C. A. BARBER 



147 



ranL't'ot variations in an^'batcli ot t!io sanio ])ar<'ntafr<' is nion' <ii' less constant, 

 the tliiiuipst cant- Ix'inir. niuirlily. hall tlic diameter of tlie thickest. Th<' 

 thickness was th'teiniined hy s|H'ea(lin<: out 20 canes. seh'Ctitiff the avera^f 

 and measuring' it at themiddh'. The accdnipanyint: }th(it(iirra])h (Pl. XX^'TT1) 

 ^ives some iih'a of the \-ariatioii in seedlinLrs ol the same parentage in this and 

 other ies]iects. 



Thickness of cdnes in seedhiif/s (ind jxirenis. 



The above mentioned are the chief characters in which voiinrr J^eedlinos 

 have been found to varv. Other minoi- difierences, sncli as h'ligth of internode 

 (usually .({reater in Indian canes), flatness of ]ea\es (whether incurved. levolute 

 or cnirn]>led). colourings due to jninoi leaf fungi or insect or other attacks, 

 and so forth, have })een noted, but tlie lesiilts are at present too ^•aglle to be 

 used in classification. The whole of the 3.400 seedlings planted out in 191J-1B 

 have been diA ided uj) into gi-ou])S according to these characters, and the 

 following classes in 500 Striped Mouiifivs seedlings (general collection) may 

 sei've as a type of the method adopted. The seedlings were examined in two 

 lots, of 300 and 200. which accounts for du])licati()n of classes. 



Class 1. — (40 seedlings) Some ]nir))le in the leaf. 

 Class 2.— (50) Plants with bright ])ink leaf sheaths. 

 Class 3. — (9) Some striping in the leaves. 



Class 4. — (30, practically all that were discoverable) Sheaths green, with 

 hardly a trace of purple or pink. 



