| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Kobe Univ. Farm, Uzurano, Kasai, 675-2103, Japan
* Corresponding author (kkawano{at}kobe-u.ac.jp)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AT, advanced yield trial • CBB, cassava bacterial blight • CIAT, Centro Internacional de Agricultura Tropical • CIAT/HQ, CIAT headquarters location in Colombia • CIAT/Thai, CIAT Thailand location • ICA, Instituto Colombino Agropecuario • PT, preliminary yield trial • SE, superelongation disease • RPT, replicated yield trial • RT, regional yield trial • SRT, single-row trial • ST, seedling trial
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The CIAT headquartered in Cali, Colombia, initiated a cassava breeding program in the early 1970s with the objectives of improving yield potential and tolerance to diseases and insect pests, and adverse soil and environmental conditions. This program activity was expanded to Asia in the early 1980s in the form of an applied breeding program in close collaboration with national programs. The Department of Agriculture (Thailand) and CIAT established a collaborative cassava breeding program that distributed the advanced breeding materials to many national programs in Asia.
The goal was defined as the establishment of a cassava breeding program with a global perspective that would generate economic benefits targeted to the less privileged in the rural sector. During the whole period of development of the program, the following operating principles or processes were closely adhered to: (i) establishment of breeding methodology; (ii) generation of useful breeding materials; (iii) distribution of advanced breeding materials to national programs; (iv) establishment of competent national cassava breeding programs; (v) development of improved cultivars; and (vi) dissemination of cultivars.
Within these procedures there were three distinct technical phases performed under specific institutional arrangements, each of which was critical to accomplishing our goals: first, germplasm collection and evaluation, which formed the most important part of the basic breeding at CIAT headquarters in Colombia; second, generation of advanced breeding materials in the applied breeding effort in the CIAT/Thai collaborative program with the Department of Agriculture, Thailand; and third, varietal selection and dissemination through CIAT collaboration with national programs. After 30 yr of these activities, many cultivars have been released in many countries, mainly in Asia. The new cultivars are now planted on more than one million hectares (Puspitorini et al., 1998; Rojanaridpiched et al., 1998; Kawano, 2001; Kim et al., 2001; Lin et al., 2001; Mariscal et al., 2001). The economic benefits resulting from the increased productivity is in the order of one billion US dollars. The target population of small farmers in the poorer rural areas of the tropics captured a large proportion of these economic benefits.
These achievements were the result of 30 yr of coherent efforts to use improved cassava production technology to alleviate rural poverty. In this paper, I describe the critical biological and social factors that made this program a success. I am in a unique position to do this because I have been involved with all aspects of the program since its inception. There are quantitative data on yield components at every step of the selection process for the duration of the program. In addition, varietal dissemination in many countries is well documented, and hence the economic benefits of the adoption of improved cultivars can be quantified relatively accurately. The objectives of this paper are to record the progress of an international cassava breeding program across the past 30 yr and identify biological and social factors that made the whole process successful.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Single-row trial. For the initial evaluation of the germplasm collection, single-row entries were established at CIAT/HQ by planting nine 20-cm stem cuttings per accession. The distance was 1 m between plants and 1.4 m between rows. Six and 10 mo after planting, three inside plants per accession were harvested. For the evaluation of hybrid materials thereafter, seven stem cuttings were planted per clone and the five central plants were harvested 11.5 mo after planting. Single-row trials (SRTs) were conducted in the three evaluation sites.
Replicated yield trial. Thirty cuttings of each genotype were planted in 5- by 6-m plots at 1- by 1-m distance on 1-m ridges with two replications. Eleven-and-one-half months after planting, the nine central plants in each plot were harvested, eliminating border effects. Replicated yield trials (RPTs) were conducted in the three evaluation sites.
In all trials, Colombian cultivars MCol 113 from Valle, MCol 22 from Magdalena, and Llanara from Meta were used as controls. In ICA/Carimagua, 100 kg ha-1 of N, 200 kg ha-1 of P2O5, 200 kg ha-1 of K2O, 500 kg ha-1 of dolomitic lime, and 12 kg ha-1 of Zn were applied. Fresh root yield, biomass (fresh total plant weight), harvest index (fresh root weight/biomass), root dry matter content (following the method described by CIAT, 1975, and Kawano et al., 1987) were determined for SRT and RPT in all locations.
Generation of hybrid populations. Hybridization started in 1973 at CIAT/HQ using local cultivars from the three evaluation sites. All cassava cultivars are highly heterozygous, and their genotype is kept intact through successive generations of vegetative propagation. Selected genotypes from the initial germplasm evaluation soon entered the hybridization scheme and selected hybrids from the subsequent evaluations were massively included in this scheme. Pollination was relatively easy and there was no noticeable cross incompatibility; however, inbreeding depression was severe in most cases where selfing was attempted (Kawano, 1980). The majority of hybrids were produced by controlled pollinations and a smaller proportion of hybrids were from open pollinations. Forty- to sixty-thousand F1 seeds (genotypes) from 200 to 400 parental combinations were produced annually.
Evaluation and selection of advanced populations. Stem cuttings from all germplasm accessions were planted in SRT at CIAT/HQ in 1973: 1950 accessions completed the evaluation, and 230 clones were selected. The selected clones were subsequently planted to RPT. These 230 clones were also planted in SRTs at ICA/Caribia and ICA/Carimagua in 1974, and selected clones from these trials were subsequently planted to RPT in each location.
Seedling trials were started in 1973 and the selected F1 plants were planted subsequently to SRT at CIAT/HQ in 1974. Selected clones from this SRT were planted to RPT at CIAT/HQ and SRTs at ICA/Caribia and ICA/Carimagua in 1975. New F1 hybrids were planted to seedling trial (ST) every year thereafter, and the cycle was thus established: hybridization, plant selection at CIAT/HQ, clonal selection at three locations, hybridization using selected genotypes. The whole scheme is a form of large-scale recurrent mass selection. On average, 30000 F1 seeds were sown in ST (average of 75% germination and survival), 3000 were selected to SRT, and 500 were selected to RPT at CIAT/HQ and SRTs at ICA/Caribia and ICA/Carimagua every year after 1976. The mean selection rate between evaluation stages was 10 to 20%.
CIAT/Thailand
Selection Sites
All the early stages of evaluation were conducted at the Rayong Field Crop Research Center, located in the eastern lowland plain of Thailand. Rayong Center represents a typical cassava-growing environment in the lowland tropics in Thailand. In addition to Rayong, advanced yield trials (ATs) were conducted at Khon Kaen Research Center (drier climate) and at Mahasarakarn Experiment Station (higher soil fertility), both located in northeastern Thailand. In addition to these three locations, regional yield trials (RTs) were conducted at Banmai Samrong Experimental Station in Central Thailand, and Kalasin and Roi Et Experiment Station in northeastern Thailand. The six locations represented the variability of major cassava-growing areas in the seasonally dry to semiarid lowland tropics in Asia. The edaphoclimatic data for these locations were given by Kawano et al. (1998).
Selection Scheme
Seedling trial. F1 seeds were sown in plastic bags of 
5-cm diam. and 8-cm depth. Forty-five-day-old seedlings were transplanted at 1-m plant spacing to a field prepared with ridges spaced 1.5 m apart at Rayong.
Single-row trial. Twelve stem cuttings from each of the plants selected from the ST were planted at Rayong in a single row with 1 m between rows and 1 m between plants within rows. The leading local cultivar Rayong 1 was planted in every 20th row as a check entry.
Preliminary yield trial. Fifty stem cuttings from each of the clones selected in the SRT were planted on ridges in a 5- by 10-m plot with a spacing of 1- by 1-m (10000 plants ha-1) at Rayong. The clonal entries were randomized with two replications.
Advanced yield trial. Clones selected from the preliminary yield trial (PT) were planted in the same manner as in the PT but with four replications in each of the three locations.
Regional yield trial. Seven to nine clones selected from the AT were planted in the same manner as in AT at the six locations.
All trials were conducted every year with changing genotypic entries from 1982 to 1998. All the trials were planted in April/May and were harvested 11 mo later. The standard recommended rate of fertilizer (50 kg ha-1 each of N, P2O5, and K2O) was applied one month after planting in all trials. A single plant in the ST, 10 plants per clonal row in the SRT, and 24 central plants free from border effect in each plot in the PT, AT, and RT were harvested to determine root yield, biomass, and harvest index on a fresh weight basis. In the 1982-1983 trial, biomass and harvest index data were not taken. A 5-kg sample of fresh root from each row (SRT) or plot (PT, AT, RT) was taken and used for the determination of root dry matter content by the specific gravity method (CIAT, 1975; Kawano et al., 1987).
Generation of advanced materials. Cassava breeding in Thailand began during the 1970s using locally available materials and a small number of introduced clones from Java, the Virgin Islands, and CIAT/HQ. Introduction of a large number of genotypes in the form of hybrid seeds from CIAT/HQ started in the late 1970s. Selected genotypes from these materials have been systematically incorporated in the hybridization program to produce a new hybrid population each year. Annually, 20000 to 40000 F1 hybrid seeds were produced mainly by controlled pollination.
Evaluation and selection of advanced materials. The evaluation and selection of advanced materials proceeded in a similar manner described for CIAT/HQ. The numbers of annual entries in the ST, SRT, PT, AT, and RT were 12000 to 25000, 1200 to 2400, 100 to 160, 16 to 20, and 7 to 9, (depending on the year), respectively.
Selection experiment. To obtain selection data unbiased by directional selection, a total of 3122 randomly chosen seedlings from 69 families from 41 cross parents of Thai and American origin were transplanted to the 1987-1988 ST at Rayong, aside from the regular breeding population. From the surviving 3047 plants, 62 plants with wide phenotypic variation were advanced as clonal entries to ST and PT without selection. This population facilitated comparisons of the same genotype through ST, SRT, and PT.
Comparison of released cultivars. Six cultivars (Rayong 3, Rayong 60, Rayong 90, Kasetsart 50, Rayong 5, and Rayong 72), which have been selected at Rayong Center (except for Kasetsart 50, whose early stage selection was conducted by the Kasetsart University cassava breeding program located in Sriracha, Thailand) during the past 17 yr, have been officially released by the Department of Agriculture and widely adopted by farmers. These six cultivars and Rayong 1 were simultaneously evaluated in a separate RT for two years (1994-1995 and 1995-1996) to analyze the change in varietal characteristics during the period of release.
Collaboration with Other National Programs
CIAT encouraged and supported national agricultural research institutions in their efforts to establish and strengthen their own cassava breeding program based on materials from CIAT/HQ and CIAT/Thai. Many institutions especially in Asia responded to this and developed breeding programs modeled after the CIAT/Thai scheme.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
25 local cultivars were planted in the ST in 1973, 190 plants selected from this ST were evaluated in the SRT in 1974, and 37 selected clones from this SRT were evaluated in the RPT in 1975 together with 22 clones further selected from the RPT planted in 1974 for the 230 selected germplasm accessions. The yield data from this RPT represented the year 1975. While the original 1950 germplasm accessions were evaluated in the SRT, seeds from open-pollinations and hand-pollinations were collected. Some 2100 seedlings from open-pollinations of 93 accessions and 4200 seedlings from hand-pollinations of 330 cross combinations were planted in 1974, 1455 selected plants were evaluated in the SRT in 1975, and 445 clones selected from this SRT were planted in RPT in 1976. The yield data from this RPT represented the year 1976. The 230 germplasm accessions selected from the 1973 SRT were also planted in 1974 as cross parents for hybridizations. Approximately 20000 seedlings produced mainly by controlled pollination in this hybridization scheme were planted in the ST in 1975, selected plants in the SRT in 1976, and 437 clones selected from this SRT were planted in the RPT in 1977. The yield data from this RPT represented the year 1977. From there onward, the same cycle was repeated. Hence, the 1973 data represent the yield level of the original unselected germplasm population, the 1974 data represent that of the population after 12% selection, the 1975 data represent that of an open-pollinated population of local cultivars after 3.4% selection, the 1976 data represent that of a hybrid population of highly diverse unselected germplasm accessions after 7% selection, and the 1977 data and those thereafter represent that of 2% selection of a hybrid populations produced in a hybridization scheme using highly selected cross parents from a diverse germplasm base.
The same control clones were used in all the yield trials; thus, comparison across years of the mean of above-mentioned yield data relative to the mean of the control should give an unbiased perspective of breeding progress in physiological yielding ability of breeding materials (Fig. 2 , top). The mean yield of the three control clones fluctuated considerably from year to year depending on the climatic condition; yet, it did not show a marked long-term decline. There was a doubling of fresh root yield in the first 5 yr. This major advance was due more to improvement in harvest index (some 65%) than to biomass (20%) (1.65 x 1.20 = 1.98) (Fig. 2b). The improvement in root dry matter content was modest (4%) during the same period (Fig. 2a). However, at this time, we were not emphasizing root dry matter content as much as in the later years.
|
CIAT/Thailand
Selection of Advanced Clones
Rayong 1, probably a farmers' selection from a naturally occurring seedling produced by introductions to Thailand many years ago, was virtually the only cultivar in Thailand during the 1960s and 1970s. The excellent adaptation of Rayong 1 to the Thai environment supported the highly successful Thai cassava processing industry. A breeding program that started in 1971 based on open-pollination of Rayong 1 did not produce any commercial cultivars (Boonsue and Sinthuprama, 1975). Controlled hybridization started in 1975 using Rayong 1 and a small number of introduced clones at the beginning and an increasing number of clones obtained from the CIAT/HQ seed introductions thereafter. A comprehensive breeding and selection scheme was developed at the Rayong Field Crop Research Center, Department of Agriculture, Thailand, and was further strengthened by the direct participation from CIAT in 1983 (Kawano et al., 1986; Sinthuprama and Tiraporn, 1986). Annually, seven to nine clones selected through ST, SRT, PT, and AT were planted in RT and incorporated into the hybridization scheme. The seven clonal entries in RT planted in 1982 and harvested in 1983 were the selections from the initial hybridizations. Rayong 1 was used as a control in every yield trial in Thailand. Comparing the mean yield of all the entries in RT relative to Rayong 1 across years would give an accurate perspective for the progress made by the CIAT/Thai collaborative cassava breeding program.
There has been a highly significant increase in the mean dry yield of RT resulting in a 50% overall improvement during the 15 yr from 1982 to 1997. Of this improvement, 30% corresponded to the increased fresh root yield and 15% to the increased root dry matter content (1.30 x 1.15 = 1.50) (Fig. 3a) . Of the 30% improvement in fresh yield, 25% corresponded to the enhanced biomass (Fig. 3b). However, there was only a slight change in the mean harvest index of the population (Fig. 3b). Thus, the increased biomass was the major factor for the population improvement and harvest index played no significant role there. This is in sharp contrast to the process of improvement previously obtained at CIAT/HQ (Fig. 2).
|
2000 to 150 in PT (<10% selection). Only 1 to 2% of the original breeding materials are selected to the PT stage. In this scheme, effective selection of potentially superior genotypes at SRT is crucial. At the SRT of the selection experiment, yield was phenotypically much more closely associated with biomass (r = 0.85; test of statistical significance is not given because yield and biomass within the same trial are mathematically interrelated) than with harvest index (r = 0.30), tempting the breeders to base the selection for yield on biomass. However, the regression coefficient of yield in PT on biomass in SRT was low (0.06), suggesting that high biomass at an early evaluation stage (SRT) will not guarantee high yield in plot trials (PT). On the other hand, the regression of yield in PT on harvest index in SRT was highly significant (Fig. 4b) . It was much higher than the regression of yield in PT on yield itself in SRT (Fig. 4a), suggesting that in SRT indirect selection for yield through harvest index is more effective than direct selection by yield itself.
|
Characteristics of Released Cultivars
There was a clear tendency of lower biomass for cultivars released in the early years when compared with the later released cultivars (Fig. 5)
. In contrast, there was little change in the harvest index, which has presumably always been close to the optimum. This may have occurred partly because we had attained a near-ideal harvest index already with Rayong 3 and the breeding materials provided by CIAT/HQ in the beginning of the 1980s and the advanced breeding materials generated by CIAT/Thai thereafter maintained harvest index near the optimum (Cock et al., 1979).
|
Vietnam shows this varietal development success more dramatically than any other country. During most of the 1970s and 1980s, agricultural research in Vietnam was isolated from progress outside the country. During this period, cassava varietal improvement research in Vietnam was little more than the maintenance and evaluation of local cultivars. The introduction into Vietnam of advanced cassava hybrid clones from CIAT/Thai began in 1989. This led to an immediate improvement in the yield level of yield trial entries at Hung Loc Agriculture Research Center, Institute of Agricultural Sciences of South Vietnam (Fig. 6) .
|
Summarizing the results of 56 RTs conducted in five countries in 1995 and 1996, the dry yield advantage of two CIAT/Thai clones (Rayong 60 and Kasetsart 50) over the best local cultivar of each location ranged from 5% in the Philippines to 99% in South Vietnam and the overall average advantage was 29% (Hershey et al., 2001).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The physiological yield basis of cassava is uncomplicated and selection using simple yield components (harvest index, biomass, and root dry matter content) is highly efficient. In relatively unselected populations with large genetic variation, harvest index was more important than biomass as a selection criterion within a given genetically mixed population and also in upgrading a population as a whole. The importance of harvest index was well recognized (Cock et al., 1979; Cock and El Sharkawy, 1988) and was first practiced in the CIAT/HQ breeding program. This practice soon became widely adopted by many national cassava breeding programs in Asia with which CIAT collaborated closely (Tan, 1987; Limsila et al., 1992; Mariscal and Bacusmo, 1995; Ngoan et al., 1995; Poespodarsono and Widodo, 1995; Tian and Lee, 1998).
Formation of a population creates a framework whose potential is open-ended. Selection within that population operates within a framework where the potential is predetermined. When breeding accomplishes a certain level and faces a new challenge, a new operational principle may be necessary. The overall progress of the CIAT/Thai breeding was achieved through the enhancement of biomass rather than harvest index of the breeding population, while within a given population selection for harvest index was invariably more important than for biomass. Thus, the factors for successful selection within a population can be radically different from those across populations. There is no guarantee that one operational principle at one breeding stage functions equally well at different stages. Progress in one context may not necessarily translate to a success in another context. A clear understanding of the nature of the different operational stages, and the identification of appropriate operational principles for each stage are the most important factors for the overall success.
The great initial germplasm variation seen in the 1973 CIAT/HQ SRT (Fig. 1) set the basis for the overall progress during the following 30 yr. Few large-scale international breeding programs start with this magnitude of genetic variation and, in retrospect, we were very fortunate to have started the breeding program giving equal importance to all the germplasm materials instead of starting with a certain predetermined population. The 20th Century masterpiece A Hundred Years of Solitude by the great Colombian writer Gabriel Garcia Marquez begins with a sentence, "The world was so new, many things did not have even a name," alluding to the fact that undefined happenings in the beginning may decide the long-lasting structure of the future.
The great diversity of evaluation and selection environments very likely rendered wide adaptability to the breeding materials while maintaining a large genetic variability in the population. Particularly, the selection input from the most stressful environment of ICA/Carimagua must have been crucial in securing tolerance to poor acid soils and disease resistances in the breeding populations, which were to be brought to Asia through seed transfer. The CBB and SE resistances obtained in the ICA/Carimagua selections were quantitatively inherited, slow disease-growing type (Kawano et al., 1983; Umemura and Kawano, 1983) and the resistances appear durable both in America and Asia until now. The breeding populations thus generated and further improved at CIAT/Thai were undoubtedly the source of successful cultivar selections for cassava-growing environments in Asia. Besides, as Asia is far from the center of origin and diversification of cassava, the number of biotic constraints is much less in Asia than in the Americas. Diseases such as SE, concentric ring spot (Phoma spp.), and dry root and stem rot (Diplodia manihotis Sacc.), and pests such as cassava hornworm (Erinnyis ello L.), mealy bugs (Phenacoccus spp.), and lace bugs (Vatiga manihotis and V. illudens) cause serious damage to cassava in Latin America but they are not known in Asia (Lozano et al., 1981). The end use of the product in Asia is largely limited to raw materials for starch and animal feed production; thus, the quality requirement is less complicated in Asian cassava. Here again, technology developed in the center of crop origin and diversification had better chances of success outside the center of origin (Jennings and Cock, 1977).
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 25, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. E. Halsey, K. M. Olsen, N. J. Taylor, and P. Chavarriaga-Aguirre Reproductive Biology of Cassava (Manihot esculenta Crantz) and Isolation of Experimental Field Trials Crop Sci., January 16, 2008; 48(1): 49 - 58. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Jaramillo, N. Morante, J. C. Perez, F. Calle, H. Ceballos, B. Arias, and A. C. Bellotti Diallel Analysis in Cassava Adapted to the Midaltitude Valleys Environment Crop Sci., May 6, 2005; 45(3): 1058 - 1063. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Moyer, J. O. Fritz, and J. J. Higgins Trends in Forage Yield and Nutritive Value of Hay-Type Sorghum spp. Agron. J., September 1, 2004; 96(5): 1453 - 1458. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
