266 Sturtevant. 



Failure to recognize tliese facts has led to mucli confusion. The 

 frequency of the use of the term "unit character" where gene or "unit 

 factor" is obviously what is meant is an illustration. This term "unit 

 character" is au unfortunate one, as it implies the conception that 

 characters are in some way caused by single genes — a view wliich is 

 not tenable in the light of our modern knowledge, and is, I feel 

 certain, really held by few, if any, geneticists today. It focuses the 

 attention upon the somatic characters of organisms rather than upon 

 their germinal constitution; whereas the present study of genetics is 

 tending in exactly the opposite direction. 



In conlusion, I wish to express my gratitude to Professor T. H. 

 Morgan, under whose direction the work was carried out, for his many 

 valuable criticisms and suggestions. Professor E. B. Wilson has furnished 

 several helpful suggestions on the historical and cytological sides of the 

 problem. I have also been greatly helped by numerous discussions with 

 Mr. H. J. MULLEE and others. 



Summary. 



There are, in Drosophila ampelophila, three groups of genes such 

 that the members of one group show linkage to each other, but not to 

 the members of the other two. 



In two of these groups (the third being not yet sufficiently studied) 

 the genes may be arranged in linear series on the basis of the strength 

 of the linkage shown by the various combinations. 



These facts are taken as strong evidence that the genes involved 

 are located in the chromosomes, each group being in a different chi-omo- 

 some pair. The results are then explained on the liasis of Moeg.\n's 

 application of the chiasmatj^i^e of Janssens to the phenomena of linkage. 



Further evidence is derived from experiments involving three or 

 more pairs of genes in the same group. It is found that as many as 

 three cross-overs may occur in one chromosome pair, but that the 

 occurrence of one break tends to prevent another one. This influence 

 is found to decrease when longer distances are used. 



A general survey of the question of the material liasis of heredity 

 leads to the conclusion that the evidence points to the chromosomes as 

 the beai'ers of Mendelian genes. 



