figs)
None prtJtnl
A. heterandromorphu1f1
None present
9
3
4
.,7
I
tOtal crop), TI1at A. heterandromorphum females were
able ro locate and pollinate the figs of such an
isolated host is indicative of the effectiveness of the
tree's volatile arrraaanr::s and the host-finding abiliry
of the wasps. This is even more impressive when
one considers the small size of the pollinating wasps
and that they are probably shOrt-lived (Kjellberg et
ai, 1988). In contrast, the low numbers of alien fig
wasps trapped on the F. ill tea during ir::s receptive
female phase can be considered as background noise
resulting from chance arrivals at the tree, rather chan
a breakdown in the specificity of arrraaion.
Once agaonines land on a fig it appears that
short-range stimuli, probably including the surface
chemisay of the lig, stimulate them to search for
the osoolar opening. Our laboratory investigations
indicated that the surfaces of receptive F. illtea figs
are recognized by females of both C. capen!iI and
E. stllckmbergi. These stimuli are thus not species
specific. Nevertheless, the failure of E. baijnalhi to
antennate the surface of F. iUlea figs shows that the
surface stimuli they present are not the same as
those of ir::s normal hose.
The osciole is generally considered to act as a
filter which prevenr::s nonadapted fig wasps from
entering the "wrong" figs (Janzen 1979). That che
ostiole aCtS as a barrier is demonstrated by the anatomy of the heads and bodies of agaonids, which
show numerous adaptations ro facilitate entry intO
rhe figs (Ramirez 1974). There is also evidence of
convergence in head shape berween agaonines and
sycoecines, another group of fig wasps chat penerrates the fig via [he osoole (van Noort and Comp-
ron, pefS. obs.). However, despite the evidence for
adaptations relatcd ro the penetration of osrioies of
specific fig species, the osriole of F. IUlea figs did
not act as a barrier ro females of E. Jllickmbergi,
C. capenJiJ, P. harbarliJ and S. cydOJligma, all of
which successfully penecraced the figs.
In other studies, Michaloud ( 1988) used a light
ro attract several species of nOCturnal agaonines, and
induced Agaon paradoxllm Dalman to enter figs of
F. nalaienJiJ leprielirii (Mig.) Berg, a cree which is
normally pollinated by Alfonliella fimbria/a Watersron. "Mistakes" made by agaonines enreringthe
wrong figs were also reported by Ramirez (1970).
Clearly, wasps adapted to enter the figs of one host
FiclIJ are not precluded from entering the figs of
other species and the filtering effect of the osriole
may not be as effective as previously imagined. From
our observations of the fig wasps that colonize F.
llilea, it appears that the long range, FicIIJ-specific,
arrractantS released by the figs (van Noort et al.
1989, Ware ef al., pers. comm.) form the basis of
host specificity in agaonines and that features of the
figs themselves have, at most, a secondary role in
determining pollinator specificity.
During 1989 and 1990, six species of fig wasp
(fWO pollinators, three other gall formers and one
putarive parasitoid) successfully reprod';'ced in the
Grahamstown F. tulea. Two of these are normally
associated with F. tlltea, three with F. fhonningii
and the host of one is indeterminate (this study;
Compton 1990). However, although they frequently entered the' figs, the fwO species normally asso-
ciared with F. 11lr failed to rcproduce. TI1US, wasps
from F. Ihonningii (subgenus Urosrigma, section
Galoglychia) were able to reproduce successfully in
the closely related F. llilea (subgenus Urostigma,
section Urostigma); whereas, those from the more
distantly related F. 11lr <subgenus Sycomorus) could
not.
Because of itS isolated location, the figs of the
Grahamsrown F. IIiJea remained unpollinated, and
therefore receprive, for an extended period. This
seems to have facilitated the incidental colonizarion
of itS figs by alien pollinators. While this increased
the likelihood of fig hybrid producrion, other natural
barriers prevenring hybridization had not been altered. However, few naturally occurring fig hybrids
have been recorded and in these cases at least one
parent cree was an incroduced species (Ramirez
1988). \X'hy have natural Ficlls crosses been so
rarely recorded? One possibility is that hybrids are
relarively common, but difficulr ro identify in the
field (Ramcharun et al. 1990). AlternariveIy, the
weakness of the hybrid seedlings recorded in this
study could be a general reason why FiclIl hybrids
fail to reach maturity.
ACKNOWLEDGMENTS
We would like t~hank
C. Zachariades both for his help
in harvesting the figs and, tOgether with P. E. Hulley,
for providing valuable comments on the manuscripc. The
FRO bursary suppore to ABW is gracefuily acknowledged. .
LITERATURE CITED
BERG, C. C. 1988. Classification and distribution of Fial!. Experienria 45: 605-611.
BOUCEK, Z. 1988. AIIJlra/ia11 Chalcido~
(H),mL11op:era). C.A.B. Intemational.
- - - , A. \'1/'ATSHAM, A....'O J. T. WlEBES. 1981. The fig wasp fauna of the re<eptades of FiclI! Ihon11ingii. Tijdschr.
Encomol. 124: 149-231.
BRONSTElN, J. L. 1987. Maintenance of species-specificity in a Neouopica1 fig-pollinator wasp murualism. Oikos
48: 39-46.
CoMPTON, S. G. 1990. A collapse of host specificity in some African fig wasps. Sth. Afr.]. Sci. 86: 39-40.
- - - , />NO R. NEFDT.
1990. The figs and fig wasps of FiellJ bllrn-davyi. Mire. Insc. Allg. Boe. Hambg. 23a:
441-450.
- - - , A.'10 H. G. ROBERTSON.
1988. Complex interactions between murualisms: ants cending homoprerans
protet:t fig seeds and pollinarors. Ecology 69: 1302-1305.
GALlL, J. 1977. Fig Biology. Endeavour 1: 52-56.
JANZEN, O. H. 1979. How to be a lig. Annu. Rev. Ecol. Syst.lO: 13-51.
KJEllBERG, F., B. OOUMESCHE, A..'It) J. L. BRONSTEIN. 1988. Longevity of a fig wasp (Bla!lOphaga punul. Proc. K.
Ned. Akad. Wet. Ser. C. BioI. Med. Sci. 91: 117-122.
.
MICHALOUD, G. 1988. Aspects de la reproduction des figuiers monoiques en forer Equatoriale Africaine. Unpublished
Ph.D. Thesis. Academie de MonrpelLier Universite des Sciences ec Tet:hniques du Languedoc.
- - - , S. MIOlALOvo-PELUTIER, J. T. \'I/'IEBES, AND C. C. BERG. 1985. The co-occurrence of tWO pollinating
species of fig wasps and one species of fig. Proc. K. Ned. Akad. Wet. Ser. C. BioI. Moo. Sci. 8S, 93-119.
NEFOT, R. J. C. 1989. Interactions between fig wasps and their host figs. Unpublished Ph.D. Thesis. Rhodes
University, GrahamstO\1m, South Africa.
151
B - Biotropica - 24(3) .. . 5285 ... Gal. 7~
NEWTON, L. E., AND 1.oMo, A.. 1979. Pollination of Ficul vogtlii in Ghana. Bot. J. Linn. Soc. 78: 21-30.
RAMCHARUN, S., H. BAlJNATI-I, AND J. V. VAN GREUNING. 1990. Some aspeCtS of the reproductive biology of the
Ficus nalalmJis complex in southern Africa. Mitt. lnst. Allg. Bot. Hambg. 23a: . 451-455'.
RAMIR.EZ B., W. 1970. Host specificity of fig wasps (Agaonidae). Evolution 24: 680-<591.
- - - . 1974. Coevolution of Ficul and Agaonidae. Ann. Mo. Bot. Gard. 61: 770-780.
- - - . 1988. F. microcarpa L., F. bmjamina L. and other species introduced in the New \'<'orld, their pollinators
(Agaonidae) and other fig wasps. Rev. BioI. Trop. 36: 441-446.
UUNUERG, S. A. 1985. The phylogeny of the genus Apocrypla Coquerel in rel:uion to itS hosts, Cn-aloJolm Mayr
(Agaonidae) and Fio/J L. Verhand. K. Ned. Akad. Wet. 83: 149-176.
VAN GREUNING,]. v. 1990. A synopsis of the genus Ficus (Moraceae) in southern Africa. S. Afr.]. Bor. 56; 599630.
.
VAN NooRT, 5., A. B. WAllE, AND S. G. CoMPTON. 1989. Pollinator-specific volatile attractants released from the
figs of F. buru-davyi. S. Afr. ]. Sci. 85: 323-324.
WIEBES,;. T. 1979. Coevolution of /igs and their insect pollinatots. Annu. Rev. Eeol. Sysr. 10: 1-12.
- - - , AND S. G. CoMPTON. 1990. Agaonidae (Hymenoptera Chalcidoidea) and FicUJ (Moraceae): /ig wasps and
their /igs. VI (Africa concluded). Pree. K. Ned. Akad. Wer. 93: 203-222.
152
CHAPTER 8
NON-POLLINATING FIG WASP
Paper 12: African fig wasp parsitoid communities. In Parasitoid Community Ecology (Eds Hawkins, B.A.
and Sheenan, W.). In press (S.G. Compton, J.-Y. Rasplus and A.B. Ware)
153
AFRICAN FIG WASP PARASITOID COMMUNITIES
S.O. Compton, J.-Y. Rasplus and A.B. Ware
WHAT ARE FIG TREES AND FIG WASPS?
Fig trees are a group of approximately 750 speCIes placed in the genus Ficus (Moraceae), and
characterised by their unique inflorescence - the fig. Around lOS Ficus species are found in Africa,
where they range in size from small shrubs to huge rainforest emergents (Berg, 1990). The term 'fig
wasps' is sometimes applied to all the hymenopterans that develop inside figs, but more often is restricted
to certain chalcid wasps (Hymenoptera, Chalcidoidea), belonging mainly to a single family, the
Agaonidae (Boucek, 1988). All agaonid species are associated exclusively with fig trees.
The few
detailed studies of parasitoid fig wasps have found that they are actually 'entomophytophagous' (Zerova
and Fulsov, 1991) inside galls produced by other species, feeding initially on plant tissue and only later
destroying the larvae of their hosts (Abdurahiman and Joseph, 1978a,b).
Most of the interest shown in Ficus biology has centred on the mutualistic interaction between the trees
and the pollinating fig wasps (Agaonidae. subfamily Agaoninae).
Fig wasp parasitoid communities
nonetheless also offer many interesting avenues for research, due to such features as their complexity,
the replication provided by the communities centred around each of the hundreds of Ficus species, and
their predominantly tropical distribution, which sets them apart from the better-known temperate
parasitoid communities.
Here we first describe the fig wasp communities associated with two African fig trees, emphasising the
consequences of the trees' unusual phenological characteristics and the unique structure of their
inflorescences on host accessibility to parasitoids. We then describe how homopterans can adversely
effect the fig wasp panlsitoids, through their attraction of predatory ants.
Finally, we expand our
perspective and discuss geographical int1uences on the composition of the local parasitoid communities
154
found on the two trees and then review what is known of the factors influencing species richness among
African fig wasp communities in general.
THE FIG ENVIRONMENT
Interactions between fig wasp parasitoids and their hosts are greatly influenced by the morphology of
figs, because their structure governs host accessibility. Usually spherical in shape, fully developed figs
of Afrotropical species vary in size from only about 8 mm diameter in F. aJltandronarum bernardii to
larger than a cricket ball in F. sycomorus form sakalavarum. Each fig is lined on its inner surface by
hundreds or thousands of unisexual flowers. Agaonines transport the pollen into the fig via the ostiole,
a bract-lined tunnel. Pollination occurs while the agaonines are galling some of the female flowers and
ovipositing down their styles. In monoecious Ficus species the majority of the flowers have ovules that
are accessible for oviposition (Bronstein, 1988; Nefdt, 1989), whereas in dioecious species the flowers
in figs of 'female' trees have very long styles that prevent successful oviposition (Verkerke, 1987).
Consequently the pollinators fail to reproduce and these figs produce only seeds.
Oviposition by a few non-pollinating fig wasps also takes place after entry through the ostiole, but most
species, including all the putative parasitoids, use their long ovipositors to reach the ovules from the
outside, through the walis of the figs. Not surprisingly, parasitoids associated with trees that produce
smaller figs also have shorter oviposrtors than species attacking hosts that develop in larger figs
(Compton, unpublished). The latter have some of the longest ovipositors, relative to their body size, of
any hymenopterans (Compton and Nefdt, 1988).
Fig wasp life cycles are closely integrated with the developmental cycle of the figs. The first potential
colonisers of a new fig crop are certain species belonging to the subfamily Epichrysomallinae which gall
fig primordia during the 'pre-floral' stage (Galil, 1977), before individual flowers have differentiated.
These galled figs develop into grossly distorted structures incapable of supporting most other fig wasps,
apart from some parasitoids specifically associated with the epichrysomalids (Compton and van Noort,
in press).
155
Most galling fig wasps only utilise figs that are at the next stage of development, the 'female' stage. At
this time the figs are 'receptive' and draw their specific species of pollinator to the trees through the
release of volatile chemicals, which are not attractive to other agaonines (van Noort et ai. 1989; Ware
and Compton, in press; Ware et ai., in press). Parasitoids probably use other cues to find the trees,
because they tend to arrive at the trees during the following 'interfloral' stage, when the pollinator larvae
are present (Compton and Dallas, unpublished). After the progeny of the various wasp species complete
their development within the figs they emerge together during the 'male' phase, when the female
pollinating wasps of the next generation collect the pollen prior to dispersing. After mating is completed
the male agaonines chew a communal exit hole, through which the female wasps escape. The males of
many non-pollinating species are also capable of producing exit holes, but this does not appear to be the
case with at least one parasitoid, Apocrypta guineensis, and this can lead to mass mortalities of adult
females in heavily-parasitised figs where few if any male pollinators were present (C. Zachariades, pers.
comm.).
Figs vacated by pollinators become attractive to fruit eating vertebrates and any wasps that have not
completed their development by this time risk being eaten by birds, fruit bats etc. In strongly seasonal
climates, such as those experienced in the Cape province of South Africa, fig development times are
extended during the winter period and can last several months, whereas in the summer wasp generations
cycle within a few weeks.
TWO EXAMPLE COMMUNITIES
The Trees
Among African fig trees, F. burtt-davyi and F. sur are the two species with distributions that extend the
furthest south. F. burtt-davyi (subgenus Urostigma, section Galoglychia) is a monoecious species with
an exclusively southern African distribution extending from Mozambique to the southern Cape Province
(van Greuning, 1990). It can grow as a strangler of other trees (Compton and Musgrave, submitted),
as a shrub on coastal sand dunes, or as a rock-splitter growing out from bare rock faces. The figs of
156
F. burtt-davyi are small, reaching a maximum diameter of about 15 mm at maturity and are produced
in the leafaxils.
F. sur is also a monoeclOus species, but belongs to subgenus Sycomorus.
It has a much wider
distribution than F. bunt-davyi, extending from the Cape northwards throughout the less arid regions of
the continent (Berg, 1990). F. sur is often found in riverside vegetation, where it can reach a far larger
size than F. burtt-davyi. The figs are also larger, reaching over 30 mm at maturity, and containing
around 3000 flowers. They are typically borne on leafless branches growing out from the old wood. On
certain trees a few of the fig-bearing branches are produced below ground level, resulting in 'geocarp'
figs projecting from the soil surface.
The wasps
The fig wasp community associated with F. bunt-davyi around Grahamstown (eastern Cape Province,
South Africa) consists of the pollinator (Elisabethiella baijnathi), three other ovule-gallers (Phagoblastus
sp., Otitesella uluzi and O. sesquianellata) and two parasitoids, Sycoscapter sp. (= Sycoryctes sp.) and
Philotrypesis sp. E. baijnathi and Phagoblastus females lay their eggs from the interior of the fig, while
the other species oviposit from the outside. Both parasitoids will attack all the gal1er species, although
the pollinator may be the preferred host. The four phytophagous fig wasps do not reproduce on any
•
other tree species in the Grahamstown area, whereas the parasitoids cannot at present be distinguished
from congeners which develop in the figs of F. thonningii, and may turn out to be associated with both
trees.
In the Grahamstown area the species which form the F. sur fig wasp community are all specifically
associated with this tree. The pollinator of F. sur in Grahamstown is always Ceratosolen capensis, while
the non-pollinating fig wasps comprise the parasitoid Apocrypta guineensis together with the gall-forming
Sycophaga cyclostigma (which enters the figs to oviposit, like the pollinator) and three Apocryptophagus
spp.
157
A. guineensis is catholic in terms of its host insect requirements, and individuals have been reared from
all the potential host species.
Uniquely among the species in either of the two Grahamstown
communities, more than a single individual of A. guineensis sometimes emerges from the very large
galled ovules produced by one of the Apocryptophagus species. The fig wasp community associated with
F. sur in West Africa is more complex. At the Ecological Station at Lamto, in Ivory Coast, where this
is by far the most common Ficus species, 11 fig wasps species have been recorded. These comprise two
species of pollinators (c. capensis and C. jZabellatus), five gall formers (Sycophaga cyclostigma, three
Apocryptophagus species and an epichrysomalline, Acophila sp.) and four parasitoids (A. guineensis, two
Sycoscapter spp. and a eurytomid, Sycophila sp.). A survey of the other 15 Ficus species in the Lamto
area (Rasplus, unpublished) found that these wasps were generally associated only with F. sur. The two
Sycoscapter spp. parasitoids were exceptional in that were also reared from related Ficus species (F.
sycomorus and F. vallis-choudae).
C. jZabellatus appears to be a genuine second pollinator of F. sur, a situation which has also been
recorded from other African fig trees (Michaloud et ai., 1985). Apocryptophagus sp.l (a species close
to A. gigas) forms large galls that protrude into the central cavity of the figs and can completely occlude
it.
Oviposition by this species, and Acophila sp., occurs before pollinator entry (Figure 1).
Apocryptophagus sp. 2 oviposits at about the same time that pollination is occurring, while the third
species in the genus oviposits at a later stage (Figure 1).
Among the parasitoids, oviposition by Sycophila occurs slightly later than that of the Acophila sp. (Figure
2). Like Apocrypta species, it is probably entomophytophagous, exploiting the gall tissue made available
by Acophila. Oviposition by the Sycoscapter species occurs somewhat later. The oviposition period of
A. guineensis is unusually broad, and consequently this species must be exploiting galls containing host
larvae of greatly varying sizes.
158
Sycophaga
Ceratoso/en
4
6
1 1;I!iP~
Apo. sp.l
Apo. sp.2+3
Ac:ophi/a
&
10
12
14
16
18 20 22 24
fig diameter (mm)
FJgUre 1. A comparison of the sizes of F. sur figs probed by gall-making fig wasps.
Number of
Apoc:rypta
8
10
~
I_~:
Sycosc:apler 1
Sycoscapler 2
Syc:ophi/a
12
14
16
1&
20
22
24
26
fig diameter (mm)
Figure 2. A comparison of the sizes of F. sur figs probed by parasitoid fig wasps.
159
Spatial structuring of resources within figs
The figs of F. burtt-davyl and F. sur enlarge considerably after pollination (Baijnath and Ramcharun,
1983; Baijnath and Ramcharun, 1988), and ovules become progressively more distant from the periphery
of the figs (Nefdt, 1989). Variation in the ovipositor lengths of those fig wasps that oviposit from the
outside of the figs might therefore be expected to reflect the timing of their oviposition, with those
species with longer ovipositors utilising hosts in older figs. Alternatively, variation in ovipositor lengths
might also reflect differential exploitation of hosts at varying depths in the figs (Bronstein, 1991).
In the F. burtt-davyi community the ovipositors of the species that oviposit from the outside of the figs
show a progressive increase in length that corresponds to the periods when they oviposit (Compton and
Nefdt, 1990; Compton, in prep.). Individual ovipositor lengths among the parasitoids associated with
F. sur are highly variable (Figure 3). Ovipositor lengths again reflect the timing of oviposition by the
two Sycoscapter species, but the ovipositors of A. guineensis are shorter than would have been predicted
(Figures 2 and 3).
8
6
4
Sycoscapter 1
2
5
6
7
8
9
10
11
12
13
Oyipositor length (mm)
Figure 3. Variation in the ovipositor lengths of parasitoids associated with F. sur.
Sycoscapler sp. I, 9.2
+
0.4, n = 13; Sycoscapler sp .2, 10.3
160
+
Mean ovipositor lengths (+ S.D.) were:
1.3, n = 12; Apoclypta guineensis 7.1
+
1.1, n = 20.
Contrary to earlier ideas (Janzen, 1979), style length variation within the figs of African Ficus species
is unimodal (Nefdt and Compton, in preparation), with no separation into discrete long- and short-styled
flowers. The ovipositors of the E. baijnathi females that pollinate F. bunt-davyi are longer than the
majority of the styles and most ovules are therefore available for oviposition (Compton and Nefdt, 1990).
In contrast, the C. capensis females that pollinate F. sur have relatively shorter ovipositors, and the
longer-styled flowers are consequently unavailable (Nefdt, 1989).
Flowers with longer styles have ovules closet to the outer surface of the figs, and therefore any larvae
they contain are potentially easier to reMh by parasitoids probing from the outside of the figs.
Conversely, if the parasitoids' ovipositors cannot reach them, larvae developing in the shortest-styled
flowers may occupy 'enemy-free space' (Jeffries and Lawton, 1984) and be immune from attack (G.
Michaloud, in Kjellberg and Valdeyron, 1984). Interestingly, E. baijnathi females preferentially oviposit
into the shorter-styled flowers in F. bunt-davyi figs, but as the density of wasp foundresses increases,
so progressively longer styled flowers are used (Figure 4).
Consequently, when they are at higher
densities the pollinator larvae may be more accessible to probing parasitoids.
0.8
o FEMALES
•
E
E
MALES
0.7
(f)
.c
"-'
01
C
0.6
Q)
r
1
Q)
'0.4
i
I
t
0.3 -l-,- - - - . - - - , r - - - , - - - - - - . - - - - - ,
5
4
3
2
1
o
Number of foundresses
Figure 4. Variation in the mean 'style lengths of flowers occupied by Eiisabezhiella baijnazhi progeny in relation to the number of
foundress females entering the figs. The distribution of female progeny changes with increasing density, with more wasps closer
to the periphery of the figs, where they are potentially easier to reach by parasitoids ovipositing from the outside of the figs.
161
The parasitoids associated with F. burtt-davyi have ovipositors of sufficient length to reach all the ovules
in the figs (Figure 5) and utilization of hosts within the whole range of style lengths has been confirmed
(Nefdt, 1989). This is despite the often tortuous routes taken by the ovipositors en route to the ovules
(Compton and Nefdt, 1988) and shows that host larvae in the shortest-styled flowers do not occupy
'enemy-free space' in the sense that they are immune from attack from parasitoids. A lack of immunity
is also evident when rates of parasitism in F. burtt-davyi flowers with different style lengths are
considered, as larvae in the shorter-styled flowers are just as likely to be attacked as those at the
periphery of the figs (Figure 6; Nefdt, 1989).
Within F. sur figs, the depth of the fig wall and the thickness of the zone containing the ovules are highly
correlated with overall fig diameter (Figure 7). These changes in the depths that the parasitoids have to
probe is reflected in the lengths of their ovipositors, which correspond closely to the depths they have
to penetrate (Figure 7). This suggests that spatial partitioning of host utilisation by the parasitoid species
is likely to be absent, as is the case with the wasps in F. burtt-davyi figs (Nefdt, 1989).
10,
J
8~
SyCOryles OVIPOSITOR LENGTH
61
41
J
2j
111111
I
10,
8~
>0
Z
W
::l
1111
Philotrypesis OVIPOSITOR LENGTH
J
6J
J
4i
1
2j
aw
II
LL
gal
i
DISTANCE TO OVULES
1201
100
1
::1
,,!
20~:
I l~
ll.
I
. '._
._..
Lo
I
.l1~H4-
2.0
3.0
4.0
LENGTH(mm)
Figure 5. Frequency histograms indicating the lengths of Sycoscapler (top) and PhiiOirypesis (middle) ovipositors in relation to
the distanct: they must travel from the outside of the figs to reach the ovules of F. bllrtt-davyi (hollom). The distance from the
ovules was measured using figs at the 'inter-floral' phase. Philolrypesis oviposits slightly earlier than Sycoscapler during this
period.
162
.wo
CI)
ll.
~
150
...J
«
b
I-
100
50
0.7
0.9
1.5
1.1
STYLE LENGTHS (mm)
Figure 6. Frequency histograms showing the distribution of flowers of varying style lengths within the figs of F. bunr-davyi (top),
the numbers of those flowers containing fig wasps (centre) and the relative numbers of parasitoids (bottom). The fig wasps present
were Eiisabethieila baijnathi, Otitesella sesquianellata and Otitesella uJuzi (gallers), together with Sycoryc!es sp. and Philorrypesis
sp. (parasitoids).
width (mm)
14-
o fig-wall
12
.. flowers
10
7" fig-wall
~
Sycoscapter 1
~
Sycoscapter 2
~
•I-
-.
Apocrypla
8
•
••
I
•e•_.•
.-
l
0
6
0
0
4
2
0
5
10
15
20
25
30
40
35
fig diameter (mm)
Figure 7. Changes in the distance that parasitoids must probe with increasing diameter of F. sur figs. Also indicated are the
potential distances that the parnsitoids can probe. in relation to the size of the figs at the times when they oviposit. Each block
delines one standard deviation from the mean ovipositor length and the mean diameter of figs that were probed by each species.
163
Wasp longevity in relation to tree Phenologies
c:=:::=J-l
1 0
2 l:':····.:.;·,: :j
I·:··.······:··:···:):·::····:·]
~
!: .... ;". >:::.' '.~
n::::::J
8
9
>-{: . ': :··,·:····::·:·:<:/:l
.~
..... :.:. :..... .
. . -::,. ::.:.\,-,,:.,:,.
10
11
E::J
I':::·:(/'{··e::-!
12 --( '.,.,.,:, i""'''''' ,,·,,1
13~
14 - "
15-~m
I.:;::::: .:.:.'::.,.;::. " .... '''''\••:.!
!:.
~
..... ,.. !
16~
o
1:4;}~Vil-It{,'·.
:;;:t:A}·:;::,:j
II::]
..:: .. ",:.<:. ".'
17
18
19
20
a:
~
· .. ·",..
,:.".'::::':':'::-1
1
22
~
25 - - - - - - - - i.........".:--.....
26 --------<c::l
ro
:2
Z
ill
o
oo
ill
ill
a:
l-
~ =~i4*k?MBj;:"'
28~
=
-~J
ill
-~
J--,
27~
"0--
29
=
. ~ . =:{~¥g>i;f'}t)M,Bh\!
~ ===-=--=--=-------..-.. ~.
32-~=·
=~'
33----------------34
35 -1' if::: /:;':'}\iY;,t:l
36
37--------------38-~!;:SjZrl
39-~=
40-c~
41=
42
r:::::::J----t':;::'::: : . . . '..:,:::,/.);::.:;:,;::-::j
43 -:::'::;'i.:::·': . .;:.:::,,(.\.,::;::..,
0--,..'.·. . '. .,. . ., '., , ,'. ,
440
45'
',>:':.:,'."...".;.!
:~
~
46
----Cl
•
-~
49
50-~c=DJ_
".,.'..... ,....
i:;l-----------------
(:.:i •.
51 ------------i,'::::: ..::.1------------{
52 --i
APRIL
1985
.\:;~*/?
::;
5
15 20 2:5 30 35
SURVEY NUMBER
c.:::J 50-500
Crop size 0 >500
10
40
45 50
=<50
MAY
1987
Figure 8, The fruiting patterns of 52 F. bum-davyi trees growing in Grahamstown, The small crops of very short duration (for
example on trees 26 and 49) aboned at an early stage of development.
The fruiting phenologies of 52 F. burtt-davyi trees growing as rock-splitters in Grahamstown are
summarised in Figure 8.
Figs were present on a proportion of the trees throughout the two year
sampling period, with an overall average of 14.02 (27.0%) bearing figs at anyone time. Thus, although
164
there was some seasonal variation in the numbers of crops, with peaks during spring and autumn (Figure
9), figs were available continuously for colonisation. Crop sizes ranged from just a single fig to several
tens of thousands. On each tree fig production was synchronised and only two of the crops
«
2 %) were
sufficiently asynchronous for wasps to be able to immediately oviposit on the same tree that they had
emerged from.
20
18
Cf)
F. burtt-davyi
16
0...
0 14
a:
0
LL
12 '
0 101
a:
ill
CO
8-
2
6-
:::;J
Z
42:
_1
0- ,- -,-,
i5
__
1~5
'1'0'
1-'-20
25
'3035
4cf45'
JAN
JULY
JULY
JAN
1985
SURVEY NUMBER
50
MAY
1987
Figure 9. The numbers of F. burtt-davyi trees bearing figs in Grahamstown over a two year period. Some figs are present in the
area throughout the year, but the abundance of fruiting trees tends to decline during mid-summer and mid-winter.
1
~=}-{J]h
-.r~
~=
21
--'----'--'-,.J.,IQ-l_,_j~it',
i-[]:~
3 li i t I
I
4 c:::::J---;
5ni-1
~
.
~
~
[J'-Ir:~
7
~
9-~=
,
' ,
10-~
11-----------------------=>-----
+1
tl} 12,
g:
"I
8~-=
Z
8
~
13 - - , - !
r-r
15 ~
16
L
0--;
1------=----=-,'----...:..--------,--'
H"
:---0_.:,-,_----'rlL..L-j
I ;rr-:-:---rr'1
~LJ
o=:J-'D
~
h==rri'rll
LLJ,-J
17 [j-C{~
18 /',D--}-D-;
"llil!'
JULY
5
r,-----;
i
'il~-.:
,,:1111:
10 15 20 25 30 35 40 45 50
1985
Periods when
sell pollination
was possible
AUGUST
SURVEY NUMBER
CO
1987
Crop size
>500 50-500 <50
Figure 10. The fruiting pat1ems of 18 F. sur trees growing around Grahamstown. Vertical lines within the bars indicate periods
when wasps were emerging while unpollinated figs were present on the same tree.
165
20
F. sur
18
Cf)
16
!l..
0
14
()
ll..
12l
cc
0
cc
ill
aJ
~
:::J
Z
I
I
!
10 -i
I
I
8
-I,
6
j
i
,
i
4!
l
I
2~
I
0
1
I
JULY
,
I
5
'
1985
'110'
I'
'1'5' , ,
JAN
'20' , , 1~5
JULY
L
I
J
L
30
SURVEY NUMBER
I
,
,
I
5'0
AUG
1987
Figure 11. The numbers of F.sur trees bearing figs around Grahamstown over a two year period. Figs are present in the area
throughout the year, with no clear seasonal trends in abundance. Count numbers 8 and 9 are slight under-estimates, resulting from
certain trees being inaccessible due to flooding.
Fruit production among 18 F. sur trees in the same area showed a rather different pattern (Figure 10).
On most trees the figs were present for a much greater proportion of each year (mean crops per sampling
period was 10.6,=
58.9% of the trees) and there were no obvious seasonal patterns in fruiting
~
frequencies (Figure 11). Fruit production within crops was also highly asynchronous on many trees,
providing frequent opportunities for self pollination (Figure 9).
The distance that female fig wasps can disperse between trees, and their chances of successfully doing
so, will depend on their longevity. In the laboratory, when kept at moderate temperatures and high
humidities, adult females of pollinating fig wasps survived at most three days (Table 1). The availability
of sugar solution did not increase longevity, suggesting that adults do not feed. Adult females of the
other galling species survived rather longer, but sugar only extended the lifespan of O. uluzi (S.
cyclostigma inexplicably survived longer if only water was present). Survivorship patterns were different
among the parasitoids, all of which lived for extended periods only if sugar was available (Table 1).
166
Table l. The longevities of adult female fig wasps maintained at 20'C and 75-80% relative humidity on diets of either distilled
water or a 10% sucrose solution. Mean longevities for PhiioErypesis sp. and Apocrypra guineensis with sugar are underestimates
as the trial were terminated after 40 and 60 days respectfully.
WI
Water
Z
Sugar
P
Species
N
.
Mean
(days)
Range
N
Mean
(days)
Range
6
!!iiiiii!i
GalJers
Elisabelhiella baijnathi
24
1.25
1-2
25
1.40
1-2
1.10
ns
Cerazosolen capensis
20
2.15
1-3
20
2.15
1-3
-0.03
ns
Phagoblastus sp.
16
3.44
2-5
17
4.06
2-8
0.68
ns
Sycophaga cyclostigma
20
8.40
4-10
20
5.75
2-10
-3.85
***
Otitesella uJuzi
26
3.88
2-6
30
14.9
2-34
5.43
......
ApocrypEophagus sp. 1
10
9.90
7-12
15
8.27
2-15
-1.21
Philotrypesis sp.
18
1.5
1-5
22
24.86
1-40
5.02
......
Sycoscapzer sp.
15
4.73
1-9
25
20.64
1-30
4.06
....*
Apocrypla guineensis
20
3.35
2-6
14
39.57
3-60
4.54
***
ns
Parasitoids
M
ns
= P > 0.05; .. *.. = P < 0.001
The longevities of adult fig wasps appear to correspond with different oviposition strategies. Those
species which enter the figs to oviposit must lay all their eggs within a day or so after entry, as the
flowers soon begin to deteriorate (Greef and Compton, personal observations). They are pro-ovigenic
(Table 2), with short adult life spans and do not feed. Apocryptophagous sp.l, despite ovipositing from
the outside of the figs, appears to have a similar strategy, and seems to oe adapted for rapid oviposition.
In contrast, the three parasitoids (and Jhe galler, O. uluzi) are syn-ovigenic (developing their eggs
progressively), with extended life spans, feed on sugar sources and appear adapted for slower rates of
oviposition. This is presumably a reflection of the greater difficulties they experience in host finding.
As the gaps between F. bunt-davyi crops on a single tree typically extend for several months, the adult
wasps cannot normally colonise figs on their natal trees, but have to fly off in search of other fig-bearing
trees in the area. The situation is different with F. sur, where wasp populations can often cycle on
individual trees, without any repeated need for dispersal.
167
Table 2. Egg loads of galler and parasitoid fig wasps associated with F. bum-davyi and F. sur. 'Internal' ovipositing species lay
their eggs after entering the figs, while 'external' species lay their eggs from the outside of the figs. Females of syn-ovigenic
species contained both mature and developing eggs, whereas pro-ovigenic females contained only mature eggs.
Species
Oviposition
N
Number of eggs
Mean
eM
SynlPro-ovigenic
Range
i
Gallers
Elisabethiella baijnathi
Internal
20
79
67- 94
Pro-ovigenic
CeralOsoien capensis
Internal
20
238
180-370
Pro-ovigenic
Phagoblastus sp.
Internal
20
88
59-121
Pro-ovigenic
Sycophaga cyclostigma
Internal
20
124
96-158
Pro-ovigenic
Otilesella uluti
External
9
91
65-119
Syn-ovigenic
Apocryplophagus sp.l
External
20
310
210-360
Pro-ovigenic
Philotrypesis sp.
External
10
25
15- 36
Syn-ovigenic
Sycoscapter sp.
External
10
41
34- 51
Syn-ovigenic
Apocrypla guineensis
External
10
20
8- 40
Syn-ovigenic
Parasitoids
Interactions with ants
A complex mutualism involving ants and Hilda patruelis (Tettigometridae), a honeydew-producing
homopteran, develops on trees belonging to several African Ficus species, including F. sur (Compton and
Robertson, 1988, 1990). Ants are attracted on to the figs by the honeydew', where they then disturb wasps
that are trying to oviposit through the fig .wall, capturing some of them. This results in lower rates of
parasitism by A. guineensis, which can be more or less excluded from individual figs or even whole trees
where ant densities are highest. External-ovipositing ovule-gallers like Apocryplophagus spp. are also
affected, but fig wasps that oviposit from the inside of the figs, such as the pollinators, are relatively immune
from the ants, because they spend little time on the fig surface. Consequently, the presence of H. patruelis
leads to reduced levels of parasitism of the tree's pollinators, together with reduced ovule-destruction, and
an indirect mutualism between the tree and the ants is established.
In contrast to F. sur, F. burtt-davyi is rarely colonised by H. patruelis. As alternative attractants for ants
are also uncommon, the parasitoids can probe the figs with much reduced risks of predation. The H.
168
patruelis - ant combination occurs on a large proportion of F. sur trees and others in Ficus subgenus
Sycomorus throughout Africa (Cushman et ai., in prep.). The frequent presence of ants on trees belonging
to this subgenus has not influenced fig wasp species richness- they have just as many associated wasps as
other monoecious fig trees (Compton and Hawkins, 1992). However, among the drosophilid flies that also
breed in the figs, it has resulted in changes in courtship behaviour that improve the chances of escaping from
the ants (Lachaise and McEvey, 1990). Similar selection pressures for ant avoidance are likely to be
operating on the parasitoids which utilise species such as F. sur, but comparisons of features such as probing
times or mobility between species such as A. guineensis and the parasitoids from F. burtt-davyi have not
been made.
Regional scale influences on community richness
Table 3. Species richness of local and sub-regional fig wasp communities associated with F. sur in southern Africa. 'North· includes
Zimbabwe, Zambia and Malawi. Local parasitoid richness is significantly lower in the Cape than in Natal (Mann-Whitney,Z = 2.88,
P < 0.01) and the Transvaal (Mann-Whitney, Z = 1.98, P < D.OS), but not elsewhere. Local gaUer richness in the Cape was not
significantly different from the three other subregions.
Subregions
Number of
Samples
Regional Pools
Gallers
Parasitoids
Mean Local Richness
Gallers
Parasitoids
North
10
6
3
2.80
1.10
Transvaal
14
7
S
3.36
1.57
Natal
8
6
4
3.7S
1.88
Cape
IS
S
3.00
1.00
0,
*
Grahamstown is situated close to the southern edge of the range of F. sur, and several species found further
north are absent from this sub-region. A total of twelve fig wasp species have been recorded from F. sur
in southern Africa, of which 11 were collected in Transvaal, 10 in Natal and just six in the Cape Province
(Table 3). Of the six species which fail to reach the far south of the continent, five are putative parasitoids.
This is not due to undersampling in the Cape, as sample-recruitment curves (Figure 12) suggest that no new
species are likely to be collected there. Within South Africa there is thus a north to south decline in the
species richness of the communities in the different sub-regions, but this simple latitudinal pattern does not
appear to extend further north into the tropics, where sub- regional richness may even decline (Figure 12).
169
The variation in the sizes of the sub-regional communities is reflected in the species richness of local fig
wasp communities found on individual trees (Table 3). Local communities in the Cape are significantly
depauperate in parasitoid species compared with the other regions of South Africa, but are not depauperate
in gallers.
F. burtt-davyi has a much smaller distribution than F. sur, but habitat-related differences in fig wasp
community composition can 'be detected even within a localised area of the eastern Cape. In the coastal
forests around Alexandria the ovule-galling Phagoblastus is far more common than around Graharnstown,
and an additional Sycoscapter parasitoid is also present. This increased sub-regional species pool has a
corresponding influence on average local community richness (the wasps colonising individual trees), which
is significantly higher in the forests (Table 4). The rarity / absence of certain species around Graharnstown
may reflect its more extreme climate, with hotter summers and colder winters than the coastal areas,
although they are only about 80 km apart.
PARASITOIDS
GALLERS
NORTH
8,
8,
6,
6;
0 ••
I
~
4j
41
••••
••
2; ~ .......•••
!
I
!
21
;1
:/
TRANSVAAL
••
•••••••••••
Cf.l"4i.
W
!
I
C3 2 i!
W ,!
~
~
:.
,d
••••••••
, I
NATAL
u..
8,
0:
W 6
••••••••
CD
4; ••••••••
~4j
:::J
i
6;
21 ••••
_ _ _ _ _ _ _ _k -_ __ _
(j')
0
.,
8,
:
Z 2{
: i
!
I
2i
J
I
I'
il
CAPE
8;
61
••••••••••••
2: •• •
2
I
4;
2j
: /
o
6,
4
6
8 10 12 14
i,···············
o
2
4
6
8 10 12 14
COLLECTIONS
Figure 12. Sample recruitment curves for collections of wasps from figs of F, sur in four sub-regions of southern Africa. 'North'
includes Zimbabwe, Zambia and Malawi. Transvaal, Natal and Cape are provinces of South Africa. The flattening of the curves
suggests that all the species associated with F. sur in the Cape have been collected.
170
Table 4. A comparison of the species richness of local fig wasp cornrnuruties associated with F. bunt-davyi in forest and inland areas
of the eastern Cape (South Africa).
•
MAiM
w·..
Alexandria Forest
MannWhitney
Grahamstown
P
Z
N
Mean
Range
N
Mean
Range
Gallers
10
3.40
2-4
30
2.50
1-4
-2.84
Parasitoids
10
2.30
1-3
30
1.53
0-2
-2.81
5
1W&
....
....
H
.... = P < 0.01
AFRICAN FIG WASP CO:MMUNITlES IN GENERAL
As with F. sur and F. burtt-davyi , African fig wasp communities are typically composed of a single species
of pollinating wasp, with larvae that develop inside galled ovules (Verkerke, 1989), together with other wasp
species that gall the ovules, and their parasitoids. In rare instances the communities may also include a
second species of pollinator, as in F. sur, (Michaloud et al., 1985) or wasps that gall the vegetative parts
of the figs (Compton and van Noort, in press; Rasplus, unpublished).
Table S. The distribution ofparasitoid genera within the southern African subgenera and subsections of Ficus.
Subgenera
Sycidium
Sycomorus
Subsections
Urostigma
Urostigma
Galoglychia
Plaryphyllae
Chlamydodorae
Caulocarpae
Wasp genera
Apocrypta
+
Watshamiella
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Sycoscapler
+
Philozrypesis
+
Onnyrus
Eurytomidae
(various)
+
+
+
+
So far as is known, there is a general consistency in trophic relationships within the various taxonomic
groups of fig wasps (Compton and van Noort, in press), but this may partly reflect the small number of
species that have had their biology investigated. In southern Africa, gallers are found in Agaoninae,
171
Sycoecinae, Epichrysomallinae, Sycophaginae and Otitesellinae (all Agaonidae), while parasitoids are found
within the Sycoryctinae (Agaonidae), the Eurytomidae and Ormyridae.
The mix of putative parasitoids
potentially associated with each fig species is largely independent of a tree's taxonomic affiliations because
most of the genera that contain parasitoids are widely distributed among the taxonomic subdivisions of Ficus
(Table 5). Apocrypta is an exception as it is restricted to subgenus Sycomorus (Ulenberg, 1985), where it
'replaces' Philotrypesis.
Host tree specificity is well developed among the pollinating fig wasps, with each Ficus species generally
having its own unique species of agaonine (Wiebes and Compton, 1990). Tree specificity is also well
developed among the gall-forming sycoecine wasps (van Noort, 1991), epichrysomalline wasps (Rasplus,
unpublished) and in the parasitoid genus Apocrypta (Ulenberg, 1985). Equivalent data for other parasitoid
groups is not available. Similarly, the extent of host insect specificity among parasitoid fig wasps is largely
unknown, although an association between epichrysomalline and eurytomid fig wasps is evident (Compton,
in press). Eurytomids have not been recorded from Ficus species that do not also support epichrysomalline
fig wasps, and this relationship also extends to individual crops or figs.
In the southern African fig wasp communities analyzed by Compton and Hawkins (1992) and Hawkins and
Compton (1992) the total fig wasp faunas associated with different Ficus ranged from about 3-30 species,
with the numbers of putative parasitoid species varying between 1 and 18. Parasitoid: galler ratios varied
from about 3:1 to 1:3, with phytoagu~
species outnumbered parasitoids in many of the communities.
Factors influencing the species richness of the gallers in the communities included ecological factors such
as the size of the trees and the habitats where they occur, but species-area effects were not significant
(Compton and Hawkins, 1992). The numbers ofparasitoid species were strongly correlated with the number
of gallers in each community, and thus presumably the diversity of potential hosts. Dioecious fig species
also supported fewer wasps than monoecious species.
Only sub-sets of the total fig wasp faunas associated with each Ficus species form the local communities
found on any individual crop. Nonetheless, as many as 18 species, 11 of them putative parasitoids, have
been reared from one F. thonningii crop, with up to nine species (five parasitoids) occupying a single fig
172
(Compton, unpublished).
The major factor determining parasitoid community richness at the level of
individual crops was the size of the regional pool associated with that particular tree species (Hawkins and
Compton, 1992). Local and regional diversities were linearly related, with no evidence of saturation of local
communities.
Latitudinal gradients in local community species richness were also present among the
parasitoids, with marginally fewer species present at more tropical latitudes. No equivalent gradient was
detected among the gall-forming groups.
DISCUSSION
How do fig wasp parasitoid communities compare with the better-known north temperate systems that are
also based around endophytic hosts? One noticeable feature is that fig wasp parasitoid: host ratios are
markedly lower than in parasitoid communities centred on hosts that gall or mine trees (Askew, 1975; Askew
and Shaw, 1986) and they are more typical of those found in early successional communities centred on
'unapparent' herbs (Askew, 1980; Hawkins, 1988; Hawkins et al., 1990; Tscharntke, 1992). Another
'early-successional' feature of fig wasp parasitoids may be their high host plant specificity, at least in the
best-studied genus, Apocrypta. This is against the general pattern, where parasitoid communities on trees
are dominated by generalists (Askew, 1980; Hawkins et ai., 1990; Rasplus, this volume) and could explain
the lack of saturation in fig wasp communities (Hawkins and Compton, 1992).
The parasitoid faunas
associated with F. bunt-davyi and F. sur ilonetheless show that not all fig wasp parasitoids are necessarily
tree specific.
Early successional communities contain host plants with low apparency, that are relatively difficult to detect
by parasitoids (Askew and Shaw, 1986).
Despite their often large stature, fig trees may also be
exceptionally unapparent to searching parasitoids.
This is because the trees can be at low densities,
especially in tropical forests (Gautier-Hion and Michaloud, 1989), and at anyone time only a fraction of
them are bearing figs, and hence offer potential hosts. The problem of host finding is especially acute for
parasitoids associated with dioecious FicLis species, where at times of the year figs on most of the trees may
contain no hosts at all (Kjellberg et al., 1987; Nair and Abdurahiman, 1984). Furthermore, on trees with
173
phenologies like that of F. bunt-davyi, only one parasitoid generation can be produced before dispersal is
required again.
There is a second, and quite different, possible explanation for the low parasitoid:galler ratios in fig wasp
communities, which would also explain the apparent prevalence of entomophytophagous parasitoids. Seeds
of most plants are rich in 'secondary compounds', many of which are toxic and may have a defensive
function (Janzen, 1969). Fig seeds, however, are unlikely to contain any such defensive compounds, because
the trees are totally reliant on pollinating fig wasps, the larvae of which also feed on the seeds.
Consequently, fig seeds may be unusually easy to eat.
Price (1991) has suggested that parasitoids are less likely to be regulating their host populations in
communities like those of fig wasps where ratios of parasitoid species to host species are low. The results
of a life table study of the wasps from F. bunt-davyi agree with this prediction (Compton and Robertson,
in prep.). Average rates of parasitism of E. baijnathi in Grahamstown tend to be less than 10% and key
factor analysis suggests that parasitoids are a minor factor in comparison with the mortalities that occur
during the movement of adults between trees. Pollinator parasitism rates were perhaps slightly higher at
Lamto, where about 25 % of the emerging adults were A. guineensis.
In concluding a review of fig wasp parasitoids it is perhaps prudent to emphasise just how little is known
about them. In particular we lack such basic information as whether they are all genuinely parasitoids, how
host specific they are or even how many species we are dealing with. Abdurahiman and Joseph (1978b) and
Joseph (1984) have shown that phytophagous fig wasps have enlarged acid glands, the contents of which are
presumably used to gall the ovaries, whereas in parasitoid species these glands are reduced.
Direct
observations on the biology of even a majority of the species in a continent as under-studied as Africa is
unlikely ever to happen, and this anatomical difference may provide the 'short-cut' that is required, once the
species have been clearly delimited.
Systematic treatments of three African phytophagous groups are
available or in preparation, covering the Agaoninae, Sycoecinae and Epichrysomallinae (by J. T. Wiebes,
S. van Noort and J.-Y. Rasplus respectively), but Ulenberg's (1985) revision of Apocrypta remains the only
detailed coverage of any of the parasitoid groups. This lack of basic taxonomic information remains the
174
major obstacle impeding community level studies of these fascinating insects.
ACKNOWLEDGEMENTS
Thanks to Brad Hawkins and Pat HuBey for their comments and to Helen Dallas, Rory Nefdt and Costas
Zachariades for use of their unpublished results.
REFERENCES
Abdurahiman, U.C. and Joseph, K.J. (1978a). Biology and Behaviour of Apocrypta bakeri Joseph
(Torymidae), cleptoparasite of Ceratosolen marchali Mayr (Agaonidae). Entomon 3: 31-36.
Abdurahiman U.C. and Joseph, K.J. (1978b). Cleptoparasitism of the fig wasps (Torymidae: Chalcidoidea)
in Ficus hispida L. Entomon 3: 181-186.
Askew, R.R. (1975). The organisation of chalcid-dominated parasitoid communities centred upon endophytic
hosts. In Evolutionary Strategies of Parasitic Insects and Mites (Ed. Price, P.W.), pp. 130-153.
Plenum, New York.
Askew, R.R. (1980). The diversity of insect communities in leaf- mines and plant galls. 1. Anim. Ecol. 49:
817- 829.
Askew, R.R. and Shaw, M~W.
(1986). Parasitoid communities: Their size, structure and development. In
Insect Parasitoids (Eds Waage, J: and Greathead, D.), pp. 225-264. Academic Press, London.
Baijnath, H. and Ramcharun, S. (1983). Aspects of pollination and floral development in Ficus capensis
Thunb. (Moraceae). Bothalia 14: 883-888.
Baijnath, H. and Ramcharun, S. (1988). Reproductive biology and chalcid symbiosis in Ficus burtt-davyi
Hutch. (Moraceae). In Modern Systematic Studies in African Botany (Eds. P. Goldblatt, P. and
Lowry, P.P.), Allen Press, Inc., Lawrence, Kansas, U.S.A.
Berg, C.C. (1990). Annotated check-list of the Ficus species of the African floristic region, with special
reference and a key to the taxa of southern Africa. Kirkia 13: 253-291.
Boucek, Z. (1988). Australasian Chalcidoidea (Hymenoptera). C.A.B. International, U.K.
Bronstein, J.L. (1988). Mutualism, antagonism and the fig - pollinator interaction. Ecology 69: 1298-1302.
175
Bronstein, J.L. (1991). The nonpollinating wasp fauna of Ficus penusa: exploitation of a mutualism? Oikos
61: 175-186.
Compton, S.G. (In press). An association between epichrysomallines and eurytomids (Hymenoptera,
Chalcidoidea) in southern African fig wasp communities. 1. Em. Soc. sth. Afr.
Compton, S.G. and Hawkins, B.A. (in press). Determinants of species richness in Southern african fig wasp
assemblages. Oecologia.
Compton, S.G. and Nefdt, R.J.C. (1988). Extra-long ovipositors in chalcid wasps; some examples and
observations. Antenna 12: 102-105.
Compton, S.G. and Nefdt, R.J.C. (1990). The figs and fig wasps of Ficus burtt-davyi. Mitt. Illst. AUg. Bot.
Hamburg 23a: 441-450.
Compton, S.G. and Robertson, H.G. (1988). Complex interactions between mutualisms: ants tending
homopterans protect fig seeds and pollinators. Ecology 69: 1302-1305.
Compton S.G. and Robertson, H.G. (1991). Effects of ant- homopteran systems on fig-figwasp interactions.
In Ant- Plant lmeracrions (Eds Huxley, C.R. and Cutler, D.F.), Oxford University Press, Oxford.
Compton, S.G. and van Noort, S. (In press). Southern african fig wasps (Hymenoptera: Chalcidoidea):
resource utilization and host relationships. Proc. K. Ned. Akad. Wet.
Galil, J. (1977). Fig biology. Endeavour 1: 52-56.
Gautier-Hion, A. and Michaloud, G. (1989). Figs: are they keystone resources for frugivorous vertebrates
throughout the tropics? A test in Gabon. Ecology 70: 1826-1833.
1
Hawkins, B.A. (1988). Species diversity in the third and fourth trophic levels: Patterns and mechanisms.
1. Anim. Ecol. 57: 137-162.
Hawkins, B.A. and Compton, S.G. (1992). African Fig wasp Communities: undersaturation and latitudinal
gradients in species richness. 1. Anim. Ecol. 61: 361-372.
Hawkins, B.A., Askew, R.R. and Shaw, M.R. (1990). Influences of host feeding-niche and foodplant type
on generalist and specialist parasitoids. Ecol. Em. 15: 275-280.
Janzen, D.H. (1969). Seed-eaters versus seed size, number, toxicity and dispersal. Evolution 23: 1-27.
Janzen, D.H. (1979). How to be a fig. AnJl. Rev. Ecol. Syst. 10: 13-51.
Jeffries, M .J. and Lawton, J.H. (1984). Enemy free space and the structure of ecological communities. BioI.
1. Linll. Soc. 23: 269-286.
176
Joseph, K.J. (1984). The reproductive strategies in fig wasps (Chalcidoidea: Hymenoptera)- A review. Proc.
Indian nat. Sci. A cad. B50: 449-460.
Kjellberg, F., Gouyon, P.-H., Ibrahim, M., Raymond, M. and Valdeyron, G. (1987). The stability of the
symbiosis between dioecious figs and their pollinators: a study of Ficus carica L. and Blastophaga
psenes L. Evolution 41: 693-704.
Kjellberg, F. and Valdeyron, G. (1984). Mini-symposium, figs and fig insects. Centre Louis Emberger,
Montpellier.
Lachaise, D. and McEvey, S.F. (1990). Independent evolution of the same set of characters in fig flies
(Lissocephaia, Drosophilidae). Evol. Ecol. 4: 358-364.
Michaloud, G., Michaloud-Pelletier, S., Wiebes, J.T. and Berg, C.C. (1985). The co-occurrence of two
pollinating species of fig wasp and one species of fig. Proc. K. Ned. Akad. Wet. C 88: 93-119.
Nair, P.B. and Abdurahiman, U. C. (1984). Population dynamics of the fig wasp Kradibia gestroi (Grandi)
(Hymenoptera Chalcidoidea, Agaonidae) from Ficus exasperata YaW. Proc. K. Ned. Akad. Wet.
C 87: 365-375.
Nefdt, R.J.C. (1989). Interactions between Fig Wasps and Their Host Figs. Unpublished M.Sc. thesis,
Rhodes University, Grahamstown, South Africa.
Price, P.W. (1991). Evolutionary theory of host and parasitoid interactions. BioI. Cant. 1: 83-93.
Tscharntke, T. (1992). Coexistance, tritrophic interactions and density dependence in a species-rich
parasitoid community. J. Anim. Ecol. 61: 59-68.
,
Ulenberg, S.A. (1985). The phylogeny of the genus Apocrypta Coquerel in relation to its hosts Ceratosolen
Mayr and Ficus L. Verhand. K. Ned. Akad. Wet. 83: 149-176.
van Greuning, J. V. (1990). A synopsis of the genus Ficus (Moraceae) in southern Africa. Sth. Afr. J. Bot.
5: 599-630.
van Noort, S. (1991). The Systematics and Phylogelletics of the Sycoecinae (Agaonidae, Chalcidoidea,
Hymenoptera). Unpublished Ph.D. thesis, Rhodes University, Grahamstown, South Africa.
van Noort, S., Ware, A.B. and Compton. S.G. (1989). Pollinator- specific volatile attractants released from
the figs of Ficus bunt-davyL Sth. Afr. J. Sci. 85: 323-324.
Verkerke, W. (1987). Syconial anatomy of Ficus ruperiijolia (Moraceae), a gynodioecious tropical fig. Proc.
K. Ned. Akad. Wet. C 91: 461-492.
177
Verkerke, W. (1989). Structure and function of the fig. Experientia 45: 612-622.
Ware, A.B. and Compton, S.G. (In press). A breakdown of pollinator specificity in an African fig tree.
Biotropica.
Ware, A.B., Kaye, P.T., Compton, S.G. and van Noort, S. (In press). Fig volatiles: their role in attracting
pollinators and maintaining pollinator specificity. Pi. Syst. Eval.
Wiebes, J.T. (1989). Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae): fig wasps and their
figs, IV (African Ceratasalen). Proc. K. Ned. Akad. Wet. C 92: 251-266.
Wiebes, J.T. and Compton, S.G. (1990). Agaonidae (Hymenoptera Chalcidoidea) and Ficus (Moraceae):
fig wasps and their figs, VI (Africa concluded). Proc. K. Ned. Akad. Wet. 93: 203-222.
Zerova, M.D. and Fulsov, V.N. (1991). The palaearctic species of Eurytoma (Hymenoptera, Eurytomidae)
developing in stone fruits (Rosaceae, Prunoideae). Bull. Ent. Res. 81: 209-219.
178
CHAPTER 9
SYNOPSIS
179
These studies encompassed many aspects which govern the interactions between figs and their fig wasps.
In this section the individual research projects making up the thesis are placed in perspective and avenues
for future research are explored.
Survey results of figs and their associated pollinators showed that each Ficus species was generally
associated with a single pollinating wasp species (Papers 1 and 2). Exceptions where more than one
species of pollinator was associated with a 'single' fig species are discussed in Paper 1. The. situation
was resolved for one of these discrepencies when F. sakalavarum was reclassified as a distinct species
with its own specific pollinator wasp species (Paper 2).
In order to maintain this specificity, fig wasps must be able to differentiate between their host Ficus and
congeneric species. Biological evidence for such recognition of Ficus by their pollinators is presented
in Papers 3 and 4. Elisabethiella baijnathi was attracted to receptive figs of its host tree, F. burtt-davyi,
even when visual contact was excluded by surrounding the figs with cotton bags. These experiments
confirmed the volatile nature of the attractants and showed that they were only present when the figs were
ready to be pollinated.
Other parts of the host plant, pollinated figs and figs of other species did not
attract E. baijnathi. The arrivals of fig wasps at the trees of two conspecific Ficus species over a two
year period confirmed that these wasp species were only attracted to their host trees when they were
bearing figs that were ready to be pollinated. The two species of pollinating fig wasps were only trapped
at bagged figs of their respective host trees confirming both the volatile nature of the attractants and their
specificity in attracting only their specific pollinator.
The chemical basis for such species-specific volatile attractants was examined in Paper 5 where
charcoal-trapped fig volatiles were analyzed by gas chromatography. Not only did the figs of each Ficus
species examined present a unique volatile profile, but additional components were recorded only at the
time when figs became attractive to their pollinators.
These additional compounds, alone or
combination with the other volatile components, probably form the basis of the fig wasp attraction.
180
III
Perception of the volatiles emanating from the figs which are ready to be pollinated will be influenced
by environmental conditions and the pollinators will have to adopt appropriate behaviours in order to find
the volatile source. In the first of the two papers examining fig wasp dispersal behaviour (Paper 6), fig
wasp departures from their natal tree and their arrivals at trees bearing re.ceptive figs was examined.
Ambient temperatures were found to influence the timing of fig wasp emergence from their natal figs.
The lowest temperature at which the pollinating fig wasps began to emerge from their natal figs was
found to be related to the critical take-off temperature of the wasps. E. baijnathi females arriving at a
new host fig avoided figs that already contained a conspecific foundress. In Paper 7 the dispersal of the
wasps was examined. Air movement influenced both the fig wasps departing from their natal trees and
those arriving at trees bearing receptive figs. On departure from their natal trees, the wasps flew upward
and were then carried with the wind. On arrival at a host tree bearing figs ready to be pollinated, the
wasps approach the tree from downwind and close to the ground.
Chemosensory receptors of insects are generally found on the antennae.
Paper 8 examines some
techniques for preparing fig wasp for examination under scanning electron microscopy. In Paper 9 the
occurrence of elongated multi porous plate sensilla was examined. Although elongation of the mutiporous
sensilla is common among male chalcids, among female chalcids it may uniquely occur among some
species of pollinating fig wasp.
Elongation results in increased sensilla surface area and may have
evolved in order to detect the minute quantities of volatiles emanating from figs ready to be pollinated.
Two cases where more than one pollinator species was recorded from a single Ficus specIes was
investigated. In Paper 10 we examined the biology of the 'cuckoo' of F. sycomorus, C. galili, which
exploited the mutualism between F. sycomorus and its pollinator C. arabicus by utilising the ovules
without pollinating the figs. In the second case examined, three pollinator wasp species were found to
simultaneously pollinate the figs of a single F. lutea (Paper 11). The small number of its normally
associated pollinator that were present are thought to have travelled long distances to find this host tree
and as a result a large proportion of the crop remained unpollinated. The two other pollinating fig wasp
species were shown to have been "incidental" arrivals and had not been attracted to the tree. Although
hybrid seeds from the two fig crosses did germinate, the seedlings did not grow beyond the cotyledon
181
stage of development and this may indicate post germination weakness / inviability.
Pollinator fig wasps represent only one species member of the fig wasp community associated with each
Ficus species. In Paper 12 the consequences of the structure of the fig and the trees' phenologies on the
biology of these non-pollinating fig wasps were examined. The influence of ants and homopterans on
the populations was discussed as was community species composition.
Future Research
4-hexene-l-01 acetate
i
i
-~
F. burtt-aavyi
4-hexene-l-01 2cetate
I,t
I
I
[~I
I
I
,J
,
o
\
I
~L-,J/l
I
F. burtt-davyi
I
. 1 Ii
d
10
20
RESPONSE TlME (minutes)
Figure 1. Gas chromatogram of 4-hexcne-l-ol acetate and its coelution with the volatiles from receptive F. bUrIl-davyi figs.
182
Both the biological and the chromatographic evidence showed that volatiles emanating from the figs when
they are ready to be pollinated are responsible for attracting the pollinators. The next logical phase of
research should be to isolate, identify and synthesise the compound(s) concerned. An attempt was made
to identify the additional volatile compound present in the chromatograph of receptive figs of F.
burtt-davyi.
High resolution GC-MS analysis (done at Oxford University) identified the volatile in
question as 4-hexene-l-ol acetate. The compound was synthesised, but unfortunately did not coelute
when rerun with the original sample (Figure 1). As the equipment available at Rhodes University was
not suitable for this type of analysis this avenue of research was abandoned.
Although aspects of fig wasp biology outside the figs were investigated, little is known about how far
the wasps can travel when in search of receptive figs. Indications are that they do not usually venture
far (Paper 7) although small numbers may travel long distances (Paper 11). A mark - release - recapture
program, perhaps using fluorescent dyes or powders, should be able to determine at what distance
volatiles are perceived.
The volatile(s) attracting the wasps to their hosts are thought to be detected by the multiporous plate
sensilla positioned on the antennae of the wasps. Once the volatile attractants have been synthesised,
electroantennogram studies would be able to confirm that the function of these sensilla is the perception
of these volatiles.
The apparent breakdowns of wasp host choice may be an indication that cryptic tree species andlor wasp
species are involved.
For example,
morphological variation within species may account for the
differences between Elisabethiella stuckenbergi and E. socotrensis, both of which are found in figs of
southern African F. natalensis subspecies natalensis. If the wasps prove to be distinct species then there
may be two cryptic Ficus species.
species status.
Analysis of the wasp mitochondrial DNA could determine their
Similarly, using chloroplast DNA, fig trees of the difficult "thonningii / natalensis"
complex could potentially be assigned to definite species.
183
Perhaps the central question of fig biology is: Have figs and fig wasps co-evolved? The generally
observed one Ficus species / one agaonine species relationship is certainly highly suggestive the two
groups have coevolved but conclusive proof is still needed. By determining the phylogenies of both figs
and their pollinating fig wasps independently, possibly using DNA restriction fragment polymorphism
techniques they could be compared.
If the phylogenies produced in this manner could be shown to
mirror one another, then figs and their associated fig wasps could be said to have co-evolved.
184