A range of populations of cereals and herbage grasses, mainly of Lolium, Phieum and Triticum, have been tested in the greenhouse for response to lowtemperature vernalization. Leaf number before heading was used as the measure of reproductive development.

A close correlation was found between the climatic or agronomic origin of each population and its response to low temperature. In the cereals, response to low temperature was found in one ecological group only, namely, the winter wheats from western Europe. In Lolium the duration of the life cycle is closely related to the response to vernalization. The annual strains, such as Westernwolths or Wimmera ryegrasses, head rapidly with little or no response, while most of the perennial strains have an obligatory requirement for low temperature. In Phieum none of the populations tested exhibits any marked response to, or requirement for, low temperature, but the rapidity of heading, as measured by leaf number, varies from strain to strain and shows a high correlation with persistency. Most of the weed species tested are short-lived annuals and show little or no response to low temperature.

For each population, the local climate and farming system determine the optimum life cycle, and response to low temperature forms one important mechanism whereby such a life cycle can be attained.

The value of these methods for varietal classification and identification was assessed by comparing the behaviour of the same material in two or more seasons or under two or more kinds of equipment. The type of lighting used to provide long-day or continuous illumination, and the date of sowing, have a differential effect only in Phalaris and Phleum. In the other material, while variations in sowing date produce slight differences in the leaf number, the interaction between strain and sowing date is small, and good discrimination between population samples is possible. It is important to keep the greenhouse temperature between 50 and 70° F. to avoid natural vernalization or devernalization of material. The use of probit diagrams for comparing truncated distributions is discussed.

It is suggested that the methods can also be used to assess the genetic variation within populations. In the present results, the effect of the breeding system on the storage of variation is well shown. In self-fertilizing species, little or no variation is carried within lines, or, in the one case tested, between stocks of the same variety; it can, however, be carried between lines in a local population. In outbreeding species, each population carries considerable variation, both as genetic differences between individuals and in the heterozygous condition within individuals.