Crop Science 40:1103-1108 (2000)
© 2000 Crop Science Society of America
SEED PHYSIOLOGY, PRODUCTION & TECHNOLOGY
Ripening Habit of Buckwheat
Hideyuki Funatsukia,b,
Wakako Maruyama-Funatsukib,
Kaien Fujinoc and
Masamichi Agatsumab
a Hokkaido Natl. Agric. Exp. Stn., Hitsujigaoka, Sapporo, Hokkaido, 082-8555 Japan
b Upland Agric. Res. Ctr. Hokkaido Natl. Agric. Exp. Stn. Shinsei, Memuro, Hokkaido, 082-0071 Japan
c Fac. Agric. Hokkaido Univ., Hokkaido, 060-8589 Japan
funazki{at}cryo.affrc.go.jp
 |
ABSTRACT
|
|---|
In buckwheat (Fagopyrum esculentum Moench), the diverse ripening times of seeds and the remarkable seed shattering can make it difficult to determine the optimum harvest time. To provide basic information for the improvement of the ripening habit of buckwheat by breeding and cultural methods, we monitored the changes in the numbers of mature (brown) and immature (green) seeds at each raceme of buckwheat plants in the field during the course of seed maturation. A determinate line was used in addition to an indeterminate one (wild type), since the former had previously been reported to display more uniform ripening. Seed weights were also determined to estimate yields. Seed discoloration proceeded from the main stem to branches, and from basal to apical racemes regardless of genotypes or years. Seed shattering progressed in a similar pattern. In contrast to these consistent habits, differences in maturation speed were seen between years. Seed discoloration progressed slightly slower in 1997, while seed shattering began much earlier, compared with that in 1998. Consequently, in 1997, seed shattering began before all seeds had matured in most plants, while in 1998, this phenomenon was observed in fewer plants. Seed yields accordingly increased slowly until reaching maximum levels, followed by rapid decreases in 1997, relative to 1998. Growth habit (determinate vs. indeterminate) appeared to have little influence on ripening. These results are the first to clearly demonstrate that seed shattering can start before all seeds mature in buckwheat, regardless of growth habit. And that the rates of seed ripening and seed shattering are independent and vary among years.
Abbreviations: SDP, seed discoloration proportion • TID, time of initiation of seed discoloration • TIS, time of initiation of seed shattering • TTD, time of termination of seed discoloration
|
INTRODUCTION
|
|---|
BUCKWHEAT IS ONE OF THE MOST widely cultivated crops in the world. It has been used as a material for various foods in the particular cultures of many countries. In addition, the nutritional value of buckwheat has recently been cited (Namai, 1992) along with its physiological mechanism for Al tolerance (Ma et al., 1997). However, the crop has several limitations with respect to its agronomic characteristics, such as low yield and high degree of lodging, which must be improved for the expansion of cropping area and production.
One problem with buckwheat production is the difficulty in determining harvest time, which has been attributed to two characteristics, its indeterminate growth habit and its high degree of seed shattering (Marshall and Pomeranz, 1982). It is thought that mature seeds start to abscise before the later-formed seeds become ripe (Kreft, 1989a). The problem could be overcome by the improvement of seed shattering resistance and uniformity in ripening among seeds.
Genetic analyses of seed shattering have not been reported in buckwheat, and it is not yet known whether there is genetic variation for this character, with the exception of the report on a mutant in which increased seed shattering resistance was observed (Alekseeva et al., 1988). However, the lines developed from this mutant display low fertility and low grain quality (Alekseeva and Malikov, 1992). Thus, it would take a long time to breed elite cultivars using this genetic resource.
Because of the difficulty in using a breeding approach to limit seed shattering, some researchers have attempted to develop cultural methods for minimizing yield losses. Miyamoto (1983) and Gubbels and Campbell (1985) attempted to determine the optimum harvest time for seed yield by sampling buckwheat populations several times during the ripening stage. In both studies, the maturation of buckwheat seed was judged by the color of the seed coat (hull). Gubbels and Campbell (1985) suggested that greatest yield is obtained when 75% of seeds turn brown, but Miyamoto (1983) did not find significant yield difference during the periods when the proportion of mature seeds was between 50 and 95%. In these studies, exact estimates of yield variation with time was hindered by the relatively large experimental errors, which were generated by sampling different populations instead of observing the same plants.
Regarding the improvement of simultaneous ripening, it is not clear whether it is necessary to improve the character at the plant level or to make the population genetically homogeneous. In buckwheat, there is quite a high level of genetic heterogeneity even within a cultivar (Inuyama et al., 1994) because of its allogamous genetic character. In addition, to screen for simultaneous ripening as well as shattering-resistant plants, it is necessary to know when to score plants and which part of a plant to examine (Fukuta et al., 1994). To establish an effective breeding strategy and selection criteria, therefore, it is important to define the developmental differences that exists among racemes on a plant, when seed shattering starts, and how these factors influence yield. There is currently little information on the ripening habit of buckwheat available in the literature in contrast to the intensive studies on its flowering habit (Sugawara, 1973; Asako et al., 1980).
Variants with determinate growth habit have been found in a wide range of buckwheat populations (e.g., Kreft, 1989b; Martinenko and Fesenko, 1989; Funatsuki et al., 1996) since the gene for determinate growth habit in buckwheat was first identified (Fesenko, 1968). This plant type is assumed to display simultaneous ripening (Kreft, 1989a; Adachi, 1991), by analogy with other crops such as soybean [Glycine max (L.) Merr.], in which determinate varieties develop more synchronously than do indeterminate ones (Saito and Hashimoto, 1980). If determinate buckwheat lines have such an advantage, they will be readily used for breeding, since determinate lines have been developed with various genetic backgrounds and proved agronomically promising (e.g., Kreft, 1989b; Martinenko and Fesenko, 1989; Funatsuki et al., 1997, unpublished). Kreft (1989b) reported the simultaneous ripening as well as increased shattering resistance in a determinate line, while Fesenko and Martinenko (1995) reported no apparent improvement for these characters; however, these studies compared lines with quite different genetic backgrounds. Therefore a more precise study needs to be conducted to reveal the potential benefits of the ripening habit of a determinate line.
As a step toward improving the ripening habit of buckwheat, we conducted this study to obtain basic information on seed maturation dynamics. The objectives were (i) to clarify the ripening pattern of buckwheat at the single plant and population levels and (ii) to compare the ripening pattern of an indeterminate vs. a determinate line with similar genetic backgrounds.
|
Materials and methods
|
|---|
Two genotypes of buckwheat were used. Kitawasesoba is the current leading cultivar in the Hokkaido district of Japan (Inuyama et al., 1994) and displays an indeterminate growth habit. Kitawase D is a determinate line developed from a variant of Kitawasesoba (Funatsuki et al., 1996). Both genotypes had similar flowering dates in 1997 and 1998 (17 July 1997 and 13 July 1998).
In 1997 and 1998, experiments were conducted at the research farm of the Hokkaido National Agricultural Experiment Station in Memuro, Hokkaido. The strains were seeded in a soil similar to Cinebar silt loam (medial, mesic Humic Haploxerand) on 9 June 1997 and on 4 June 1998. Fertilizer was applied prior to planting at the levels of 20, 35, and 39 kg ha-1 of N, P, and K, respectively, in 1997. In 1998, the levels were reduced to 6, 11, and 12 kg ha-1 to provide a different growth environment. Weather conditions during the growth period are shown in Table 1 . A rate of 150 seed m-2 was used in rows 2 m long and spaced 60 cm apart.
Plots were arranged in a randomized, complete block design with three replications in 1997 and four in 1998. Observations on randomly selected plants whose height was within ±10 cm of the mean in the plots were made twice a week during the ripening seasons, except that the intervals between observations were 1 wk after 11 September in 1997 and 10 September in 1998. Five plants per plot were used in 1997 and three in 1998. To measure 1000-seed weight, all plants except for lodged or damaged ones were harvested in alternate rows once or twice a week from 18 August to 18 September in 1997 and from 17 August to 24 September in 1998 (Fig. 1)
.
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 1 Relative yield and seed maturation proportions for two buckwheat genotypes during the ripening season. Relative yields, represented as open or filled bars, were calculated as percentages compared with the maximum yields during the ripening period for each genotype. Discoloration proportions were represented as lines with triangles or squares. LSD (0.05) in relative yields were 4.5 for Kitawasesoba in 1997, 4.2 for Kitawase D in 1997, 2.6 for Kitawasesoba in 1998, and 3.2 for Kitawase D in 1998
|
|
Seeds of each raceme on main stems and branches were observed and the numbers of immature (green) and mature (brown) seeds were recorded. Racemes were numbered from the base, while branches were numbered from the top according to their position (Fig. 2 and 3) . Branches 1 and 2 were grouped as B1, while Branches 3 and higher were grouped as B2. Seeds in Racemes 1 to 5 are pooled as were those in the basal or proximal (lower) racemes (-L). Seeds in Racemes 6 or greater were pooled as apical or distal (upper) racemes (-U). Time of initiation of seed discoloration (TID) for a raceme was defined as the day when seed discoloration was first seen in the raceme. Time of termination of seed discoloration (TTD) was taken as the day when the last immature seed in the raceme became brown. Time of initiation of seed shattering (TIS) was recorded as the day when the first mature seed abscised. These parameters were expressed relative to the day the first mature seed was observed on the plant. Statistical analyses on TID, TTD, and TIS were conducted on the data from 10 plants per genotype each year, which were pooled from all replications, and avoiding plants that had lodged or had damaged branches. Since no significant differences for any parameter were found among blocks, each plant was treated as a replication. Seed discoloration proportion (SDP), which was employed as an index for the degree of maturation of a population by Miyamoto (1983) and Gubbels and Campbell (1985), was expressed as [(number of brown seeds)/(number of total seeds)]100. Total number of seeds was recorded as the sum of mature seeds and immature ones that would give rise to grains if harvested (
20 d or more old after flowering). Population yields were calculated from the 1000-seed weight at harvest (excluding damaged or lodged plants from the same plot) and the total seed number.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 3 Grouping of racemes for sampling purposes. MS-L, five basal racemes on the main stem; MS-U, additional apical racemes on the main stem; B1-L, five proximal racemes on primary Branches 1 and 2; B1-U, additional apical racemes on the same branches; B2, racemes on Branches 3, 4, etc.; SB, racemes on secondary branches
|
|
|
Results
|
|---|
Time of seed maturation varied among raceme positions (Table 2) . Seed ripening started in the basal racemes on the main stem (MS-L, Fig. 3) and progressed to the apical racemes (MS-U). Seed discoloration was first seen in the proximal racemes on upper branches (B1-L), which occurred at roughly the same time as in the apical racemes on the main stem (MS-U), and then progressed to distal racemes (B1-U) and lower branches (B2), and finally to secondary branches (SB). The TTD and TIS were much more uniform among raceme groups compared with TID, although the termination of seed discoloration and the initiation of seed shattering proceeded in a similar manner.
Although these general patterns were consistent for all plants, several quantitative differences were found in TID and TIS between years and between genotypes when the data for main stems and upper branches were compared (Tables 2 and 3)
. Although there were relatively small differences in TID, seed discoloration, in general, was delayed in 1997 relative to 1998. This delay probably was caused by the lower temperatures in mid August (Table 1). In TIS, larger differences were observed between years. Plants grown in 1997 started to shed seeds 7 to 10 d earlier than in 1998 regardless of genotype or position of raceme. The TTD was similar in both years. However, in 1997 seed shattering started in most plants when some seeds were still immature, while in 1998 most plants retained all seeds until they became ripe. In addition, there was a tendency for seed shattering on main stems to begin later in the determinate line.
View this table:
[in this window]
[in a new window]
|
Table 3 Analyses of variance for times of seed discoloration initiation (TID), times of seed discoloration termination (TTD), and times of seed shattering initiation (TIS) of proximal racemes on main stem (MS-L) and proximal racemes on upper branches (B1-L)
|
|
The ripening habit for the population reflected that at the plant level (Fig. 4)
. First, mature seeds were formed on basal racemes on main stems, followed by an increase in the number of mature seeds on primary branches and of upper racemes on main branches. Secondary branches in Kitawase D and upper racemes on primary branches in Kitawasesoba contributed to the final increase, although the proportions of SB (Kitawase D) and B1-U (Kitawasesoba) racemes were not large in comparison with other racemes. In 1997, the maximum values for the total number of seeds were obtained between 4 and 8 September in both genotypes, and considerable decreases in seed number were already observed by 11 September. In 1998, more than 90% of the maximum seed number was obtained by 31 August, but large reductions were not observed until 17 September. These results indicate that in 1998 seeds became mature more rapidly and were retained longer on the plants, which is in agreement with the observations at the plant level (Table 2 and 3).
View larger version (61K):
[in this window]
[in a new window]
|
Fig. 4 Distribution of mature seeds at various raceme postions in two buckwheat genotypes during the ripening season. Areas represent basal racemes on main stem (MS-L), apical racemes on main stem (MS-U), proximal racemes on upper primary branches (B1-L), distal racemes on upper primary branches (B1-U), racemes on lower primary branches (B2), and racemes on secondary branches (SB). See also Fig. 3. SB for Kitawasesoba and MS-U and B1-U for Kitawase D are absent since no mature seeds were observed in Kitawasesoba and none of Kitawase D plants bore more than five racemes on the main stem or branches
|
|
The progress of SDP in Kitawasesoba and Kitawase D populations, which is considered an indicator of maturation, are shown in Fig. 1. No clear genotypic difference was observed, especially in 1998, in which the SDP in both genotypes changed synchronously. This result may be explained by the observation that in Kitawase D, seeds on branches, which matured later, constituted a larger proportion of the total seed number, even though most seeds on main stems matured earlier (Fig. 4).
Relative yields of the populations are also shown in Fig. 1. In 1997, the maximum yield was obtained around 1 September and 28 August, when the SDP was 70% for Kitawasesoba and 60% for Kitawase D. Yields subsequently decreased due to increased seed shattering. However, in 1998, more than 95% of relative yield was maintained during the period with 70 to 97% of SDP. These tendencies were consistent in both genotypes.
|
Discussion
|
|---|
The ripening habit of buckwheat was closely associated with its flowering habit, which has been described in detail in previous studies (Sugawara, 1973; Asako et al., 1980). For raceme position, the initiation of seed ripening proceeded in a manner similar to the initiation of flowering. This result reflects the fact that flowers that bloom earlier in a raceme tend to have a higher probability of seed set than those that bloom later on the same raceme (Sugawara, 1973; Asako et al., 1980). Compared with TID, TTD values were much more similar among raceme groups. This may also be related to the flowering habit in which the number of flowers diminishes on the more apical or distal racemes (Sugawara, 1973; Asako et al., 1980). Our study also clearly demonstrates that the difference in developmental state among racemes sometimes results in seed shattering before the completion of seed ripening throughout the plant, which has been presumed by many researchers (Marshall and Pomeranz, 1982; Kreft, 1989a).
Some differences in maturation speed were seen between years. The analyses of TID and TTD in addition to SDP indicate that seed maturation progressed more rapidly in 1998, suggesting that it is affected by environmental conditions such as temperature. On the other hand, seed shattering started later in 1998, suggesting that the two characters associated with maturation do not always coincide. Furthermore, the difference seen in TIS between years suggests that the physiological states of the plant and/or reproductive morphology play important roles in determining TIS, since there was no apparent disadvantage in 1997 in regard to wind and rainfall from late August to early September (Table 1), which could directly cause seed shattering. Furthermore, seeds on main branches tended to remain longer that those on branches, and environments were quite different between years (e.g., the higher fertilizer level and the higher temperatures during the vegetative growth period and lower temperatures during the ripening period in 1997). These findings may provide basic information for the establishment of selection methods in breeding. In addition, they suggest the possibility of retarding seed shattering in buckwheat by improvement of cultural methods. This approach is also important because of the long time required to achieve the introduction of genetic factors conferring shattering resistance to buckwheat. Strengthened shattering resistance should provide an ideal ripening pattern as illustrated in the results in 1998 when maximum yields were obtained at the stages with more than 95% of SDP.
In contrast to a previous work (Kreft, 1989a), our study using lines with similar genetic backgrounds showed little or no advantage for the determinate genotype in terms of uniformity of seed ripening. This may be due to the fact that in determinate plants, more seeds are set on branches, which are developmentally younger than those on the main stem. The characteristic of buckwheat that inflorescence development of flowers on racemes is indeterminate even in determinate plants may also be involved. Although some combinations of determinate growth and certain genetic backgrounds or growth environment might result in simultaneous ripening, our findings suggest that the introduction of determinate growth habit does not necessarily improve the ripening habit of buckwheat. This conclusion is supported by the observation by Fesenko and Martinenko (1996), who found no apparent difference in ripening pattern between a determinate cultivar and other Russian indeterminate ones.
Although a direct comparison is not possible as harvest losses and the influence of damaged plants were not considered, our results are in general agreement with those of the previous studies (Miyamoto, 1983; Gubbels and Campbell, 1985). When SDP was between 50 and 95% (Miyamoto, 1983), yields were estimated to be more than 90% of the maximum for both cultivars in both years, with the exception of the case of Kitawasesoba in 1997, when harvest at 75 of SDP (Gubbels and Campbell, 1985) would have given more stable yields. However, our data demonstrate that the relationship between SDP and yield varies considerably between years and also show that use of the indicators established in the earlier studies (Miyamoto, 1983; Gubbels and Campbell, 1985) may lead to yield loss in some cases. Although Gubbels and Campbell (1985) obtained similar results, they did not point out the possible sources of yield loss. The continuous observation of ripening plants employed in our study instead of harvesting different plots enabled a more accurate estimation of population yields as a function of time. Our results suggest a limited use for SDP as a common indicator to obtain maximum yields, and the absolute requirement of the improvement of seed shattering by breeding or the development of a new cultural method.
In conclusion, we demonstrated that seed shattering can begin before all seeds have ripened, at least in the two genotypes investigated. In addition, the speed of seed ripening and the time of the initiation of seed shattering appear to be independent and vary among years. Based on the two genotypes examined, altering growth habit from indeterminate to determinate would have little influence on the ripening habit of buckwheat. This study is, to our knowledge, the first report on the ripening habit of buckwheat and may provide helpful information for further breeding and cultivation studies.
|
ACKNOWLEDGMENTS
|
|---|
The authors wish to thank the technical assistance of N. Murakami and T. Yamada. Financial support from Hokscitec is also gratefully acknowledged.
Received for publication March 19, 1999.
|
REFERENCES
|
|---|
- Adachi, T. 1991. Recent trends in genetics and breeding of buckwheat. (In Japanese.) Comm. Inst. Genet. Ecol. Tohoku Univ. 15:1–8.
- Alekseeva E.S., Malikov V.G., Falendysh L.G. Green-flower form of buckwheat. Fagopyrum 1988;8:79-82 (In Russian with English summary.).
- Alekseeva, E.S., and V.G. Malikov. 1992. The results of buckwheat green floret form. p. 264–270. In L. Rufa, Z. Mingde, T. Yongru, L. Jianying, and Z. Zongwen (ed.) Proc. 5th Intl. Symp. on Buckwheat. Taiyuan, China. 20–26 Aug. 1992. Agriculture Publishing House, Taiyuan, China.
- Asako Y., Ujihara A., Matano T. Relation between the position of flowers and their flowering of fruiting in common buckwheat. Hokuriku Crop Sci. 1980;15:27-30 (In Japanese.).
- Fesenko N.V. A genetic factor responsible for the determinant type of plants in buckwheat. Genetika 1968;4:165-166 (In Russian with English summary.).
- Fesenko, N.V., and G.E. Martinenko. 1996. Contemporary buckwheat breeding work in Russia. p. 3–5. In T. Matano and A. Ujihara (ed.) Proc. 6th Intl. Symp. on Buckwheat. Ina, Japan. 25–27 Aug. 1995. Shinshu Univ. Press, Nagano, Japan.
- Fukuta Y., Yano M., Kobayashi A. Genetic analysis of shattering-resistant mutant lines induced from an indica rice (Oryza sativa L.) variety, "Nan-jing 11". Breed Sci. 1994;44:325-331.
- Funatsuki H., Suvorova G., Sekimura K. Determinate type variants in Japanese buckwheat lines. Breed Sci. 1997;46:275-277.
- Gubbels G.H., Campbell C.J. Visual indicators for determining the optimum windrowing time for buckwheat. Can. J. Plant Sci. 1985;65:207-209.
- Inuyama S., Honda Y., Furuyama S., Kimura M., Kasano H. The breeding and characteristics of a buckwheat cultivar, "Kitawasesoba". (In Japanese with English summary.) Res. Bull. Hokkaido Natl. Agric. Exp. Stn 1994;159:1-10.
- Kreft, I. 1989a. Ideotype breeding of buckwheat. p. 3–6. In N.V. Fesenko (ed.) Proc. 4th Intl. Symp. on Buckwheat. Orel, Russia. 11–15 July 1989. All-Russian Res. Inst. Legumes and Groat Crops, Orel, Russia.
- Kreft I. Breeding of determinate buckwheat. Fagopyrum 1989;9:57-59 b.
- Ma J.F., Zheng S.J., Hiradate S., Matsumoto H. Detoxifying aluminum with buckwheat. Nature (London) 1997;390:569-570.[Medline]
- Marshall H.G., Pomeranz Y. Buckwheat: description, breeding, production, and utilization. Adv. Cereal Sci. Technol. 1982;5:157-210.
- Martinenko, G.E., and N.V. Fesenko. 1989. Selection of determinant buckwheat varieties. p. 334–342. In N.V. Fesenko (ed.) Proc. 4th Intl. Symp. on Buckwheat. Orel, Russia. 11–15 July 1989. All-Russian Res. Inst. Legumes and Groat Crops, Orel, Russia.
- Miyamoto, H. 1983. Relationships between seed yield, shattering and harvesting time in buckwheat. (In Japanese.). Rep. Hokkaido Branch Jpn. Soc. Breed. Crop Sci. Soc. Jpn. 23:40.
- Namai, H. 1992. Strategies for sustainable conservation and efficient utilization of buckwheat genetic resources in the world. p. 93–104. In IBPGR (ed.) Buckwheat genetic resources in East Asia. Int. Crop Network Series. 6. IBPGR, Rome.
- Saito, M., and K. Hashimoto. 1980. Classification, distribution and agronomic characteristics of cultivars. (In Japanese.) p. 37–62. In M. Saito and T. Okubo (ed.) Ecology and culture technology of soybean. Nosan Gyoson Bunka Kyokai. Tokyo.
- Sugawara K. Study on buckwheat. Ann. Rep. Fac. Edu. Iwate Univ. 1973;33:1-96 (In Japanese.).
This article has been cited by other articles:
|
|
|
|
 
M. Quinet, V. Cawoy, I. Lefevre, F. Van Miegroet, A.-L. Jacquemart, and J.-M. Kinet
Inflorescence structure and control of flowering time and duration by light in buckwheat (Fagopyrum esculentum Moench)
J. Exp. Bot.,
July 1, 2004;
55(402):
1509 - 1517.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
