426 TRANSMISSION 



In the study of variability it is worse than useless to study a 

 few scattered individuals here and there. What we seek is a 

 measure of what may be called the average tendency to deviate 

 from type. Some individuals deviate but little, others more, 

 and still others very much, and we seek a measure of this non- 

 conformity to type. To find this we must study groups of in- 

 dividuals sufficiently large to be representative of their race. 

 This brings us back to the frequency distribution and what it 

 can teach as to variability. 1 



impression that if the law of heredity were infallible all offspring would be of a 

 common type, and that any departure from the type of the race, variety, or breed 

 is to be regarded as by so much a failure of heredity and a concession to variation. 



The truth is that all transmission is heterogeneous in the sense that the indi- 

 viduals of any race, whether parents or offspring, belong not to a fixed type but 

 to a frequency distribution similar to the one now under discussion. The idea of 

 type thus arises out of the distribution, and it constitutes a convenient base from 

 which to reckon deviation. 



The chief conception to rest in the mind of the student at the present stage of 

 matters is that, whatever the parentage, the offspring will constitute a distribution 

 extending through a considerable range, and that the parent itself also belonged 

 and was drawn from some portion of a frequency distribution not very different 

 from that of the race in general. 



Variability is, therefore, not the opponent of heredity but its inevitable 

 accompaniment in transmission, and our problem is to devise methods of accu- 

 rately measuring and expressing its range and extent in any particular instance. 



1 No apology is made for introducing the so-called statistical method of study 

 at this point ; first, because it is the only reliable method of attacking problems 

 in transmission, and second, because this method is everywhere coming into use 

 among careful students. The reader is urged, and the student should be required, 

 not to evade this portion of the subject because the method of treatment may 

 happen to be unfamiliar. On the other hand, he is urged to familiarize himself 

 not only with the method of work but with the point of view involved. If he will 

 do this, both variability and later on correlation and heredity in general will come 

 to have a new meaning, and one far more rational and comprehensive than the 

 hazy notions evolved from the unsystematic study of isolated individuals. The 

 principles involved are for the most part simple, and in this elementary treatise 

 every effort will be made to treat the subject from the standpoint of the non- 

 mathematical reader. 



For the convenience of those who may care to pursue a little further some of 

 the more strictly mathematical conceptions involved, an Appendix has been pre- 

 pared by Dr. H. L. Rietz, of the mathematical department of the University of 

 Illinois. 



The statistical method of study of problems in heredity, as distinct from the 

 strictly biological, was introduced by Dr. Francis Gallon of England (see Natural 

 Inheritance, 1889), and afterward much extended by Karl Pearson and others 

 (see especially Grammar of Science and Philosophical Transactions of the Royal 

 Society). It is now coming into such common use that a quarterly journal 



