98 GENETICS [Bot. Absts. 



tions are due in part to Castle's use of data that are not significant because of the small number 

 of individuals involved, or because particular characters were sometimes not distinguishable, 

 etc. In part also they are due to his use of data derived from different experiments and hence 

 not safely comparable, where the result turns on very small differences, because of the varia- 

 tion of crossing over values due to genetic differences, environment, age, and differential 

 viability. — Where data are derived from same experiment (i.e., where genes are all followed 

 in the same individuals) arrangement of loci is always in a straight line provided that distances 

 involved are short enough to allow no double crossing over. Since entire X chromosome of 

 Drosophila can be mapped by combining overlapping segments which are themselves straight 

 lines, arrangement of loci is necessarily represented by a single straight line. — In Castle's 

 diagrams the genes which are located on the basis of sufficient data are all arranged approxi- 

 mately in a line — only the line (if he had drawn it) would be curved. This curvature is due 

 merely to existence of double crossing over. Apparent distance between widely separated 

 loci is less than distance obtained by summating the intermediate segments, because double 

 crossing over is not detected when only two loci are followed. Where double crossing over is 

 followed, the total distance is always sum of component distances; i.e., arrangement is a 

 straight line. — Occurrence of double crossing over between widely separated loci explains 

 why observed crossover values have not exceeded 50 per cent. — Castle is unwilling to admit 

 existence of double crossing over; but his attempt to explain small size of smallest classes by 

 single break fails, because combinations of characters impossible on Castle's scheme have 

 already been reported. Castle's own hypothesis to account for double crossing over is incon- 

 sistent with his representation of distances as proportional to crossover values. Moreover, 

 even if Castle's scheme had successfully weathered double crossing over, it would fail to 

 deal with triple crossing over, of which many cases have been found. — Alexander Weinstein. 



667. Tedin, H. Vaxtforadling. [Plant improvement.] Den mindre jordbrukarens hand- 

 bok XXXVII-XXXVIII. Stockholm, 1919. — Contains information about the theory and 

 importance of genetical science, written for farmers. — K. V. Ossian Dahlgren. 



668. Thompson, J. Arthur. [French rev. of: Newman, H. H. The biology of twins 

 (mammals). (Biologie des jumeaux (Mammiferes).) 186 p., 56 fig. University of Chicago 

 Press: Chicago, 1917.] Scientia 25: 511-513. 1919. 



669. Whiting, P. W. Genetic studies on the Mediterranean flour-moth, Ephestia kiihn- 

 iellaZeller. Jour. Exp. Zool. 28: 413^45. 2 pi., 1 fig. July 5, 1919. 



670. Whitney, David D. The ineffectiveness of oxygen as a factor in causing male 

 production in Hydatina senta. Jour. Exp. Zool. 28: 469-492. July 5, 1919.— Food of these 

 rotifers was in all cases green flagellate Chlamydomonas. All rotifers were reared in mass 

 cultures from which random collections were taken to determine sex ratio. Fewer males were 

 produced in sunlight than in darkness, reversing results of former experiments. Reversal is 

 attributed to fact that in former experiments Chlamydomonas was artificially kept active, 

 whereas in present experiments it was allowed to settle on glass and become less available 

 for food. In some experiments oxygen was measured and found to be more abundant in sun- 

 light (2 to 15 cc. per liter) than in darkness (2 to 8 cc. per liter). In other experiments in 

 which oxygen was allowed to dissolve from the air in some cultures but was excluded from 

 others, same proportion of male-producers was obtained in each. General conclusion is that 

 oxygen is not a factor affecting sex. Paper contains brief criticism of methods and conclu- 

 sions of Shull and of Shull and Ladoff. — A. Franklin Shull. 



671. Witte, H. tiber weibliche Sterilitat beim Timotheegras (Phleum pratense L.) und 

 ihre Erblichkeit. [On female sterility in timothy (Phleum pratense L.) and its inheritance.] 

 Svensk Bot. Tidsskr. 13 1 : 32-42. 2 fig. 1919.— In a pedigree of an isolated Phleum pratensi 

 at Svalof author has found several plants which did not produce any seeds, as the female 

 organs were rudimentary. Exceptionally however a single seed could be found. The normal 

 plants produced in average 4895 =*= 260 seeds, and the male ones only 1.7 ±0.92 seeds per indi- 



