DEPARTMENT OF EXPERIMENTAL EVOLUTION. 117 



rain-forest. In general the tissue fluids of the parasite show a higher osmotic 

 concentration than do those of the leaf-tissue of the host tree." 



"Miscellaneous studies of variation and correlation in plants. — A study of 

 the distribution of abnormalities in the inflorescence of Spircea and an investi- 

 gation of the relationship between the number of stamens and pistils in the 

 flowers of Ficaria have been pubUshed." 



"Studies on variation, correlation, and selective death-rate in garden beans. — 

 These investigations have been continued along Unes indicated in earher 

 reports. A general summary is now being prepared of studies on the influence 

 of the depauperization of the ascendants, on the characteristics of the 

 descendants, of the correlation between the weight of the seed planted and 

 the characters of the plant produced, and of the correlation between the char- 

 acters of the plant and the characters of the pods and seeds which it bears, 

 and of some other problems. 



"In the teratological phases of the work, a paper on the correlation of the 

 characters of the second and third node has been published in cooperation 

 with Mr. Avery. A detailed investigation of the green weight, dry weight, 

 and tissue-water content of the leaves of the seedlings showing the sHghtest 

 structural variation has been pubUshed. A similar investigation of seedlings 

 which are trimerous at the first and second nodes has also been completed. 

 These studies show that variations are not merely structural, but that phy- 

 siological differences are associated with the structural variations. These 

 results will eventually be considered in relation to the problems of the selective 

 death-rate." 



"Miscellaneous studies in statistical theory. — As in previous years, consider- 

 able attention has of necessity been given to statistical methods. A paper 

 on the correlation between a component and between the sum of two or more 

 components and the sum of the remaining components of a variable, and 

 another paper on further applications of a coefficient measuring the correla- 

 tion between a variable and the deviation of a dependent variable from its 

 probable value have been pubUshed. The formulae developed in both of these 

 papers have had or are shortly to have application in the solution of problems 

 of economic importance." 



"Variation, correlation, and probable errors of plot tests. — If genetic work is 

 to be of importance in practical agriculture it must not result merely in the 

 formulation of the laws by which breeds of animals and plants showing new 

 characters or new combinations of characters may be secured, but it must 

 provide the means by which these strains, breeds, or varieties may be tested 

 with respect to their economic performance. If such means of critically 

 testing the various strains are not provided, the multiplication of forms can 

 result only in confusion, and not in progress. The plot test has been in the 

 past and is at present the chief method of determining the relative values of 

 a series of strains. Notwithstanding the enormous economic importance of 

 such work and the great annual financial outlay in the making of such tests, 

 our knowledge of the theory of plot tests is far from satisfactory. 



"During the past several years various problems of plot tests have been 

 considered at this Station. At present the work is being continued with the 

 cooperation of the Office of Western Irrigation Agriculture and the Office of 

 Dry Land Farming of the United States Department of Agriculture." 



"Biometric studies on egg-production. — The studies on egg-production out- 

 lined in earlier reports have been continued and extended on data furnished 

 by the Poultry Department of the Connecticut Agricultural College and 

 Agricultural Experimental Station at Storrs. In connection with Messrs. 

 Blakeslee, Kirkpatrick, and Warner an experiment station bulletin has been 



