dufrenoy: metaphanic variation in grasses 



535 



largest families, containing 297 guinea-pigs. These contained, 

 between 32 and 45 guinea-pigs each. Naturally, the coefficients 

 varied greatly but all but one were positive and the weighted 

 average came out + 0.256 ± 0.036, showing a substantial cor- 

 relation. This, however, does not use all the data. A calcula- 



table i 



Constants Used in Calculations 



Total (560 pigs)... 

 24 family means. . . 

 Average family (de- 

 duced) 



Average 8 families 

 with 297 pigs 



by 



+ 0.375 ±0.024 

 +0.630 ±0.083 



+0.308 ±0.026 

 +0.256 ±0.036 



tion directly from the correlation and standard deviations of the 

 total population and the array of means, by the method suggested 

 in this paper, gives + 0.308 =•= 0.026 as the average within a 

 family. This agrees reasonably well with the first calculation, 

 uses all the data and involves very much less labor. Table 1 

 shows the constants used in calculating the average correlation 

 within families and the averages derived from the 8 largest 

 families as a check. o- b and <r y are the standard deviations, in 

 grams, of birth weight and year weight respectively. 



GENETICS. — A new case of metaphanic variation in grasses 

 and its significance. J. Dufrenoy, Station Biologique 

 d'Arcachon. (Communicated by K. F. Kellerman.) 



In his recent work on the Hybrids of Zea ramosa and Zea 

 tunicata, G. N. Collins described these two mutants of maize as 

 being ancestral forms. In crossing them "the hope was enter- 

 tained that their combination might bring to light still other 

 ancestral characters." This hope was not fulfilled, but important 

 data were obtained and the conclusions brought forth seem to 

 have a wide bearing. 



