92 



JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



The numbers with asterisks are omitted in calculating the means, 

 as the foliage had become partly flaccid, and they are undoubtedly too 

 high. It was on these two occasions that I first reahzed the important 

 difference between evaporation from dying or dead matter and transpira- 

 tion from an actively growing or living plant. Subsequent experiments 

 have proved that evaporation proceeds much more rapidly when life is 

 enfeebled or extinct than when the plant is alive. In the latter case 

 this pm^ely physical process is to some extent kept in check, while 

 transpiration, so to say, takes its place. 



Taking the mean of the three series, the results are as follows : — 



R. Y. G. B. V. Ci. 



5-09 4-46 4'82 4-70 666 5-20 



These give decided maxima under red, violet, and clear glass, with 

 minima under the central portion of the spectrum. 



Box. 



Six small box plants, well rooted, were grown in pots as described, 

 with the following results after twenty-four hours' exposure in each 

 case. 



Eange of temperature oT^-GS*^. 



No. 



R. 





G. 



B. ' 



V. 



Cl. 



I. . 



•75 



•65 



-76 



-81 



1-00 



•74 



II. . 



1-37 



1-37 



1-38 



1-23 



1-52 



1-14 



III. . 



1-30 



•88 



1-21 



1-04 



1-48 



1-17 



IV. . 



1-71 



1-96 



2-63 



2-47 



2-82 



1^43 



V. . 



3-48 



3^64 



3-72 



3-54 



3-15 



2-40 



VL . 



2-58 



2-91 



2-33 



2-83 



2-67 



2-35 



Total 



11-19 



11-41 



12-03 



11-92 



12-64 



9-23 



Mean 



1-36 



1-90 



2-00 



1-98 



210 



1-54 



In tills, evergreen with a thick cuticle to the leaves, the differences 

 are not so^ pronounced; but violet glass still shows a maximum. 



Experiments with palms also showed the necessity of making a 

 prolonged series with the same plants ; for the totals might even taken 

 alone lead us to suppose that the yellow glass was most favourable of 

 all the colours; but variations of sunlight, coupled with slight varia- 

 tions of temperature, may account for it. Thus, when the yellow 

 glass gave a loss of "97 gramme, and red only "70, the temperature 

 ranged from 63° to 66° with the form.er, and only from 61° to 64° 

 with the latter on June 28, 1883. Again, yellow^ gave a loss of '88 gr. 

 on July 16, with a temperature ranging from 57° to 62° ; but on the 

 day before the red glass gave a loss of '75 gr., while the temperature 

 ranged from 59° to 64°; hence, in all probability, it was a duller day. 



These irregularities, however, become eliminated when so large 

 a series of observations as eighteen for each colour are taken ; for 

 I consider the proof of their elimination to reside in the general 



