26 Method of conducting the Canal Surveys. 



that as no survey is made along the inner edge of the towing path, 

 any changes or variations in it cannot be so easily detected and rec- 

 tified. These offsets likewise, owing to the great difference in level 

 of the surface of the towing path, and the ground on which the out- 

 lines are situated, particularly in places where there are high em- 

 bankments or deep excavations, would be subject to very great in- 

 accuracy, which combined with the difficulty of reducing them to 

 any regular system, would occasion many irreconcilable discrepan- 

 cies between the measures upon the offsets and those upon the out- 

 lines, and render the precise location of the boundaries a matter of 

 corresponding uncertainty. In the mode as pursued, the accuracy 

 or inaccuracy of the offsets does not in the least affect the location of 

 the base line, and by means of the measures upon it, and the uniform 

 mode of describing the offsets, the bearings and distances of the out- 

 lines can be calculated, if required, with much greater precision than 

 they could possibly be measured, and when so calculated, the differ- 

 ent parts of the survey, will have the additional merit of a perfect 

 agreement with each other, a desideratum which in the other method 

 must be pronounced to be practically unattainable. 



Another consideration of much importance in favor of this mode 

 is found in the facilities afforded for recording the field notes and 

 representing the whole by means of sketches end diagrams in such a 

 manner as to avoid all liability to mistake or confusion and present- 

 ing at the same time a very tolerable map of the survey. The check 

 likewise which the mode of sketching exercises over the measures 

 with the chain — the one keeping pace in all cases with the other, and 

 both under the immediate and constant supervision of the survey, or 

 (each chain-distance on the base line being represented by its cor- 

 responding space in the field-book,) combined with the practice of 

 requiring a separate account from each of the chainmen, rendered an 

 error in the reckoning almost impossible. 



In the other mode the frequent obstructions to be encountered up- 

 on the outlines and the constant necessity of deviating by offsets from 

 a direct course, would add very much to the liabihties to error, and 

 although the measures, upon the two outlines if the cross measures 

 were repeated often enough, would serve to detect any errors or 

 omissions of integer chains upon each, yet no evidence would be af- 

 forded, upon which of the lines it occurred, and an attempt to cor- 

 rect without an actual re-survey would be as likely to increase as to 

 remedy the evil ; add to this, the discrepancy that would unavoidably 



