ITEMS FROM ROMANIA

S.C.A.-AGRICULTURAL RESEARCH STATION

Turda, 3350, str. Agriculturii 27 Jud Cluj, Romania.

Breeding procedure applied for improvement of winter wheat varieties at ARS Turda. [p. 113-115]

V. Moldovan, Maria Moldovan, and Rozalia Kadar.

Wheat breeders rely on classical methods to produce new, improved varieties. Reliance on classical-breeding procedures seems to be sound from a theoretical point-of-view, because genetic variation available in common wheat and related species has not yet been exhausted. As long as such usable variability exists, the opportunity for further significant increases genetic progress in wheat by breeding will exist. Classical breeding consists of two important steps. First, crosses are planed to achieve reasonable odds for desired genetic recombinants. Second, the selection process presumably is designed for success in identifying agronomically superior phenotypes within the hybrid populations.

Therefore, classical breeding has numerous variations that are dictated largely by the philosophy of the breeder, the wheat crop specificity, the breeding objectives and priorities, and different economic constraints. The most productive wheat-breeding programs rely upon numbers, both in hybrid population sizes and number of hybrid combinations. The number of crosses can be relatively large, and it reflects, in part, the inability of breeder to predict with certainty the worth of specific hybrid combinations on the basis of known attributes of the parental varieties. Nevertheless, our experience has shown that a smaller number of well-planned crosses may provide better hybrid material than a large number of poorly planed crosses. The size of a manageable hybrid population is estimated from the number of traits that are considered as breeding objectives with which the breeder is concerned. Establishing realistic priorities at a reasonable number is very important, because each added objective reduces the rate of expected genetic progress for each of the others. Changes in priorities also can affect the achieving desired genetic progress.

Wheat breeding procedure at ARS Turda. The objectives that are emphasized in our wheat-breeding program have been reported previously (Ann Wheat Newslet 45:119). We now report on the wheat-breeding procgram used at ARS, Turda, as one way to produce new, improved varieties. This breeding and selection procedure can be summarized as follows:

1. Three to four hundred of crosses are made annually. Our experience has shown that we can maintain adequately as many as 400 hybrid combinations with available land, experimental equipment, financial resources, and personnel. Careful consideration is given to the choice of parents from an available core collection, which is annually refreshed with new accessions. Intercrosses of parental varieties or breeders' lines are almost exclusively made by hand during heading in the field. The majority of crosses involve two parents (A x B), but every year a number of three-way crosses ((A x B) x C) and single backcrosses ((A x B) x A) or ((A x B) x B) are made. Because genetic variability and diversity can be increased by multiple crosses (((A x B) x C) x D) and convergent methods ((A x B) x (A x C)), they also are taken into consideration. Lately, recrossing selected lines within a hybrid population has been used in a manner similar to recurrent selection.

2. Grow F1 hybrids in the field. The F1s are evaluated for agronomic and disease resistance traits. The few that are of poor quality are discarded, reflecting the inability to predict the worth of progeny from known attributes of the parental varieties used in the crosses.

3. Hybrid populations are used as bulks through the F2 generation. Bulked hybrids are grown in the field near the normal rate of seeding in order to evaluate them under conditions that approximate plant competition in commercial production. Based on agronomic and disease-resistance evaluations, some populations of no value are discarded. From the most promising F2 hybrid populations, we use head selection. In a limited number of crosses of special interest, the first selection is delayed until the F3, F4, or F5 generation. However, F2 selection is most common for us. Head selections are made essentially with attention on such attributes as maturity, plant height, disease reaction, and plant type. The number of heads selected per population averages 50 to 100, although in some valuable populations it can be 100 to 300. Larger numbers exceed our capability for adequate evaluation.

4. The head-row nursery is usually comprised of 25,000 to 30,000 entries, including reselections. For the most part, these nurseries are represented by F3 progenies from selected heads. This process of selection is repeated, within superior rows, generation after generation, until row progenies are sufficiently uniform and stable. In our program, no more than 2 or 3 reselections are made, so the last generation of selection is F4-F5. Simply inherited, observable and measurable morphological traits can be readily monitored in the selection process to assure reasonable uniformity and nonsegregation. This is easy in hybrid combinations with parents that are similar in plant type, maturity, plant height, and other readily observable traits. However, varieties selected in an early generation are likely to conserve a substantial amount of heterogeneity, especially for complex traits such as the yield. This heterogeneity can contribute to yield performance as well as to stability of performance in an array of environments. After this screening process, in which apparent agronomic value and disease resistance are assessed, the best lines, approximately 5 %, are advanced.

5. Preliminary observation plots, without replications, include those approximately 750 to 1500 lines retained as selections from the head-row nursery. Lines remain in observation plots for only 1 year, in which they are more comprehensive evaluated. At maturity, heads of each plot are harvested for a sprouting test in artificial conditions in a mist cabinet. Sufficient grain is available for small-scale, bread-making quality tests. Lines also will be entered in a special disease nursery where they will be evaluated using artificial epidemics. Based upon of these complex evaluations, approximately 25-30 % promising lines are retained.

6. Observation yield trials are organized as randomized, block-design experiments with three replications. The number of trials is limited to the 250-300 promising lines retained from preliminary observation plots. They remain in observation yield trial for 1 year and continue to be evaluated for agronomic performance, disease resistance, and bread-making quality. At this stage, we have the first opportunity to precisely quantify yield performance and compare them with the check varieties. On this basis, only truly superior lines are advanced. The number of lines does not exceed 100-150 annually.

7. Competition yield trials, which are randomized block design experiments with more than three replications, are organized 1-3 years and comprise those 100-150 lines retained from observation yield trials. Evaluation for agronomic performance, disease resistance, and a large-scale, bread-making quality analysis are continued. Each year doubtful lines are discarded. After 1 to 3 years of evaluation, valuable lines (10-15 annually) may be introduced in regional yield trials at nine Agricultural Research Stations, which are located in different ecological areas. Performance and stability evaluations, simultaneous with local evaluation and seed increase, are continued. Those lines of value are advanced into the Official Yield Trials.

8. Official Yield Trials at the State Institute for Variety Testing and Registration (ISTIS) give an indispensable and independent assessment of new lines over a wider range of environments. After 3 years of testing for agronomic and technological value (VAT) and distinctness, homogeneity, and stability (DOS), the new lines that have sufficient merit have been entered in the Romanian Official Catalogue of Varieties, named, and released to growers. Continued station and regional performance trials are concurrent with foundation seed increase.

Outlook. Consistent with the assumptions that the useful genetic variability in wheat has not been exhausted and such variability for yield is mostly additive and can be fixed in true-breeding lines, we use variations of classical-breeding procedures to manipulate this variability for varietal improvement.

Our breeding procedure does not differ widely from the pedigree-selection method. In particular, it differs in the volume of breeding material that is managed in the system. We do not pretend that this system is the best in all circumstances, nor do we suggest that it could be applied with equal success to others wheat-breeding programs.

Using the system presented here, we have developed and released for commercial production, nine improved varieties of winter wheat since 1971 (Table 1). Five of these still maintain good acceptance and popularity and are listed in The Romanian Official Catalogue of Varieties for 2002.

Since 1971, an average of one new variety every 3 years has been released to wheat producers from our program. Of these nine varieties, seven were from single crosses, Transilvania was selected from a three-way cross, and Ariesan had its origin in an backcross. Note that Turda 2000 is the result of a single cross made between two of our own varieties (Apullum/Ariesan). The genetic progress achieved in previous breeding cycles can be used for continued improvement of wheat genotypes in a manner similar to that for broad-sense recurrent selection.

The variety Ariesan, released in 1985, is most extensively grown in central and northeastern Romania. We believe that a breeding procedure where some heterogeneity is conserved into wheat varieties, especially for complex traits such as the yield, that buffers against environmental change contribute yield stability are factors favoring the long life and wide acceptance of Turda varieties among the wheat growers of central Romania. At the same time, such varieties could be variable for a number of resistance genes, thus simulating a multiline variety.

We follow a liberal variety-release policy. The five varieties listed in The Romanian Official Catalogue of Varieties represent 15 % of the total wheat varieties cultivated in Romania. They constitute nearly one-fourth of the total Romania wheat acreage. It is in our own best interest that our varieties are replaced with another of our improved varieties. For this reason, keeping pace with the advances in wheat productivity achieved in the other programs is important. Thus, our wheat-breeding program also encompasses genetic studies, disease-resistance investigations, end-use quality studies, and breeding methodology. All are involved in the task of improvement of winter wheat varieties.