1911.] The Mechanism of Carbon Assimilation. 



Ill 



interchanged as in the film experiments. The maximum difference of 

 temperature recorded under these conditions was o- 06, an amount too small 

 to affect the conclusions. 



Conclusion. 



The present paper is concerned only with the initial stages of the 

 assimilation process, and therefore no reference has heen made to the 

 synthesis of sugars or of starch. The particulars in which the conclusions 

 given in Part I require modification have already been noticed — the 

 exclusive localisation of catalase in the chloroplasts is abandoned, and also 

 the dependence of the post-mortem bleaching of chlorophyll on the presence 

 of carbon dioxide. 



Finally, there now appears to be ample justification for considering that 

 the primary products of the photolysis of aqueous carbon dioxide are 

 formaldehyde and hydrogen peroxide ; that the evolution of oxygen is due 

 to decomposition of the latter substance by catalase ; and that up to this 

 point the process is entirely non-vital, and can be reconstructed in vitro. 



In a paper published in 1907, A. J. Ewart (' Koy. Soc. Proc.,' B, vol. 80, p. 30) has 

 criticised most of the experiments and all the conclusions recorded in Parts I and II. As 

 it is impossible to answer all the objections within the limits of a short paper, replies to 

 a few of the more important are briefly indicated below, (i) The experiments upon 

 which Ewart bases his opinion that formaldehyde was not produced as a decomposition- 

 product of carbon dioxide under the conditions named in Parts I and II are quite 

 valueless for that purpose, because he used the reagent (decolourised rosaniline) in such a 

 way that a colouration would inevitably be produced in the material to be tested. 

 Since the sulphur dioxide (used to keep the rosaniline decolourised) escapes from the 

 solution on warming, or when a large surface (compared with the volume) is left exposed 

 to the air, the method employed and described by Ewart (pp. 30 — 31) is clearly 

 inadmissible. It may be as well to state here that all the materials — gelatine, petroleum, 

 ether, etc. — used for the experiments described in Part II were tested with the same 

 specimen of SchifPs reagent which was afterwards used to detect formaldehyde, and were 

 only employed if found to be initially free from that substance. In view of the more 

 recent experimental work of Schryver referred to in this paper, it seems unnecessary to 

 discuss the subject at greater length here, (ii) The phenomenon of the bleaching of 

 chlorophyll, and its explanation, have already been dealt with (p. 104). (iii) With regard to 

 the production of hydrogen peroxide, Ewart is mistaken in supposing that we were 

 " unaware that the absence of hydrogen peroxide from living cells has been definitely 

 established " : ou the contrary, it was expressly stated that an enzyme was present whose 

 function was to decompose that substance as fast as it might be formed. Since 

 chlorophyll is itself attacked (bleached) by hydrogen peroxide, the latter has also escaped 

 detection when the enzyme has been destroyed, (iv) The extra-cellular evolution of 

 oxygen.— The experiment described in Part II has been misrepresented in important 

 particulars by Ewart (p. 34), and in his attempt to repeat it the conditions of the original 

 ■experiment were not observed, (v) The simultaneous production of formaldehyde and 

 hydrogen peroxide is objected to (p. 35) on the ground that these two substances under 

 certain conditions interact, and form carbon dioxide and hydrogen (or, ultimately, carbon 



