86 THE NEW WORLD OF SCIENCE 



average of eight locations showed it to have the map co- 

 ordinates, x=350930 meters and 7=234700 meters; subse- 

 quent survey showed that the middle of the battery actually 

 had the coordinates, x=35O92o and y=2347io. The error 

 amounted to 16 meters (in x) too far to the East and 10 

 meters (in y) too far to the South or about 19 meters actual 

 error on the ground. This battery had eight gun pits, six of 

 which had been recently occupied ; nearby were deep, safe dug- 

 outs and the whole position was well designed and executed; 

 it had never been shelled by our artillery as it had never been 

 listed as an active battery. About two hundred meters away 

 from this very active battery was an emplacement which 

 showed up clearly on the airplane photographs but which an 

 examination of the ground, after the position had been cap- 

 tured, showed had not been active within at least a year. This 

 inactive battery was listed by our artillery and it was assumed 

 that the locations reported by sound-ranging of the really active 

 but concealed battery were incorrect locations of the visible 

 but inactive battery. Had either the artillery or the sound- 

 rangers had a proper confidence in the accuracy of the sound 

 location the German gunners would have had need of their 

 deep safe dug-outs and our own lines would have been shelled 

 less. Instances like the above were not common, however, and, 

 generally speaking, the artillery made full and efficient use of 

 the data supplied by sound-ranging. The surveys brought to 

 light full confirmation of the theoretical considerations on the 

 accuracy of the individual locations ; thus, errors in range were 

 always greater than errors in line and errors of both kinds 

 were less when the gun was opposite the middle of the sound- 

 ranging base than when it occupied a flanking position. Obser- 

 vations on guns lying more than one kilometer outside the 

 perpendiculars erected on the ends of the chord of the base 

 may be quite worthless as regards the determination of the 

 range though still quite accurate as regards the determination 

 of the line. This was predicted from geometrical considera- 

 tions and in consequence it was the practice for some time 



