72 Mackinder : The Advancement of Geographical Science. 



plants themselves form the required environment of the minor 

 species associated with them. I commend to you the study of 

 these maps themselves, for they will give you a far better idea 

 of the nature and value of this kind of botanical geography 

 than any mere description of mine. Admirable examples of the 

 same kind of work are the memoirs and maps of the late 

 Robert Smith, published in the sixteenth volume of the ' Scottish 

 Geographical Magazine ' under the title of ' A Botanical Survey 

 of Scotland.' Results of this nature, I may point out, are, 

 however, comparatively useless unless the different parts of 

 the country are mapped according to a more or less uniform 

 scheme ; hence the value of the lead which such a conference 

 as this may give to local societies. 



The distribution, however, of plant associations is of com- 

 paratively little value when studied alone. We require for its 

 interpretation a knowledge of the local land forms and drainage 

 systems, of local drift geology, of local climate, and many other 

 local data which can be expressed upon maps. The geog- 

 raphical method of research is to construct with scrupulous 

 care separate maps of each of these orders of phenomena, and 

 then to compare them, when correlations of distribution will 

 leap to notice, and will suggest fresh inquiries. It is obvious 

 that for the study of the causes of local distribution we must 

 often go to historical records, whether embodied in documents, 

 in place names, or in archaeological relics. My suggestion is 

 that the distribution of all these things should be systematically 

 studied upon the map. It is true, no doubt, that maps are 

 attached to many special studies, botanical, geological, or archae- 

 ological ; but the research which I am suggesting treats the 

 comparison of a large number of such maps as its main material, 

 and is not satisfied with having them as incidental illustrations 

 in books of non-geographical aim, and with having them pre- 

 pared according to different methods, and therefore without 

 facilities for comparison. In other words, the object is to 

 have a complete analysis of each district from a geographical 

 standpoint. 



We already have examples of the kind of work which I am 

 indicating, although, as being product in each case of one 

 man's research only, they have not and cannot have the 

 thoroughness and richness which would ensue from the com- 

 bined and prolonged endeavour of one of your societies. Dr. 

 H. R. Mill has described a small part of Sussex in his ' Fragment 



Naturalist, 



