46 



The [(H'rxal ()i- I Ikukditv 



is never grown comnierciall\-. Bloom- 

 field is in the center of the tobacco- 

 growing region and the soil is jierfectly 

 adapted to it and heavily fertilized. 

 The soil at New Ha\'en is a i)oor gra\'el- 

 ly loam, onl\- moderateh' dressed with 

 manure and chemicals. 



The means of the dift'erent selections 

 compared with the average mean show 

 only a variation of 1.8 leaves per plant, 

 and as only about one hundred and 

 fifty ]3lants were counted the results 

 appear very uniform. 



STANDARD DEVIATIONS. 



As a measure of variability the stand- 

 ard deviations of the selections were cal- 

 culated. A consideration of them shows 

 the variation to be very uniform and 

 little, if any, afifected by difTerent con- 

 dition of environment. We may also add 

 that these same selections have now 

 been grown at the three different locali- 

 ties for two later generations without 

 any evidence that environment very 

 strongly modifies the number of leaves 

 per plant. 



Some observations have also been 

 made upon the effects of different en- 

 vironments on plant height. Two se- 

 lections were grown both at Bloomfield 

 and in New Haven from seed of the 

 same self-fertilized plants. These se- 

 lections were very variable, due to the 

 fact that they represent the second 

 generation of a cross between two 

 varieties. While difference in environ- 

 ment has very strongly modified the 

 heights of plants, a consideration of the 

 Standard Deviation convinces one that 

 such modifications of this character 

 as have a])i)eared ai)])ear alike in all 

 jjlants of a selection. 



Although no statistical studies can l)e 

 given showing the elTects of environ- 

 ment on the same selection for leaf 

 area, we have some observations on 

 several pure lines of tobacco varieties 

 grown under cultural conditions difTer- 

 cnt from those under which the varieties 

 are normally grown. 



These are pure lines of a Connecticut 

 Havana type grown under shade, of a 

 small-leaved Sumatra tyi)e grown in the 

 oi^en, and a Broadleaf tyi)e grown on a 

 poor sandy soil. By means of a plani- 

 meter the area of the bottom, middle 

 and top leaves of one huiidrccl and fifty 



plants of each of these pure lines was 

 taken. Of course, the area of leaf is a 

 character which is greatly modified by 

 cultural and en\'ironmental conditions. 

 The results, however, show about the 

 same variability for the three pure lines 

 given. Observations of these pure lines 

 when growing showed no variation 

 which could be considered as a breaking 

 up of type. 



VARIATION DUE TO CROSSING. 



The study of the heredity of separate 

 characters of tobacco shows that such 

 characters as shape and size of leaf, 

 size of stalk, and leaf number are in a 

 large measure inherited indei)endently. 

 In brief, the results show that the first 

 generation of a cross between two pure 

 lines is of an intermediate nature in 

 those characters by which the i)arents 

 differ. In the second generation there 

 is a segregation and recombination of 

 characters and often new forms are 

 produced. Some of the second genera- 

 tion forms breed true in the third gen- 

 eration, others breed true for some 

 characters and are variable in the re- 

 mainder, and others are as variable as 

 the second generation itself. The 

 length of time which it takes to produce 

 a uniform type will depend largely on 

 the number of varieties which can be 

 grown in F2 and the niunber of row 

 tests which can be grown in F3. 



The general conclusions from this 

 work are that environment is of great 

 im])ortance in any s>'stem of tobacco 

 breeding, and quantitative characters 

 and especially quality of cured leaf are in 

 a large measure dependent on this 

 feature. Change of environment, how- 

 ever, does not cause a breaking up of 

 tyi)e, and whatever variations occur 

 due to environment ajipear alike in all 

 l)lants of a particular tyi)e. 



Heredity is the second important 

 factor, and ]30or tyj^es will give unfav- 

 orable results even under the best 

 en\'ironmental conditions. Any system 

 of tobacco breeding must take both 

 heredity and environment into account. 



The only known means of jjroducing 

 variability as a source of new types is 

 by crossing. The number of new fonns 

 which will appear due to a i)articular 

 cross will de])end on the numl)er of 

 germinal characters l)y which the i)arent 

 plants dilTcr. 



