348 



B. ORCHARD 



effect (54%) in 1959 and the least (25%) in i960. The effect in i960 was 

 not significant, probably because of the large initial differences between 

 plots but there was however a significant, positive correlation between L 

 and plant moisture content at the first sampling. Treatment C had very- 

 similar maximum effects upon leaf area index in different years (30, 36 and 

 35% respectively). After watering, the decrease in L from treatment B 

 disappeared within 21 days. The high relative leaf growth rates (Fig. 5) 

 during the recovery period were transient and the B plots never attained a 



D 



a 



Q. 



40 



30 



20 ■ 



Expl. I. 1958 



t T 



y B.Watered 

 C. Started 



I I I L. 



Expt.2. 1959 



J I I 1 1- 



Expt 3 I960 



I July I Aug I Sept 1 July I Aug I Sept Oct I Nov I Dec I 



Fig. 6. Leaf number per plant, (i960 treatments B, C, not shown). Symbols as Fig. i. 



significantly greater L than the A plots. Recovery from the BL drought in 

 i960 was incomplete and temporary, as the leaf area decreased at the final 

 harvest to that of treatment C. The water appHed during the eleventh week 

 of treatment C in 1958 checked the decrease in L. 



In 1958 and 1959, both treatments B and C decreased the area per leaf 

 but only C decreased the leaf number significantly (Fig. 6). Leaf number 

 also increased rapidly after watering and occasionally sHghtly exceeded that 

 of treatment A, as in 1959 when with treatment B leaf number made a 

 transient but just significant contribution to recovery. 



