4 20 TRANSMISSION 



SECTION I TYPE 



What now is our conception of type ? If ten-inch ears will 

 not produce ten-inch ears, but something else, and not only 

 something else but a considerable variety of lengths ; and if 

 what we get extends both above and below the parent, then we 

 arrive at once at a double conception as to type ; that is to say, 

 the type of the offspring is not the same as that of the parent. 

 The type of the parent is very definite, representing an ideal ; 

 but if the offspring is distributed both above and below that 

 ideal, some being better and some not so good, then a close 

 analysis of the real character of that offspring becomes necessary 

 in order to make any just comparison between the two, or to 

 arrive at any adequate conception of type in a mixed population, 

 even in one arising from a selected ancestry. 



A concrete case will serve best to illustrate the principle 

 involved. In the year 1906 some Learning corn was raised on 

 good ground from seed ears exactly ten inches in length. A 

 random sample 1 of this crop, consisting of 327 ears, gave the 

 following distribution as to length : 



One ear was 3.0 inches long, one was 4.0 inches, two were 

 5.0 inches, three were 5.5 inches, nine were 6.O inches, eight 

 were 6.5 inches, twelve were 7.0 inches, nineteen were 7.5 

 inches, thirty-two were 8.0 inches, forty were 8.5 inches, 

 sixty-seven were 9.0 inches, sixty-three were 9.5 inches, 

 thirty -eight were 10 inches, twenty-one were 10.5 inches, eight 

 were n.o inches, two were 11.5 inches, and one was 12.0 

 inches long. 2 



1 By a " random sample " is meant a sufficient portion of the whole, taken so 

 much at random as to fairly represent the entire crop, or "population," as the 

 technical phrase goes. 



2 Measurements might be taken at quarter inches with a seemingly higher 

 degree of accuracy, but repeated trials show that the same final results follow 

 whether measurements are taken at the quarter inch or at the half inch. The main 

 point is that the numbers shall be sufficient and that the sample shall be repre- 

 sentative. Judgment must dictate as to the accuracy of the sample, but the num- 

 ber depends upon the degree of reliability desired. This matter will be fully 

 discussed under the subject of probable error, but experience shows that in studies 

 with corn excellent results can be had with from 200 to 300 ears, and very fair 

 results may generally be had with half that number. 



