CHANCrS IN FOREST ARECA IN NFAV ENGLAND 445 



Although the change in the amount of woodland in New England 

 has been observed by many, it does not seem to have been shown statis- 

 tically very well hitherto. The writer has recently extracted from 

 census reports some interesting information on this point, and as the 

 results are obtained in a rather roundabout way. the method will be 

 explained in detail, so that readers may judge for themselves how much 

 confidence can be placed in it and perhaps at the same time think of 

 improvements. 



The first census of the United States (which was the first of similar 

 scope in the world) was taken in 1790. A few estimates of the popu- 

 lation of the New England colonies had indeed been made before that, 

 but if we disregard them and assume that the population of each in- 

 creased uniformly from the time of first settlement to 1790, we will 

 probably be near enough to the truth. The first six decennial censuses 

 dealt chiefly with population, but that of 1850 gave the acreage of im- 

 proved and unimproved land in farms in every state and county, 

 improved land being that cleared and used for grazing, mowing or 

 tillage, or lying fallow, and unimproved being mostly woodlots. Sim- 

 ilar data were given in i860; but in 1870, under the direction of Gen. 

 Francis A. Walker, the noted economist, an improvement was made by 

 separating the woodland from other unimproved land. (In 1880, again 

 imder General Walker, the improved land was also divided, into tilled 

 and untilled. which however has little to do with our problem. ) 



In 1890 and 1900 the farm land was divided into improved and un- 

 improved, as in 1850 and i860, but the definition of improved land 

 was narrowed to exclude abandoned fields and rocky pastures. Finally, 

 in 1910, the woodland was separated again, as in 1870, and it is to be 

 hoped that this distinction will be maintained in future censuses. In 

 recent decades some of the individual States have taken censuses mid- 

 way between the Federal ones, giving similar data : but these are com- 

 paratively inaccessible and will be ignored for the present, in order not 

 to complicate matters too much. 



To get the total forest area of any state (or county) at any census 

 period from these data, we may proceed as follows: First deduct the 

 area assumed to be originally treeless (5 jier cent) and tiiat occupied 

 by cities, towns, etc. No statistics of the latter are given in the Gov- 

 ernment census reports; but if we allow a fifth of an acre per inliab- 

 itant we will ])rol)al)ly be not far wrong, for although that may be too 

 little for rural coninuniities. wIutc liouses liave large yards, it is cer- 

 tainly too much for the cities, where the population is more congested. 

 It is better to err on the side of exaggeration, however, to make allow- 



