| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
0 Department of Plant Biology, Botany Division, University of Cordoba, Avda. San Alberto Magno s/n, E-14004 Córdoba, Spain
ABSTRACT
Cleistogamous capitula formed by Centaurea melitensis display a number of morphological and functional changes with respect to chasmogamous capitula that ensure self-fertilization. Because no studies have hitherto addressed the evolution of cleistogamy in Asteraceae, it was considered useful to ascertain whether these changes are attributable to one or more of the heterochronic processes reported in the literature. Bivariate allometric analyses were performed, and changes were represented graphically using Gould's clock models for size, shape, and age of several capitulum and floret structures. Results suggest that the partially paedomorphic appearance of cleistogamous with respect to chasmogamous capitula is attributable to three processes: (1) early onset of floral development (predisplacement), (2) decreased growth rate of the whorls studied (except gynoecium width) and (3) early offset time (progenesis). The latter appears to play the most significant role in the origin of the cleistogamous capitulum.
Key Words: allometry • Asteraceae • Centaurea • cleistogamy • heterochrony
Heterochronic evolutionary mechanisms may be defined as changes in the rate and/or timing (onset-offset) of developmental events between ancestral and descendant ontogenies (Gould, 1977
). These relatively simple processes may play a major role in morphological evolution, giving rise either to paedomorphosis, i.e., retention of the ancestral juvenile shape in the descendant adult, or peramorphosis, i.e., extension of descendant ontogeny relative to the ancestor (Alberch et al., 1979
).
Using the terminology formulated by Alberch et al. (1979)
, heterochronic processes involve any of six possible mechanisms: neoteny (decrease in the rate of shape development, K
), postdisplacement (later onset time, 
), progenesis (earlier offset time, ß), acceleration (increase in the rate of shape development, K), predisplacement (earlier onset time, ), and finally hypermorphosis (later offset time, ß). The first three give rise to paedomorphosis and the rest to peramorphosis. Two further mechanisms, proportional dwarfism and gigantism, result from changes in the rate of size development, Ks, although the descendant offset shape is not affected (isomorphosis).
Reilly, Willey, and Meinhardt (1997)
recently proposed that the terms "neoteny" and "progenesis" be replaced by "deceleration" and "hypomorphosis," respectively. Chief among the reasons cited for this modification are the fact that the proposed terms are less ambiguous and that they stand in logical opposition to the existing terms "acceleration" and "hypermorphosis." Both sets of terms will be used in this paper, the more recent modifications appearing in parenthesis.
All these heterochronic processes may affect a whole organism, global heterochrony, or only single traits, dissociated heterochrony (Gould, 1977
; McKinney and McNamara, 1991
). The heterochronic processes taking place may be identified by means of analysis of size, shape, and age (timing) (Alberch et al., 1979
; McKinney, 1988
). The methodology formulated by Alberch et al. (1979)
is based on animals, which follow a determinate growth pattern. Most plants, however, develop according to an indeterminate modular pattern in which every vegetative meristem is semiautonomous and potentially reproductive (Guerrant, 1982
).
A number of studies have addressed heterochrony in plants, equating the development of plant modules with whole animals (Guerrant, 1982
; Jones, 1992, 1993
; McLellan, 1993
; Kampny, Dickinson, and Dengler, 1994
; Zopfi, 1995
). Specifically, work has been done on the phylogenetic derivation of autogamous (including cleistogamous) flower forms from allogamous progenitors (Lord, 1982, 1984
; Lord and Hill, 1987
; Guerrant, 1988
; Lord, Eckard, and Crone, 1989
; Gallardo, Domínguez, and Muñoz, 1993
). In particular, the derivation of the cleistogamous flower from the chasmogamous flower is generally accepted and involves a system that is of special interest when studying the evolution of floral morphology (Lord and Hill, 1987
).
The present study of Centaurea melitensis L. (Asteraceae) examines the possible heterochronic origin of small cleistogamous capitula (initial cleistogamous, iCL, and final cleistogamous, fCL) from chasmogamous (CH) capitula via one or more of the heterochronic processes reported in the literature (Gould, 1977
). Particular attention is paid to those processes resulting in paedomorphic derivatives, since cleistogamy habitually entails a number of traits giving the capitulum a juvenile appearance (Darwin, 1877
; Lord, 1981
).
For this purpose, the following were performed: (1) bivariate allometric analyses (studies of size-shape; Gould, 1966
; McKinney and McNamara, 1991
); (2) growth equations for various floral structures; (3) measure of the duration of various development events, and (4) graphical representation in the form of clock models of heterochronic changes (Gould, 1977
) in the size, shape, and age of the floral structures under study.
Because the functional reproductive unit in this case is the capitulum, an analysis of the extent to which capitular traits have been modified in the evolutionary shift from chasmogamy to cleistogamy was considered of particular interest.
MATERIALS AND METHODS
Study species
Centaurea melitensis L. (Asteraceae) is a cleistogamous herb species rarely reaching 1 m in height, in which a single plant forms discoid capitula varying in terms of morphology, showiness, placement on the plant, and ontogenic timing of appearance (Porras, 1998
). This species presents three capitulum types considered to be extremes within a broad spectrum: initial cleistogamous capitula (iCL), the first capitula produced, between one and four in number, with a mean size of 12.5 x 5.6 mm, and located towards the base of the plant or in the axils of the main branches; final cleistogamous (fCL) the smallest capitula produced by the plant (9.6 x 3.2 mm), developed towards the end of the flowering period, on the main axis, or on first- and second-order branches (in both types, all florets are bisexual and structurally adapted to obligate self-fertilization); and chasmogamous capitula (CH, 17.4 x 8.6 mm), developed during the middle phase of the flowering period, located near to the apex of the plant, with an outermost row of sterile florets; the remaining florets are fertile hermaphrodite, potentially adapted to cross-fertilization (Porras, 1998
).
Allometry
Analysis of development in size and shape of the different floral structures throughout ontogeny requires the use of destructive methods, since some whorls remain embedded within others, thus rendering it impossible to monitor the development of organs in a single flower (or, in this case, capitulum). Samples of previously identified capitulum types at varying stages of development (from 4.5 mm long to fertilization) were therefore fixed in FAA (formalin-acetic acid-70% ethanol; Johansen, 1940
). All samples were collected from 27 (9 x 3 populations) plants grown in greenhouse conditions from seeds obtained in spring 1995 from three natural populations in Spain: (1) Córdoba, Lucena-Cerro Acebuchoso; (2) Seville-between Coripe and Morón; (3) Jaén-between Luque and Alcaudete, vicinity of the River San Juan.
Sample size was N = 30 for initial cleistogamous (iCL), N = 40 for final cleistogamous (fCL), and N = 50 for chasmogamous (CH) capitula.
A total of eight characters were measured (in millimetres) for each capitulum (Fig. 1) using a calibrating gauge, a dissecting microscope, or a microscope with graduated eyepiece depending on the size of the organ examined. Since florets on each iCL or fCL capitulum tend to be similar, characters were measured on one randomly selected floret per capitulum. On CH capitula, however, two types of fertile flower develop: the outermost fertile florets and the rest, which are larger (Porras, 1998
). In this analysis one floret from the outer area was studied per capitulum. It was considered appropriate to use the outermost florets for reference purposes, because it seems more likely that CL florets derive from these (Porras, 1998
).
|
simple allometry equation y = axb generally provides a valid model for most size-shape studies (Kuhry and Marcus, 1977
). Logarithmic transformation of this equation yields a linear model logy = blogx + loga.
To estimate the slope of this regression (b) least squares regression (LSR) was used, even though the assumption of independence in one variable is violated. This method was used for two reasons: (a) analyses allowing comparison of slopes (analysis of covariance) are based on LSR, and (b) in cases of complex allometry (such as that considered here), where the ratio of variable "x" to variable "y" is not constant (McKinney, 1988
) the relationship is best fitted by a polynomial curve. No curvilinear models are available for other statistical techniques such as major axis regression (MAR), and reduced major axis (RMAR).
Since the best fit (R2) was obtained with quadratic regressions, for each whorl (involucre, corolla, androecium, and gynoecium) a width vs. length equation was produced (log-log scale). However, despite a reduced R2 value, a linear model was used to facilitate comparison of slopes (growth rates) for the different capitulum types studied. Slope homogeneity was tested by analysis of covariance (procedure proc GLM; SAS, 1988
). Comparison of elevation (adjusted means) was performed where no significant difference was recorded between slopes, in order to distinguish between a single regression line and a pair of parallel lines with different elevations.
Growth equations (variable age)
In spring 1995, daily increase in length, from the base to the spine apex of the largest bracts, was monitored in a group of tagged capitula (iCL, N = 28; fCL, N = 31; CH, N = 27) on plants grown in greenhouse conditions (nine pots for each of the three natural populations described earlier). Measurements were made with a 0.1-mm precision gauge, and monitoring began (T = 0) at a length of ~4.5 mm. Measurements were not made in smaller plants, since handling could cause damage that might impair normal capitulum growth. Monitoring stopped when growth ceased.
Mean involucre length values as a function of time elapsing were recorded, and an estimation of the logistic curve [y = k/(1 + be-aT), where y is involucre length; k, the ceiling size of y; b, the time shift parameter; a, growth parameter; and T, the time in days] that best fitted the growth observed was made for each capitulum type (Guerrant, 1982
). The procedure NLIN method DUD (SAS, 1988
) was used for the calculation of this equation.
In spring 1996, samples from greenhouse plants (same populations) were fixed in FAA to determine the size and timing of two development events: pollen mother cell (pmc) meiosis (tetrad phase) and anther dehiscence for all capitulum types (pmc meiosis, iCL, N = 30; fCL, N = 23; CH, N = 30; dehiscence iCL, N = 30; fCL, N = 29, CH, N = 30). Capitulum and corolla length (one corolla per capitulum) were measured. Since CH florets do not develop simultaneously, uniformity was achieved by always taking measurements on the second row of fertile florets at meiosis and anther dehiscence.
The age of the capitulum samples used for this allometric analysis was estimated from the logistic growth curves produced earlier, thus enabling comparison of organ growth rates between capitulum types.
Anther dehiscence in CH capitula is followed by rapid growth of the corolla tube, stamen filaments, and style, giving rise to anthesis. In contrast, involucre growth ceases gradually. It was therefore considered preferable to include for study CH capitulum samples only for the period from 4.5 mm in length to anther dehiscence. Similarly, only iCL and fCL capitula showing no sign of fertilization were included. Capitula exceeding the maximum size of the logistical growth equations were also excluded, giving a sample size for iCL of N = 24, for fCL of N = 31 and for CH of N = 40.
The same statistical criterion was used as for bivariate analysis. Depending on the fit, either quadratic or cubic polynomial equations were employed.
Clock models
Clock models of heterochronic change (Gould, 1977
) were used to represent the size, shape, and age of capitular/ floral whorls at three developmental stages: (a) anthers showing clear differentiation of thecae, upper appendage, and lower appendages; (b) pmc meiosis, and (c) anther dehiscence. Stage (a) was determined on capitula already used for allometric analysis, while a specific sample was used for (b), and (c), as indicated earlier.
Clocks for each whorl were calibrated by placing at the midpoint of the semicircular scale the values recorded for the size (inner scale) and shape (outer scale) either of CH capitulum involucre or floral characters at anther dehiscence (i.e., ancestral characters). The age at which dehiscence took place was similarly placed at the midpoint of the graduated baseline. The size of each whorl was estimated as length plus width; shape was expressed as the width:length ratio for each whorl.
Both size and shape were estimated from polynomial growth equations, except for involucre length at T = 0 in CH, whose value was obtained from the logistic curve. Semicircular scales for size and shape were plotted as a function of time elapsing between T = 0 (capitulum length 4.5 mm) and anther dehiscence of second-row fertile florets.
RESULTS
Bivariate allometric analysis
Throughout development, and for all capitular/floral whorls (involucre, corolla, androecium and gynoecium) except the involucre of CH capitula, the relative growth in width was smaller than that of length. Due to this progressive elongation of organs, the value of regression slopes obtained (Table 1, b1 in linear equations) was less than 1, i.e., negative allometry. In CH capitula the involucre tended to thicken, thus yielding positive allometry (Table 1). The width:length ratio (slope) was not constant, although it was considered constant to facilitate comparison of capitulum types. The ratio in fact varied slightly throughout development, while for involucre (CH and fCL) and gynoecium it tended to increase, in corolla and androecium it decreased (Table 1). Growth of the C. melitensis capitulum thus displays complex allometry.
|
Although there were no significant differences between slopes for iCL and fCL, the equations cannot be considered the same since elevations are not homogeneous (Table 1).
Table 1 shows analyses of slopes and, where appropriate, elevations of allometry equations for corolla, androecium and gynoecium. Interestingly, overall analysis of the three floral types revealed no significant differences in slope values for androecium (Table 1). However, slope analysis including only CH florets of the same size as CL yielded significant differences between CH and CL, although not between iCL—fCL.
Differences observed between slopes and elevations highlight a dissociation between size and shape of CL characters with regard to the CH development pattern. CL floret corolla and androecium have the same shape, but at smaller size, as CH florets. CH stamens thus tend to be relatively wider than CL stamens in capitula of the same size.
Finally, and unlike the corolla and the androecium, the gynoecium displays a delay in shape with respect to size. At equivalent lengths, pistil width is greater in CL florets than in CH florets.
Growth models
From the logistic growth equations for involucre length of the various capitulum types (Fig. 2, Table 2) it was possible to estimate growth rates throughout development of the floral structures studied by bivariate allometric analysis. Growth rates, like allometry, vary throughout ontogeny so length is best fitted by polynomial equations (quadratic or cubic, as appropriate) (Table 3). In some cases, such as width of corolla, androecium, and gynoecium of iCL capitula, R2 values were greatly increased by polynomial rather than linear fitting. In other cases, such as corolla width in fCL capitula and androecium width in iCL, the assumption of linearity may not be valid for slope determination due to the low value of R2.
|
|
|
The rate of involucre width increase, though similar for all capitula, is significantly different (Table 3). fCL capitula are narrower, while the growth rate of CH capitula tends to increase towards the end of development.
In each floret type studied, the corolla, androecium, and gynoecium display a very similar growth pattern in terms of both length and width. However, the growth rate of gynoecium width tends to increase towards the end of development, while that of corolla and androecium width tends to diminish until growth ceases.
Growth rates of corolla, androecium and gynoecium were always larger, in terms of both length and width, in CH florets than in CL florets (iCL and fCL) (Table 3; linear equations). However, gynoecium width displays a greater growth rate in CL.
Differences between slopes (rates) for corolla length were not significant, although more than one line was involved, since there were significant differences for elevations between CH and CL (Table 3).
With regard to androecium length and gynoecium length, significant differences in slopes were found only between iCL and CH. However, elevations differed between CH and fCL capitula (Table 3). This implies that after the same development time, these whorls are shorter in CH florets.
Finally, there was a significant difference in growth rate (slope) for width (corolla, androecium and gynoecium) between CH and CL (Table 3), with equally significant differences in elevations between iCL and fCL for gynoecium width.
Gould clock models
The age scale on the clock models shows that flower formation and development in CL capitula generally takes place ahead of CH capitula (Fig. 3). Time elapsing between the different events recorded (clear anther differentiation, pmc meiosis, and anther dehiscence) decreases from CH 
iCL fCL (Figs. 2, 3).
|
At pmc meiosis and dehiscence, corolla, androecium, and gynoecium are appreciably smaller in iCL and fCL than in CH. The iCL and fCL whorls are generally similar in size, the latter being slightly smaller. At both pmc meiosis and dehiscence, the fCL involucre is narrower than that of the CH capitula. Since throughout ontogeny it never reaches a similar shape, the hands representing shape on this clock model could not be placed (Fig. 3). The shape of the involucre in iCL and of the other floral structures studied at pmc meiosis and dehiscence in CL corresponds to shapes at earlier stages of CH. There is, however, a certain dissociation between size and shape. Involucre, corolla, and androecium display an acceleration of shape relative to size, i.e., a given shape is observed in CH at slightly larger sizes. In contrast, gynoecium shape at anther dehiscence is somewhat delayed relative to size.
DISCUSSION
In C. melitensis the shift from chasmogamy to cleistogamy entails a number of morphological and functional changes that ensure self-fertilization in cleistogamous capitula (Porras, 1998
). Can these structural changes be attributed to a heterochronic process? And if so, which of the various heterochronic mechanisms may have triggered the structural and functional differences observed?
Involucre
At first sight, the results obtained here would seem to indicate that cleistogamous capitula (iCL and fCL) at adult stage (anther dehiscence) are more juvenile in appearance and smaller in size than chasmogamous (CH) capitula, suggesting a heterochronic process leading to paedomorphosis. Yet some of the observations in fact suggest the involvement of several heterochronic processes (i.e., predisplacement, or decrease in Ks) rather than just one [such as progenesis (hypomorphosis)] in the production of this final paedomorphic appearance, as analyzed below.
At the different ontogenic stages studied (anthers clearly distinguishable, pmc meiosis, and dehiscence) the involucre decreases in size from CH iCL fCL (Fig. 3). Involucre shape in iCL is similar to that found in earlier developmental stages in CH capitula, while fCL capitula are relatively narrower.
To identify the cause of these variations in size and shape, it is first necessary to ascertain whether there has been any change in one or more of the control parameters [ or onset time, ß or offset time, and K or rate of change in shape (K) and/or size (Ks)] cited by Alberch et al. (1979)
as regulating the ontogenic trajectory.
We found (1) a clear decrease in the time taken to reach the development stages studied in CL relative to CH (final development gain in iCL= 15.02 d and in fCL= 19.93 d), (2) decreased size of capitula and other floral characters, and (3) juvenile appearance of CL capitula. All this suggests that changes are related to a decrease in ß (offset time) in CH capitula, perhaps prompted by an increase in sexual maturation relative to CH. However, other findings cannot be accounted for solely by the decrease in ß. These include the fact that the involucre in fCL capitula never attains the shape observed in CH capitula, the smaller number of florets per capitulum in CL (Porras, 1998
), and the slight dissociation between size and shape in iCL capitula (Fig. 3).
These findings may be partially explained by the early onset of floral primordia () in CL capitula. Earlier onset would imply that floral primordia in CL capitula start to develop from a smaller involucre and, hence, on smaller receptacles. This hypothesis is borne out by ontogenic studies (Porras, 1998
) showing that CH capitula start to develop fertile primordia when the involucre is ~4 mm in length and receptacle diameter is ~310 µm. CL capitula, however, initiate floral development at sizes of ~2 mm and 210 µm, respectively (iCL). The smaller space available on the receptacle and early truncation of the centripetal development of new floral primordia in CL capitula relative to CH would account for the lower number of florets per capitulum (CH = 90.17 florets, iCL = 16.77 florets, and fCL = 8.86 florets). As a result of these changes, less time would be invested in the initial phases of floral development. The decrease in the time taken by each CL capitulum from T = 0 to the stage where anthers become clearly distinguishable (Fig. 3, iCL= 7.50 d and fCL= 9.34 d ahead of CH) can thus be seen as reflecting this modification of .
This, however, would assume that there is only a moderate difference between capitular types in the time that elapses between initiation of floral primordia and the time that anthers are clearly distinguishable. Since the size of floral organs is similar by this stage (Fig. 3) and the growth rate (Ks) is actually smaller (except for pistil width) in CL florets (Table 3), an earlier onset time is the only feasible explanation.
Growth in involucre width of fCL capitula, which recorded the earliest onset time and produced the smallest number of florets, is negatively affected from the earliest stages of development onwards. This would explain why their shape (narrower at the same size) is never attained by CH florets at anther dehiscence.
The involucre size:shape ratio in iCL capitula, with a slightly later onset time and a larger number of florets than fCL, is scarcely affected relative to CH (Fig. 3). The minor size/shape dissociation observed for these capitula with respect to the CH development pattern can be accounted for by a growth rate slightly greater in terms of length and slightly smaller in terms of width (Table 3).
Corolla, androecium, and gynoecium
As already indicated, when anther morphology was clearly distinguishable, the size and shape of floret characters were similar for the three floral types (CH, iCL, and fCL). Thenceforth, however, differences in both size and shape between CL (iCL and fCL) and CH florets start to become apparent. At pmc meiosis, and especially at anther dehiscence, CL florets are smaller than CH florets, and the corolla, androecium, and gynoecium are paedomorphic with respect to CH (Fig. 3).
Analysis of the possible changes in control parameters (, ß, and K) for floral characters reveals an equally complex situation. Relative to CH florets, there is a decrease in ß due to early offset in CL florets, together with earlier onset of floral development () (Fig. 3).
Unlike the involucre, however, the growth rate (Ks) both in length and width of the corolla, androecium, and gynoecium is significantly lower in CL florets (iCL and fCL), with the exception of pistil width.
A proportionately greater decrease in the growth rate (Ks) for width relative to length in corolla and androecium (Table 3) may partially account for the fact that at similar sizes these whorls are narrower in CL capitula. This gives rise to the dissociation of size and shape evident in the clock models (Fig. 3) for corolla and androecium. The complexity of development, with varying growth rates, makes it difficult to determine to what extent there is a modification of the rate of change in shape (K) and to ascertain the involvement of any additional heterochronic process.
Although overall gynoecium growth (ovary, style, and stigma) was monitored, it should be stressed that differences in growth in length between CL and CH florets are mainly due to style development. In CL florets, the style generally attains a length of 1.5–2 mm, while in CH it reaches 10–12 mm (Porras, 1998
). Differences in ovary and stigma length are relatively minor, final ovary size in fact being very similar for all floret types. Therefore, decrease in the growth rate (Ks) of gynoecium length in CL is due to a less developed style with stigmatic lobes remaining at the height of the anther base in order to facilitate self-fertilization. Gynoecium width development was also characterized by ovary growth; as indicated earlier, ovary size was similar in all capitula. Since less time is invested in floral development (decrease in ß), there must be an increase in the growth rate (Ks) of this organ relative to CH in order for that size to be reached (Table 3).
The changes occurring in CL (iCL and fCL) capitula relative to CH can be summarized as follows. (1) Modification of the ontogenic trajectory of CL capitula starts in the earliest stages of development with early onset time () of bud formation, coinciding with a reduction in the number of florets produced. (2) Offset time (ß) occurs earlier due to a reduction in the time taken to reach sexual maturity. (3) The growth rate (Ks) in involucre length increases slightly, although that of width decreases. Involucre shape in fCL is different, and new relative to CH. (4) Only the ovary retains a size and shape similar to those of CH florets. An increase in Ks is necessary in order for this to occur. (5) Other whorls (corolla and androecium) are smaller, with a slower growth rate in terms of both length and width. Style and stigma are also smaller.
The literature contains no studies of the evolution of cleistogamy in Asteraceae, in which the capitulum is the reproductive unit. This is therefore the first attempt to determine, at least in C. melitensis, the evolutionary changes taking place in structures such as the involucre or the set of flowers contained within it, giving rise to small cleistogamous capitula.
Although this entails greater complexity with respect to the study of cleistogamous flowers, many of the changes taking place in CL capitula are similar to those reported in cleistogamous species of other families; the literature contains a number of references to precocious sexual maturation in CL flowers (e.g., Harlan, 1945
; Langer and Wilson, 1965
; Schemske, 1978
; Lord, 1979
). This finding is reported to a greater or lesser degree in cleistogamous species such as Astragalus cymbicarpos (Gallardo, Domínguez, and Muñoz, 1993
), Collomia grandiflora (Minter and Lord, 1983
; Lord and Hill, 1987
; Lord, Eckard, and Crone, 1989
), Lamium amplexicaule (Lord, 1979, 1980, 1982, 1984
), and Viola odorata (Mayers and Lord, 1983a, b
), highlighting its progenetic origin.
As in C. melitensis, additional heterochronic processes are reported, which in some cases even lead to the development of new forms, such as the corolla in Lamium amplexicaule (Lord, 1982
) or the anthers in Collomia grandiflora (Lord, Eckard, and Crone, 1989
). A study of Astragalus cymbicarpos (Gallardo, Domínguez, and Muñoz, 1993
) also reports dissociation of size and shape for three whorls (corolla, androecium, and gynoecium), although in that case shape lags behind size—a situation not observed here in the corolla and androecium of C. melitensis.
Adaptive significance of cleistogamy
What, finally, is the adaptive purpose of all these modifications in floral development that ultimately give rise to cleistogamy? Cleistogamy is widespread in angiosperms, and affects various species belonging to unrelated families (Lord, 1981
), which, however, share a common feature: they tend to be associated with unpredictable, heterogeneous habitats (Schemske, 1978
; Wilken, 1982
). The unpredictability of C. melitensis habitat has been reported elsewhere (Porras, 1998
). In adverse ecological conditions, a reduction of the time invested in floral development and an increase in pollination–fertilization efficiency, which in turn allows the energy cost of reproduction to be reduced, afford obvious advantages for the plants concerned. The heterochronic processes taking place in the cleistogamous capitula of C. melitensis and in cleistogamous species in general may be a response to such requirements. The early onset of floral development and the precocious sexual maturation of flowers in CL capitula imply a considerable reduction in the time taken for each capitulum to develop.
Rapid maturation [via progenesis (hypomorphosis)] may therefore be the character selected, since it affords an immediate ecological advantage (Gould,1977
). Juvenile morphology might therefore be a consequence linked to that selection process.
Other authors, including Arroyo (1973)
, Lloyd (1965)
, and Guerrant (1988)
, associate the reduction in the time required for flower and seed production with the evolution of autogamous species.
Although the reduction in energy cost implied by the formation of CL rather than CH capitula has not been addressed, it is likely to be substantial, given the reduction in size of all floral structures except the ovary (Porras, 1998
). This size reduction may in turn be associated with a change in the breeding system. The move towards cleistogamy enables a reduction in size, particularly of the corolla and the androecium (Lord, 1981
), since the pollination mechanism is more efficient (Porras, 1998
). This energy saving could therefore also represent an advantage for the plant in habitats where availability of resources may become a factor limiting development.
FOOTNOTES
LITERATURE CITED
Alberch, P., S. J. Gould, G. F. Oster, and D. B. Wake. 1979 Size and shape in ontogeny and phylogeny. Paleobiology 5: 296–317.[Abstract]
Arroyo, M. T. K. 1973 Chiasma frecuency evidence on the evolution of autogamy in Limnanthes floccosa (Limnanthaceae). Evolution 27: 679–688.[CrossRef][ISI]
Darwin, C. 1877 The different forms of flowers on plants of the same species. Appleton, New York, New York, USA.
Gallardo, R., E. Domínguez, and J. M. Muñoz. 1993 The heterochronic origin of the cleistogamous flower in Astragalus cymbicarpos (Fabaceae). American Journal of Botany 80: 814–823.[CrossRef][ISI]
Gould, S. J. 1966 Allometry and size in ontogeny and phylogeny. Botanical Review 41: 587–640.
———. 1977 Ontogeny and phylogeny. Belknap Press of Harvard University Press, Cambridge, UK.
Guerrant, E. O. 1982 Neotenic evolution of Delphinium nudicaule (Ranunculaceae): a hummingbird-pollinated larkspur. Evolution 36: 699–712.[CrossRef][ISI]
———. 1988 Heterochrony in plants: the intersection of evolution, ecology, and ontogeny. In M. L. McKinney [ed.], Heterochrony in evolution: a multidisciplinary approach, 110–130. Plenum, New York, New York, USA.
Harlan, J. R. 1945 Cleistogamy and chasmogamy in Bromus carinatus Hook. & Arn. American Journal of Botany 32: 66–72.[CrossRef][ISI]
Huxley, J. S. 1932 Problems of relative growth. Methuen, London, UK.
Johansen, D. A. 1940 Plant microtechnique. McGraw-Hill, New York, New York, USA.
Jones, C. S. 1992 Comparative ontogeny of a wild cucurbit and its derived cultivar. Evolution 46: 1827–1847.[CrossRef][ISI]
———. 1993 Heterochrony and heteroblastic leaf development in two subspecies of Cucurbita argirosperma (Cucurbitaceae). American Journal of Botany 80: 778–795.[CrossRef][ISI]
Kampny, C. M., T. A. Dickinson, and N. G. Dengler. 1994 Quantitative floral development in Pseudolysimachion (Scrophulariaceae): intraspecific variation and comparison with Veronica and Veronicastrum. American Journal of Botany 81: 1343–1353.[CrossRef][ISI]
Kuhry, B., and L. F. Marcus. 1977 Bivariate linear models in biometry. Systematic Zoology 26: 201–209.[CrossRef]
Langer, R. H. M., and D. Wilson. 1965 Environmental control of cleistogamy in prairie grass (Bromus unioloides H. B. K.). New Phytologist 64: 80–85.[CrossRef][ISI]
Lloyd, D. G. 1965 Evolution of self-compatibility and racial differentiation in Leavenworthia (Cruciferae). Contributions from the Gray Herbarium 195: 3–134.
Lord, E. M. 1979 The development of cleistogamous and chasmogamous flowers in Lamium amplexicaule L. (Labiatae): an example of heteroblastic inflorescence development. Botanical Gazette 140: 39–50.[CrossRef]
———. 1980 An anatomical basis for the divergent floral forms in the cleistogamous species Lamium amplexicaule L. (Labiateae). American Journal of Botany 67: 1430–1441.[CrossRef][ISI]
———. 1981 Cleistogamy: a tool for the study of floral morphogenesis, function and evolution. Botanical Review 47: 421–449.[ISI]
———. 1982 Floral morphogenesis in L. amplexicaule L. (Labiateae) with a model for the evolution of the cleistogamous flower. Botanical Gazette 143: 63–72.[CrossRef]
———. 1984 Cleistogamy: a comparative study of intraspecific floral variation. In R. A. White and W. C. Dickison [eds.], Contemporary problems in plant anatomy, 451–494. Academic Press, Orlando, Florida, USA.
———, K. J. Eckard, and W. Crone. 1989 Development of the dimorphic anthers in Collomia grandiflora: evidence for heterochrony in the evolution of the cleistogamous anther. Journal of Evolutionary Biology 2: 81–93.
———, and J. P. Hill. 1987 Evidence for heterochrony in the evolution of plant form. In R. A. Raff and E. C. Raff [eds.], Development as an evolutionary process, 47–70. Alan R. Liss, New York, New York, USA.
Mayers, A. M., and E. M. Lord. 1983a Comparative flower development in the cleistogamous species Viola odorata. I. A growth rate studio. American Journal of Botany 70: 1548–1555.[CrossRef][ISI]
———, and ———. 1983b Comparative flower development in the cleistogamous species Viola odorata. II. An organographic study. American Journal of Botany 70: 1556–1563.[CrossRef][ISI]
McKinney, M. L. 1988 Heterochrony in evolution: a multidisciplinary approach. Plenum, New York, New York, USA.
———, and K. J. McNamara. 1991 Heterochrony: the evolution of ontogeny. Plenum, New York, New York, USA.
McLellan, T. 1993 The roles of heterochrony and heteroblasty in the diversification of leaf shapes in Begonia dregei (Begoniaceae). American Journal of Botany 80: 796–804.[CrossRef][ISI]
Minter, T. C., and E. M. Lord. 1983 A comparison of cleistogamous and chasmogamous floral development in Collomia grandiflora Dougl. ex Lind (Polemoniaceae). American Journal of Botany 70: 1499–1508.[CrossRef][ISI]
Porras, R. 1998 Biología reproductiva y evolutiva en Centaurea melitensis L. (Asteraceae). Ph.D. dissertation. Universidad de Córdoba, Spain.
Reilly, S. M., E. O. Willey, and D. J. Meinhardt. 1997 An integrative approach to heterochrony: the distinction between interspecific and intraspecific phenomena. Biological Journal of the Linnean Society 60: 119–143.[CrossRef]
SAS. 1988 SAS procedures guide, release 6.03 edition. SAS Institute, Cary, North Carolina, USA.
Schemske, D. W. 1978 Evolution of reproductive characteristics in Impatiens (Balsaminaceae): the significance of cleistogamy and chasmogamy. Ecology 59: 596–613.[CrossRef][ISI]
Wilken, D. H. 1982 The balance between chasmogamy and cleistogamy in Collomia grandiflora (Polemoniaceae). American Journal of Botany 69: 1326–1333.[CrossRef][ISI]
Zopfi, H. J. 1995 Life history variation and intraspecific heterochrony in Rhinanthus glacialis (Scrophulariaceae). Plant Systematics and Evolution 198: 209–233.[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
