166 STUDIES IN INDIAN SUGARCANE SEEDLINGS 



incomplete, and it is possible that some of the irregulai'ities noted may be 

 due to the presence of crosses in the general collections. 



The following is the list of seedlings examined : — 



(1) 1911-13. A small lot of Cheni seedlings (about 18) which were very 

 carefully studied, but were grown under somewhat abnormal conditions. 



(2) 1912-14. About 500 Chittan seedlings, over 300 Karun, 50 Kaludai 

 Boothan and 13 Poovan, all general collections in the field, that is without any 

 knowledge of possible local crossings : 50 selfed Saretha seedlings and 65 

 crosses between Shnharchynia and Sacchannn Sjxwlanenm. 



(3) 1913-15. About 80 crosses between Velhii and Saccharion Narenga. 



These seedlings were studied with regard to the following possible correla- 

 tions. The leaf width at maturity was compared with the sucrose in the 

 juice, with the thickness of the cane, the total weight of the seedling and the 

 number of canes and shoots, as indicating tillering power. Other characters 

 compared with sucrose were the length of leaf, the leaf module (length divided 

 by width), thickness of cane, length of cane and cajie module (length divided 

 by thickness), and colour of cane. In a number of these comparisons a distinct 

 correlation has been observable, but it is important to remember that this 

 does not mean that a uniform series was obtained. The greatest variations 

 occurred in all the series, and it was only on taking them as a whole and dividing 

 them into classes, that the tendency was definitely established. AVhile, for 

 instance, the general tendency is for narrow leafed seedlings in any series to 

 have richer juice, this does not preclude the occasional occurrence of excellent 

 seedlings among the broader leafed classes. It is possible that these irregu- 

 larities may be due to faults in analysis or in observation, as suggested above, 

 but they may be intrinsic, and we have at present no means of determining 

 their cause. Naturally, the greatest reliance is placed on the larger series 

 {Chittan and Karun) and it is not certain what is the lowest number of seedlings 

 in which true correlations may be expected. Probably this varies with the 

 difi'erent characters under consideration but, taking the whole series of obser- 

 vations, it is considered that 40 or 50 seedlings Avill usually be sufficient, but 

 that smaller numbers than these are liable to errors, w hich. liowever, can usually 

 be detected on inspection by the occurrence of isolated a))Jiormal seedlings 

 in their wrong place. This has been easily observable in the small Poovan 

 lot, where one seedling, with 20*40% sucrose in the juice, towered above the 

 rest and, to a certain extent, dominated the results according to its position. 



In order to make the general method of calculation clear, it has been 

 thought advisable to give the details of ojie correlation studied, and I have 



