1909.] Vegetable Assimilation and Respiration. 41 



Templet, 20"1 sq. cm. 



19'7 sq. cm.") 



19-5 „ I . , 



iQ-« ^Eectangles printed together in same frame, 



ly D „ I 



19-6 „ J 



Maximum difference from average +0"1 sq. cm., i.e. +0'5 per cent. 



The greatest error revealed here is O'o per cent. ; and even this is likely 

 to be an excessive estimate of the errors of the templet method itself, since 

 changes of area from decrease of tension and evaporation of water may vary 

 somewhat from piece to piece. The estimate thus includes errors in the 

 method adopted to test the templet method, as well as the real errors of the 

 latter. 



It may be concluded, therefore, that the greatest error involved in cutting 

 pieces from leaves with templets as small as 40 and 20 sq. cm. is less than 

 O'o per cent, of that area ; and in comparing two individual pieces the 

 maximum error is less than 1 per cent. Such errors, therefore, only account 

 for a small part of the differences measured in asymmetry tests, which reach 

 maxima of 4 per cent, or more. 



The magnitude of the error may be expected to vary witli the character of the 

 venation. If outstanding veins are excluded altogether, the error from extensibility is 

 reduced to a minimum. The only other sources of error are the sloping of the cutting 

 instrument, and inaccui'acy in following the edge of the templet : such errors are 

 reducible with careful cutting within very small limits. All the errors depend upon 

 the perimeter, and so are less in proportion for larger areas. 



Where outstanding veins were few, templets as small as 10 sq. cm. have been used, and 

 the results of the asymmetry tests with Paulownia imperialis* warrant the inference 

 that the use of larger templets may be a distinct disadvantage from the point of view of 

 asymmetry if veins are included that are at all prominent. These asj'mmetry deter- 

 minations include the errors of measurement of area, but the differences are far too 

 great to be accounted for by these errors, and it may safely be concluded that any 

 possible disadvantage from the relatively greater perimeter, when small templets are used, 

 is not to be compared with the advantage of reducing asymmetry by avoiding veins.t 



* Pp. 17 and 18. 



t For small leaves, Muller felt that the chance of the cutting error becoming serious if 

 small templets were employed made some other method desirable Hoc. cit.). He took 

 sunprints of half-leaves, and estimated the area by cutting out the print and comparing 

 its weight with the weight of a known area of the .same paper. He found that when 

 equal areas were cut out of different sheets from the same packet, the maximum error 

 was equivalent to 0"52 sq. cm. in an area of 40 sq. cm. This is equivalent to 1'3 per cent, 

 of the total area. It would seem, therefore, even taking his own estimate of the 

 maximum error in the templet method, 1-5 per cent., that the only advantage of his 

 printing method lies in the use of entire half-leaves, instead of smaller areas cut from 

 them ; but his estimate is excessive, and the use of small templets is in reality much more 

 accurate. 



