138 GENETICS [BoT. Absts., Vol. VII, 



tively warm. Winter begins the last of October and winter temperatures are sometimes 

 — 40° to — 50° or lower. The most trying times for trees are ( 1 ) Spring ; severe frosts after buds 

 swell are frequent and kill all growth. (2) Late autumn or early winter when trees are badly 

 frozen while foliage is still on. (3) Winter. Long-continued spells of severe cold. All these 

 factors require closer investigation. It has been found difficult and practically impossible 

 over most of the area to bring commercial varieties of apples to bearing age. — For the past 

 30 years many varieties have been tested at Experiment Stations and by individuals in Mani- 

 toba, Saskatchewan and Alberta, but few trees have borne and, when they have, they have not 

 lived long afterwards. Seedlings of Pyrus baccata from seeds obtained in Russia were dis- 

 tributed to Brandon, Manitoba, and Indian Head, Saskatchewan, in 1890. These proved 

 hardy and fruited well in 1894. Crossing in the hope of increasing size and quality while still 

 retaining hardiness was begun in 1894. With Pyrus baccata and Pyrus prunifolia as pistillate 

 parents and 21 varieties as pollen parents, 800 seedlings were grown from the crosses. These 

 were distributed in 1902 to various places at altitudes from 740 to 4200 feet. At the lower 

 elevations in southern Manitoba most of the trees proved hardy and, had the fruit been larger, 

 would have been satisfactory; but at higher altitudes only two of the 17 best proved hardy in 

 exposed places. These two are named Columbia and Osman. Columbia is from the cross 

 Pyrus baccata X Broad Green, a summer Russian variety. Osman is from the cross Pyrus 

 baccata X Osimoe, which is also a Russian variety ripening in August. In 1904 Dr. Saunders 

 recrossed the best of the first generation crosses, using 18 varieties of apples, still with the hope 

 of increasing size and quality and retaining hardiness. There were 407 trees obtained from 

 these second crosses the first of which fruited in 1910. Most of the trees were grown at Ottawa 

 and fruits have been described from 223 of them. The fruit of 171 or 76.68 per cent was under 

 two inches in diameter, or little or no larger than fruit from Fi crosses. The fruit of 48 or 

 21.52 per cent was between 2 and 21 inches and three bore fruits between 2J and 2^ inches. 

 One bore fruit 2.62 inches in diameter. Fruits of most second generation crosses have long 

 slender stems, thin tender skin and the crisp breaking flesh characteristic of crabs. From 

 the very few Fi generation which proved hardy under the most trying conditions, it is 

 doubtful if any second generation seedlings will prove hardy in the high altitudes, but 

 several have fruited at Morden, Manitoba, and it is hoped that some of them will extend 

 the area over which apples may be grown. — Details are given of some of the second gener- 

 ation crosses and it is shown that Pioneer (Pyrus baccata X Tetofsky) X Mcintosh gives 

 a higher percentage of comparatively large apples and higher percentages of fruit of good 

 quality and high color than when Ontario, Northern Spy, or Cranberry Pippin are used 

 as pollen parents. — Results lead to the conclusion that in breeding for quality it is necessary 

 to choose as one of the parents a variety possessing the desired character developed to the 

 highest possible perfection. From experience to date, production of varieties hardy enough 

 to survive the test of the Canadian prairies, by methods thus far used is regarded as doubtful. 

 It is proposed in future work to use Pyrus baccata as the male rather than as the female 

 parent as it is believed a large percentage of the Fi crosses will be apples of marketable size and 

 at least a few of them may prove hardy. — C. S. Crandall. 



915. Marshall, F. R. Some experiences in breeding range sheep. National Wool 

 Grower 8: 35-37. Jan., 1918.— A description of some of the work of the Bureau of Animal 

 Industry at Laramie, Wyoming. Comparison is made of the lambing and shearing records 

 of Rambouillet, Lincoln-Merino crossbreds and Corriedales under range conditions. The 

 records indicate that very fine wool is usually associated with shorter staple and lighter 

 fleece. — G. F. Finlay. 



916. McEwEN, Geo. F., and Ellis, L. Michael. The functional relation of one variable 

 to each of a number of correlated variables determined by a method of successive approxima- 

 tion to group averages: A contribution to statistical methods. Proc. Amer. Acad. Arts Sci. 

 55:95-133. Dec, 1919.— The authors attempt to develop a method of multiple regression 

 independent of any assumption as to form of regression lines. Material is grouped with 

 respect to each independent variable, and simultaneous equations formed expressing group 



