Plant Syst. Evol. 219:225-241 (1999)
Plant Systematics and Evolution © Springer-Verlag 1999 Printed in Austria
Reproductive conflicts of Palicourea padifolia (Rubiaceae) a distylous shrub of a tropical cloud forest in Mexico Paola Sabina Contreras and Juan Francisco Ornelas Departamento de Ecologfa y Comportamiento Animal, Instituto de Ecologfa, A. C., Xalapa, Veracruz, M6xico Received October 28, 1998 Accepted December 1, 1998
Abstract. We studied a population of the distylous Palicourea padifolia (Rubiaceae) in a cloud forest remnant near Xalapa City, Veracruz, M6xico to explore possible asymmetries between floral morphs in the attractiveness to pollinators, seed dispersers, nectar robbers, floral parasites, and herbivores, w e first assessed heterostyly and reciprocal herkogamy by measuring floral attributes such as corolla length (buds and open flowers), style and anther heights, stigma and stamen lengths and the distance between the anther tip to the stigma lobe. We then estimated floral and fruit attributes such as flower size, anther height, number and size of pollen grains, fruit size, seed size, nectar production, and flower and fruit standing crops to assess differences between floral morphs in attracting and effectively using mutualistic pollinators and seed dispersers. Also, floral parasitism and nectar robbing were assessed in this study as a measure of flower attractiveness to antagonists. The system seems to conform well to classical heterostyly (e.g. reciprocal stamen/style lengths, pollen and anther dimorphism, intramorph incompatibility) yet, there were several tantalizing differences observed between pin and thrum morphs. Thrum flowers have longer corollas and larger but fewer pollen grains than pin flowers. Both morphs produced the same total number of inflorescences, developed the same number of buds, and opened the same number of flowers per
inflorescence during the flowering season. Nectar production and sugar concentration were similar between floral morphs but the reward was not offered symmetrically to floral visitors throughout the day. Nectar concentration was higher in pin flowers in the afternoon. The numbers of developing, fully developed, and ripe fruits were the same between floral morphs, however, fruits and seeds were larger than those of thrums. The incidence of fly larvae was higher among thrum flowers and damage by nectar robbing was the same between floral morphs. Fruit abortion patterns of flowers manually pollinated suggest intra-morph sterility (self and intramorph incompatibility). There were no differences between morphs in fruit and seed set per flower following legitimate pollination although thrums were more leaky than the pins (intramorph compatibility).
Key words: Palicourea padifolia, herkogamy, hummingbirds, pollination, Rubiaceae, sexual selection. Heterostyly, a floral polymorphism characterized by the reciprocal positioning of stigmas and anthers between morphs, has arisen independently several times among angiosperms and occurs in approximately 27 angiosperm families (Barrett 1992, Ree 1997). Traits commonly associated to the heterostylous
226
E S. Contreras a~d J. E Ornelas: Heterostyly in Palicourea padifolia
plants include, (1) reciprocal herkogamy (herkogamy is the spatial separation of anthers and stigma within a flower), (2) self and intramorph incompatibility, and (3) an array of ancillary floral polymorphisms (Barrett 1990). The loci controlling self-incompatibility in heterostylous species does not occur in some heterostylous species and the presence of incompatibility is unknown for most heterostylous species (Ganders 1979a). The diversity of floral displays, pollen and nectar rewards, and expressions of sexuality in higher plants offer exciting evolutionary problems (Andersson 1994). The traditional explanation for most such characters is that these floral traits are favored by natural selection through increased fecundity or other advantages; however, it is likely that they are also favored by sexual selection (Andersson 1994). The form and strength of sexual selection, particularly male (pollen) competition over mates, apparently is one of the main variables that explain the diversity of sex habits among higher plants (Lloyd 1982). There is evidence that sexual as well as natural selection have contributed to the evolution of conspicuous flowers and several other traits in higher plants (Andersson 1994), and that sexual selection arguments can be invoked to explain the differences between the sexes such as the production of more flowers and inflorescences in male plants (Stephenson and Bertin 1983), the amount, size and shape of pollen grains (Campbell and Waser 1987), pollinator attraction (Waser and Price 1983, Feinsinger et al. 1991), nectar production (Lloyd and Yates 1982, Devlin and Stephenson 1985, Mitchell 1993), and flower and fruit phenologies (Lloyd and Yates 1982, Stephenson and Bertin 1983). Because it has been hypothesized that floral morphs may exhibit gender specialization and gain differential reproductive success through male and female function on the basis of sexual selection and sex allocation theory (Beach and Bawa 1980, Barrett 1992, Domfnguez et al. 1997), the objective of the present
study was to describe the reproductive biology of Palicourea padifolia (Roem. and Schult.) C.M. Taylor and Lorence to understanding the functional significance of heterostyly in this species. We also described flower and pollen dimorphism, phenological changes of flower and fruit availability, patterns of nectar production, and levels of nectar robbery and floral parasitism with the purpose of identifying possible sexual conflicts between floral morphs.
Materials and methods Study species. The monophyletic Palicourea Aublet (Rubiaceae) is represented by 200 species of shrubs and small trees commonly found in the understory and subcanopy of moist to wet tropical montane forests in the Neotropics (Taylor 1996, 1997). A close relationship between Palicourea and Heteropsychotria Steyerm. (Subg. Psychotria) has been suggested because they share the same fruit color (Taylor 1996). The two are distinguished by their corollas, well-developed tubes, swollen and usually gibbous at the base, and internally glabrous except for a ring of dense pubescence that end the basal swelling that contains the nectar (Taylor and Lorence 1985, Lorence 1990, Taylor 1996). Nearly all Palicourea species are distylous, which appears to be the ancestral condition for the genus, and then lost in some of the monomorphic species of the Caribbean Islands (Vuilleumier 1967; Taylor 1993, 1997). Palicourea padifolia is a distylous small tree of 2-7 m high or sometimes even larger. It can be recognized by its short-petiolate leaves, short stipules with the lobes short or elongate, and reddish-purple, pyramidal inflorescences with yellow flowers with long, tubular corollas (ca. 1015 ram), opposite-branched panicles with flowers occurring in cymose clusters of the terminal inflorescence (Taylor 1989, Ree 1997, Contreras 1998). Two-seeded fruits (ca. 7 x 8 mm) are green when immature, ripening to fleshy and purplishblack (Ree 1997, Contreras 1998). Territorial hummingbirds and bees have been observed visiting flowers of P. padifolia (Ree 1997). Study site. Fieldwork for this study was conducted on a single population of Palicourea
R S. Contreras and J. E Ornelas: Heterostyly in Palicourea padifolia
I
1 cm
I
1 cm
Fig. 1. Drawings of the thrum (top) and pin (bottom) flowers of Palicourea padifolia
padifolia located in the Parque Ecoldgico Francisco Xavier Clavijero near the city of Xalapa (19°30 ~N, 96°57 t W), Veracruz, M~xico. Elevation at the 29ha cloud forest remnant is 1225m a.s.1 with irregular topography and soils derived from volcanic ashes with good internal drainage and high quantities of organic matter (Zolfi 1987, Tolome 1993). Mean annual rainfall is 1492.4mm, with a minimum of 44.8mm in December and a maximum of 273.4 mm in June. Mean annual temperature is 18 °C, with a maximum temperature of 20.4°C in May and a minimum of 14.9°C in January. The original vegetation at Xalapa and its surroundings, described by Rzedowski (1978) as cloud forest, has been heavily fragmented and few isolated remnants of the original vegetation remain today along with an agrarian landscape composed mostly by coffee plantations (Tolome 1993, Williams Linera 1993). Palicourea padifolia Jis a common
227
species of the understory where canopy is dominated by arboreal species such as Carpinus caroliniana (Betulaceae), Cinnamomum effusum, Ocotea sp. (Lauraceae), Liquidambar macrophylla (Hamamelidaceae), Quercus germana, Q. xalapensis (Fagaceae), Turpinia insignis (Staphyleaceae), Clethra mexicana (Clethraceae), Eugenia xalapensis (Myrtaceae), Lonchocarpus sp. (Leguminosae), Meliosma alba (Meliosmaceae), IIex tolucana (Aquifoliaceae) and Oreopanax xalapensis (Araliaceae) (Williams Linera 1993). The species may produce flowers and fruits throughout the year, but flowering is typically known to be interrupted periodically by breaks lasting 1-4 months (Contreras 1998). During the study period (early April 1996 to late September 1997), the population at Clavijero was observed to be flowering primarily from March to August, with fruiting (ripe fruits) occurring until mid-October (Contreras 1998). Flower buds open at dawn and the flowers stayed open until the end of the following day, but pollen shedding occurred between 03:00h and 05:00h (Contreras 1998). Corolla and style abscising from the ovary typically occurred from late afternoon to end of day (Contreras 1998). Floral morphology and heterostyly. A total of 1420 flowers (677 pin and 743 thrum) was collected from different inflorescences of 15 plants of each morph in May and June of 1996 and measured while still turgid using a digital caliper (Flower Ultra-call II) precise to the nearest 0.01 ram. Corolla length, anther height (from the corolla base to the distal anther's tip), style height (from the base of the ovary of the stigma lobes' divergence), and the stigma-anther separation (distance between the anther's tip to the stigma's lobe), were estimated to determine heterostyly according to Sobrevila et al. (1983), Hamilton (1989), Richards and Barrett (1992), P~rez-Nasser et al. (1993), Richards and Koptur (1993), Riveros et al. (1995), and Pailler and Thompson (1997). Morphological differences between pin and thrum flowers were tested with a one-way ANOVA (Zar 1984) to each of the floral attributes above mentioned. Reciprocal herkogamy. We collected 1269 flower buds of different ages and buds about to open (pin : 606, thrum= 663) to determine reciprocal herkogamy. Data of the corolla length and stigma height were transformed to natural logarithm and graphed to obtain the developing
228
E S. Contreras and J. F. Oruelas: Heterostyly in Palicourea padifolia
trajectories of pin and thrum flowers (Richards and Koptur 1993). We also determined reciprocal herkogamy by measuring stigma length, stamen length, and the stamen-style difference in open flowers of our sample described above. Morphological differences between pin and thrum flowers were tested with a one-way ANOVA (Zar 1984) to each of the floral attributes above mentioned. The effectiveness of heterostyly as an outcrossing promoter must depend strongly on the degree of matching between pin and thrum flowers of heterostylous species. The degree of matching between floral morphs, defined here as a perfect morphological matching of the stigma of one type of flower with the height of the anthers of the other type of flower, was estimated by calculating the average height of the anther of a floral morph and then subtracting it from the average height of the stigma of the alternate morph. To scale for absolute size, the product was then divided by the highest average height. The standard deviation was then calculated and the positive and negative values were compared to the ranges who proposed for eleven Rubiaceae species (Niessen et al. 1990). Flower phenology and flower standing crop. We selected 30 P. padifolia plants (15 of each morph) in April of 1996, and marked 30 inflorescences of each one with permanent flagging tape. We then monthly censused the number of buds and open flowers per inflorescence every month during both years of the study. One-way ANOVA (Zar 1984) and post hoc means comparisons were conducted to test for differences between floral morphs in the number of buds and open flowers per inflorescence in 1996 and 1997. Pollen morphology. A total of 615 flowers was also collected in 1997 from 20 plants (10 of each morph) to determine the number and size of anthers and pollen grains. Anthers per flower were counted with the aid of a handlens and then measured with digital calipers precise to the nearest 0.01 ram. Pollen grains were counted under a microscope calibrated at 10x in four quadrats of 3-4 anthers using a hematocytometer and then measured in gm (Wolfe and Barrett 1989). Differences between floral morphs in the number and size of anthers and pollen grains were determined by using a one-way ANOVA (Zar 1984). Nectar production and floral visitors. Inflorescences of 20 plants (10 of each morph) were
bagged in 1997 with mosquito-netting 12 h before bud opening. Nectar production was then estimated the following day at 2h intervals following standard procedures (Kearns and Inouye 1993). Nectar volume per flower ( n = 340 pin, n =285 thrum) was estimated by using calibrated micropipets (10 gl) and a ruler, and nectar concentration (BRIX) with a hand-refractometer (American Optical 10431). Differences in nectar volume and nectar concentration between floral morphs were analyzed by using a two-way ANOVA (Zar 1984). Post hoc tests were then conducted to contrast mean differences among time intervals. Daily observations (8:00 to 12:00 am) were also conducted from April to July of 1996 and 1997 to determine the most common foral visitors. Nectar robbery and dipteran parasitism. Buds and flowers collected to determine reciprocal herkogamy were examined for two kinds of floral damage: (1) floral parasitism as those buds containing a larva, and (2) nectar robbery as those flowers with holes at the base of the corolla. The data were analyzed with a one-way ANOVA (Zar 1984) to determine if one of the floral morphs was more vulnerable to damage than the other. Fruit and seed morphology. A total of 507 ripe fruits ( n = 2 5 2 pin, n = 2 5 5 thrum) was collected in 1996 from several plants to estimate differences in morphology between floral morphs. Fruits were first measured (length a~nd width) using digital calipers precise to the nearest 0.01 mm and then weighted (wet weight) with an analytical balance precise to the nearest 0.01 g. Fruit pulp was removed and weighted (wet pulp weight) and then dried our for ca. 20min. at 250 °C and weighted again (dried pulp weight). Seeds were measured and weighted similarly and the number of seeds per fruit were also counted. Morphological differences between floral morphs were tested with a one-way ANOVA (Zar 1984) to each of the morphological attributes we measured. Fruit phenology and fruit standing crop. We marked 30 infructescences with permanent flagging tape from 30 P. padifolia plants (15 of each morph) in April of 1996. We then censused each infructescence every other week to estimate the number of fruits classified as developing fruits classified as developing fruits (immature, green fruits), unripe (fully developed, but still green) and ripe fruits (from purple to black), along the fruiting season of 1996, to determine any asynchrony in
R S. Contreras and J. E Ornelas: Heterostyly in Palicourea padifolia fruit development and availability. Differences in total fruit production and fruit standing crop between floral morphs were analyzed by using a two-way ANOVA (Zar 1984). Post hoc tests were then conducted to contrast mean differences in time. lntramorph incompatibility. To document self and intramorph incompatibility, we bagged a total of 135 buds with mosquito netting one day before they opened. Manual crosses included: (1) autogamy (pollen of the same flower), (2) geitonogamy (pollen of a flower on the same inflorescence), (3) intramorph xenogamy (pollen of a flower of a different plant but the same floral morph), (4) intermorph (pollen of a flower of the opposite floral morph), and (5) a control (non manipulated flowers; Bawa and Beach 1983). After manual pollination, flowers were tagged and followed for three months until fruit development. The results of this experiment were then analyzed with a ANOVA with repeated measures (Zar 1984).
Results Floral morphology and heterostyly. Palicourea padifolia is distylous. The style of the pin morph is exserted beyond the corolla tube and is longer than the style of the flowers of the thrum morph. Length of stigmas appear to differ in two morphs. Stigmas of pins are separated, those of thrums are oppressed to
229
each other (Fig. 1). Anthers are inserted middle in the tubular corolla in pin flowers, whereas in thrum flowers are positioned high in the corolla tube. In Table 1, we show mean values for each of the morphological attributes and Fig. 2 shows the resulting frequency distribution (bimodal) of each of the attributes. Thrum flowers had significantly longer corollas than pin flowers (ANOVA, d . f . = 1, F = 1020.6, P>.0001, n = 1 4 2 0 ; Fig. 2), and is not measured in most heterostylous species. Pin flowers had significantly longer styles than thrum flowers (ANOVA, d.f. = 1, F = 15087.3, P<.0001, n = 1420; Fig. 2). Thrum flowers had significantly longer stamens than pin flowers (ANOVA, d.f. = 1, F = 5458.2, P<.0001, n = 1 4 2 0 ; Fig. 2). Stigma-anther separation (distance) was significantly wider in thrum flowers than in pin flowers (ANOVA, d.f.=l, F=676.8, P<.0001, n=1420; Fig. 2). Reciprocal herkogamy. Deviation values from perfect matching between floral morphs (reciprocal herkogamy) ranged from - 0 . 3 7 0 for pin anther flowers vs. thrum stigma flowers and 0.016 for thrum anther flowers vs. pin stigma flowers. These values suggest a morphological mismatch (stigma-height polymorphism) between pins and thrums of P
Table 1, Floral measurements (mm) and results of one-way ANOVA's (see text) for dimorphism in Palicourea padifolia. A sample size of 677 pin flowers and 743 thrum flowers was used for this comparison Pin Floral attribute Heterostyly Corolla length Anther height Style height Stigma-anther separation Herkogamy Stigma length Stamen length Stamen-style difference
Thrum
Mean
(SD)
Mean
(SD)
P
14.114 11.48 15.91 4.77
(1.35) (1.08) (1.20) (0.97)
16.90 15.65 8.38 6.23
(1.05) (1.03) (1.10) (1.13)
< .0001 < .0001 < .0001 < .0001
16.62 12.11 4.91
(1.32) (0.99) (1.26)
8.23 15.04 5.78-
(1.32) (1.06) (1.35)
NS < .0001 < .0001
P.S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia
230 • Pin
[] Thrum
evolution of reciprocal herkogamy according to the model of Lloyd and Webb (1992) (see also Table 1).
A) Corolla length
400 "1
Floral phenology and flower standing crop. Palicourea padifolia flowered in 1996
300 ~ 200"
L
100 -
B) Anther height 400 300'
¢0
200' 100'
400 300' 200' 100'
1.5
2.0
2.5
3.0
D) Stigma-antherseparation 400 300" 200" 100"
0.5
1.0
1.5
2.0
Fig. 2. Differences between floral morphs in (A) corolla length (B) anther height, (C) style height, and (D), stigma-anther separation. Data in mm were In transformed (n = 677 pin flowers, n : 743 thrum flowers)
padifolia, according to the range (-0.333 to 0.333) proposed to Rubiaceae by Niessen et al. (1990). The developing trajectories of pin and thrum flowers shown in Fig. 3 (stigma height and bud length data transformed to natural logarithm) suggest the evolution of a stigmaheight polymorphism as a initial step in the
from March to August and in 1997 flowering was delayed until early May and lasted until August. In 1996, thrum plants exposed mostly to the sun started developing inflorescences in December (during fruiting), but only one pin plant with inflorescences was observed in the same date at Clavijero. Variation in the total number of buds produced per inflorescence and the number of flowers that open every day per inflorescence (flower standing crop) was statistically the same in both floral morphs in 1996 (ANOVA, d . f . = l , F = 1 . 9 7 1 , P>.05, n = 89) and 1997 (ANOVA, d.f. = 1, F = 0.157, P>.05, n = 1 2 0 , Fig. 4). However, total number of buds and flower standing crop decreased from one year to another (Fig. 4). In 1997, an inflorescence produced significantly fewer flowers (ANOVA, d.f. = 1, F = 5 7 . 0 3 , P < .0001, n = 240; Fig. 4) and opened fewer flowers per day (ANOVA, d.f. = 1, F = 2.603, P<.05, n = 2 4 0 ; Fig. 4). The interaction between morph and time variables was not statistically different for the total number of buds produced (ANOVA, d . f . = 1, F = 0 . 1 5 3 , P > .05, n = 240) neither in flowers opened per day (ANOVA, d . f . = l , F = 0 . 3 2 8 , P>.05, n = 2 4 0 ) . In 1996, the flowering peak was reached at June and then decreased by midJuly. In 1997, the highest flower production for both morphs occurred earlier, in May, and then decreased by mid-June. Pollen morphology. The number of anthers varied from 4 to 9 (meanei~ = 4.9 40.25, n : 1316, meanTHRU~ = 4.9 4- 0.26, n = 1352), but differences between morphs were not statistically significant (ANOVA, d.f. = 1, F = 0.420, P > .05). The anthers are longer in thrums than in pins (Fig. 5). Pollen grains in anthers of thrum flowers were significantly longer than those of pin flowers (ANOVA, d.f. : 1, F = 17.1, P<.0001; Fig. 6), however, anthers of pin flowers had more pollen grains
R S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia
231
3.0
Pin
2.8 •© ,"~
2.6
.:5/
hdBltg •
2.4
TM
•
•
0
2.2
"m r~
2.0 1,8 o 1.6
t
2.0
-
2.1
t
•
2.2
t
2.3
-
t
-
2.4
o ~
-
2.5
t
2.6
.
t
2.7
Thrum -
I
-
2.8
t
.... 3.0
2.9
Ln Bud Length Pin []
Fig. 3. Developmental trajectories of pin and thrum flowers showing a stigma-height polymorphism as a initial stage in the evolution of reciprocal herkogamy in P. padifolia
N Thrum
A) 70
a
"A
60 q..,
~
-l-
f/3
a
a
50 40 30 -L
20
4 rnm
3 mm
10
Fig. 5. Drawings showing the differences of the thrum (left) and pin (right) anthers of P. padifolia
0
B) ~
3
a
than longer anthers of thrum flowers (ANOVA, d.f. = 1, F = 5 . 5 3 , P < . 0 5 ; Fig. 6). ~ +
2
Nectar production and floral visitors.
0
Nectar is produced from early morning to late afternoon for one day. Floral morphs produced the same nectar volume (ANOVA, d.f. = 1, F = 2 . 2 8 , P > . 0 5 , n = 7 0 ) , and at different times of day (ANOVA, d . f . = 1, F = 0 . 6 0 9 , P > .05; Fig. 7). Sugar concentration in nectar of P. padifolia was also the same in both floral morphs (ANOVA, d.f. = 1, F = 0.07, P > .05, n : 70) and at different times of day (ANOVA, d . f . = l , F = 0 . 9 1 7 , P > . 0 5 , n = 7 0 ) . Post-hoc contrasts showed that thrum flowers have
b
Z
1996
J 1997
Fig. 4. Bud (A) and flower (B) production of P.
padifolia in 1996 and 1997. Data indicate means ± standard error (n = 89 in 1996, n = 120 in 1997). ** <0.001 *** <0.0001
R S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia
232 1.0
PIN - - o - -
THRUM
-
1.6. 0.8
©
""~
"~
"-"
1.4.
0
1.2.
0.6
1
.8
0.4
.6.
~D "-" O
A)
0.2 .4.
.2
i
i
i
i
i
i
i
B)
io
4 14.
"~ ©
I~l
3
~0 0
2
-~
12.
~,
lO, 9.
o
6i
1
4
Pin
Thrum
I
i
i
i
i
i
i
i
7:00
9:00
11:00
]3:00
15:00
17:00
19:00
Hour of Day
Fig. 6. Pollen dimorphism in (A) size in pm and (B) the number of grains (number of pollen grains per quadrant) in P. padifolia. Data indicate means ± standard error (n = 300 quadrants in pin flowers, n = 312 quadrants in thrum flowers
Fig. 7. Nectar volume and concentration in pin and thrum flowers of P. padifolia. Data indicate means±standard error (n=340 pin flowers, n = 285 thrum flowers). ** <0.001 *** < .0001
nectars with a higher concentration of sugars only at 1500h ( F = 4 . 2 5 , P<.0001, n = 7 0 ; Fig. 7). The main visitors to P. padifolia flowers at our study site were hummingbirds. Eleven hummingbird species were observed on P. padifolia flowers: the Wedge-tailed Sabrewing, Campylopterus curvipennis (Depee), the Violet Sabrewing, C. hemileucurus (Depee), the Green Violet-ear, Colibri thalassinus (Swainson), the Green-breasted Mango, Anthracothorax prevostii (Lesson), the White-bellied Emerald, Amazilia candida (Bourcier and Mulsant), the Azure-crowned Hummingbird, A. cyanocephala (Lesson), the Berylline Hummingbird, A. beryllina (Depee), the Buff-bellied Hummingbird, A. yucatanensis (Cabot), the Amethyst-throated Humming-
bird, Lampornis amethystinus (Swainson), the Magnificent Hummingbird, Eugenes fulgens (Swainson), and the Bumblebee Hummingbird, Atthis heloisa (Lesson and De Lattre). A. cyanocephala, A. beryllina, and A. heloisa were the most frequent visitors (>2 visits/ hour), whereas the visits of other hummingbird species were sporadic (< 1 visit/hour). A. cyanocephala and A. beryllina exhibited territorial behavior; between foraging bouts they were frequently observed defending patches of P. padifolia by vocalizations and aggressive attacks on other foragers that attempted to feed nearby. After visiting 2030 flowers of one floral morph in one foraging bout, individuals of A. cyanocephala were commonly observed to perch nearby, and then they would clean off their bills to remove
R S. Contreras and J. F. Ornelas: Heterostyly in Palicouma padifolia pollen loads by rubbing their bills against a twig or a branch. The same behavior was observed but less commonly to A. beryllina (Omelas et al. unpubl, data). Campylopterus spp., C. thalassinus, A. prevost& L amethystinus, and E. fulgens exhibited non-territorial behavior, acting as territory-parasites, and visiting patches of P. padifolia along traplines. A. heloisa seemed a marauder, in the sense of Feinsinger and Colwell (1978), but during its hovering visits, it is the only bird that touched the flower's reproductive organs with the upper bill and forehead; the pollen loads were deposited onto the other hummingbird species along their bills. Activity levels of hummingbirds are highest between 07:00h and 10:00h, and decline rapidly until 12:00h, following closely the amount and quality of nectar available. Less frequently, we observed diurnal nonidentified solitary bees and butterflies visiting flowers of P. padifotia. We also observed two passerine species feeding on buds and flowers of P. padifolia: the Common Bush-tanager, Chlorospingus ophthalmicus (Du Bus) and the Black-headed Saltator, Saltator atriceps (Lesson). These birds arrive commonly in groups of four shortly after hummingbirds and foraged differently: the Common Bush-tanagers pull off the buds and/or flowers, squeeze them presumably for nectar (nectar robbers) and then dropped them to the ground, whereas the Black-headed Saltators pull them off and then gobbled them entirely (herbivores).
Nectar robbery and dipteran parasitism. There were no significant differences between floral morphs in damage by nectar robbers (ANOVA, d.f. = 1, F = 0.074, P > .05), but flowers were more perforated than buds (ANOVA, d . f . = l , F=11.06, P<.0001; Fig. 8). Flower buds of both floral morphs had the same larvae incidence (ANOVA, d.f.= 1, F=0.075, P>.05), but the incidence of a non-identified fly larvae was significantly higher among thrum flowers than pin flowers (ANOVA, d.f. = 1, F=5.28, P<.05; Fig. 8). Fruit and seed morphology. Fruits and seeds were significantly different in size
233 Buds
[]
Flowers []
0.03 exl) ° v,,,~
0.02
~+ ~
0.01
2: 0.2
+
0.1
~D
©
0.0
Pin
Thrum
Fig. 8. Nectar robbery (A) and floral parasitism (B) in flowers and buds of P. padifolia. Data indicate means±standard error (n= 1269 buds, n = 1369 flowers) (length and width) between floral types but no differences were found in weight (Table 2). Pin flowers produced larger fruits with wider seeds than those of thrum flowers in 1996. Fruit phenology. The total number of fruits per infructescence between morphs was not statistically different (ANOVA, d.f. = 1, F=0.78, P>.05; Fig. 9), but there was a significant time effect (ANOVA, d.f.= 1, F = 17.24, P<.000t). Post-hoc contrasts showed a higher production of fruits in September in thrum plants (F=3.93, P<.0001; Fig. 9), and a higher production of fruits in October and December in pin plants (F=3.89 and 4.31, P<.0001, respectively; Fig. 9). The number of developing fruits per infructescence between morphs was not statistically different (ANOVA, d.f. = 1, F = 0.24,
E S. Contreras and J. E Ornelas: Heterostyly in Palicourea padifolia
234
Table 2. Fruit and seed measurements and results of one-way ANOVA's (see text) for dimorphism in Palicourea padifoIia. A sample size of 252 fruits of pins and 255 fruits of thrums was used for this comparison Pin
Thrum
Fruit attribute
Mean
(SD)
Mean
(SD)
P
Fruit length (mm) Fruit width (ram) Fruit weight (g) Pulp weight (wet) Pulp weight (dry) Seed number Seed 1 Length (mm) Width (mm) Weight (g) Seed 2 Length (ram) Width (mm) Weight (g)
8.95 7.40 0.42 0.32 0.04 1.97
(1.13) (1.33) (0.37) (0.12) (0.01) (0.20)
8.74 6.96 0.41 0.30 0.04 1.94
(1.29) (1.12) (0.26) (0.14) (0.02) (0.26)
< .05 < .0001 NS NS NS NS
3.96 2.88 0.04
(0.63) (0.79) (0.02)
4.04 2.74 0.04
(0.73) (0.62) (0.02)
NS < .05 NS
3.90 2.73 0.06
(0.73) (0.81) (0.25)
4.18 2.72 0.06
(2.81) (0.77) (0.31)
NS NS NS
P>.05; Fig. 9), but there was a significant time effect (ANOVA, d.f. = 1, F = 9.98, P < .0001). Post-hoc contrasts showed a higher number of developing fruits in July in thrum plants (F=3.03, P<.0001; Fig. 9). The number of fully developed fruits per infructescence between morphs were not statistically different (ANOVA, d . f . = l , F = 0 . 2 5 , P>.05; Fig. 9), but there was a significant time effect (ANOVA, d . f . = 1, F = 3 3 . 5 4 , P<.0001). Post-hoc contrasts showed a higher number of fully developed fruits in thrum plants in October (F = 7.39, P < .0001) and in September in pin plants ( F = 7 . 1 6 , P < .0001; Fig. 9). Lastly, the number of ripe fruits per infructescence between morphs was not statistically different (ANOVA, d.f. = 1, F = 0 . 2 0 , P > .05; Fig. 9), but there was a significant time effect (ANOVA, d.f. = 1, F = 16.54, P < .0001). Post-hoc contrasts showed no significant differences between morphs over time. I n t r a m o r p h incompatibility. Experiments of manual pollination showed that both floral morphs respond differently to different
types of pollination. Out of the total of flowers pollinated manually (71 pin flowers, 64 thrum flowers), more than 60% of the flowers pollinated with pollen of a flower from the opposite floral morph (intermorph flowers) initiated fruit development by August and less than 17% of the flowers from autogamous, geitonogamous, and intramorph crosses, and the control initiated and developed fruits. The same pattern was maintained in September (54% flowers of intermorph crosses, 13% flowers of autogamous, geitonogamous, and intramorph crosses, and the control flowers), but by October the pattern changed dramatically (18% and 0.9%, respectively; Table 3). Time after pollination had a very significant effect in increasing fruit abortion over time (ANOVA, d.f.----3, F = 342.8, P < .0001), but a more interesting result was the significant pollination type by time interaction, that is, the patterns of fruit abortion over time differed depending on the pollination type (ANOVA, d . f . = 12, F = 6 . 5 , P<.005), suggesting self and intramorph incompatibility.
P. S. Contreras and L F. Ornelas: Heterostyly in Palicourea padifolia Pin - -o- -
Thrum
:
235
interesting differences observed between pin and thrum morphs. We found that corollas of A) # total fruits 30 P. padifolia have an average length of 14 to 16 mm and the anther number was statistically //,,I -'v~. ~ . . . . 20 the same in both floral morphs, a characteristic also observed in heterostylous Rubiaceae 10 (Richards and Koptur 1993), however, thrum flowers were significantly longer than pin flowers. Longer corolla tubes on thrum flowers 10 B) # initiated fruits have been observed in several families with distylous species (Baker 1956; Levin 1968; 8 *** Ornduff 1976, 1980; Ganders 1979; Sobrevila + % et al. 1983; Feinsinger and Busby 1987; 4 ~'~, Hamilton 1989; Richards and Koptur 1993; 2 Riveros et al. 1995; Pailler and Thompson 1997; Ree 1997), however, it is not a general] ¢3 ity to all distylous species, since there are 15 t *** C) # developed fruits species in which the corolla length of pins is greater that of thrums (Hamilton 1989, Riveros et al. 1995, Stone 1995, Pailler and /,, Thompson 1997) or the same in both morphs (Ganders 1979b). It has been proposed that the corolla elongation is somehow correlated to stamen elongation in heterostylous plants 6 D) # ripe fruits / ~ _ _ ~ _ . . ~ . 5J (developmental constraint) (Dulberger 1992) because the direct attachment of anthers to the 4 -t [ corolla tube may allow for selection on corolla ~" it % /i III ~% length to put the anthers in the right position in a heterostylous species (Pailler and Thompson 1997). Therefore, the sexual dimorphism in • • i • i • i , J , J . i flower size observed in P. padifolia may be 12Jul llSept 10Oct 30Oct 7Nev 13Nov 2lDec associated with the evolution of distyly and Date not involved in the evolution of male fitness in dioecious plants. This hypothesis is supported Fig. 9. Fruit standing crop measured per infructesby the higher significant positive correlation cence as (A) total number of fi'uits, (B) number of between corolla length and stigma-anther initiated fruits, (C) number of developed fruits, and (D) number of ripe fruits of P. padifolia in 1997. separation in thrums (Spearman Rank CorrelaData indicate means ! standard error (n = 179 pin, tion, r = 0 . 4 1 , P<.0001), and a lower but 177 thrum). ** <0.001 **** <0.0001 positively significant correlation observed in pins (Spearman Rank Correlation, r = 0 . 1 9 , P < .0001). This positive correlation may have Discussion been of fundamental importance in the evoluin P. padifolia. Traits commonly Distyly tion of thrums with longer corollas than the associated to the heterostylous plants were pins as pointed out by Pailler and Thompson observed in this system (reciprocal herko(1997). gamy, pollen and anther dimorphism, intraDifferences in corolla length between morph incompatibility) yet, there were some floral morphs need to be investigated as a
/•.
,o]
R S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia
236
Table 3. Intramorph incompatibility results from hand self- and cross-pollination of Palicourea padifolia. N = number of individuals, n = number of flowers (June). Note that many fruits of both morphs never make it (see text) Number of fruits developing (%) Cross
N
June
August
September
October
5 5 5 5 5 25
16 (100) 17 (100) 14 (100) 14 (100) 10 (100) 71 (100)
5 2 3 11 0 21
(29.5)
5 (31.2) 1 (5.8) 3 (21.4) 9 (64.2) 0 18 (25.3)
0 0 0 3 (21.4) 0 3 (4.2)
5 5 5 5 5 25
15 (100) 13 (100) 13 (100) 14 (100) 9 (100) 64 (100)
4 (26.7) 1 (7.7) 2 (15.3) 7 (50) 1 (11.1) 15 (23.4)
2 (13.3) 0 1 (7.6) 6 (42.8) 1 (11.1) 10 (15.6)
0 0 0 2 (14.2) 1 (11.1) 3 (4.6)
Pins
Autogamy Geitonogamy Intramorph Intermorph Control Total
(31.2) (11.8) (21.4) (78.6)
Thrums
Autogamy Geitonogamy Intramorph Intermorph Control Total
possible reproductive conflict because the attractiveness of a longer corolla to mutualists and antagonists can be asymmetrical. Webb and Lloyd (1986) suggested that the avoidance of pollen-stigma interference is a general problem of flower architecture among outcrossing flowering plants. Self-interference may occur when stamens restrict access to stigmas or remove incoming pollen, or when self pollen causes stigma clogging that reduces male or female fitness (Lloyd and Webb 1986, Webb and Lloyd 1986). Self-interference is avoided with the spatial separation of pollen presentation and its receipt, i.e., herkogamy, and is presented in different classes as mechanisms to maintain an effective pollination (Lloyd and Webb 1986). Heterostyly may be seen as the only way to solve this selfavoiding conflict by more precise pollination that minimizes self-fertilization and self-interference because its features encourage cross fertilization (Lloyd and Yates 1982, Lloyd and Webb 1986, Webb and Lloyd 1986). In this sense, Lloyd and Webb (1992) assumed that the ancestors of heterostylous populations had
approached herkogamy and that the first step in the evolution of heterostyly is the evolution of a stigma-height polymorphism rather than an anther-height polymorphism as observed in P. padifolia. A factor that contributes to the reciprocity of anther position in most distylous species is the longer corolla tube of thrums, which because the anthers are attached directly onto the corolla of this species, causes the anthers to be placed at a level equivalent to the stigma in pins (Pailler and Thompson 1997). However, our data suggest that the reciprocity of anther position is at best apparent in P. padifolia because of the significant difference between morphs in the stigma-anther separation. The consequences of this floral asymmetry into how pollen is removed and transferred by short- and longbilled hummingbirds awaits to be investigated. Anderson (1973) proposed that the evolution of heterostyly can be facilitated by the apparently widespread occurrence of selfcompatibility and the elaborate development of protandry (i.e. elongation and maturation of the style are delayed with respect to
R S. Contreras and J. E Ornelas: Heterostyly in Palicourea padifolia dehiscence of the anthers) in extant Rubiaceae suggesting that the whole family may have descended from self-compatible ancestors. Recently, Lloyd and Webb (1986, 1992) suggested that approach herkogamy (i.e. the pollinator contacts the stigma first) is the ancestor mechanism of heterostyly, and that physiological incompatibility arose after the establishment of reciprocal herkogamy. According to BalTett's model (1992), a bimodal distribution in anther and stigma height, besides the difference in the corolla length should be observed in heterostylous plants. These characteristics were observed in the studied population indicating that P. padifolia is morphologically distylous. Also, our study suggests reciprocal herkogamy in P padifolia in which the anthers are placed near the throat of the corolla and the stigma deep within the tube, so pollen in positioned in such a way that is expected to be contacted before the stigma (thrum flowers). Several authors have proposed models to explain the evolution of heterostyly. Charlesworth and Charlesworth (1979) suggested that the floral condition antecedent to distyly was monomorphism with stigma and anthers at the same level within a flower and that diallelic self-incompatibility arose first; Baker (1966) suggested that distyly evolved through selective pressures to avoid selffertilization and inbreeding depression. Ganders (1979a) criticized the Charlesworth's model and he argued that most self-compatible, monomorphic species show some degree of stigma-anther separation so, he suggested that heterostyly is characterized as a mutual heteromorphism in anther and stigma lengths, so that the style of one morph approximately corresponds in position to the anthers of the alternate morph. The expected separation between anthers and stigmas of both morphs according to Charlesworth's model was observed to be the same in P. padifolia (reciprocal herkogamy). Reproductive conflicts in P. padifolia. Much of the morphological foundation for the
237
development of functional dioecy seems already to be present particularly in distylous plants. A single-step origin of dioecy is conceivable by the simultaneous loss of ovule function of the short-styled (thrum) form and of pollen production of the long-styled (pin) form, suggesting a sexual conflict in which one of the morphs, or both, are loosing one function. If this is true, morphological changes among distylous plants such as the typical corolla elongation in one of the morphs among and/or physiological changes such as an increment in the nectar production be one of the floral morphs, will modify pollinator preferences in such a way that the most attractive floral morph will be more visited and pollen transferred asymmetrically. This disassortative mating consequently will affect seed production and plant fecundity. However, the existence of such floral asymmetries and reproductive conflicts among heterostylous plants (Ganders 1979a, Kohn and Barrett 1992), and its impact in the genetic evolution of heterostyly awaits been investigated. The attractiveness of longer thrum flowers can be evaluated by looking at possible asymmetries in flower size, pollen transfer, and nectar production. Also, the attractiveness can be evaluated by looking at the floral parasites and nectar robbers to determine whether the imbalance observed in P. padifolia represents reproductive conflicts (Wiess 1996). P. padifolia thrums have longer corollas, larger pollen grains, and fewer pollen grains, a condition that is common in distylous plants (Ganders 1979a, Ree 1997, this study). This has lead to the suggestion that thrum corollas are larger to increase their attractiveness to pollinators (Ganders 1979a), and one would expect stigmatic pollen loads to differ between morphs. Ree (1997) observed pollen flow symmetry in a population of P. padifolia in Costa Rica and explained it as to be the consequence of longer stigmatic surface in thrums to which pollen can adhere (differential access between morphs). Furthermore, the imbalance imposed by the corollas of longer
238
R S. Contreras and J. F. Ornelas: Heterostyly in
thrums can be compensated by pins encouraging pollinators to visit more frequently their flowers, providing a better nectar reward as shown by our data. Preliminary observations on focal plants (33.6 h) suggest that hummingbirds visit more frequently pins than thrums (Pin = 4.3 4- 3.2 visits/h, n = 11 plants; Thrum = 2.7 4- 3.4 visits/h, n = 7 plants) and probe more flowers per visit of pins than thrums (Pin = 19.8 4- 16.2 flowers per visit, n = 38 visits; T h r u m = 18.7 4- 16.0 flowers per visit, n = 16 visits), but these differences were not statistically different (Mann-Whitney U test, P > .05). This data suggest that hummingbirds are visiting more frequently pin flowers because they offer a richer resource, however, the difference in the nectar production/concentration between morphs observed in P. padifolia can also be explained as a differential in nectar evaporation (nectar evaporation would be higher among thrum flowers because their corolla tubes are larger). The pollen flow symmetry observed by Ree can be selectively modified by developing and aborting fruits at different rates. We observed that many fruits of both floral morphs never make it in the manual pollination experiment (Table 3), and we cannot distinguish such loss of fruits as being selective abortion and/or inviability of the seeds (perhaps from inbreeding). We plan to examine exactly where the incompatibility reaction of self and illegitimate pollination occurs in this system. The larvae incidence was higher in thrum flowers probably because of their higher attractiveness to floral parasites. This has been reported to our knowledge only for dioecious plants in which male flowers have larger corollas and are more vulnerable to galls (Wolfe 1997). Longer thrum flowers would offer more quantity of resources than shorter pins (e.g. larger pollen grains), so that the fly larvae will have a higher probability to mature in thrum flowers (Weiss 1996). Further studies should address this asymmetry more carefully by knowing the natural history of this dipteran
Palicourea padifolia
and then presenting various degrees of attractiveness of buds and flowers to be chosen by the female fly to determine whether larval infestation has important fitness consequences in P. padifolia. Thrum P. padifolia plants start producing flowers earlier in the season than pin plants (thrum plants produced inflorescences in December meanwhile pin plants start production the next month (see also Dulberger 1975). This pattern is also observed in some dioecious plants in which males begin flowering earlier in the season than females (Lloyd and Webb 1977). Another evidence of reproductive conflicts between floral morphs in P. padifolia is the difference in flower production between sexes in distylous plants increase when the flowering intensity raises (Lloyd and Webb 1977). The number of buds initiated in pins was much higher compared with the number of flowers finally produced, and this phenomenon was inverse in thrums in which bud initiation was relatively low although the final flower number exceeds that produced by pin morph. The flowering peak in both years was in June, however, the number of buds and flowers produced per inflorescence in the first year was higher compared with the second year. A possible cause of this variation is age difference among chosen plants we selected for this part of the study. However, the differences in the quantities of flowers produced from one year to another are probably due to plant differences in the reproductive effort that depend on several factors as the probabilities of a plant or shoot will flower, the number of flowers per inflorescence, the number of inflorescences in flowering individuals or shoots, and the effort per flower (Lloyd and Webb 1977). In 1998, P. padifolia produced a higher number of inflorescences in flowering pin individuals than in thrum individuals, and this floral asymmetry was maintained over flowering time (Ornelas et al. unpubl, data), meaning that pin plants produced ca. 60 more flowers than thrum plants. However, this floral asymmetry is reversed and
R S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia more pronounced in the total number of fully developed fruits since thrum plants developed a significantly higher number of fruits by the end on the fruiting season. The higher reproductive success of thrum plants (more fruits) can be correlated to the morph differences in pollen size, but also it can be explained as the result of pollinator competition where longer corollas are more attractive to pollen dispersers. Also, these fluctuations can be a consequence of phenomena such as E1 Nifio in 1998, however, we believed that the floral asymmetries we found represent truly reproductive conflicts that have to be addresses in a spatial and temporal scales. The patterns of pollen deposition, the density of mites, differential evaporation of nectar between morphs, selective abortion of illegitimate offspring, viability of seeds and rates of germination between morphs, droughts and larval parasitism, and the behavior and effectiveness of pollinating hummingbirds varying in bill morphology are currently investigated in an ongoing study to determine whether the fecundity of pins exceeds that of thrums. Comments by Jose Gpe. Garcfa Franco, Victoria Sosa, Billie L. Turner and two anonymous reviewers greatly improved an earlier version of this paper. We thank to the Departamento de Ecologfa y Comportamiento Animal of the Instituto de Ecologfa, AC for the support and the facilities for conducting this work. Thanks to Antonio Andrade-Torres, Jos~ Bonilla and Lorena Ldpez de Buen for helping during field work; to Carlos Lara, Carlos M. Contreras, Jos6 G. GarcfaFranco and Victoria Sosa for sharing valuable information; to Edmundo Saavedra who prepared for us Figs. 1 and 3.
References Anderson W. R. (1973) A morphological hypothesis for the origin of heterostyly in the Rubiaceae. Taxon 22: 537-542. Andersson M. (1994) Sexual selection in plants. In: Andersson M. (ed.) Sexual Selection. Princeton University Press, Princeton, NJ, pp. 396-432.
239
Baker H. G. (1956) Pollen dimorphism in the Rubiaceae. Evolution 10: 23-31. Baker H. G. (1966) The evolution, functioning and breakdown of heteromorphic incompatibility systems I. The Plumbaginaceae. Evolution 20: 349-368. Barrett S. C. H. (1990) The evolution and adaptive significance of heterostyly. TREE 5: 144-148. Barrett S. C. H. (1992) Heterostylous genetic polymorphisms: model system for evolutionary analysis, ln: Barrett S. C. H. (ed.) Evolution and function of heterostyly. Springer, Wien New York, pp. 1-24. Bawa K. S., Beach H. (1983) Self-incompatibility systems in the Rubiaceae of a tropical lowland wet forest. Amer. J. Bot. 70: 1281-1288. Beach J. H., Bawa K. S. (1980) Role of pollination in the evolution of dioecy from distyly. Evolution 34: 1138-1142. Campbell D. R., Waser N. M. (1987) The evolution of plant mating systems: multilocus simulations of pollen dispersal. Amer. Nat. 129: 593-609. Charlesworth B., Charlesworth D. (1979) The maintenance and breakdown of distyly. Amer. Nat. 114: 499-513. Contreras R S. (1998) Biologfa reproductiva de Palicourea padifolia Roemer and Schultes (Rubiaceae): un arbusto distflico de un bosque mesdfilo de monafia. Tesis de Licenciatura (Biologfa). Universidad Veracruzana, Xalapa, Veracruz. Devlin B., Stephenson A. G. (1985) Sex differential floral longevity, nectar secretion, and pollinator foraging in a protrandrous species. Amer. J. Bot. 72: 303-310. Domfnguez C., Avila-Sakar G., V~zquez-Santana S., Mfirquez-GuzmLn J. (1997) Morph-biased male sterility in the tropical distylous shrub Erythroxylum havanense (Erythroxylaceae). Amer. J. Bot. 84: 626-632. Dulberger R. (1975) S-Gene action and the significance of characters in heterostylous syndrome. Heredity 35: 407-415. Dulberger R. (1992) Floral polymorphisms and their functional significance in the heterostylous syndrome. In: Barrett S. C. H. (ed.) Evolution and function of heterostyly. Springer, Wien New York, pp. 41-77. Feinsinger R, Busby W. H. (1987) Pollen carryover: experimental comparisons between morphs of Palicourea lasiorrachis (Rubiaceae), in a
240
P.S. Contreras and J. F. Ornelas: Heterostyly in Palicourea padifolia
distylous, bird-pollinated, tropical treelet. Oecologia 73: 231-235. Feinsinger P., Colwell R. K. (1978) Community organization among neotropical nectar-feeding birds. Amer. Zool. 18: 779-795. Feinsinger P., Tiebout H. M., Young B. E. (1991) Do tropical bird-pollinated plants exhibit density-dependent interactions? Field experiments. Ecology 72: 1953-1963. Ganders E R. (1979a) The biology of heterostyly. New Zealand J. Bot. 17: 607-635. Ganders F. R. (1979b) Heterostyly in Lithospermum cobrense (Boraginaceae). Amer. J. Bot. 66: 746-748. Hamilton C. W. (1989) Variations on a distylous theme in Mesoamerican Psychotria subgenus Psychotria (Rubiaceae). Mere. N. Y. Bot. Gard. 55: 62-75. Kearns C. N., Inouye D. W. (1993) Techniques for pollination biology. University Press of Colorado, Niwot, CO. Kohn J. R., Barrett S. C. H. (1992) Experimental studies on the functional significance of heterostyly. Evolution 46: 43-55. Levin D. A. (1968) The breeding system of Lithospermum caroliniense: adaptation and counteradaptation. Amer. J. Bot. 102: 427-441. Lloyd D. G. (1982) Selection of combined versus separate sexes in seed plants. Amer. Nat. 120: 571-585. Lloyd D. G. (1986) The avoidance of interference between the presentation of pollen and stigmas in Angiosperms. I. Dichogamy. New Zealand J. Bot. 24: 135-162. Lloyd D. G., Webb C. J. (1977) Secondary sex characters in plants. Bot. Rev. 43: 177-216. Lloyd D. G. (1992) The evolution of heterostyly. In: Barrett S. C. H. (ed.) Evolution and function of heterostyly. Springer, Wien New York, pp. 151-175. Lloyd D. G., Yates J. M. (1982) Intrasexual selection and the segregation of pollen and stigmas in hermaphrodite plants, exemplified by Wahlenbergia albomarginata (Campanulaceae). Evolution 36: 903-913. Lorence D. (1990) A phylogenetic list of the genera of Rubiaceae in Mexico. Acta Bot. Mex. 12: 1-7. Mitchell R. J. (1993) Adaptive significance of lpomopsis aggregata nectar production: obser-
vation and experiment in the field. Evolution 47: 25-35. Niessen K., Wells S., McDade L., Bakarr M., Boyle B., Eisen M. (1990) Heterostyly in eleven species of Rubiaceae. In: Loiselle B. (ed.) Tropical Biology 90-3. O.T.S. Organization for Tropical Studies (OTS), Costa Rica, pp. 26-29. Ornduff R. (1976) The reproductive system of Amsinckia grandiflora, a distylous species. Syst. Bot. 1: 57-66. Ornduff R. (1980) Heterostyly, population composition and pollen flow in Hedyotis caerulea. Amer. J. Bot. 67: 95-103. Pailler T., Thompson J. D. (1997) Distyly and variation in heteromorphic incompatibility in Gaertnera vaginata (Rubiaceae) endemic to La Reunion Island. Amer. J. Bot. 84: 315-327. P6rez-Nasser N., Eguiarte L. E., Pifiero D. (1993) Mating systems in the Rubiaceae of a tropical lowland wet forest. Amer. J. Bot. 70: 12811288. Ree R. H. (1997) Pollen flow, fecundity, and the adaptive significance of heterostyly in Palicourea padifolia (Rubiaceae). Biotropica 29: 298-308. Richards J. H., Barrett S. C. H. (1992) The development of heterostyly. In: Barrett S. C. H. (ed.) Evolution and function of heterostyly. Springer, Wien New York, pp. 85-124. Richards J. H., Koptur S. (1993) Floral variation and distyly in Guettarda scabra (Rubiaceae). Amer. J. Bot. 80: 31-40. Riveros G. M., Barria O. R., Humafia E A. M. (1995) Self compatibility in distylous Hedyotis salzmannii (Rubiaceae). Plant Syst. Evol. 194: 1-8. Rzedowski J. (1978) La vegetaci6n de M6xico. Editorial Limusa, M~xico, D. F. Ramfrez N., Xena De Enrech N., Sobrevila C. (1983) Reproductive biology of Palicourea fendleri and P. petiolaria (Rubiaceae) heterostylous shrubs of a tropical cloud forest in Venezuela. Biotropica 15: 161-169. Stephenson A. G., Bertin R. I. (1983) Male competition, female choice and sexual selection in plants. In: Real L. (ed.) Pollination biology. Academic Press, Orlando, FL, pp. 110-149. Stone J. L. (1995) Pollen donation patterns in a tropical distylous shrub (Psychotria surrensis: Rubiaceae). Amer. J. Bot. 82: 1390-1398.
E S. Contreras and Jo F. Ornelas: Heterostyly in Palicourea padifolia Taylor C. M. (1989) Revision of Palicourea (Rubiaceae) in Mexico and Central America. Syst. Bot. Monogr. 26: 1-102. Taylor C. M. (1993) Revision of Palicourea (Rubiaceae: Psychotriae) in the West Indies. Moscosoa. 7: 201-241. Taylor C. M. (1996) Overview of the Psychotriae (Rubiaceae) in the Neotropics. Opera. Bot. Belg. 7: 261-270. Taylor C. M. (1997) Conspectus of the genus Palicourea (Rubiaceae: Psychotriae) with the description of the some new species from Ecuador and Colombia. Ann. Missouri Bot. Gard. 84: 224-262. Taylor C. M., Lorence D. H. (1985) Lectotypification of Palicourea geleottiana M. Martens (Rubiaceae) and a new name for this common species. Taxon 34: 667-669. Tolome J. (1993) Cafda de hojarasca y comportamiento fenoldgico de las especies arbdreas del bosque mesdfilo de montafia del Parque Ecoldgico Francisco X. Clavijero. Tesis de Licenciatura (Biologfa). Universidad Veracruzana, Xalapa, Veracruz, M6xico. Vuilleumier B. S. (1967) The origin and evolutionary development of heterostyly in the Angiosperms. Evolution 21: 210-226. Waser N. M., Price M. V. (1983) Pollinator behaviour and natural selection for flower colour in Delphinium nelsonii. Nature 302: 422--424. Webb C. J., Lloyd D. G. (1986) The avoidance of interference between the presentation of pollen
241
and stigmas in angiosperms. II Herkogamy. New Zealand J. Bot. 24: 163-178. Weiss M. R. (1996) Pollen-feeding fly alters floral phenotypic gender in Centropogon solanifolius (Campanulaceae). Biotropica 28: 770-773. Williams-Linera G. (1993) Vegetaci6n de bordes de un bosque nublado en el Parque Ecol6gico Clavijero, Xalapa, Veracruz, M~xico. Rev. Biol. Trop. 41: 107-117. Wolfe L. M. (1997) Differential flower herbivory and gall formation on males and females of Neea psychotroides, a dioecious tree. Biotropica. 29: 169-174. Wolfe L. M., Barrett S. C. H. (1989) Patterns of pollen removal and deposition in tristylous Pontederia cordata L. (Pontederiaceae). Biol. J. Linnean Soc. 36: 317-329. Zar J. H. (1984) Biostatistical analysis, 2nd edn. Prentice Hall, New Jersey. Zol~ M. G. (1984) Estudio de la vegetacidn de Ios alrededores de Xalapa, Veracruz. Tesis de Licenciatura (Biologia). Universidad Veracruzana, Xalapa, Veracruz, M6xico.
Addresses of the authors: Paola Sabina Contreras and Dr. Juan Francisco Ornelas, Departamento de Ecologfa y Comportamiento Animal, Instituto de Ecolog/a, A. C., Apartado Postal No. 63, 91000 Xalapa, Veracruz, M~xico, (e-mail: ornelasj @ecologia.edu.mx)