mination of the limits of the successive strata and even the geologic correlations of the well 

 logs lack seriously in precision. Further difficulty results from the fact that the wells sunk for 

 reasons entirely aside from the determination of underground topography only by remote 

 accident happen to be at critical points. In theory, it is wrong even to assume that they have 

 done so, yet it is virtually impossible so to assume and yet get any results. 



The distribution of wells is most uneven. In one locality they are reasonably close 

 together and immediately alongside over a wide area there may be no wells at all. Thus, in 

 Brooklyn practically no wells have been sunk in the Harbor Hill moraine on account of the 

 difficulty and high cost of sinking wells in masses of boulders; nor have they been sunk in 

 numbers in the Gravesend region and other areas south of the moraine where much of the 

 water is salt; nor in Jamaica bay where, in general, water is not needed. In eastern Suffolk 

 county the wells are few and deep wells even fewer for various reasons. Several of these 

 areas contain important points of underground topography such as the limits of strata, deep 

 valleys and possible points of intercommunication between aquifers normally separated by 

 heavy clay beds. 



In drawing the contours, these matters have been considered and an effort made to show 

 the general accuracy of the results by using different contour intervals, broken or dotted 

 lines. Thus, it is known that the North Shore bays are deeply eroded in the older strata, a 

 most important point hydrologically, but thera is substantially no data as to this. To draw 

 the nearby contours across these bays as though they were not there or to draw no contours 

 in them would be misleading. Therefore, they have been partly contoured almost entirely by 

 guess. The difficulty is partly overcome by using dotted contours which indicate deep depres- 

 sions but are noncommittal as to the exact form of the eroded surfaces. When the wells are 

 badly scattered, 100-foot contours are shown and these are broken lines when they are based 

 on little or no factual information. Even when a 20-foot contour interval is used, the details 

 of the contour lines are omitted. Thus, when a contour reaches the inner valley of a stream 

 flowing across a fairly level and uniform plain, it turns upstream and follows the near valley 

 wall perhaps for miles before crossing over and returning on the other wall to a point opposite 

 the first turn. This gives a long thin loop or horn. Actually wherever a tributary stream joins, 

 there should be a similar twig on this horn. At the upper end of such a valley there are apt to 

 be several tributaries and the contours bulging downstream between these streams give the 

 familiar "oak leaf" pattern. Where only the facts of the existence of such a valley are known, 

 and neither the exact location of the main stream nor of the branches are shown by the data, 

 a relatively broad simple loop is drawn on these maps. This is the best that can be done under 

 the circumstances, and if the convention is known, should not be misleading. 



152 



