No. 1, July, 1920] * GENETICS 81 



holds that selection is effective in pure lines, at leasl in Bome cases. Genea vary in stability, 

 some yielding to selection much more readily than others. Prom results of '•■•:|>erimental 

 breeding, author comes to the belief thai selection can cause further variability. Opinion 

 is expressed thai selection is the most reasonable explanation of evolution. However, it 

 ca d in >t instil nte new lines of variation l>ut acts in modifical ion of existing types and in direc- 

 tion in which variations tend. — K. Roberts. 



534. CastlBjW. E. Inheritance of quantity and quality of milk production in dairy cattle. 

 Proc. Nation. Acid. Sci. (U. S.) 5: 128 434. Oct., L919. — The material presented is that col- 

 lected in Bowlker's herd of Fi and Fa cows from crosses of pure-bred Eolstein-Friesian and 

 Guernsey animals reciprocally mated. — Mass methods are used throughout the paper. Of 

 31 Fi cows at ages ranging from 2', to :;', years with milk yield over 7- to 12-months perio 

 the average milk production was 6612 pounds of milk, 4. OS per cent butter fat and 270 pounds 

 butter fal in the first lactation. The second lactation average was 8683 pounds of milk an 1 

 pounds of butter fat. — The data for pure-bred Holstein-Friesian herd (25 cows) ranged in age 

 from 2% to 3£ years and in the length of lactation from 8£ to 12 months. In the first lactation 

 the averages were 7673 pounds of milk and 3.4 per cent butter fat (on 8 cows) and 261 pounds 

 of butter fat (on 8 cows). The second lactation, age range 3£ to 4£ years, length of lactation 

 7-£ to 12 months, average for milk 9475 pounds and butter fat 322 pounds. The Guernsey herd 

 (8 cows) ranged in age from 2| to 3 years, length of lactation 7i to 12 months, and average in 

 milk 4617 for first lactation and for the second 5593 pounds. — No correction for increased milk 

 yield with age is made other than that of averaging the parental and Fi herd's milk production. 



Comparison of these average milk yields shows that of the Fj herd to be 1061 pounds of 

 milk less than the average yield of the pure Holsteins, but 1995 pounds more than the average 

 yield of the pure Guernseys. The average butter-fat percentage content indicates an aver- 

 age butter-fat content of 4.08 per cent, which is slightly less than intermediate between the 

 3.4, the butter-fat percentage of the pure Holsteins and 5.0, the butter-fat percentage of the 

 pure Guernsey. In those Fi cows which have a second lactation the approach to the higher 

 milk yield of the pure Holstein-Friesian is even more striking. — In butter-fat production 

 (milk yield X butter-fat percentage) the Holsteins averaged 322 pounds, the Guernseys 280 

 pounds, the Fi 363 pounds. — These results seem to be considered as due to an increased vigor 

 of the Fi hybrids due to the out-cross. — Comparisons were made of the milk yield and butter 

 fat of Fi cows coming from the cross Holstein cow X Guernsey bull and its reciprocal. The 

 milk yield and butter fat were found to be approximately equal in the two crosses, variations 

 occurring in both directions. From this the conclusion is made, that it seems unlikely that 

 any sex-linked factors are concerned in the case. — John W. Gowen. 



535. Castle, W. E. Are genes linear or non-linear in arrangement? Proc. Nation. Acad. 

 Sci. [U. S.] 5:500-506. Nov., 1919. — Rebuttal to criticisms made by Morgan and his asso- 

 ciates to conclusions of a former paper by Castle on same subject. Position taken by Castle 

 and here reaffirmed is (1) that forces linking genes together are possibly molecular rather than 

 mechanical; (2) reconstruction of sex-chromosome data of Morgan and Bridges proves arrange- 

 ment cannot be linear; (3) values of crossovers greater than 50 per cent necessitated by linear 

 hypothesis have never been observed and are logically impossible; (4) non-linear hypothesis 

 is simpler because it eliminates secondary hypothesis needed on linear hypothesis to har- 

 monize greater with lesser crossover values, particularly the hypothesis of double crossing 

 over. Little new data are presented in this discussion, attention being devoted chiefly to 

 extensions and further exposition of material presented previously in support of above 

 propositions. — John W. Gowen. 



536. Chace, E. M., and C. G. Church. Notes on California and Arizona grapefruit. 

 California Citrograph 3: 200-201. 5 tables, 2 diagrams. July, 1918.— On maturing and com- 

 position of fruit; includes analyses of nine "off types" of Marsh pomelo, in comparison with 

 "standard types," showing inferiority of certain bud-variation strains. — Hoirard B. Frost. 



BOTANICAL ABSTRACTS, VOL. IV, NO. 1 



