252 Double Cross-Grain 



average (Table V) amplitude remains approximately constant at the 

 successive transverse le\els in the longitudinal series, its value ranging 

 between 15 and 18 degrees, but on the other hand the average value of 

 the successive swings increases with age up to a maximum of 25 degrees 

 which is followed by a slight subsequent decrease in value. The figures 

 of Table VI show that the same is true of the transverse series. Although 

 both period length and amplitude reach their maximum value at about 

 the same time, yet the two do not appear to be very closely correlated, 

 for tbe subsequent decrease in the period length is so much greater than 

 the decrease in the amplitude that the ratio of period length to amplitude 

 steadily decreases with age except for the last period of the transverse 

 series which shows a slight decrease. 



On comparing together the average period length and the average 

 amplitude, there are indications tbat in the transverse series the longer 

 periods are correlated with bigger amplitudes, but no such correlation is 

 apparent in the longitudinal series. 



The correlation of period length and amplitude with width of ring 

 is much more indefinite and is only recognisable in so far as all three 

 tend to increase with age up to a maximum which, is followed by a 

 greater or smaller subsequent decrease. 



CHLOROXYLON SWIETENIA. 



The curves of the longitudinal and transverse series are given in 

 Figs. 5 and 6 respectively, but the data from which they have been 

 constructed are not printed here. 



The numbering of the rings in the two series is practically identical as 

 the several zones of narrow rings afforded reliable points for comparison. 



Due to the uniformity in the rate of growth, the proportional width 

 of the rings remained the same along the different radii of the transverse 

 disc, so the ring widths were only measured along three of the sticks. 

 In the longitudinal series the ring widths were fully measured along one 

 stick from the middle of the series. 



An examination of the longitudinal and transverse series of curves 

 shows that the course of the grain in Chloroxylon Swietenia is very similar 

 to what it is in Shorea robusta. The periodic changes in the inclination 

 of the grain are continuous both longitudinally in a radial plane and 

 tangentiallv at a transverse level. The inclination of the grain at corre- 

 sponding points of a period does not remain the same in adjacent 

 sticks of the transverse series or at the different transverse levels in the 

 longitudinal series, a striking instance occurring at the 100th ring of the 



