p. Groom 23 



of inaccuracy as to the exact amount of water present in the board at 

 the time of the records. Both these objections are shown in the suc- 

 ceeding remarks to be groundless as regards the essential results obtained. 



Rate of loss of water during measurement and recording of curves. 

 The board was first left in very damp warm air for nearly 24 hours 

 (vi. 21-22) ; on June 22nd during measurement and so forth from 10 a.m. 

 to 11.45 a.m. it lost water at the rate of 1-7 grammes per hour. In 

 this case the water was wholly hygroscopic and in the cell- walls. On 

 other days in summer in the dry laboratory air the board lost water 

 during measurement etc. in most cases at the rate of 7-2 grammes per 

 hour (vi. 22 from 2.15 to 3.5 p.m., and from 3.5 to 3.30 p.m.; vi. 23 

 from 11.15 a.m. to 12.5 p.m.; vi. 25 from 11.15 a.m. to 12.30 p.m.), 

 though on one day (vi. 24 from 11.15 a.m. to 12,15 p.m.) the rate was 

 8'1. In these cases it may be assumed that the loss of water was largely 

 at least at the expense of that present in the lumina of the wood, and 

 at the surface of the sides. These figures render possible an approxi- 

 mately accurate estimate of the amount of water present in the wood at 

 the middle of the time occupied in recording measurements and curves 

 on any day. 



In regard to the estimates of the amount of water in the board and 

 the length the following cases may be discussed. On June 24th 

 observations were made in the following order: at 11.15 a.m. the board 

 was weighed (881-7 g.), then curves and measurements were taken 

 until 12.10 p.m., when the board was once more weighed (874-3) and 

 a second set of curves and measurements were immediately taken. 

 When these last were concluded the approximate weight would be 

 874-3-7-3 -= 867 g. The board was then submerged until 2.30 p.m., 

 when curves and measurements were made before it was weighed 

 (873-69) at 3.30, so that at 2.30 p.m. its approximate weight was 880-9. 

 But whether these corrected or the recorded weights be accepted the 

 changes in linear dimensions are not proportional to the change in the 

 amount of water contained. During this phase these dimensions are 

 determined rather by the exact distribution than by absolute amount 

 of the moisture, probably because the water is largely passing from 

 the lumina into the cell-walls whether the wood be soaking or drying. 



Again on June 28th, 29th, and July 1st, to begin with the measure- 

 ments and curves were taken, then the board was weighed, after which 

 a second set of measurements and curves were made. This second set 

 therefore represents the wood in a drier condition (11, 7-3, and 12-5 

 grammes of water respectively having been lost) so that the linear 



