88 RET. G. HENSLOW ON THE 



and accurately weighed to the 500th part of a gramme. The tube 

 with the shoot or leaf thus prepared was enclosed in a box 

 covered above with the sheet of coloured glass. 



Although experimenters have so generally employed cut shoots 

 and leaves, I soon found, on making my calculations from data 

 accumulated from the weights, that although some may appe;ir 

 quite fresh to the eye for three or even more days after the com- 

 mencement of the experiment, yet the vitality of the shoot or leaf 

 had, nevertheless, been becoming enfeebled all the while, and the 

 amount of water transpired steadily decreased day by day 

 irrespective of the characters of the coloured glasses used; so 

 that finally the relative amount of loss in successive days 

 became untrustworthy. 



To ascertain the difftrences in the amount of water trans- 

 pired under the influence of difterent rays of the spectrum, pro- 

 longed and repeated exposure of the same specimen to the same 

 kind of light is necessary. The reason for this is obvious ; for 

 there are (as stated above) so many disturbing influences which 

 may materially affect the results, that unless they be reduced 

 to a minimum, the effect due to colour alone cannot be even 

 approximately ascertained. 



If, however, the experiments be carried on under conditions 

 which will reduce the above disturbing elements to a minimum, 

 then the differences due to the colour of the li^ht transmitted 

 will be the most powerful agent in the process of transpiration ; 

 and it is only by taking the mean of many experiments that the 

 above influences can he virtually eliminated. 



In attempting to do this with cut specimen?, none will be 

 found to last long enough unaffected by the lesion to give very 

 trustworthy results. Hence, although such experiments may 

 tend to corroborate those obtained by more perfect methods, yet 

 I do not think it worth while to enumerate more than a brief 

 series of my os\t\, numerous as they have been, as by themselves 

 they would not furnish a sufficiently :icctirate basis for induction. 

 They may have their use, however, in showing the somewhat 

 negative, or at least uncertain, results which are generally only 

 obtainable from cut specimens. 



The next method I adopted was to take small plants with their 

 roots carefully lifted and freed from soil, and to insert them in 

 tCBt-tubes as before. The results were even more unsatisfactory 

 than with cut specimens, for, excepting liadish (the thick root ol 



