408 



X. GENERAL REMARKS. 



namely, to consider as primary orders all those which 

 include one per cent, or upwards of the total flora. But 

 any such rule is of course arbitrary, and can be adhered 

 to only as a practical convenience. On turning to the 

 census, page 359, it will be seen that we should stop at 

 PrimidacecB, the twenty-third order, if rigidly restricting 

 comparisons to those British orders which exceed the 



