from Windsor (which is at the eastern end of the Detroit tunnel). 

 Stone walls, which are such a characteristic feature of the same 

 latitudes in New England, were therefore almost wanting. Peb- 

 bles or small rock fragments were frequent, though, in the more 

 hilly portions. 



The soil seemed to be much the same all the way from Windsor 

 to Niagara Falls, a buff or brownish loam, evidently near or 

 above the average in fertility. Fertile soil and good water 

 hardly ever go together, and many of the houses in both town 

 and country have cisterns for rain water. 



The topography along this route is not very diversified. For 

 a distance of about 65 miles at the west end (Windsor to Ridge- 

 town, in Essex and Kent Counties), and 18 miles at the east end 

 (a little west of Welland to Niagara Falls, in Welland County), 

 the surface is almost perfectly level, probably representing an 

 old lake-bottom plain, like the Maumee basin at the southwest 

 end of Lake Erie. In this flat portion drainage ditches and tile 

 drains are frequent. 



Between Ridgetown and Welland, a distance of about 140 miles, 

 the topography is mostly undulating or moderately hilly, with 

 valleys 25 or 30 feet deep along some of the streams; but there 

 are occasional level stretches, probably representing the edge of 

 the plain bordering Lake Erie. The hills are not high enough to 

 present any obstacle to railroad builders, though, and any good 

 map will show that the railroads are about as straight in the 

 undulating country as in the lake plains. Swamps and ponds, 

 such as characterize most glaciated regions, were hardly seen 

 at all. 



About Brownsville, near the southern corner of Oxford County, 

 62 miles from Ridgetown and 82 from Welland, there is a change 

 in vegetation which must be correlated with a difference in soil, 

 indicating poorer soil east of that point (as will be shown pres- 

 ently). It happens that the boundary between the Delaware 

 and Onondaga limestones (both Devonian) passes through or 

 close to Brownsville, but it is not obvious why that should make 

 a difference in the vegetation, and the change is more likely due 

 to some difference in the glacial drift, which covers the bed rock 

 to a considerable depth. Avai able maps do not throw much 

 light on this point, however. 



Streams are rather few and small in this part of southern 



