BC INSTITUTE FOR BREEDING AND PRODUCTION OF FIELD CROPS
d.d. Zagreb, Marulicev trg 5/I, 10 000 Zagreb, Croatia.
S. Tomasovic.
Abstract. Evaluation of wheats for resistance to Fusarium head blight was made in comparison with their crosses, e.g., F1 and 'F1 x F1', in two field trials under both artificial and natural (check) infection in 1993. We evaluated the most important yield components including 1,000-kernel weight, kernel weight/spike, and kernel number/spike. Kernel number/spike is the most reliable indicator of the reduction in grain-yield that is affected by Fusarium. The differences in kernel number/spike among the parents and the F1s and 'F1 x F1s' were high in both types of infection. The 1,000-kernel weight and kernel weight per spike also proved as very reliable indicators in evaluating the effect of Fusarium on yield reduction in wheat and were considerably lower under artificial infection than those recorded under natural infection.
Introduction. A severe onset of FHB was recorded on wheat spikes in Croatia in 1975. Because of the increased damage by scab, a breeding program aimed at developing cultivars resistant to this disease was initiated in 1978 at our Institute. Fusarium species cause scab and are among the major parasites in intensive wheat growing areas (Milatovic et al. 1982, Tomasovic et al. 1983, Cizmic 1986).
Materials and methods. We began a search for sources of resistance to scab both at home and abroad from 1976 to 1996. More than 20,000 genotypes were tested in conditions of both artificial and natural infection. In 1990, seven genotypes were selected based on disease severity and better agronomic performance more suited to the growing conditions in Croatia. In 1991, the seven selected sources of resistance were intercrossed in a semidiallelic method. In 1992, crosses were made among the F1 generation. Thirty-four combinations of double crosses were made. Four trials were established in 1993, two under artificial and two under natural infection, to test materials for resistance to F. graminearum. Disease severity and degree of resistance were rated on a 0-5 scale at the late-milk and early dough stages when symptoms of infection are most evident (77 late-milk stage and 83 early dough stage) (Tottman and Broad 1987). Infections were made at several times (most often 23), and the highest disease severity was recorded.
Results and discussion. Results of the ANOVAs for different properties under artificial and natural infection are given in Table 1. An F-test for 1,000-kernel weight was significant under natural infection in trial 1 but was not significant in all other cases. However, grain weight/spike had a significant difference among the genotypes (crosses) in trial 1, under both types of infection. This result was not observed in trial 2. The ANOVA for kernel number/spike show a significant difference among the genotypes under artificial infection in both trials and under natural infection in Trial 1.
| Property | Artificial infection | Natural infection | ||
|---|---|---|---|---|
| Trial 1 | Trial 2 | Trial 1 | Trial 2 | |
| 1,000-kernel weight (g) | --- | --- | ++ | --- |
| Kernel weight/spike (g) | ++ | --- | ++ | --- |
| Kernel number/spike | ++ | ++ | ++ | --- |
The 1,000-kernel weight is a very important and one of the most reliable indicators of F. graminearum infection on grain-yield reduction in wheat. In the initial crosses, the sources of resistance were Toropi; Encruzilhada (Brazil); Roazon, Bizel and Poncheau (France); and Mironovskaya 808 and Balaya-Cerkov (Russian Federation). The 1,000-kernel weight in trial 1 was lower for the initial parental lines than for the single and double crosses for both types of infection. Very similar values also were obtained in Trial 2. Differences in 1,000-kernel weight were evident in both trials. For example, the 1,000-kernel weight was considerably lower with artificial infection. In trial 1, the average reduction in 1,000-kernel weight under artificial infection for the parentals was 31.4 % that of natural infection. For the single crosses of the F1 generation, the average yield reduction was 24.2 % and for the 'F1 x F1s' the reduction was 21.3 %. In trial 2, the average reduction in 1,000-kernel weight under artificial when compared to natural infection was 21.4 % for the parentals, 19.6 % for the F, generation, and 20.5 % for the 'F1 x F1s'.
Kernel weight/spike is a result of kernel number/spike and 1,000-kernel weight and also is one of the most reliable indicators of grain yield reduction in wheat induced by F. graminearum. The results of trial 1 show that on average, the parents were more susceptible to F. graminearum relative to their F1s and 'F1 x F1s' for both artificial and natural infection. Kernel weight/spike also was lower in the parents when compared to the F1s, and 'F1 x F1s' in both types of infection. Similar values were obtained in both trials, but the differences were more pronounced. In Trial 1, the average reduction in kernel weight/spike under artificial versus natural infection was 32.9 % for the parents, 17.5 % for the F1s, and 25.5 % for 'F1 x F1s'. In trial 2, the average reduction was 35.3 % in the parents, 11.7 % in the F1s, and 17.4 % in the 'F1 x F1s'.
The effect of F. graminearum on kernel number/spike is very important. Artificial infection at the right time can considerably reduce kernel number/spike. Kernel number/spike for the parents was lower or at the same level as that for the F1s for both infection types in trial 1. The 'F1 x F1' averages were markedly higher in comparison with the F1s, and even more so than the parents. Similar results were observed in trial 2. Resistance to scab was much higher in the crosses than in the parents. When artificial and natural infections were compared, trial 1 averaged a reduction in kernel number per spike of 10.4 % for the parents, 13.7 % for the F1s, and 12.3 % for the 'F1 x F1s'. In trial 2, the values were 15. 9 %, 17.6 %, and 17.7 %, for the parentals, F1s, and 'F1 x F1s', respectively.
References.
Slobodan Tomasovic.
From the large number of varieties available, producers have a difficult time selecting the best assortment for specific growing conditions. Therefore, the value of a certain variety is most often judged from the results of large-scale trials conducted in a number of places. Although such testing has a few shortcomings (soil heterogeneity, number of varieties and sites, different levels of cultural practices, and statistical data analysis), large-scale variety trials still offer useful information about the properties of a particular variety to producers and breeders alike.
The highest average grain yield over a 3-year period and 12 locations was in the winter wheat variety Tina (7,230 kg/ha), ranking first in 1993 and 1995. Tina outyielded the well-known and well-accepted varieties Marija (7,210 kg/ha) and Sana (6,856 kg/ha). Compared to the standard variety Sana, Tina was 5.45 % (374 kg) better, and 13.21 % (844 kg) better than the standard variety Zitarka (6,386 kg/ha). The new variety Patria also produced high yields (6,970 kg/ha), which is an average from 15 locations over 3 years of testing. Patria was rated sixth in grain yield. Compared to Sana, Patria yielded 1.66 % higher (114 kg), and when compared with Zitarka, by 9.14 % (584 kg). Based on these and earlier tests, these are varieties with high-yield potential and very good grain and flour quality. Both deserve to be in wide production.
After 2 years of tests, high-yield capacity and good quality traits of grain and flour were found in the varieties Rina, Plodna, Pakra, and Alena, the leading varieties in large-scale trials. Plodna produced 7,315 kg/ha, which is an average over seven locations over 2 years (five in 1993 and twp in 1994). Rina averaged 7,120 kg/ha at 10 locations (six in 1993 and four in 1995). Alena was at 6,815 kg/ha in 11 locations (seven in 1993 and four in 1995. Pakra had an average of 6,705 kg/ha at seven locations (three in 1994 and four in 1995).
The variety Davorka also exhibited high-yielding ability and produced 8,270 kg/ha at five locations in 1,993. Among the new lines, Bc 575/89 produced 7,810 kg/ha at two locations in 1994 and was first in the large-scale trials. Among the leading materials in testing, of note is Bc 6937/90, producing 7,050 kg/ha at six locations in 1994. In 1996, Bc 6937/90 was registered by the Croatian Committee for Variety Registration under the name Mirjana.
Acknowledgment. On behalf of the Bc Institute and its Department of Cereal Crops-Botinec, the author wishes to thank the farm producers (organizations) who made the above tests and provided the data for interpretation and for many years of mutually beneficial cooperation.