J. F. Martlby 255 



Swietenia. The amplitude is greater in Chloroxylon, reaching a maximum 

 of 40 degrees as compared with 30 degrees in Shorea. 



The net result of a shorter period and a bigger amplitude is that the 

 i.i I e of change in the inclination of the grain will be more rapid, and when 

 this is combined with narrow growth-rings errors in the practical work 

 of investigation will be much more frequent than under the opposite 

 conditions of broad rings and a slow rate of change in the inclination 

 of the grain. 



In spite of these disadvantages in Chloroxylon Swietenia the Plate, 

 fig. II, of the series of curved fractures obtained by the method 

 of radial splitting, shows that the errors from this source are not so 

 serious as might be expected. 



On comparing the position of the periods with reference to tiie growth- 

 rings in the sticks of the two series it is seen that the periods in Chloroxylon 

 are much more closely connected with the growth-rings and show no 

 tendency to cut across periods of growth as in Shorea robusla. 



Exceptions to this contemporaneity of the periods are due to the 

 somewhat frequent appearance of subsidiary or union periods which 

 retain their identity through a series for short distances only, the period 

 appearing at the 44th ring in sticks A3, B, Bl and B2 if the transverse 

 series is a case in point. Many more of these subsidiary periods can be 

 recognised in the curves of the transverse and longitudinal series. 



On account of the frequent appearance of these subsidiary periods 

 it was difficult to obtain data which could be relied on to the relation 

 between period length, amplitude, width of ring and age. 



If the 15th stick of the longitudinal series is taken as representing 

 the average condition of the grain for the series it is seen that long periods 

 occur equally where the rings are broad as where they are narrow. No 

 gradual increase in period length, with age, up to a maximum followed 

 by a subsequent decrease is seen as in Shorea robusta. 



On the other hand it was found on calculating the average that the 

 biggest amplitudes were always correlated with the longest periods. 



GMELINA ARBOREA. 



The examination of Gmelina arborea showed that the grain possesses 

 the same serpentine character that is characteristic of the two species 

 already described. 



The method of radial splitting as applied to a portion of the disc 

 showed that the periods are tangentially continuous at a transverse level 

 (see Plate, III). The series of curves which were obtained from the 



17—2 



