1. For both genetical and methodological reasons, tests for quality, in the majority of breeding schemes, are not required at the early stages of the breeding process.

2. Reliability and definition of any one empirical test is not sufficient to safeguard against losses of valuable material.

3. It is suggested, therefore, that as a rule agronomically promising lines should be carried on to a stage where comprehensive tests, using a number of selected methods, can be carried out, preferably over a number of seasons. It may be questioned whether a combination of such tests as the wholemeal fermentation and the gluten swelling tests (Pelshenke, 1938a) would suffice. In the light of the available evidence a testing system might be suggested comprising baking tests together with such “indirect” and, better still, “analytical” methods as will afford accurate information of an analytical nature.

4. This review of the testing problem suggests the conclusion that empirical tests are limited in their significance and reliability, and that only an increased knowledge of the fundamentals can afford a truly scientific approach. It should be admitted that the plant breeder, with his insistence on empirical tests where fundamental knowledge was lacking, has side-tracked the cereal chemist from the narrow path of scientific approach all too often. It is questionable whether this has been to the ultimate advantage of wheat breeding.

5. Inheritance studies on quality should greatly facilitate the task of the breeder. They will become feasible and valuable only when the nature of the desirable characters and the true significance of the testing methods are clearly understood.