212 



SCIENCE 



[N. S. Vol. LI. No. 1313 



gaps may be bridged by the use of lights 

 raised to a high altitude by aircraft or pilot 

 balloons. For example, the distance between 

 the Florida reefs and Cuba is about 90 miles, 

 and the shores not high enough to permit of 

 intervisibility. From an aircraft at a height 

 of 5,000 feet or more above the middle of the 

 straits both sides would be readily visible in 

 clear weather. Suppose now that a series of 

 stations along the Florida coast had been con- 

 nected in the usual manner with the triangu- 

 lation net of the United States, and that an- 

 other series of points on the Cuban coast had 

 been connected with a triangulation covering 

 the island. A light carried by a dirigible or 

 pilot balloon above tlie middle of the straits 

 could be observed from two or more stations 

 on each shore, and its position accurately 

 fixed with respect to both systems of triangu- 

 lation. If two or three such aerial points at 

 distances of 30 or 40 miles along the axis of 

 the channel have been tied in this fashion to 

 both triangulations, a strong connection will 

 have been established between them. 



It is obviously necessary either that the 

 " aerial point " should remain fixed while ob- 

 servations are being made on it, or that the 

 observations at the different stations should 

 all be exactly synchronized. The first is im- 

 possible, but the second alternative can easily 

 be realized by using practically instantaneous 

 flashes as signals and observing them photo- 

 graphically. A quantity of flash powder suffi- 

 cient to produce a signal which could be 

 photographed from 50 miles distance could 

 probably be carried by an unmanned balloon 

 of moderate size and cost, or failing this, a 

 series of such charges attached to parachutes 

 and ignited by time fuses could be dropped 

 from a dirigible. 



The photographic records would preferably 

 be made with lenses of moderately large 

 aperture and long focus, such as are used for 

 astronomical chart work, which give a field 

 of good definition several degrees in diameter. 

 If the observation stations are several miles 

 back from the shore line, a series of reference 

 lights can be established on the shore, and 

 their azimuths accurately detei-mined in ad- 



vance. The photographs wiU then show these 

 lights as well as the distant flashes, and the 

 angular elevation and azimuth of the latter 

 can be determined directly from the plates 

 themselves, in exactly the same manner in 

 which astronomers determine the position of 

 a planet with reference to neighboring stars. 

 A number of successive flashes could be 

 recorded on one plate, provided they were 

 so spaced as to avoid confusion, with marked 

 economy both in flying time and computa- 

 tion. Clear weather would be necessary, but 

 not more so than in the case of ordinary 

 methods of observation. 



With regard to accuracy, it is well known 

 that this standard method of determining 

 angular position by the measurement of photo- 

 graphic plates is capable of very high 

 precision. For example, at the Allegheny Ob- 

 servatory with a 4-inch objective the probable 

 error of a resulting angular coordinate derived 

 from two plates was found to be =!= 0.2". The 

 apparent angular diameter of the flash as seen 

 from a distance of 50 miles would be roughly 

 1" for each foot of its actual linear diameter. 

 As settings may be made on the center of a 

 photographic image within 1 per cent, or 2 

 per cent, of its diameter, the azimuth of the 

 flash should be obtainable with sufficient 

 accuracy for purposes of primary triangula- 

 tion, particularly as the mean position deter- 

 mined from the several successive flashes on 

 one plate should be regarded as the real unit 

 of observation. Irregularities in refraction 

 are likely to be less serious than in the case 

 of rays which pass closer to the earth's 

 surface. 



This method might also be advantageous in 

 crossing wide areas of swamp or jungle. The 

 limiting distance over which it is available 

 can be determined only by actual experiment, 

 but it is likely to exceed 100 miles, which 

 would be great enough to permit the exten- 

 sion of continuous triangulation along the 

 whole chain of the West Indies. The theoret- 

 ical distance of the horizon from an altitude 

 of 20,000 feet is over 170 miles, so that if the 

 difficulties involved in producing flashes pho- 

 tographically observable at this great distance 



