| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Seed Biology, Dep. of Crop and Soil Sciences, Cornell Univ. Agric. Exp. Stn., 617 Bradfield Hall, Cornell Univ., Ithaca, NY 14853-1901
* Corresponding author (rlo1{at}cornell.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
10%) was not a major factor limiting seed set. Eighty percent of the flowers failed to initiate seeds even under controlled growing conditions when hand pollinated. Approximately 20% of megagametophytes appeared to be defective at anthesis. The proportion of ovules containing a proembryo at 24 h after pollination (HAP) paralleled the proportion initiating seeds, indicating that fertilization is a limiting process. Pollen tube remnants were observed near the micropyle in 65% of the unfertilized ovules at 24 HAP. Lack of fertilization accounts for most of the reduction in seed set, appears to be caused by failure of the pollen tube to grow into the micropyle, and increases with plant age at pollination. Flowers that form late in the plant cycle contribute little to seed set and may be ignored in yield management decisions.
Abbreviations: HAP, hours after pollination • WAS, weeks after seeding
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Reduced seed number can be caused by: (i) insufficient pollination, (ii) sexual selection of male and female gametophytes, and (iii) resource limitation of the maternal plant (Bawa and Webb, 1984). The male component of fitness has been assessed in Campsis radicans (Bertin, 1982), Calathea ovandensis (Horvitz and Schemske, 1988), Cassia fasiculata (Lee and Bazzaz, 1982), Agave mckelveyana (Sutherland and Delph, 1984), and Lithopsermum caroliniense (Weller, 1980). In some cases the evidence indicates that excess flowers contribute to male fitness by acting as pollen donors (Agave), while in other cases (Cassia, Lithospermum, Calathea, Campsis) it does not. Sutherland (1987) pooled published observations of 445 species to test hypotheses related to male fitness. He concluded that excess flowers that fail to form mature fruits do not contribute to female fitness (fruit or seed set) because they abort regardless of their pollination history. They may act as pollen donors, however, and thus contribute to male fitness. Conditions for pollen competition exist in nature, and pollen competition may enhance offspring vigor (Winsor et al., 2000). With multiple donor pollinations in buckwheat, Björkman et al. (1995b) observed certain donors to be favored, probably through speed of pollen tube growth, but donors had equal success in single donor pollinations. Further, greater resource allocation to seeds did not reduce the quality of the pollen.
In regard to the female component of fitness (Stephenson, 1981), the developing fruit or seed may suppress the initiation or development of subsequent seeds by competing for the maternal resources allocated to developing seeds, even within an inflorescence (Medrano et al., 2000). This may be a general mechanism whereby plants regulate reproductive output. Lloyd (1980) and Lloyd et al. (1980) proposed that many plants perform what is, in effect, a cost-benefit analysis to determine how best to partition maternal resources among offspring. The earliest opening flowers within an inflorescence are more likely to set seeds than flowers that open later (Medrano et al., 2000).
Selective abortion reduces seed number in some wild species (Willson and Price, 1980; Bawa and Webb, 1984; Guth and Weller, 1986; Mikesell, 1988) and some domestic species (Cooper et al., 1937; McGregor, 1981; Nakamura, 1986; Furukawa and Bukovac, 1989; Duthion and Pigeaire, 1991; Franz et al., 1991). Abortion has often been reported indirectly as the number of ovules minus the number of mature seeds, and therefore, the point at which abortion occurs is not certain.
In distinguishing between prefertilization abortion of flowers and postfertilization abortion of fruits and seeds, Bawa and Webb (1984) emphasized that prefertilization abortion is a misnomer because zygote and endosperm must be present before they can abort. The term, however, pinpoints a potentially significant limitation to seed set, the number of fertile flowers. If the primary limitation to seed set can be assigned to pre- or post-pollination events, the mechanisms responsible can be more likely assigned to sexual selection in the former case, or to some other cause, such as resource limitation, in the latter.
Buckwheat is advantageous for this study because it is an hermaphroditic, self-incompatible plant (Stevens, 1912b; Morris, 1952) that contains one ovule per ovary (Stevens, 1912a; Mahony, 1935; Obendorf et al., 1993), thereby simplifying assessment of abortion. In the greenhouse, flowers are open and turgid for 1 d when they are pollinated, and the cycle from seed to seed is complete in 12 wk (preliminary observations). A healthy buckwheat plant may produce 4000 flowers, of which 1% result in mature seed (preliminary observations).
Buckwheat flowers are incomplete (lacking a corolla), perfect, and heterostylous (Stevens, 1912b; Mahony, 1935; Morris, 1952). Half of the plants have pin-type flowers with long styles and short filaments, and half of the plants have thrum-type flowers with short styles and long filaments. Flowers within each type are self- and cross-incompatible. Only legitimate cross-pollination, pin by thrum or thrum by pin, results in fertilization in most cultivars (Morris, 1952). Under field conditions, buckwheat normally is pollinated by insects, mainly (>95%) honey bees (Björkman, 1995c), resulting in 15 legitimate pollen grains delivered to each flower type (Björkman, 1995b). Buckwheat flowers produce a mature nondehiscent fruit (achene) (Stevens, 1912a; Obendorf et al., 1993) which herein is referred to as a seed. Factors reported to have altered seed set and yield in buckwheat have been reviewed (Adachi, 1990; Namai, 1990, 1991; Björkman, 1995a,b,c; Björkman et al., 1995b).
The objectives of the present work were to describe the temporal pattern of seed set in buckwheat when grown under controlled greenhouse conditions, and to evaluate whether seed set is limited by post-zygotic events such as abortion or resource limitation, or by pre-zygotic events such as flower abortion or abnormal megagametophyte development.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4 WAS. Ten groups of 10 plants, five thrum- and five pin-type, were hand pollinated by legitimate cross-pollination until the plants were senescent, typically 12 to 13 WAS. Pollen was collected from the appropriate flower type with a camel's hair brush and applied to the stigmas of all receptive flowers in a selected, tagged raceme. Pollination, in the absence of hand pollination, was kept to a minimum by separating plants such that there was little contact between them. There were no natural pollinators in the greenhouse, and the rate of pollination in the absence of hand pollination was nil. Four or five racemes were tagged on each plant in a cohort (a group of 10 plants with the same seeding date). All open and apparently normal, nectar-producing flowers on each tagged-raceme were pollinated on a given day each week until the plants were showing clear signs of senescence, usually 12 to 13 WAS. Approximately three-fourths of the tagged inflorescences were axillary. Terminal clusters of inflorescences were tagged after individual racemes separated enough to be distinguishable. To test hypotheses related to resource allocation and/or selective abortion, pollen (13 grains per stigma) was applied in such a manner that seed set would not be limited by lack of pollen.
Data Collection and Analysis
The number of flowers that were open and judged to be receptive at the time of pollination was recorded. The number of pollinations resulting in seed set on each raceme was determined at 1 wk after pollination by observing the number of new seeds present with a fully-elongated, green pericarp (Obendorf et al., 1993). Ovaries which elongated but failed to fill were considered aborted. At 7 d after pollination, aborted ovaries typically are discolored and reduced in axial and radial expansion when compared with set seeds. To determine the effect of seed load on seed set and abortion, subsequent newly opened flowers were pollinated during the 6- to 8-wk flowering phase. More than 4000 observations were entered into a relational database for analysis. Ten replicates of a group of 10 plants were sampled for the data points in Fig. 1
. Two replicates of each age group were taken for data in Table 1, and three replicates for Tables 2 and 3. Proportional data were treated with the arcsine transformation before statistical evaluation. Results (mean and SD or SE of the mean), reported herein as proportions, were back transformed from arcsine values.
|
|
|
|
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Effectiveness of Hand Pollination
The mean number of pollen grains per stigma after hand pollination was 13.5 (SD = 9.2, range = 0-38, N = 152 stigmas) or 40 pollen grains per ovule (each ovary has three stigmas and one ovule). This number is probably a slight underestimate since part of the stigma was obscured when scoring. The mean number of pollen grains on each stigma was 16.0 for pin-type flowers (Npin = 91, SD = 9.8) and 9.9 for thrum-type flowers (Nthrum = 61, SD = 6.5) after hand pollination. There was no significant difference between pin- and thrum-type flowers in the proportion that set seed using a two-tailed t-test (
= 0. 01, Npin = 4882 and Nthrum = 4608). Also, there was no significant difference between pin- and thrum-type flowers that set seed following pollination at successive WAS. Therefore, data for both flower types were pooled.
Seed Set in Relation to Plant Age at Pollination
Buckwheat seedlings grew very rapidly and began to flower at 4 WAS. Seed set, after hand pollination, reached a maximum at 5 WAS and declined as the plants aged (Fig. 1A). To determine the effect of previously set seed on a raceme on the ability of subsequent flowers to initiate seeds, the proportion of flowers that set seed on all racemes was compared with the proportion of flowers that set seed when one or more seeds were already present on the same raceme. Except for Week 12, there was no significant decrease in the proportion of flowers that set seed when one or more seeds were already present on the same raceme (Fig. 1A).
To test the effect of seed set on subsequent flower production, the number of open, nectar-secreting flowers per raceme when no seeds were present was compared each week with the case where one or more seeds were present on the raceme. The average number of flowers on an individual raceme was maximum at 5 WAS and declined gradually as the plants aged (Fig. 1B). If one or more seeds was present on a raceme, the average number of subsequent flowers was not significantly different from a raceme bearing no seeds on plants of the same age.
To determine the effect of abortion on seed set, the proportion of seeds aborted as a function of plant age at pollination was observed. Abortion was defined as the cessation of growth of seeds that apparently had been initiated and had developed for 2 to 6 d, as judged by length of ovary (Obendorf et al., 1993), degree of discoloration, and increase in volume. There was no significant increase in the number of seeds that aborted when two or more seeds were on the raceme at the time of pollination (Fig. 1C). At 13 WAS, the proportion of seed aborting increased to 1.8. This high ratio represented abortion of some immature seeds at more advanced stages of seed development during the period of gradual senescence of the whole plant.
The proportion of flowers resulting in seed set was compared with the proportion when corrected for losses due to abortion (Fig. 1D). New flowers continued to form for several weeks into the period of gradual plant senescence in the greenhouse (Fig. 1B), but seed initiation at this time (10–12 WAS) was low, and virtually all seeds that initiated growth were aborted.
Seed set among individual plants growing in controlled greenhouse conditions was highly variable. A coefficient of variability (CV = 100 x SD/mean) was routinely calculated for the proportion of flowers that set seed each week; the CV averaged 110% (range = 58-155%, N = 13 wk). Of 414 racemes examined on 102 plants, the mean number of seeds per raceme, resulting from pollination once each week, was 2.9 with a variance of 0.8 during the first 4 wk of flower production, that is, Weeks 5 to 8.
Megagametophyte Development
As a first approximation, low seed set in buckwheat could not be attributed to either abortion or competitive effects, but instead to lack of seed initiation. Preliminary microscopic examination of methacrylate–embedded ovaries (not shown) indicated that few (3 of 20) contained a complete megagametophyte at anthesis. To examine seed initiation processes quantitatively, cleared ovules from plants of increasing age were examined at anthesis for the presence and condition of the megagametophyte. Ovaries were harvested from plants at 4 to 9 WAS, fixed, cleared, and examined using differential interference contrast optics for the presence and condition of megagametophyte structures. The criteria for scoring components were straightforward. If the membrane-bound structures of the megagametophyte were intact, the megagametophyte was scored as viable (Fig. 2A–C)
. When the membrane system of the megagametophyte was not intact, or if either the egg or central cell was disorganized, the megagametophyte was scored as non-viable (Fig. 2D). It was necessary to change the plane of focus to visualize all cells of the megagametophyte; the egg was frequently obscured by being below or above the plane of a synergid (Fig. 2C). The proportion of flowers that contained visibly defective egg sacs ranged from 0.12 to 0.26 (Table 1). There was no significant difference (Tukey's HSD0.05) among WAS in the proportion of flowers with nonviable megagametophytes. When pooled across WAS, the average proportion of flowers containing megagametophyte components (central cell, egg cell, or synergids) that were judged to be abnormal was 0.192 (SD = 0.167, N = 663) (Table 1). This is a conservative estimate of nonviable megagametophytes because, if there was any doubt concerning the appearance of a megagametophyte component, it was scored as viable.
|
|
0.10 and did not vary significantly across all ages of plants. The proportion of ovules that were judged not to be fertilized at 24 HAP increased as plants aged (Table 2).
Pollen Tube Remnants
In incompletely cleared samples, the degenerated remains of pollen tubes could be distinguished in the stylar canals to the region of the micropyle (Fig. 3D). The appearance of the degenerated remains of pollen tubes was seldom as clear as in Fig. 3D. If the discolorized material was present at the base of the stylar canal and/or near the micropyle, it was judged to be pollen tube remnants. Visible evidence of pollen tube growth to but not into the micropyle was observed in 65% (111 of 171) of the cases judged not to be fertilized at 24 HAP (Table 3). Visible evidence of pollen tube remnants was present in 49% (40 of 81) of fertilized ovaries.
Graphical Summary
A graphical summary of results is presented in Fig. 4
using data from Tables 1 and 2. At 4 to 9 WAS, 20% of the ovules had defective egg sacs and were female sterile, and 80% of the ovules had normal egg sacs. At 5 to 9 WAS, 24 to 55% of the ovules were fertilized, 10% of the embryos aborted, and 7 to 45% formed a proembryo at 24 h corresponding to seed set. A large proportion, 45 to 76% of ovules, were not fertilized, including 20% with defective components in the egg sac (Fig. 4). After 2 to 3 wk of flowering (6–7 WAS), lack of fertilization appeared to be a major factor limiting seed set in buckwheat.
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4 weeks with 14-h photoperiods during seed set, and harvest occurs soon after the fall equinox. The Mancan cultivar was relatively insensitive to photoperiod; first flowers appeared at 4 WAS in the greenhouse at various times of the year, even when natural photoperiods exceeded 14 h. By contrast, photoperiod-sensitive Japanese cultivars exhibited delayed flowering at 14-h and 18-h photoperiods compared with 10-h photoperiods (Lachmann and Adachi, 1990). The experimental conditions used in our study were similar to environments typical of commercial buckwheat production in New York State, and the patterns for flowering and seed set replicated those observed in the field (Björkman et al., 1995a). The effective period of seed set for Mancan buckwheat under greenhouse conditions was confined to the first 5 wk of flowering. Senescence became evident at 9 to 10 WAS, as judged by loss of chlorophyll in leaves and cessation of vegetative growth. Although flower production continued during senescence, seed set essentially ceased after 10 WAS due to the low number of seeds initiating growth and a high frequency of abortion.
Experiments testing limitation of seed set by availability or allocation of resources usually require manipulation of resources to determine if there is an increase or decrease in seed set. Initial observations of greenhouse-grown plants showed that seed set in Mancan buckwheat was low and variable among individual plants (3–52% of flowers pollinated), even under controlled conditions. This observation provides evidence that low seed set was not due to resource limitation, but more likely was due to genetic determinants. Competition for resources (and effects of resource limitation), however, may be manifest between developing seeds on the same racemes, even on plants in controlled environments. Three hypotheses concerning resource limitation of seed set were tested at the level of the individual nutritional units (racemes). The hypothesis that previously set seed could reduce the number of subsequently formed seeds on the same raceme was rejected (Fig. 1A). The hypothesis that the number of flowers available for fertilization on a raceme could be reduced by the presence of developing seeds was also rejected (Fig. 1B). The hypothesis that seed abortion is increased when developing seeds are already present on the same raceme was rejected (Fig. 1C).
Namai (1990) observed 40% seed set with one, 70% with three or five, and 80 to 90% with 10 or more compatible pollen grains per stigma (three stigmas per flower). Björkman (1995a) observed universal penetration of egg sacs by pollen tubes with 10 or more pollen grains per flower, but seed set increased with pollen loads up to 30 grains per flower. In the present experiments, seed set probably was not limited by pollen availability because an average of 13.5 pollen grains were applied per stigma (40 per flower) by legitimate hand pollination.
There is evidence that implicates prefertilization abortion as a factor in the regulation of seed production in many species (Herr, 1971; Guth and Weller, 1986; Franz and Jolliff, 1989). Guan and Adachi (1992)(1994) suggested that defective embryo sacs and proembryo abortion are major causes for failure of fertilization, which result in low and unstable seed set, especially at high temperatures. In an effort to identify possible causes for low rates of seed initiation in greenhouse-grown buckwheat, megagametophyte structure was surveyed across the time course of maximum reproductive activity (Weeks 5 through 9). About 20% of anthesis stage buckwheat flowers contained nonfunctional megagametophytes, as judged by gross visual examination of their components. Additional lesions could render megagametophytes nonfunctional, but these may not be visually manifested in the present study. Therefore, at least 20% of flowers were effectively female sterile on Mancan buckwheat grown in a greenhouse.
In an effort to determine more precisely whether the apparent lack of fertilization of flowers (ovules) was due to lack of delivery of male gametes or to abortion early in development, cleared ovules were examined at 24 HAP. Even when abortion was defined as the loss of proembryos as early as 24 HAP, abortion accounted for a relatively small (10%) and consistent proportion of lost seed set. Since the proportion of flowers containing recognizable proembryos at 24 HAP decreased with increasing age of the maternal plant at pollination, the temporal pattern of the proportion of flowers containing proembryos is in general agreement with the temporal pattern of overall seed set. These two findings support the preliminary appraisal that embryo abortion does not play a major role in the limitation of seed set in Mancan buckwheat under our growing conditions.
The temporal pattern of ovules that showed no evidence of having been fertilized gives further support to the assessment that lack of seed initiation was the major cause of low seed set. Following the peak period of seed set at 5 wk, set seed declined with plant age to 10 wk, while the proportion of flowers that showed no signs of fertilization increased from 45 to 76% in one experiment (Table 2) and 52 to 91% in another experiment (Table 3). Thus, at this finer resolution, the primary limitation to seed set still appears to be lack of embryo initiation.
Late in the process of scoring cleared ovules at 24 HAP, it became evident, in some less-than-optimally-cleared tissue samples, that the remnants of pollen tubes could frequently be observed in the basal stylar canal near the micropyle. This incidental information is limited but of considerable interest as stochastic information. It is limited because the use of cleared tissue is not the best way to get information on the progress of pollen tube growth. The preservation of pollen tube remnants was of low quality and quite variable as indicated by the fact that a high proportion of ovules scored as fertilized had no evidence of a degenerated pollen tube in the micropyle. It follows that the large proportion (65%) of unfertilized ovules that had visible remnants of pollen tubes, is a conservative estimate of the incidence of pollen tube penetration to the micropyle. It is probably not a coincidence that a large proportion of pollen tubes grow the length of the style only to stop in a well-defined region at the entrance to the micropyle. Of the 81 ovules containing proembryos, 40 also had degenerated pollen tubes in the micropylar region. This is consistent with the idea that there may be a mechanism to actively exclude pollen tube penetration once one has penetrated the micropyle. During the first 2 wk of flowering, Björkman (1995a) noted that all egg sacs were penetrated by pollen tubes, but at later stages of flowering the frequency of fertilization appears to be much lower (Fig. 4). Mahony (1935) described the fertilization process in detail. Morris (1952) recorded 33 to 40% normal seed set plus an additional 30% aborted seeds presumably with early flowers in compatible crosses. In field experiments, Björkman et al. (1995a) also observed 25 to 65% seed set within 2 wk of first flower and a rapid decline in seed set with later flowers. Greater reproductive success, that is, seed set, also occurs in the earliest opening flowers within an inflorescence in other species (Medrano et al., 2000).
In summary, seed set was highly dependent on plant age at the time of pollination. Although buckwheat plants produce flowers continuously from 4 WAS to plant senescence at 12 to 13 wk (even within an individual raceme), most of the seeds are set after pollination of flowers appearing between 5 and 7 WAS when the seed abortion rate was low. The high rate of seed abortion between 9 and 12 wk had little effect on seed set because few seeds initiated development during this period. Seed set was not limited by the availability of pollen in this study or the absence of pollen tube growth in the stylar canal. Because 20% of the egg sacs were abnormal and 10% had aborted proembryos, defective egg sacs and proembryo abortion were not major causes of low seed set in the present work. Lack of fertilization appears to be a major cause of low seed set, especially after 2 to 3 wk of flowering. Failure of the pollen tube to penetrate the micropyle apparently prevented fertilization and seed set. The cause of the failure is unknown. Shrinkage of ovules within ovaries was evident in unfertilized ovules and those with aborted embryos at 24 HAP. Perhaps misalignment of the micropyle relative to convergence of the three stylar canals prevented pollen tube penetration of the micropyle.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication August 2, 2000.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Horbowicz and R. L. Obendorf Fagopyritol Accumulation and Germination of Buckwheat Seeds Matured at 15, 22, and 30{degrees}C Crop Sci., May 27, 2005; 45(4): 1264 - 1270. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
