GENE INTERACTION IN HETEROSIS 321 



wild forms has been somewhat arbitrarily classified into two great groups — - 

 the S. spontaneum group and the S. robiistum group. Each of these groups 

 comprises a diversity of tyj)es which differ among themselves in morphology 

 and in chromosome number. The members of the sponlanenni group have 

 slender stalks; they are often strongly stoloniferous. The members of the 

 robust U7n group have hard, woody stalks, sometimes of good diameter; sto- 

 lons, if present, are not strongly developed. 



The original cultivated varieties likewise may be classified into two great 

 groups. The first of these comprises a number of slender varieties which ap- 

 pear to be indigenous to India, and which have been lumped together under 

 the name S. Barberi. Certain of the Barberi varieties bear a striking resem- 

 blance to the wild spontaneums of that region. 



The New Guinea region is the home of a group of large-stalked tropical 

 cultivated varieties of the type which Linnaeus named S. officinarum. The 

 wild form most closely resembling S. officinarum and possibly ancestral to 

 it is S. robustum, which is indigenous to that region. 



In the closely related genus Sorghum, the difference between varieties 

 having pithy stalks containing but little sugar, and varieties with sweet 

 juicy stalks, has been shown to be determined by a single major gene. In 

 Saccharum the change from the dry, pithy, low-sucrose stalks of the wild 

 forms to the juicy, high-sucrose stalks of the cultivated varieties appears to 

 have been brought about by several, but perhaps by no more than three or 

 four major gene changes. 



The cultivated and wild forms also differ in genes for stalk size. In crosses 

 between the two, the genes responsible for the slenderness of the wild forms 

 show a high degree of dominance. 



A striking feature of this multiform genus is the prevalence of inter- 

 fertility among its members. Widely divergent forms can be crossed without 

 undue difficulty. The resulting hybrids are rarely completely sterile; they 

 are often highly fertile. The explanation is presumably to be sought in the 

 polyploidy which is characteristic of both the wild and the cultivated forms. 

 They range in chromosome number from 24 to 80 or more pairs. It appears 

 that once the minimum chromosomal complement needed to produce a func- 

 tional zygote has been supplied, there is considerable latitude in the number 

 and in the assortment of chromosomes that can be added without impairing 

 the viability, or even the fertility of the hybrids. 



Since the breeder is as yet unable to create superior genes at will, he is 

 obliged to content himself with developing new combinations of the genes 

 available in whatever breeding material he may be able to assemble. The 

 sugar cane breeder is fortunate in having in the wild relatives of sugar cane a 

 reservoir of genes for disease-resistance and hardiness. Those are traits that 

 had to some degree been lost in the course of domestication. Considerable 



