PARNELL, RASTGASWAMI AYYANGAR, AND RAMIAH 79 



The setting inside such a bag is perfectly normal and the bag may be left 

 on till harvest when it is tied tightly round the bottom, after removing the 

 bamboos, and the whole plant is pulled up and dried. 



This method of selfing has proved very satisfactory, requiring little labour 

 and causing no loss of seed. Only in one or two cases, out of a very large 

 number of plants, have wrong types appeared which may have been due to 

 crossing. It is possible, though not certain, that these were due to accidents 

 of another nature. 



Waterproof paper bags, one over each head, were tried, but, in spite of 

 liberal ventilating holes, the moist atmosphere developed inside very 

 seriously interfered with setting. 



In all cases where the progeny of an uuselfed plan!; are included in a table 

 this fact is indicated by the letter " N " after the reference number of the 

 parent. Such plants have been used freely in tables for simple ratios where 

 a small amount of crossing would only slightly affect the ratio. 



III. CHARACTERS INVESTIGATED. 



(a) Size of outer glumes. 



The spikelet of rice is single flowered and possesses two outer or sterile 

 glumes, an inner fertile glume and a pa lea. In all references to the spikelet, 

 the outer glumes will be referred to as such. The fertile glume and palea are 

 hooked together at their edges to form one common structure, the husk ; they 

 are of similar texture and appearance and are affected by the same characters 

 and will, therefore, be referred to together as inner glumes. 



The outer glumes are usually small and inconspicuous, about i to 4 the 

 length of the inner glumes (Plate I, figs. 2-6). In one of the varieties growing 

 ai Coimbatore, Rakki Pakshi Bhatta, the outer glumes are approximately 

 equal in length to the inner glumes (Plate I, fig. 1). In India this character 

 appears to be met with in only a few rare varieties in different parts. 



In a plot of a short-glumed variety, which was showing considerable 

 variation due to previous crossing, several long-glumed plants were found. 

 Eight of these plants were selected and their progeny, though showing segrega- 

 tion for other characters, were all pure long-glumed. On the other hand, of 

 the short-glumed plants from the same plot, selected for variations in other 

 characters, several gave families showing segregation for this character. There 

 were two very definite groups, short-glumed and long-glumed, in the ratio of 

 about 3 : 1 respectively. The individuals of the dominant group were all 

 typically short-glumed with no distinction between heterozygotes and 



