Previous Page Table of Contents Next Page

Irrigated Rice
for the Temperate
Climate in Chile

Santiago I. Hernaiz-L.[13]
José Roberto Alvarado[13]
Marc Châtel[14]
Dalma Castillo[13]
Yolima Ospina[15]


Santiago I. Hernaiz-L.


Current varieties in Chile have reached their maximum productivity possible under field conditions. Hence, to increase yield potential, the national rice programme decided to incorporate population improvement by recurrent selection. During the 1996/97 cropping season, a site-specific population (PQUI-1) was developed from a gene pool introduced from CIRAD (GPIRAT-10). In the 2000-01 season, a second site-specific population (PQUI-2) was obtained, using PQUI-1\CH\3\1 as base population. The first recombination was done at CIAT, Colombia. The Chilean PQUI-1 population was improved, using two methods. The first was based on selecting male-sterile plants in two ecosystems, as represented by Chillán at 36º34'S and 72º02'W, and Colchagua at 34º33'S and 71º24'W. The second method was based on selecting fertile plants and testing them under controlled laboratory conditions for cold tolerance at germination. Taking into account the broad variability shown by the populations, fertile plants were selected and their progeny evaluated. Evaluation of the first fixed lines obtained showed yields were 6% to 9% higher than those of the commercial variety Diamante-INIA.


Las variedades actuales de Chile han alcanzado la máxima productividad factible de lograr en el campo, por lo que para elevar el potencial de rendimiento se decidió incorporar el mejoramiento poblacional por selección recurrente. Durante la siembra 1996/97 se desarrolló una población de sitio especifico (PQUI-1) a partir de un acervo genético introducido del CIRAD (GPIRAT-10). En la siembra 2000/01 se creó la segunda población de sitio específico denominada PQUI-2, que se obtuvo utilizando como base la población PQUI-1\CH\3\1 y su primera recombinación se hizo en el CIAT, Colombia. La población chilena PQUI-1 se mejoró utilizando dos métodos: el primero basado en la selección de plantas androestériles en dos ambientes diferentes (Chillán 36º34’S, 72º02’W y Colchagua 34º33’S, 71º24’W), y el segundo basado en la selección de plantas fértiles sometidas a una prueba para determinar tolerancia al frío en la germinación, en condiciones controladas de laboratorio. Teniendo en cuenta la buena variabilidad que presentaron las poblaciones durante su manejo, se seleccionaron plantas fértiles y se evaluaron sus descendencias. La evaluación de rendimiento de grano de las primeras líneas fijas provenientes de la selección de plantas mostró que estas líneas presentaron un rendimiento entre 6 y 9% mayor que la variedad comercial Diamante-INIA.


The current development of rice farming in Chile was possible thanks to technologies developed mostly by the rice programme of the Instituto de Investigaciones Agropecuarias (INIA). Progress was also made in the crop’s genetic improvement for both industrial and culinary agronomic traits. Current varieties now bring together the most important requisites demanded by both farmers (good yield levels) and industry (grain type and industrial quality). Many of the varieties, however, were recently observed to have reached their maximum productivity feasible in the field.

Historically, the improvement of autogamous plants has concentrated on generating variability by using methods such as artificial mutation and hybridization (crosses and backcrosses between two or more parents). The progenies, in general, are selected and directed towards homozygosis through conventional methods such as pedigree, mass selection, and modified mass selection. Population methods are almost never used because of the difficulties in carrying out a large number of manual crosses (Fehr, 1987). Although the newly developed commercial varieties represent an advance for farmers, Cuevas-Pérez et al. (1992) report that the genetic base has narrowed. The different rice breeding programmes in Latin America are using few sources of variability, thus potentially limiting future progress in both productivity and quality.

To find a solution to this problem, INIA-Chile, and other Latin American institutions have, since the mid-1990s, sought alternative ways to broaden and diversify the genetic base of rice. To do so, the decision was made to incorporate into rice breeding projects other methodologies, including population improvement by recurrent selection, which has proved responsive to specialist concerns.

This method was proposed by Fujimaki (1979) and implemented by Châtel et al. (1997). It permits the development and later improvement of gene pools that group together and recombine various genotypes that differ from those now being used. The method consists of evaluating individuals within a genetically heterogeneous population and selecting them as fulfilling the breeding programme’s objectives. A set of these individuals is then recombined.

Although this tool is normally used with populations of open-pollinated plants, it can be applied to populations of self-pollinating plants such as rice by using a recessive nuclear gene for male sterility, known as ms. Discovered by Singh and Ikehashi (1981) in a mutant of ‘IR36’, this gene allows transforming the autogamous population into an allogamous one. The methodology makes it possible to progressively increase, in a genetically variable population, the genetic value of one or more agronomic traits under selection. From the improved populations, we can obtain promising improved lines with genetic bases that are most likely to differ from those of current commercial varieties.

Methodology used for population improvement in Chile

Population improvement begins with forming populations, each made up of varieties and lines that bring together the traits desired by the programme and that are adapted to the ecosystems for which they are being developed. A population should have sufficient variability to allow the recombination of favourable genes, thus permitting the maximum possible recombinations among the alleles, as well as increasing their frequencies.

The Chilean population improvement project has used two recurrent selection methods: mass selection of S0 male-sterile plants, and the application of a cold evaluation trial to improve the efficiency of selecting plants with genes for tolerance of low temperatures (Hernaiz- L. et al., 2000). The first method was used when population work began, leading to the development of two populations adapted to different ecosystems. The second method was carried out, once the third recombination of the populations was completed.

Mass selection of male-sterile plants

From a population of S0 plants that segregates for fertile (MsMs or Msms) and male-sterile plants (msms), male-sterile plants are chosen that fulfil the objectives of the genetic improvement programme. Seed is harvested from selected plants (which had been pollinated by fertile plants) and mixed in equal proportions. Their sowing will give rise to a population composed of 50% fertile and 50% male-sterile S1 plants. The harvest of grains from the male-sterile plants will represent the recombinant population (Châtel and Guimarães, 1995).

It should be noted that the methodology has a disadvantage in that selection can only be carried out with male-sterile plants and that the contribution of fertile plants can be either positive or negative for the traits being improved.

Selecting S1 progeny

Within a population, fertile S0 plants are chosen on the basis of having either the heterozygous (Msms) or homozygous genotype (MsMs). These plants are evaluated as families to later select and recombine the best-performing S1 plants. Recombination is carried out with seeds from S1 plants that may be fertile homozygotes (MsMs), fertile heterozygotes (Msms) or male-sterile homozygotes (msms). The seed harvested from the male-sterile plants represents the recombination of individuals present in the population, and the mixture of that seed will form the base of the improved population (Châtel and Guimarães, 1995).

Cold evaluation trial under controlled conditions

The trial is carried out with the recombinant population, thus taking advantage of the existing variability. The trial begins with selecting fertile plants from the population under study, continuing until seed is harvested from each plant. The seed is divided into two lots, one of which is kept and the other used for the cold evaluation trial.

For the trial, each selected plant contributes 150 seeds, which are distributed in three replications, with three checks, in a randomized complete block design. The experiment is then evaluated for tolerance of low temperatures in growth stage Z00, according to the Zadoks scale (Quiroz, 1982) corresponding to dry seed (Castillo, 2001). The methodology involves the following steps:

After exposure to low temperatures, the number of normal seedlings is counted, that is, those seedlings that have developed a coleoptile and radicle as if they were growing under optimal conditions

Tolerant materials are those whose percentage of normal seedlings is equal to or higher than that of the check with the smallest response (Castillo, 2001)

Methodology used for developing population lines

The principal objective of a genetic improvement programme, regardless of the method used, is to obtain and evaluate plant material for developing future lines and varieties. Fertile plants can be selected for line development from a population, whatever its phase of improvement.

For line development based on the variability available in the fertile plants of populations being improved, two selection methods are used: conventional selection by pedigree and anther culture in the laboratory.

Pedigree method

Once the fertile S0 plant is selected, its S1 progeny is planted, which segregates for fertile and male-sterile plants. Of the progeny of S1 plants, fertile plants that meet the criteria previously established by the programme are selected. The S2 seed is again sown individually, the plants selected and the S3 seed harvested. This process is repeated until segregation disappears. The ms gene is eliminated by taking out male-sterile plants in each stage and eliminating segregating lines.

Anther culture

Anther culture permits the production of homozygous lines from segregating populations, thus fixing genotypes more rapidly than any other method of progeny selection. This method significantly reduces the time span for varietal production. It is initiated by selecting plants from a population before flowering, harvesting the stems and extracting the anthers. The anthers are placed in a medium to induce calluses and then in a regenerating medium to obtain haploid plantlets (Lentini et al., 1997).

Table 1. Lines introduced to develop the rice population PQUI-2 for Chilean conditions.


Parental materials


B 581-A6-545-2/Peta IR276-1-6-9//Kuatsu

Quila 121304


Quila 68405

Delta/Quila 29101 (Krasnodarskj 3352//Gallardo/Kuatsu)

TUC 25

Not determined


Dw/T(N)1IR151-4-19/CH101 (RR/B138-1-1/Oro)

Cinia 1014




PRA 767-5CH

PRA 523/CIRAD 403

PRA 775-1CH

PRA 622/Luluwini 22-M

PRA 741-1CH

Estrela/Long sweet glutinous rice

PRA 737-1CH

Cuibana/Long sweet glutinous rice

PRA 760-1CH

Long sweet glutinous rice/Progresso

Developing populations

Population PQUI-1

When population improvement was initiated, the gene pool GPIRAT-10, developed by CIRAD, France, was introduced into Chile. This germplasm, designated as cold tolerant, was planted, but it did not adapt to local Chilean conditions, mainly because of its long cycle. A new, site-specific, and much earlier maturing population had to be created, using as base the GPIRAT-10 population and conducting an introgression of Chilean material that was earlier maturing and cold tolerant.

The introgression of these varieties and lines enabled the programme to create a new population in which the principal defect of GPIRAT-10 was improved, that is, its lack of earliness. Hernaiz-L. et al. (2000) describe this stage of the process.

Population PQUI-2

After some years of work with population PQUI-1, we decided to create a new population that complemented the variability of the older one. The development of this new population would be very important, as it would permit the accumulation of new genes favourable to the traits being sought and become an additional source for increasing the number of lines produced. Moreover, other types of materials for the rice genetic improvement programme could be developed. The new population was called PQUI-2.

This population was developed, using as base population PQUI-1\CH\2\1 and introducing both Chilean and foreign varieties and lines that had previously been tested by the project (Table 1). PQUI-1\CH\2\1 and the lines were planted on two occasions (2 and 22 October 2000) under greenhouse conditions. When the plants were 0.15 m high, they were transplanted, at a density of 0.25 × 0.30 m, to the Chillán Experimental Rice Field (CEAC, its Spanish acronym). The population was surrounded by fences of polyethylene sheets to prevent possible contamination by foreign pollen.

Table 2. Percentage of contribution of the varieties and lines used as parental materials to develop the rice population PQUI-2, and their geographic origins.


Relative contribution(%)





Quila 121304



Quila 68405



TUC 25






Cinia 1014






PRA 767-5CH



PRA 775-1CH



PRA 741-1CH



PRA 737-1CH



PRA 760-1CH






Crossing was initiated in mid-January 2001, using Sarkarung’s method (1991), which basically consists of harvesting tillers before panicle exsertion and subsequently carrying out pollination under laboratory conditions. The F1 seed obtained as a result of these crosses is mixed in equal parts (10 seeds per cross) and planted at the CEAC in the 2001/02 season. All the plants were harvested individually and F2 seed mixed in equal parts, giving rise to the base population PQUI-2\0\0\0. Part of the seed was kept and the rest sent to CIAT, Cali, Colombia, to obtain the first recombination. Chile is continuing with the remaining cycles of recombination and improvement of this population.

In each recombination cycle, both male-sterile plants (whose seeds are mixed in equal parts to obtain the recombinant population) and fertile plants are selected to constitute segregating generations. From these, lines are extracted to form the future line nurseries for the conventional improvement programme.

The lines used for introgression to create the new population possessed great variability, and were characterized for their grain quality, plant type and cold tolerance. The final composition of PQUI-2 is presented in Table 2.

Population improvement

Population PQUI-1\0\0\0

The F1 seed obtained from the introgression of the Chilean varieties and lines to the gene pool GPIRAT-10 was sent to CIAT, for a generational advance and F2 seed was obtained. The F2 of this future population was planted in 1997/98 in Chile at Chillán (36º34'S and 72º02'W), the most southerly region of rice farming, and Colchagua (34º33'S and 71º24'W), the most northerly region with the fewest problems of cold.

The resulting plants corresponded to population PQUI-1\0\0\0, which was called PQUI-1. In contrast to GPIRAT-10, the source of male sterility and of genetic background, this population had greater genetic variability and more agronomic traits fulfilling the rice breeding programme’s objectives. Improvement was implemented as according to the methods described above.

Selecting for adaptation to two ecosystems

From population PQUI-1\0\0\0, male-sterile plants presenting early maturity and better agronomic traits for each ecosystem were selected and harvested. The seed mixture produced by male-sterile plants selected at each locality made possible the acquisition of two new selected and recombinant populations. Following the nomenclature proposed by Châtel and Guimarães (1995), these were called PQUI-1\CH\0\1 and the PQUI-1\CO\0\1.

During the 1998/99 season, the two improved populations were planted to obtain the first recombination cycle and were identified as PQUI-1\CH\1\1 and PQUI-1\CO\1\1, respectively.

For the 1999/00 harvest, the two populations PQUI-1\CH\1\1 and PQUI-1\CO\1\1 were planted to obtain the second cycle of recombination identified as PQUI-1\CH\2\1 and PQUI-1\CO\2\1, respectively. At each stage, fertile plants were selected to develop segregating lines. Figure 1 schematically describes the work carried out.

Evaluating for tolerance of low temperatures

Once the third recombination cycle of both populations PQUI-1\CH\2\1 and PQUI-1\CO\2\1 was completed, improvement for resistance to cold was begun, using the mass selection method in fertile plants.

For the 1999/00 season at CEAC, populations PQUI-1\CH\2\1 and PQUI-1\CO\2\1 were planted. In both populations, fertile plants were evaluated for tolerance of low temperatures and selected.

A total of 285 fertile plants were selected from PQUI-1\CH\2\1 and 262 from PQUI-1\CO\2\1. Seeds of each plant were separated into two lots, one to form a future population and the other to carry out the cold evaluation trial. The seeds subjected to the trial were distributed across 13 experiments for the population from Chillán, and across 16 experiments for the population from Colchagua. The statistical design used was randomized complete block with three replications of 50 seeds each, similar in form, colour and size. Checks were the varieties Diamante- INIA, Oro and Brillante-INIA.

Figure 1. Flow of germplasm material in the procedures used to recombine rice populations PQUI-1CH and PQUI-1CO.

With the seeds kept from selected plants (which had presented better tolerance of cold, compared with the checks), improved populations were developed. The mixture of seeds of those plants that presented the best tolerance of cold was planted in the field, and male-sterile plants representing the recombination of the best-performing individuals were harvested. The improved populations were called PQUI-1F and PQUI-2F, which we can identify as PQUI-1\CH\2\1,F\1 and PQUI-1\CO\2\1,F\2, respectively. The development of the work on cold tolerance is described in Figure-2.

Figure 2. Flow of germplasm material in the procedures used to improve cold tolerance in rice populations PQUI-1\CH\2\1 and PQUI-1\CO\2\1.

Line development

When one conducts genetic improvement, the main objective, regardless of the method used, is to acquire and evaluate lines for producing varieties and their recommendation for their commercial cultivation by farmers. For rice population improvement, in each planting of a population, regardless of its phase of improvement or recombination, the possibility exists of selecting fertile plants of good phenotype. Thus, the variability existing in the populations is taken advantage of during the different population cycles. In line development, two selection methods are used: the conventional selection by pedigree and anther culture in the laboratory.

Table 3. Preliminary results of the best-performing rice lines extracted from an improved population and evaluated in the yield trial carried out at the Campo Experimental de Arroz en Chillán (CEAC), Chile, 2001/02 season.

Lines & Checks

Tiller Number per plant

Grain number per panicle

1000-grain weight (g)

Yield (kg ha-1)
















































Lines developed by selecting fertile plants

The fertile plants selected in each recombination cycle of populations PQUI-1\CH and PQUI-1\CO gave rise to nine line nurseries with a total of 356 families. From these, plants were selected to obtain homozygous lines free of the male-sterility gene. During the 2001/02 season, fertile plants were harvested from population PQUI-2F, destined to form part of the line nurseries.

During the 2000/01 season, the productivity of the homozygous lines was evaluated, using Federer’s augmented block design. The results made possible the selection of 17 lines that were integrated into a yield trial. In this trial, using a randomized complete block design, they were compared with Chilean commercial varieties. The yields obtained in the evaluation of these 17 lines demonstrated that good potential exists to obtain better material than the varieties currently in use. In the 2002/03 season, 36 lines were being evaluated in two yield trials.

Table 3 shows preliminary results (without analysis) of the most outstanding lines in the 2001/02 yield trial.

Lines developed through anther culture

The first R1 anther culture lines from plants of the Chillán and Colchagua populations (PQUI-1\CH\1\1 and PQUI-1\CO\1\1, respectively) were obtained in 1999 at CIAT during their first recombination. Generation R2 was advanced, also at CIAT, resulting in 325 R2 lines, which are currently undergoing evaluation at CEAC. In the 2002/03 season, some of these lines were included in the yield trials.

Prospects for population improvement in Chile

Improvement through successive cycles of recurrent selection will continue being applied to the existing populations. Moreover, the new population PQUI-2 is expected to present variability that differs from PQUI-1, thus contributing to new advances for population improvement. Furthermore, the possibility of extracting segregating lines from those populations has already been announced, even though they are still in the initial stages of improvement. We are sure that this will permit the development of increasingly better lines for planting by Chilean farmers.


Castillo, D. 2001. Caracterización de germoplasma de arroz (Oryza sativa L.) según su tolerancia a baja temperatura en el estado de germinación. Temuco, Chile, Universidad de La Frontera. 60 pp. (IA thesis)

Châtel, M.; Guimarães, E.P; Ospina, Y. & Borrero, J. 1997. Utilización de acervos genéticos y poblaciones de arroz de secano que segregan para un gen de androesterilidad. In E.P. Guimarães, ed. Selección recurrente en arroz, pp. 125-138. Cali, Colombia, CIAT.

Châtel, M. & Guimarães, E.P. 1995. Selección recurrente con androesterilidad en arroz. Cali, Colombia, CIRAD-CA & CIAT. 70 pp.

Cuevas-Pérez, F.E.; Guimarães, E.P.; Berrío, L.E. & González, D.I. 1992. Genetic base of irrigated rice in Latin America and the Caribbean, 1971 to 1989. Crop. Sci., 32(4): 1054-1059.

Fehr, W. 1987. Principles of cultivar development: theory and technique. Vol. 1. New York, Macmillan Publishing. 536 pp.

Fujimaki, H. 1979. Recurrent selection by using genetic male sterility for rice improvement. Jpn. Agric. Res. Q., 13(3): 153-156.

Hernaiz-L., S.I.; Alvarado, J.R.; Châtel, M. & Borrero, J. 2000. Creación de la población de arroz PQUI-1 desarrollada con tolerancia a frío mediante selección recurrente. Agric. Téc., 60(2): 195-199.

Lentini, Z.; Martínez, C. & Roca, W. 1997. Cultivo de anteras de arroz en el desarrollo de germoplasma. Cali, Colombia, CIAT. 57 pp.

Quiroz, C. 1982. Código decimal de estados de crecimiento en cereales. Santiago, Chile, Estación Experimental La Platina of the Instituto de Investigaciones Agropecuarias. 20 pp.

Sarkarung, S. 1991. A simplified crossing method for rice breeding. Cali, Colombia, Fondo Latinoamericano de Arroz de Riego (FLAR) & CIAT. 32 pp.

Singh, R.J. & Ikehashi, H. 1981. Monogenic male sterility in rice: induction, identification and inheritance. Crop. Sci., 21(2): 286-289.

[13] Centro Regional de Investigación Quilamapu (CRIQ), Instituto de Investigaciones Agropecuarias (INIA), Casilla 426, Chillán, Chile. E-mail: [email protected]
[14] CIRAD/CIAT Rice Project, CIRAD-CA, A.A. 6713, Cali, Colombia. E-mail: [email protected]
[15] Rice Project, CIAT, A.A. 6713, Cali, Colombia. E-mail: [email protected]

Previous Page Top of Page Next Page