|
|||||||
|
|
|
|||||
|
|
|||||||
International Journal of Plant Sciences, Nov 1998 v159 i6 p948(1)
GYNODIOECY AND REPRODUCTIVE TRAIT VARIATION IN THREE
THYMUS SPECIES (LAMIACEAE). Domenica Manicacci; Anne Atlan; Juana Anna
Elena Rossello; Denis Couvet.
Abstract: Data on females frequency and reproductive modes in natural
populations of 2 congeneric species have been collected. The species, from the
Author's Abstract: COPYRIGHT 1998 University of
We studied the general features of the reproductive system of two sympatric
species Thymus zygis and Thymus mastichina (Lamiaceae) near
Full Text: COPYRIGHT 1998 University of
Introduction
Gynodioecy (the cooccurrence of hermaphrodite [cosexual] and female
individuals in natural populations) is present in ca. 7.5% of angiosperm
species and occurs in at least 25 mainly dicotyledon, unrelated families among
the European flora (Delannay 1978). The incidence of females in different
species varies from very low (5%) in Daucus carota (Ronfort et al. 1995), to
high frequencies (50%) in Hebe subalpina (Delph 1990b) and Sidalcea oregana
(Ashman 1994), and 60% in Thymus vulgaris (Dommee et al. 1978). Variation in
the frequency of females among populations is also a feature of many
gynodioecious species: 21%-65% in Cucurbita foetidissima (Kohn 1989), 0%-76% in
Beta vulgaris ssp. maritima (Cuguen et al. 1994), 1%-25% in Plantago lanceolata
(van Damme and van Delden 1982), and 5%-95% in T. vulgaris (Dommee et al. 1983).
The adaptive significance of gynodioecy has been widely studied since
In various species, the greater seed-set of females compared with
hermaphrodites results in part from the reallocation of resources not spent in
male function into an increased female investment (Kohn 1989; Atlan et al.
1992; Eckhart 1992; Ashman 1994). Under the "compensation law" hypothesis
(Darwin 1877), which assumes limited reproductive resources, a high female
frequency and/or a high female advantage in fecundity are expected to induce a
selective pressure for an increased allocation into male function in
hermaphrodites (Charnov et al. 1976; Maurice et al. 1993). Female frequency,
relative fecundity of females compared with hermaphrodites, and sex allocation
in hermaphrodites are thus expected to covary among natural populations and/or
species.
The gynodioecious T. vulgaris has been largely studied in southern
Females have been consistently reported in botanical descriptions of other
species of the genus Thymus (Jalas 1972). In some of these species, the
frequency of females was studied precisely. In four species that have both
clonal and sexual reproduction, very high average frequencies of females have
been estimated from the proportion of female ramets. Indeed, around 80% females
were found in Thymus nervosus and Thymus pulegioides in three populations
sampled in the
The present study provides quantitative data on the frequency of females and
reproductive parameters in natural populations of two congeneric species:
Thymus zygis and Thymus mastichina from the
Material and Methods
Thymus zygis and Thymus mastichina are insect-pollinated, perennial bushes
widespread in
Sex Ratio and Floral Morphology
The frequency of females was estimated by sexing 100 randomly sampled
individuals per population in 50 populations of T. zygis and 38 populations of
T. mastichina in June 1990 in
To quantify sexual dimorphism, the number of inflorescences and the mean
flower width were measured in situ on 10 females and 10 hermaphrodites from
three randomly chosen populations in each species. The number of flowers per
inflorescence could not be estimated since the inflorescence development was
not totally accomplished. Flower width was measured as the widest part of the
corolla from five flowers per plant in T. zygis and T. mastichina and from
three flowers per plant in T. vulgaris, using hand calipers.
Reproductive Parameters
Sexual allocation parameters were estimated as seed and pollen production
per flower (SF and PF, respectively), seed-gemination rate (GR), and the
percentage of full pollen grains (FPR; see below for details). For T. zygis and
T. mastichina, measurements were performed on 10 females and 12 hermaphrodites
randomly sampled from six populations per species. They are referred to as
populations Z1-Z6 (T. zygis) and M1-M6 (T. mastichina), were randomly chosen
among populations assayed for sex ratios, and were equally distributed north
and south of
The SF was averaged on 200 fruits randomly sampled from each plant. Seeds
rather than ovules were counted because in Lamiaceae each flower produces a
constant number of four ovules. The GR was measured using 100 seeds per progeny
in petri dishes at the optimal conditions determined for each species: 15
[degrees] C/12 h light per day for T. zygis and T. mastichina preceded by 1 wk
vernalization at 4 [degrees] C for T. zygis and scarification followed by
germination at 25 [degrees] C/12 h light for T. vulgaris. Percentage of
germination was recorded 1 mo after sowing seeds. A combined parameter for
female function was calculated as the number of germinating seeds per fruit
(GSF) by multiplying SF and GR at the plant level. The relative fecundity of
females (RFF) was calculated for each population and species as the ratio of
the average GSF in females to the average GSF in hermaphrodites. The standard
error of this ratio was calculated from the variances for GSF in females and
hermaphrodites.
The PF and the FPR were measured on two flowers per plant from 10-12
hermaphrodites per population (Dommee 1973). The PF was estimated from flowers
acetolyzed in pure sulfuric acid, and extracted pollen grains were counted
using a 1-[micro]L cell hematocytometer under x200 magnification with an
Olympus CH-2 microscope. The FPR was estimated by staining fresh flowers with
methylene blue and recording the proportion of pollen grains containing
cytoplasm. This parameter gives an underestimate of pollen viability or
germinability but was preferred to germination tests since it gives a more
appropriate estimation of the resources invested in pollen grains and is less
sensitive to experimental conditions in the laboratory. A combined parameter
for male function was calculated as the number of full pollen grains per flower
(FPF) by multiplying PF and FPR at the plant level. An estimate of the relative
male investment in hermaphrodites (RMI; Atlan et al. 1992) was calculated for
each individual as follows:
RMI 2 = 2/[Pi] x arctan (FPF/GSF X C),
where C is an allometry coefficient that corresponds to the equivalent
number of full pollen grains produced with the amount of resources needed to
produce a germinating seed. The C-value has been estimated to be ca. 1000 in T.
vulgaris (Atlan et al. 1992). This value was confirmed from the negative
correlation between male and female functions in T. zygis and taken as 1000 for
the three species in this study. The 2/[Pi] coefficient constrains the RMI to
vary from zero (total investment into female function) to one (total investment
into male function).
For each species, the effect of population (random effect), sex (fixed
effect), and their interaction were tested on flower diameter, the number of
inflorescences per plant, all the basic reproductive traits, and RMI using a
mixed model 2 type 3 ANOVA in PROC GLM (SAS 1990). To test the correlation
among different reproductive traits and the female frequency, Spearman's rank
correlation coefficients were calculated among population means within each
species.
Sex Ratio in Progenies
The sex ratio in the offspring of females and hermaphrodites was measured on
seedlings from seeds collected in the six populations of T. zygis and T.
mastichina and grown in an experimental garden at CEFE-CNRS,
Expected female frequency
= F x RFF x HMS + (1 - F) x HMF/F x RFF + (1 - F)
where F is the current female frequency; HMS is the frequency of females in
the progenies of females (male-sterile plants), HMF in the progenies of
hermaphrodites (male-fertile plants); and RFF is the relative fecundity of
females. This assumes an equal survival of both sexes, which has not been
tested in T. zygis and T. mastichina. In T. vulgaris, no significant difference
between sexes was shown in controlled conditions (Assouad 1972) or in natural
populations (Belhassen et al. 1990).
Results
Floral Morphology and Female Frequency
Our observations on Thymus zygis and Thymus mastichina confirm that the two
species are gynodioecious and that females occur at noticeable frequency in
most natural populations of both species (fig. 1). In both Spanish species, as
in Thymus vulgaris (Assouad and Valdeyron 1975), hermaphrodites have large,
protandrous flowers producing substantial amounts of pollen and seeds while
females have smaller, shorter-lived flowers with no stamens. Hermaphrodites
produce significantly larger flowers than females in all three species (mean [+
or -] SE; females = 2.91 mm [+ or -] 0.05, hermaphrodites = 3.92 mm [+ or -]
0.07, [F.sub.(1, 2)] = 55.92, P [is less than] 0.05 for T. zygis; females =
2.93 mm [+ or -] 0.06, hermaphrodites = 3.96 mm [+ or -] 0.08, [F.sub.(1,2)] =
98.91, P [is less than] 0.01 for T. mastichina; females = 3.61 mm [+ or -]
0.09, hermaphrodites = 4.98 mm [+ or -] 0.10, [F.sub.(1, 2)] = 46.15, P [is
less than] 0.05 for T. vulgaris). For each sex, T. vulgaris produced bigger
flowers than both other species (species-effect, [F.sub.(2, 6)] = 20.6, P [is
less than] 0.005), but the difference between sexes was similar in all three
species (nonsignificant sex x species interaction: [F.sub.(2, 6)] = 1.73, P =
0.26). Moreover, we did not observe any plant with flowers of intermediate
sexual phenotype or with both female and hermaphrodite flowers on the same
individual. As a result, sexual phenotypes were as easily distinguishable from
each other in T. zygis and T. mastichina as they were in T. vulgaris, and their
relative occurrence in natural populations could be rapidly quantified.
[Figure 1 ILLUSTRATION OMITTED]
The frequency of females observed in natural populations of T. zygis and T.
mastichina in
As previously observed in T. vulgaris (Bonnemaison 1980), the number of
inflorescences per plant was not significantly different between sexes in the
two Spanish species (mean = 170 and 237 for females and hermaphrodites,
respectively, in T. zygis, [F.sub.(1, 53)] = 1.51, P = 0.22; mean = 105 and 92,
respectively, in T. mastichina, [F.sub.(1, 54)] = 0.25, P = 0.62). Neither the
population effect nor the population x sex interaction was significant for the
number of inflorescences per plant in each species. Thymus mastichina produced
significantly fewer inflorescences per plant than T. zygis ([F.sub.(1, 107)] =
12.71, P[is less than] 0.001). Personal observations indicate that the number
of inflorescences per plant in T. vulgaris is similar to that observed in T.
zygis and much higher than that in T. mastichina.
Inheritance of the Sexual Phenotype
The frequency of females was quantified in open-pollinated offspring from
females and hermaphrodites (table 1). In all three species, females and hermaphrodites
produced individuals of both sexes in their progenies, except in two T.
mastichina populations (M1 and M4) where only hermaphrodites were observed in
the progenies of hermaphrodites. For all species and all populations, the
frequency of females in the progenies of females was much higher than in the
progenies of hermaphrodites and exceeded 50% in many populations. For each
population, the predicted female frequency at the following generation was
compared with the current female frequency to evaluate the stability of the
female frequency in natural populations. For T. zygis and T. vulgaris, the
expected variation in female frequency (table 1) was either positive or
negative depending on the population, with an average decrease of -0.03 in T.
zygis and -0.06 in T. vulgaris expected in one generation. In T. mastichina,
the expected deviation in female frequency through one generation was
consistently negative in all populations, averaging -0.10.
Table 1 Female Frequency in Six Natural Populations for Each of Three Thymus
Species and the Average Frequency of Females in the Offspring of Females and
Hermaphrodites
Female frequency (n)
in offspring
Female
Population frequency Females Hermaphrodites
Thymus zygis 0.51 0.63 (203) 0.18 (238)
Z1 0.17 0.53 (19) 0.03 (37)
Z2 0.29 0.47 (36) 0.10 (30)
Z3 0.41 0.74 (27) 0.13 (38)
Z4 0.52 0.35 (46) 0.05 (42)
Z5 0.77 0.89 (37) 0.56 (54)
Z6 0.83 0.82 (38) 0.22 (37)
Thymus mastichina 0.72 0.68 (206) 0.12 (200)
M1 0.55 0.68 (37) 0.00 (32)
M2 0.60 0.50 (24) 0.34 (41)
M3 0.66 0.70 (37) 0.09 (33)
M4 0.75 0.79 (42) 0.00 (21)
M5 0.81 0.66 (29) 0.14 (28)
M6 0.81 0.78 (37) 0.16 (45)
Thymus vulgaris 0.63 0.65 (1742) 0.19 (1109)
PH 0.10 0.45 (98) 0.03 (693)
M 0.45 0.54 (172) 0.14 (141)
LJ1 0.50 0.51 (258) 0.20 (117)
LJ2 0.73 0.78 (232) 0.25 (36)
LJ3 0.85 0.81 (816) 0.11 (112)
PB 0.85 0.79 (166) 0.40 (10)
Expected
deviation
of female
Population frequency
Thymus zygis -0.03
Z1 -0.05
Z2 -0.02
Z3 +0.01
Z4 -0.21
Z5 +0.08
Z6 -0.08
Thymus mastichina -0.10
M1 -0.01
M2 -0.14
M3 -0.03
M4 -0.02
M5 -0.17
M6 -0.18
Thymus vulgaris -0.06
PH +0.01
M -0.04
LJ1 -0.06
LJ2 -0.06
LJ3 -0.08
PB -0.08
Note. See "Material and Methods" for the expected deviation of
female frequency across one generation, n = number of offspring sexed.
Sexual Resource Allocation
Reproductive parameters are summarized for the three species (table 2). Thymus
mastichina showed the highest number of seeds per fruit and lowest
seed-germination rate for both females and hermaphrodites, averaging a much
lower number of GSF (mean [+ or -] SE; 0.22 [+ or -] 0.066 and 0.07 [+ or -]
0.006 in T. mastichina vs. 0.49 [+ or -] 0.048 and 0.27 [+ or -] 0.027 in T.
zygis and 0.69 [+ or -] 0.080 and 0.29 [+ or -] 0.037 in T. vulgaris for
females and hermaphrodites, respectively).
Table 2 Year of Sampling, Female Frequency, and Basic Reproductive
Parameters for Six Populations of Thymus zygis; Six Populations of Thymus
mastichina Sampled near Salamanca, Spain; and 18 Populations of Thymus vulgaris
Sampled near Montpellier, France
SF, mean
Female
Population Year frequency Females
Thymus zygis 0.51 1.08 (0.065)
Z1 1990 0.17 1.10 (0.223)
Z2 1990 0.29 1.50 (0.123)
Z3 1990 0.41 1.13 (0.127)
Z4 1990 0.52 0.98 (0.162)
Z5 1990 0.77 0.98 (0.108)
Z6 1990 0.83 0.78 (0.114)
Thymus mastichina 0.72 0.30 (0.029)
M1 1990 0.55 0.41 (0.048)
M2 1990 0.60 0.26 (0.058)
M3 1990 0.66 0.46 (0.082)
M4 1990 0.75 0.22 (0.050)
M5 1990 0.81 0.35 (0.087)
M6 1990 0.81 0.11 (0.023)
Thymus vulgaris 0.63 1.47 (0.104)
PH 1978 0.10 2.26 (0.159)
M 1986 0.45 1.50 (0.163)
LJ1 1978 0.50 1.96 (0.152)
LJ2 1978 0.73 0.73 (0.084)
LJ3 1986 0.85 1.67 (0.323)
PB 1978 0.85 0.76 (0.099)
V1 1988 0.13
V2 1988 0.23
V3 1988 0.24
V4 1988 0.40
V5 1988 0.42
V6 1988 0.42
V7 1988 0.44
V8 1988 0.50
V9 1988 0.70
V10 1988 0.77
V11 1988 0.80
V12 1988 0.88
SF, mean GR, mean
Population Hermaphrodites Females Hermaphrodites
Thymus zygis 0.65 (0.046) 0.44 (0.003) 0.39 (0.001)
Z1 0.71 (0.076) 0.10 (0.010) 0.26 (0.004)
Z2 0.89 (0.149) 0.20 (0.008) 0.18 (0.005)
Z3 0.96 (0.095) 0.57 (0.005) 0.49 (0.002)
Z4 0.26 (0.047) 0.79 (0.013) 0.43 (0.008)
Z5 0.50 (0.084) 0.75 (0.005) 0.64 (0.004)
Z6 0.56 (0.056) 0.40 (0.006) 0.37 (0.004)
Thymus mastichina 0.11 (0.009) 0.63 (0.001) 0.60 (0.001)
M1 0.11 (0.023) 0.67 (0.002) 0.61 (0.014)
M2 0.11 (0.025) 0.56 (0.013) 0.67 (0.003)
M3 0.15 (0.022) 0.70 (0.013) 0.62 (0.002)
M4 0.07 (0.018) 0.78 (0.000) 0.62 (0.012)
M5 0.07 (0.013) 0.64 (0.006) 0.50 (0.006)
M6 0.13 (0.023) 0.38 (0.007) 0.57 (0.001)
Thymus vulgaris 0.79 (0.051) 0.42 (0.001) 0.28 (0.001)
PH 1.11 (0.111) 0.61 (0.015) 0.60 (0.001)
M 0.66 (0.064) 0.20 (0.003) 0.16 (0.003)
LJ1 0.86 (0.140) 0.44 (0.003) 0.26 (0.005)
LJ2 0.77 (0.055) 0.33 (0.004) 0.21 (0.002)
LJ3 0.69 (0.183) 0.59 (0.001) 0.45 (0.010)
PB 0.41 (0.127) 0.36 (0.004) 0.10 (0.013)
V1 0.47 (0.108) 0.32 (0.003)
V2 0.33 (0.091) 0.49 (0.008)
V3 0.51 (0.090) 0.37 (0.005)
V4 0.52 (0.084) 0.48 (0.003)
V5 0.27 (0.057) 0.58 (0.011)
V6 0.56 (0.086) 0.53 (0.007)
V7 0.60 (0.077) 0.33 (0.006)
V8 0.20 (0.020) 0.40 (0.005)
V9 0.36 (0.144) 0.22 (0.001)
V10 0.44 (0.082) 0.11 (0.003)
V11 0.57 (0.093) 0.47 (0.006)
V12 0.36 (0.089) 0.55 (0.003)
Population PF, mean FPR, mean
Thymus zygis 2052 (124) 0.77 (0.001)
Z1 2673 (238) 0.82 (0.006)
Z2 2396 (254) 0.65 (0.002)
Z3 1476 (180) 0.75 (0.002)
Z4 1825 (256) 0.72 (0.003)
Z5 1291 (375) 0.88 (0.006)
Z6 2380 (349) 0.82 (0.007)
Thymus mastichina 1526 (97) 0.85 (0.000)
M1 1616 (271) 0.88 (0.001)
M2 1413 (223) 0.84 (0.002)
M3 1559 (210) 0.84 (0.004)
M4 1770 (228) 0.86 (0.003)
M5 1736 (296) 0.85 (0.001)
M6 1129 (187) 0.81 (0.002)
Thymus vulgaris 2250 (115) 0.81 (0.000)
PH
M
LJ1
LJ2
LJ3
PB
V1 1890 (378) 0.78 (0.002)
V2 1879 (180) 0.77 (0.004)
V3 3211 (235) 0.91 (0.002)
V4 1640 (299) 0.89 (0.003)
V5 1505 (390) 0.78 (0.005)
V6 2961 (178) 0.84 (0.003)
V7 3959 (412) 0.81 (0.005)
V8 1667 (301) 0.88 (0.003)
V9 1020 (373) 0.78 (0.007)
V10 2681 (162) 0.70 (0.013)
V11 1417 (220) 0.75 (0.006)
V12 2033 (522) 0.66 (0.012)
Note. Average frequencies of females for each species are estimated from 50
T. zygis populations, 38 T. mastichina populations, and 68 T. vulgaris
populations. SF = number of seeds per fruit; GR = seed-germination rate; PF =
number of pollen grains per flower; FPR = rate of full pollen grains per
flower. Standard errors are in parentheses.
Females produce more seeds per fruit than hermaphrodites for all three
species (see mean values of SF above: [F.sub.(1, 5)] = 25.3, P [is less than]
0.01 in T. zygis; [F.sub.(1,5)] = 14.7, P [is less than] 0.05 in T. mastichina;
[F.sub.(1, 5)] = 12.3, P [is less than] 0.05 in T. vulgaris), and this sexual
difference was consistently found in all populations, except populations M6 (T.
mastichina) and LJ2 (T. vulgaris) where no difference in SF was revealed among
sexes (table 2). In T. vulgaris, seeds from females germinated better than
those from hermaphrodites (GR: table 2 and [F.sub.(1, 5)] = 10.3, P [is less
than] 0.05). This was not the case in T. zygis ([F.sub.(1, 5)] = 0.82, P =
0.41) or in T. mastichina ([F.sub.(1, 5)] = 0.17, P = 0.70). On average, the
number of germinating seeds per fruit was significantly higher in females than
hermaphrodites in all three species (GSF: [F.sub.(1,5)] = 8.6, P [is less than]
0.05 for T. zygis; [F.sub.(1, 5)] = 9.6, P [is less than] 0.05 for T. vulgaris;
and [F.sub.(1, 5)] = 11.8, P [is less than] 0.05 for T. mastichina).
The overall ANOVA including all species revealed significantly higher values
in females than hermaphrodites for SF (sex effect: [F.sub.(1, 15)], i.e., a
lower value for T. zygis [1.85], an intermediate one for T. vulgaris [2.40],
and the highest for T. mastichina [3.09], although the sex x species
interaction was not significant ([F.sub.(2, 15)] = 2.3, P = 0.13). This leads to
an among-species tendency for a positive relationship between the average
frequency of females and the average relative fecundity of females (fig. 2a).
[Figure 2a ILLUSTRATION OMITTED]
Within each species, the effect of sex on GSF varied among populations (sex
x population interaction, [F.sub.(5, 116)] = 2.9, P [is less than] 0.02 in T.
zygis; [F.sub.(5, 105)] = 4.2, P [is less than] 0.002 in T. vulgaris; and
[F.sub.(5, 119)] = 4.9, P [is less than] 0.001 in T. mastichina). However,
females produced consistently more germinating seeds per fruit than
hermaphrodites in all populations but Z1 and M6, and the RFF varied among
populations from 0.6 to 7 in T. zygis, from 1.5 to 6.8 in T. vulgaris, and from
0.57 to 7.3 in T. mastichina. No correlation between RFF and female frequency
was found among populations of each species (fig. 3a-c).
[Figure 3a-c ILLUSTRATION OMITTED]
Data on reproductive parameters in hermaphrodites (table 2) reveal a
significant species effect for the female function ([F.sub.(2, 21)] = 18.7, P
[is less than] 0.001 for SF; [F.sub.(2, 21)] = 8.3, P [is less than] 0.01 for
GR; and [F.sub.(2, 2,)] = 8.4, P [is less than] 0.01 for GSF) with
hermaphrodites in T. mastichina producing fewer seeds per fruit with a higher
germination rate, and, on average, fewer germinating seeds per fruit than
hermaphrodites in both other species. In contrast, no significant species
effect was found on the male function ([F.sub.(2, 21)] = 2.4, P = 0.12 for PF,
[F.sub.(2, 21)] = 2.1, P = 0.15 for FPR, and [F.sub.(2, 21) = 1.9, P = 0.17 for
FPF), although the trend showed that hermaphrodites in T. mastichina produced
fewer pollen grains per flower with a higher rate of full pollen grains and, on
average, fewer full pollen grains per flower than hermaphrodites in both other
species. As a result, the RMI was higher in T. mastichina than in both other
species, the species effect being very close to significance ([F.sub.(2, 21)] =
3.2, P = 0.06). Among the three species, the relative male investment in
hermaphrodites increased with the frequency of females (fig. 2b).
[Figure 2b ILLUSTRATION OMITTED]
Within each species, the relative male investment in hermaphrodites varied
significantly among populations for T. zygis ([F.sub.(5, 61)] = 6.1, P [is less
than] 0.001) and T. vulgaris ([F.sub.(11, 96)] = 2.8, P [is less than] 0.003),
but not for T. mastichina ([F.sub.(5, 62)] = 1.6, P = 0.17). The correlation
between this trait and female frequency among populations was not significant
in any of the three species (fig. 3d-f).
[Figure 3d-f ILLUSTRATION OMITTED]
Discussion
Gynodioecy and Female Frequency in the Three Thymus Species
Our results for Thymus zygis and Thymus mastichina studied in
As indicated in the introduction, females have been reported in many Thymus
species, and all the seven species studied in detail (including the three
present species and T. serpyllum, T. pulegioides, T. nervosus, and T.
sibthorpii) show average female frequency higher than 50%; i.e., much higher
than in gynodioecious species from other genera (Couvet et al. 1986; Richards
1986; Kaul 1988; Delph 1990a). These seven Thymus species are not closely
related within the genus (Morales Valverde 1986). The maintenance of gynodioecy
in many separate Thymus species may be the result of common genetical and/or
physiological constraints that prevent the breakdown of this polymorphism
toward hermaphroditism or dioecy. However, it would be particularly worthwhile
to examine the ecological and genetic factors influencing the independent
maintenance of gynodioecy with high female frequencies in different Thymus species.
Highly variable female frequencies among populations of the three species
clearly illustrate that local female frequency can be much higher than the
maximal 50% predicted under nuclear determination (Lewis 1941). Additionally,
the proportion of female plants in the offspring of females (table 1) exceeded
50% in many populations and was, moreover, consistently greater than in the
offspring of hermaphrodites, indicating a strong maternal effect on the
inheritance of sexual phenotype. These results strongly support the hypothesis
that cytoplasmic factors are involved in sex determination in the three
species. Females and hermaphrodites, nevertheless, produced mixed progenies in
the three species, indicating that nuclear factors are also involved in sex determination.
A nuclear-cytoplasmic determination of sex is already known in T. vulgaris
(Belhassen et al. 1991) as well as in several gynodioecious species from other
genera (see Keyr-Pour 1980 for Origanum vulgare; van Damme 1983 for Plantago
lanceolata; and Boutin et al. 1987 for Beta vulgaris). Our results indicate a
similar determination of sex in T. zygis and T. mastichina.
By estimating the expected frequency of females in the next generation from
the current one and the sex ratio in the offspring of females and
hermaphrodites in natural populations, we predict an average negative deviation
in female frequency for the three species (table 1). This may partly result
from an underestimation of the selective advantage of females compared to
hermaphrodites, which would lead to an underestimation of the expected female
frequency at the following generation. First, the sexual difference in
seed-germination rate was estimated for early stages in controlled conditions
and may be more pronounced at later stages of plant development and/or in the
field. Second, we considered only fecundity differences between sexes and
ignored potential differences in survival rate that may be of great
consequences on female frequency (van Damme and van Damme 1986). In T. vulgaris,
no significant difference in survival rate was observed between sexes (Assouad
1972; Belhassen et al. 1990). This parameter was not assessed in T. zygis or T.
mastichina.
In the three species, the predicted deviation in female frequency varies greatly
among populations and may be either positive or negative, indicating that
female frequencies within populations may vary over time. Models dealing with
the maintenance of gynodioecy show that female frequency may be strongly
variable with time, either because of founder effects (Frank 1989) or
limit-cycle equilibrium (Gouyon et al. 1991). In T. vulgaris, female frequency
has been observed to increase with disturbance and founder events (Manicacci et
al. 1996) and decrease later (Dommee and Jacquard 1985; Belhassen et al. 1989).
The recently disturbed population LJ3 (see Manicacci et al. 1996) may be in the
latter situation since it shows the highest negative value for predicted
deviation in female frequency among the T. vulgaris populations. Among the
sampled populations of T. zygis and T. mastichina, none exhibited patches of
females, and populations with the highest female frequencies did not show
spatial structure of sexual phenotypes (D. Manicacci, personal observation).
However, a long-term study of sex-ratio variation in a large number of
populations would be relevant to understand better the dynamics of gynodioecy
in these species.
Sexual Resource Allocation
Although all three species are gynodioecious with high female frequencies,
they show some differences in terms of reproductive parameters. Thymus zygis
and T. vulgaris produce more seeds with a higher germination rate in both
females and hermaphrodites, more pollen grains, more inflorescences and flowers
per inflorescence than T. mastichina (table 2; D. Manicacci, personal
observation). Given that seeds and pollen grains are highly comparable in size
in the three species (D. Manicacci, personal observation), these results
indicate that T. mastichina has a lower reproductive output per unit vegetable
biomass than the other species. Conversely, T. mastichina produces larger
plants with long, erect branches and larger leaves than the other species. The
low reproductive effort in T. mastichina could result from a trade-off between
reproductive and vegetative investments. It could also be influenced by the
higher ploidy level of T. mastichina, since polyploidization is often
associated with increased vegetative growth (Thompson and Lumaret 1992).
Another difference among species is that in T. mastichina both sexes produce
seeds with similar germination rates, while in T. zygis and T. vulgaris,
females produce seeds with higher germination rates than hermaphrodites (table
2). Hermaphrodites of T. mastichina may have either a better tolerance to inbreeding
depression that could result from their tetraploid nature (Briggs and Walters
1984; but see discussion by Bever and Felber 1992) or a lower selfing rate
resulting from less geitonogamy being favored by the lower number of flowers
per plant (Harder and Barrett 1995). Finally, the greater investment in
vegetative biomass in T. mastichina may reduce the resource compensation at the
flower level in females.
In populations with more females and/or where the relative fecundity of
females is higher, one may expect the functional gender of hermaphrodites to
evolve toward an increased male function since more female ovules are available
in the population. This correlation was not observed within any of the three
study species, among populations (fig. 3). Likewise, other intraspecific
comparisons have seldom shown any relation between female frequency and the
relative fecundity of females compared with hermaphrodites (Glechoma hederacea,
Widen 1992; Trifolium hirtum, Molina-Freaner and Jain 1992) or between female
frequency and sex allocation in hermaphrodites (several Apiaceae, Webb 1979;
Gingidia flabellata, Webb 1981; but see Boutin et al. 1988 for only two
populations of Beta vulgaris ssp. maritima). Theoretical studies dealing with
the maintenance of gynodioecy predict temporal variation in female frequency
within populations (Frank 1989; Gouyon et al. 1991), as observed in T. vulgaris
populations (Belhassen et al. 1989; Manicacci et al. 1996). Consequently, the
evolutionary stable strategy for sex allocation in hermaphrodites may not be
observed in individual populations (Levins 1968, p. 11), resulting in a lack of
correlation. Our results on sex ratios in progenies indicate that the frequency
of females may vary across generations in the natural populations of the three
species (table 1). Thus a delayed response to selection may be a cause for the
lack of correlation within each species. In contrast, the comparison of Wurmbea
dioica populations in different regions of
In the three study Thymus species, the average relative fecundity of females
and the average sex allocation into male function in hermaphrodites both
increase with the average frequency of females (fig. 2). Stronger relationships
are found when reproductive parameters are calculated from the number of seeds
per fruit and pollen grains per flower, that is, without considering
seed-germination rate or the proportion of full pollen grains in the estimation
of female and male functions, respectively. Thus, we consider our result as an
indication that the expected correlation between sex ratio and sex allocation may
occur at the species level. However, this relationship should be considered
with caution for several reasons. First, our observations are based on only
three species. Second, differences among species are not strongly significant
for the average number of germinating seeds per fruit or full pollen grains per
flower. Similar relations have been observed in other interspecific comparisons
of sex ratio and sex allocation in hermaphrodites of gynodioecious species
(Lloyd 1976; Gouyon and Couvet 1987; Delph 1900b, 1990c) as well as in
comparisons of sex ratio and relative fecundity of females (Webb [1979] in
Apiaceae, but see Kohn [1989] for Cucurbita species). The fact that female
frequency, relative fecundity of females, and sex allocation in hermaphrodites covary
among species in accordance with the predictions derived from sex-allocation
theory (Charnov et al. 1976) indicates that these traits are modeled by
selective pressures that may influence the maintenance of the gynodioecious
reproductive system. However, if traits are strongly influenced by stochastic
events at the population level, the relationship among characters may be
difficult to detect at a local scale, as indicated by within-species results
that do not show any tendency. This may be particularly true in the case of
gynodioecy since the frequency of females can strongly vary with time. In this
context, studies at a higher level, such as the metapopulation level, and/or
among closely related species may thus be required to remove the effect of stochastic
variation and detect the effect of selective forces.
Acknowledgments
We thank M. Beltran and M. A. Reglade for technical and fieldwork
assistance, and I. Olivieri, J. Ronfort and J. D. Thompson for critical
comments on the manuscript. Financial support from the Spanish Government to
Juana Anna Elena Rossello and Domenica Manicacci and a BDI-CNRS research
scholarship to Domenica Manicacci are gratefully acknowledged.
Literature Cited
Agren J, MF Willson 1991 Gender variation and sexual differences in
reproductive characters and seed production in gynodioecious Geranium
maculatum. Am J Bot 78:470-480.
Arroyo MTK, PH Raven 1975 The evolution of subdioecy in morphologically
gynodioecious species of Fuschia sect. encliandra (Onagraceae). Evolution 29:500-511.
Ashman TL 1992 The relative importance of inbreeding and maternal sex in
determining progeny fitness in Sidalcea oregana ssp. spicata, a gynodioecious
plant. Evolution 46:1862-1874.
-- 1994 Reproductive allocation in hermaphrodite and female plants of
Sidalcea oregana ssp. spicata (Malvaceae) using four currencies. Am J Bot
81:433-438.
Assouad MW 1972 Recherches sur la genetique ecologique de Thymus vulgaris
L.: etude experimentale du polymorphisme sexuel. These doctorat es sciences.
Universite des Sciences et Techniques du Languedoc,
Assouad MW, B Dommee, R Lumaret, G Valdeyron 1978 Reproductive capacities in
the sexual forms of the gynodioecious species Thymus vulgaris L. Bot J Linn Soc
77:29-39.
Assouad MW, G Valdeyron 1975 Remarques sur la biologie du thym Thymus
vulgaris L. Bul Soc Bot Fr 122:21-34.
Atlan A, PH Gouyon, T Fournial, D Pomente, D Couvet 1992 Sex allocation in
an hermaphroditic plant: the case of gynodioecy in Thymus vulgaris L. J Evol
Biol 5:189-203.
Barrett SCH 1992 Gender variation and the evolution of dioecy in Wurmbea
dioica (Liliaceae). J Evol Biol 5:423-444.
Barrett SCH, K Helenurm 1981 Floral sex ratios and life histories in Aralia
nudicaulis (Araliaceae). Evolution 35:752-762.
Belhassen E 1989 Etude evolutive du systeme reproductif de Thymus vulgaris
L.: analyse ecologique, genetique et moleculaire de la sterilite-male. These de
l'Universite des Sciences et Techniques du Languedoc,
Belhassen E, M Beltran, D Couvet, B Dommee, PH Gouyon, I Olivieri 1990
Evolution des taux de femelles dans les populations naturelles de thym, Thymus
vulgaris L.: deux hypotheses alternatives confirmees. C R Acad Sci Paris 310:
371-375.
Belhassen E, B Dommee, A Atlan, PH Gouyon, D Pomente, W Assouad, D Couvet 1991
Complex determination of male sterility in Thymus vulgaris L.: genetical and
molecular analysis. Theor Appl Genet 82:137-143.
Belhassen E, L Trabaud, D Couvet, PH Gouyon 1989 An example of
nonequilibrium processes: gynodioecy of Thymus vulgaris L. in burned habitats.
Evolution 43:662-667.
Bever JD, F Felber 1992 The theoretical population genetics of
autopolyploidy. Oxf Surv Evol Biol 8:185-217.
Bonnemaison F 1980 Etude stationnelle de la dynamique du maintien d'un
polymorphisme genetique: cas de quatre populations naturelles de Thymus
vulgaris L. These de doctorat de troisieme cycle. Universite des Sciences et
Techniques du Languedoc,
Boutin V, R Jean, M Valero, P Vernet 1988 Gynodioecy in Beta maritima. Acta
Oecol 9:61-66.
Boutin V, G Pannenbecker, W Ecke, G Schewe, P Saumitou-Laprade, R Jean, P
Vernet, G Michaelis 1987 Cytoplasmic male sterility and nuclear restorer genes
in a natural population of Beta maritima: genetical and molecular aspects.
Theor Appl Genet 73:625-629.
Briggs D, SM Walters 1984 Plant variation and evolution.
Charlesworth D 1981 A further study of the problem of the maintenance of
females in gynodioecious species. Heredity 46:27-39.
Charnov EL, J Maynard Smith, JJ Bull 1976 Why be an hermaphrodite? Nature
263:125-126.
Chenuil A 1989 Sterilite male chez Thymus vulgaris L.: de la variabilite
moleculaire a la variabilite sexuelle. DEA, Universite Montpellier II, France.
Correns C 1906 Die rolle der mannlichen keimzellen bei der
geschlechtsbeststimung der gynodicischen pflanzen. Ber Deutsch Bot Ges
26a:686-701.
Couvet D, F Bonnemaison, PH Goyoun 1986 The maintenance of females among
hermaphrodites: the importance of nuclear-cytoplasmic interactions. Heredity
57:325-330.
Cuguen J, R Wattier, P Samitou-Laprade, D Forcioli, M Morchen, H van Difjk,
P Vernet 1994 Gynodioecy and mitochondrial DNA polymorphism in natural
populations of Beta vulgaris ssp. maritima. Genet Sel Evol 26:87-101.
Darwin CR 1877 The different forms of flower on plants of the same species.
J
Delannay X 1978 La gynodioecie chez les angiospermes. Nat Belg 59:223-235.
Delph LF 1990a The evolution of gender dimorphism in New Zealand Hebe
(Scrophulariaceae) species. Evol Trends Plants 4:85-97.
-- 1990b Sex-differential resource allocation patterns in the subdioecious
shrub Hebe subalpina. Ecology 71:1342-1351.
-- 1990c Sex-ratio variation in the gynodioecious shrub Hebe strictissima
(Scrophulariaceae). Evolution 44:134-142.
Dommee B 1973 Recherche sur la genetique ecologique de Thymus vulgaris L.:
determinisme genetique et repartition ecologique des formes sexuelles. These
d'Etat. Universite des Sciences et Techniques du Languedoc,
Dommee B, W Assouad, G Valdeyron 1978 Natural selection and gynodioecy in
Thymus vulgaris L. Bot J Linn Soc 77:17-28.
Dommee B, JL Guillerm, G Valdeyron 1983 Regime de reproduction et
heterozygotie des populations de thym, Thymus vulgaris L., dans une succession
post-culturale. C R Acad Sci Paris Ser III 296: 111-114.
Dommee B, P Jacquard 1985 Gynodioecy in thyme, Thymus vulgaris L.: evidence
from successional populations. Pages 141-164 in J Antonovics and P Jacquard,
eds. Genetic differentiation and dispersal in plants. Springer,
Eckhart VM 1992 Resource compensation and the evolution of gynodioecy in
Phacelia linearis (Hydrophyllaceae). Evolution 46: 1313-1328.
Frank SA 1989 The evolutionary dynamics of cytoplasmic male sterility. Am
Nat 133:345-376.
Gouyon PH, D Couvet 1985 Selfish cytoplasm and adaptation: variations in the
reproductive system of thyme. Pages 299-319 in J Haeck and JW Woldendorp, eds.
Structure and functioning of plant populations. Vol 2.
-- 1987 A conflict between two sexes, females and hermaphrodites. Pages
245-261 in S. Stearns, ed. The evolution of sex and its consequences.
Gouyon PH, F Vichot, JMM van Damme 1991 Nuclear-cytoplasmic male sterility:
single-point equilibria versus limit cycles. Am Nat 137:498-514.
Harder LD, SCH Barrett 1995 Mating cost of large floral displays in
hermaphrodite plants. Nature 373:512-515.
Harris W 1968 Environmental effects on the sex ratio of Rumex acetosella L.
Proc N Z Ecol Soc 15:51-54.
Jalas J 1972 Thymus L. Pages 172-182 in T Tutin et al., eds. Flora Europea.
Vol 3.
Kaul MLH 1988 Male sterility in higher plants. Monographs on theoretical and
applied genetics, no. 10. Springer. 1005 pp.
Keyr-Pour A 1980 Nucleo-cytoplasmic polymorphism for male sterility in
Origanum vulgare L. J Hered 71:253-260.
Kikuzawa K 1989 Floral biology and evolution of gynodioecism in Daphne
kamtchatica var. jezoensis. Oikos 56:196-202.
Kohn JR 1989 Sex ratio, seed production, biomass allocation, and the cost of
male function in Cucurbita foetidissima Hbk (Cucurbitaceae). Evolution
43:1424-1434.
Levins R 1968 Evolution in changing environments: some theoretical
explorations.
Lewis D 1941 Male-sterility in natural populations of hermaphrodite plants.
New Phytol 40:56-63.
Lloyd DG 1976 The transmission of genes via pollen and ovules in
gynodioecious angiosperms. Theor Popul Biol 9:299-316.
Maki M 1993 Outcrossing and fecundity advantage of females in gynodioecious
Chionographis japonica var. kurohimensis (Liliaceae). Am J Bot 80:629-634.
Manicacci D, D Couvet, E Belhassen, PH Gouyon, A Atlan 1996 Founder effects and
sex ratio in the gynodioecious Thymus vulgaris L. Mol Ecol 5:63-72.
Maurice S, D Charlesworth, C Desfeux, D Couvet, PH Gouyon 1993 The evolution
of gender in hermaphrodites of gynodioecious populations with
nucleo-cytoplasmic male-sterility. Proc R Soc Lond B 251:253-261.
Molina-Freaner F, SK Jain 1992 Female frequencies and fitness components
between sex phenotypes among gynodioecious populations of the colonising
species Trifolium hirtum All. in California. Oecologia 92:279-286.
Morales Valverde R 1986 Taxonomia de los generos Thymus (excluida la seccion
Serpyllum) y Thymbra en la peninsula iberica. Monografias del jardin botanico.
Consejo Superior de Investigaciones Cientificas, Madrid.
Opler PA, KS Bawa 1978 Sex ratios in tropical forest trees. Evolution
32:812-821.
Reichenbach HG 1857 Uber drei neuaufgestellte Labiaten-Gattungen. Ost Bot
Wochenbl 7:159-161.
Richards AJ 1986 Plant breeding systems. Allen, London.
Ronfort J, P Saumitou-Laprade, J Cuguen, D Couvet 1995 Mitochondrial DNA
diversity and male sterility in natural populations of Daucus carota ssp.
carota. Theor Appl Genet 91:150-159.
Ross MD 1970 Breeding systems in Plantago. Heredity 25:129-133.
-- 1978 The evolution of gynodioecy and subdioecy. Evolution 32:174-188.
-- 1982 Five evolutionary pathways to subdioecy. Am Nat 119: 297-318.
SAS 1990 SAS user's guide: statistics, version 6 edition. SAS Institute,
Cary, N.C.
Shykoff JA 1988 Maintenance of gynodioecy in Silene acaulis
(Caryophyllaceae): stage-specific fecundity and viability selection. Am J Bot
75:844-850.
Sun M 1987 Genetics of gynodioecy in Hawaiian Bidens (Asteraceae). Heredity
59:327-336.
Thompson JD, R Lumaret 1992 The evolutionary dynamics of polyploid plants:
origins, establishment and persistence. Trends Ecol Evol 7:302-307.
Van Damme JMM 1983 Gynodioecy in Plantago lanceolata L. II. Inheritance of
three male-sterility types. Heredity 50:253-273.
Van Damme JMM, W Van Damme 1986 On the maintenance of gynodioecy: Lewis'
result extended. J Theor Biol 121:339-350.
Van Damme JMM, W Van Delden 1982 Gynodioecy in Plantago lanceolata L. I.
Polymorphism for plasmon type. Heredity 49:303-318.
-- 1984 Gynodioecy in Plantago lanceolata L. IV. Fitness components of sex
types in different life cycle stages. Evolution 38: 1326-1336.
Webb CJ 1979 Breeding systems and the evolution of dioecy in
-- 1981 Gynodioecy in Gingidia flabellata (Umbelliferae). N Z J Bot
19:111-113.
Widen M 1992 Sexual reproduction in a clonal, gynodioecious herb Glechoma
hederecea. Oikos 63:430-438.
Domenica Manicacci,(1)(*) Anne Atlan,(2)(*) Juana Anna Elena
Rossello,([dagger]) and Denis Couvet(3)(*)
(*) Centre d'Ecologie Fonctionnelle et Evolutive/CNRS, BP 5051, 34033
([dagger]) Departamiento de Biologia Vegetal,
(1) Author for correspondence and reprints. Current address: Station de
Genetique Vegetale,
(2) Current address: Dynamique du Genome et Evolution, Tour 42, Universite
de Paris VI, 4 place Jussieu, 75251
(3) Current address: Institut d'Ecologie, Universite de Paris VI, Batiment
A, 7eme etage, 7 quai St Bernard, Case 237, 75752 Paris Cedex 05, France.