1909.] 



Vegetable Assimilation and Respiration. 



49 



very different types, the latter somewhat extreme. With the former an error 

 of + 2'5 milligrammes is + 0'8 per cent, of the dry weight of a square decimetre. 

 With the latter it is +0'3 per cent. 



When the stamping method is employed, whereby shrinkage errors are 

 entirely eliminated, the asymmetry error alone remains. In this case, the 

 maximum error from asymmetry must be less than +0-8 percent, of the total 

 dry weight for Heliantlius, less than +0"3 per cent, for Cherry Laurel, and the 

 probable error not greater than about +0'5 per cent, and +0'2 per cent, 

 respectively. 



The average degree of asymmetry to be expected in individual leaves is 

 about + 2 per cent. Apart from the possibility of reducing asymmetry 

 errors by avoiding prominent ribs, for which some experimental evidence has 

 been given, the other means of reducing them is to use a large number of 

 leaves. The probable error is thus reduced in the inverse ratio of the square 

 of the number of leaves used. To bring the average error from asymmetry 

 within the required limits, 16 leaves of Helianthus, or 100 leaves of Cherry 

 Laurel, would therefore be required. 



There are obvious practical objections to using a very large number of leaves, 

 since the time taken in cutting out stamped areas with scissors is a considera- 

 tion of real importance ; as also in cutting out pieces by means of templets, 

 or in taking photographic prints. 



Moreover, with large-leaved plants like Hdianthus anmius, bearing com- 

 paratively few leaves of which but a small proportion are perfectly sound and 

 flat, the necessity of using such a large number of leaves would be a great 

 drawback. 



One hundred leaves of Cherry Laurel, too, would be quite unmanageable by 

 the stamping method. On the other hand, with an improved templet or the 

 rotating punch, it would be possible to manipulate a number of these fairly 

 firm smooth leaves at the same time. With such leaves, which as a rule also 

 show but a small degree of shrinkage under insolation, these methods are to 

 be recommended in preference to the stamping method, a correction for area 

 changes being estimated from measurement of representative leaves. 



The accuracy with which shrinkage can be allowed for by measurement can 

 only be roughly gauged. The error remaining after approximate correction is 

 probably well within +0-5 per cent, of the total area, and therefore less than 

 one-fifth of that due to asymmetry. 



The conclusion seems unavoidable, nevertheless, that the probable error 

 from all sources can in general only with difficulty be reduced, in a five-hour 

 experiment, to so low a figure as + 0'5 milligramme per square decimetre 

 per hour. To attain to this degree of accuracy without making the method 

 VOL. LXXXII. — B. E 



