482 



On the Composition of Linseed Oil- Cake, 



never be wanting in chemical papers, wheresoever they may be 

 published, I have made a point of recording the double result. 

 It has not appeared necessary to make a second analysis for oil ;* 

 but some cases are mentioned which show that the liability to 

 error is not great. f 



In order that any differences between the varieties of cake 

 might be perceived, I have grouped them according to the dif- 

 ferent countries from which they were imported. 



The first table contains the analyses of eight specimens of 

 French cake, all of them imported this year, and the growth of 

 1848. 



Table 1. — Analyses of French Linseed Cake. 



Any remarks about the extent of variation will come better 

 when the whole evidence is before us ; in the meanwhile, the 

 mean composition of the above may be given. 



In the Table is one very high and one equally low per-centage 

 of nitrogen. These will in great measure balance each other, 

 and not greatly affect the result. The mean of eight specimens 

 of French cake is — 



* The oil was obtained by extraction with ether in the usual way, 

 I It may appear to some unnecessary and objectionable that the pages of the 

 Journal should be occupied with the detail of experiments where the results only are 

 immediately serviceable. To these it must be answered, that in all scientific journals 

 it is usual to give the most minute details, in order that the reader may possess ample 

 means of discovering error, whether of experiment, calculation, or opinion ; and this 

 amplification forms, in the eyes of the scientific man, an indispensable requisite of sci- 

 entific writing. Neither is it just to the writer that the evidences of his labour should 

 in large measure be suppressed, for it must be remembered that, when the results of a 

 research have been published in this Journal, the details are of no value to any other; 

 and as they would never meet the public eye, all trace of them would be eventually 

 lost. These remarks apply more to other instances than the present, where the double 

 results, being tabulated, occupy no additional space. 



