426 Mr, D. Thoday. Experimental Researches on [Mar. 1, 



With the smallest stamp, ditiFerences per square decimetre of —8, +4, -14 and 

 -10 milligrammes, or per cent. -I '9, +1'0, -3'4, and -2'2; with the 20 cm. stamp 

 a difference of +16 milligi'ammes, or +3'2 per cent, (and another so great that I 

 rejected it, as obviously due to mistake in weighing or in cutting out the area). The 

 average difference per cent, was ± 2'3. A comparison of this with the results obtained 

 before by the templet method with H. tuherosus* proves that the smallest stamp, 

 measuring only 111 sq. cm., has given very satisfactory results. 



Of the following experiments, the data are given in detail only for 

 Expt. 2, which is interesting as an example of the use of small areas. In 

 other cases, the o'ate of increase alone is given for each leaf ; the pieces 

 employed were usually 20, sometimes 40 sq. cm. in area, and in some cases 

 a number of pieces were taken together, so that the total area from each 

 half-leaf added up to 90 sq. cm. (Expt. 3). 



Experiment 2. — July 20, 1909. Four small leaves, collected 6.45 A.M. 

 Each half-leaf stamped at 8.50 a.m. Experimental half-leaves exposed 

 at 9.15 A.M. Experiment closed at 4.15 P.M. ; duration, 7 hours. Con- 

 tinuous bright sunshine, some haze. Maximum sun temperature, 27°'2 C. 



Table II. — Area cut from each half-leaf, ll'l sq. cm.f Asymmetry error 

 per sq. decim. per hour : average about + 1 milligramme, maximum 

 about +2 milligrammes. 



a = control half -leaf ; h = experimental half -leaf. 



Leaf. 



Dry weight 

 of ll'l sq. cm., 

 in grammes. 



Gain of dry 



weight 

 in 74 hours. 



Gain per sq. decim. 

 jjer liour, 

 in milligrammes. 



Condition of 

 experimental half- 

 leaf. 



6 



{a) 0-0340 

 (6) 0-0472 



0 -0132 



17 -0 



Turgid. 



7 



(a) 0-0356 

 (*) 0-0457 



0 -0101 



13 -0 



Turgid, then limp. 



8 



{a) 0-0300 

 (4) 0-0361 



0 -0061 



7-9 



Limp. 



9 



(a) 0-0300 

 (4) 0-0300 



-0-0000 



-0-8 



Flaccid. 



As in the first experiment, the rate of increase varied with tlie degree of 

 turgidity, and leaves in similar states gave rates which agree closely in both 

 experiments. The same agreement is to be observed in the other experi- 

 ments. In Table VI (below, p. 430) the results are all tabulated for com- 

 parison. 



* See Thoday, loc. cii., p. 10, Table IX. 



t The pieces cut after stamping witli the stamp reputed to be 10 sq. cm. are found to 

 meaHurc iros hcj. cm. on the average. 



