*4.1. Influence of Mixing*

The mixing of isolates is a generally applied method. This is done, not only to regulate aggressiveness, but it can be a necessity when a larger amount of inoculum is needed than the present methods can produce. The cited literature (16–63) is not consequent. The number of components strongly varies. Moreover, many authors seemingly feel that they are more secure when more isolates are mixed, and they think that a mixture comes closer to the variability found in nature. However, we did not find anything pertaining to this problem in the literature. Thus, it was not clear what the result of the mixing considering the aggressiveness is. As nobody, except one [40], used the aggressiveness test, the mixing was made without an aggressiveness control. The lesson is that the di fferent mixtures did not provide the same aggressiveness. For this reason, nobody can hope that a mixture automatically will have high aggressiveness. This was not an accident; the experimental results do not support such assumptions. Additionally, the reactions di ffered between DI, FDK and DON contamination. Large and significant di fferences were found between the di fferent mixtures and pure isolates. There is another problem; that the conidium concentration was adjusted to di fferent values. Of course, this could influence the e ffect of mixing. Most people used 5 × 10<sup>4</sup> or 5 × 10<sup>5</sup> conidia/mL, but (once) 10<sup>6</sup> conidia/mL was applied. Here, we remark only that to the uncontrolled mixing and uncontrolled adjusting of conidium concentration was a general rule. The e ffect of dilution was influenced by the isolate, the host variety through resistance, and the aggressiveness level, but other agents may also be present. In laboratory and greenhouse seedling tests [1], mixing of isolates did not give the mean of aggressiveness of the participating inocula aggressiveness in greenhouse experiments and the di fferent mixtures reacted di fferently. However, these greenhouse tests could not provide any information about the behavior of the mixtures for field conditions. In the best case, it was mentioned in the papers that the isolates were aggressive in earlier tests and, therefore, the authors hoped in a successful infection. Therefore, the general experimental praxis and the earlier data pressed us to investigate this problem.

The result of the mixing depends most on the aggressiveness level of the participating inocula. From the low pathogenic isolate, no highly aggressive inocula can be produced, but from an aggressive isolate from the same test tube, inocula with di ffering aggressiveness could be produced [75]. When we see the most aggressive isolate (inoculum), the mixture decreases its aggressiveness (each case compared to the most aggressive component). When we compare the aggressiveness of the mixture to the arithmetical mean of the individual components, it becomes clear that in most cases the aggressiveness is higher, with even 50–100% di fference in some cases than the mean aggressiveness level counted for the participating inocula. On the other side, the resulting aggressiveness was always lower than that of the most aggressive component. Therefore, the exact forecasting of the result of the mixing is hardly possible. This means that the most aggressive isolate can significantly overbalance the arithmetical mean, but still does not reach the level of the original aggressiveness of the most aggressive component. The system is much more complicated than it was thought earlier. The DI/DON and FDK/DON rates showed an interesting picture. The DI/DON rate was much more variable; this was found mostly when Fc. 12,375 isolate was a participant of the mixture or was alone. We do not know whether this, an instinct attribute, is for other variants of the DI/DON and FDK/DON rations, which were rather similar. It is also considerable that Fg 13.38 has a relatively high DON and FDK ratio and a very low DON content, the other three isolates perform similarly with lower DI, medium FDK and higher DON contamination with strong di fferences. This is supported also by the higher number of significant correlations from Tables 3C, 5C and 7C.

Uncontrolled mixing and diluting may result in low aggressiveness level and bad di fferentiation of the genotypes. The only advantage is that it stabilizes the variation in aggressiveness of the inocula compared to the individual components. However, it does not solve the problem of the use of one inoculum. As we do not have races at hand, the most important argumen<sup>t</sup> (the mixtures represent a

better scope of the pathogenic population) is refuted. In this test series, this view could not be supported. The question is which of the 15 the best solution is. We cannot say each inoculum behaved di fferently. Having an inoculum the resistance can be underestimated (high aggressiveness) or underestimated low aggressiveness; therefore, several aggressiveness levels are needed to reach more reliable results. On the other hand, the di fferent di fferentiation of the genotypes at di fferent aggressiveness levels needs a correction mechanism, by the use of more isolates to produce more reliable results. This has a significance in QTL analysis where the large di fferences can be evaluated well, but the chance to identify for lower or medium e ffective QTL is not very convincing. The conclusion is that the mixing does not solve the problem to produce top quality inoculum for artificial inoculation. Even the aggressiveness of the participating inocula is known, an exact forecast of the aggressiveness of the mixture is not possible—especially when we would like to know FDK and DON reaction. In resistance response, the mixtures are not better, or worse, than a single aggressive inoculum.

There is another problem. As in this test, also, the case is, the mean performance of di fferent independently used inocula produce a mean that describes much better the resistance level than any of the inocula alone. Earlier research [69–71] supports its usefulness. The use of more independent isolates is costly. Therefore, it should be used where high quality phenotyping is necessary for genetic research and other scientific purposes. It is needed to qualify new cultivars with higher or high resistance to FHB. The risk analysis needs careful work. The behavior of the most aggressive isolate resembles well to the combining ability of a given crossing partner in di fferent crosses. In plant breeding, this is a well inheritable trait. It is supposed that such research would be important to make here, to be able to select the isolates that give the highest aggressiveness if possible.

An aggressiveness test is necessary to qualify inoculum before use. No mixing, no adjusting inoculum can secure the high aggressiveness needed for e fficient work. When we want to have a mixture, an aggressiveness test can help choose the suitable inocula for mixing.
