392 BOTANICAL GAZETTE [may 



would attach very great importance to glands in a system of 

 classification." 



On the whole, the tendency of later writers has been to attach 

 less significance to glands in classification than has been done by 

 earlier writers. In technical fruit descriptions, or in systematic 

 classifications, it is evident that the value of a character as a 

 distinguishing feature between forms depends largely upon its 

 constancy of expression. Consequently, a statistical analysis was 

 undertaken with the object of determining the number and dis- 

 position of glands in certain species and hybrids available. 



Material 



Data were first collected in 191 4 in the F x generation of crosses 

 between Burbank (Prunus triflora) and Wolf (P. americana), and 

 Abundance (P. triflora) and Wolf. The gland condition was sub- 

 sequently (August 1 016) obtained in an additional number of 



interspecific hybrids. Single 



uniform 



and 400 



number 



from vigorous i-year shoots. By following this method of collec- 

 tion consistently on trees under fairly uniform growth conditions, 

 the data obtained for the different forms are as nearly comparable 

 as can be obtained under field culture. 



number 



5 



ment 



conditions 



produce vigorous vegetative growth favor gland development, 

 since on old trees or on trees subjected to unfavorable growth 

 conditions, the petiolar glands become much reduced, some- 

 times even disappearing, although normally present in the varie- 

 ties. On the other hand, position has an influence on glandular 

 development. Leaves borne at the basal position on terminal 

 growth, on fruit spurs or thorns and also in flower buds, typically 

 bear no glands at all or have them less well developed than leaves 

 borne at other points. 



The arrangement of the glands (that is, whether opposite or 

 alternate on the petiole or leaf) was not recorded. Glands occur 

 both in pairs and alternately, near together or widely separated, 



