No abstract

The registration plan of the American Society of Agronomy provides for registration of new varieties on the basis of performance, preferably at state or federal experiment stations. In order to be approved for registration a new variety als0 must be grown commercially or be of value and used in wheat breeding. Since the pub-" lication of the "Classification of American Wheat Varieties" by Clark, Martin and Ball (2),~ numerous new varieties have been developed-and grown. Applications for the registration of most of these have been received by the committee. For those which have been ap-'proved, registration nttmbers have been assigned and certificates of registration have been issued to the applicant. The varieties approved, together with their registration numbers, are as follows:

THE purpose of this paper is to set forth some results obtained in certain crosses between the hull-less types of oats and some of the hulled forms. The hull-less types belong to the species Avena nuda. The origin of these forms is not definitely known, although, according to Carleton,2 they appear to have come from central and eastern Asia. Reports are to the effect that a certain form of this oat has been cultivated in China for a thousand years or longer. These types are not generally cultivated in any other countries and perhaps due to this fact little study has been made of the various hull-less types and their possible origin. From some observations made on our material it seems quite possible that these hull-less forms may have originated through putative changes. For example, in a pure line of the variety Sixty Day certain spikelets suddenly appeared which were very similar to the true hull-less forms in having the caryopsis loosely held in the glumes and an increased number of flowers per spikelet. These seeds were tested, but did not reproduce this hull-less tendency. A large number of similar cases have been observed particularly with hybrids, although none of these

No abstract

Elsewhere in this ] OURNAL 3 there has been presented some argument for the use of _Student's method for interpreting the results of experimental work, particularly where the experiments have been so conducted as to admit of pairing. This method consists first in determining the difference between two experiments, say A and B. Then the mean difference or mean of these differences is obtained. The difference between each value and this mean is then obtained and from these values the standard deviation is calculated by the usual method.