ANALYSIS OF THE IONOSPHERE 461 



"multiple echoes"; the signal has traveled two to five times the 

 entire journey from ground to ionosphere to ground again, the surface 

 of the earth being itself a good reflector. Those marked G are due 

 to the signal spreading along the ground itself. If the sender and the 

 receiver are practically side by side, as usually is the case, the kicks 

 G occur at the instants of departure of the signals. The record is 

 moving laterally with the speed intimated by the wavy line beneath, 

 and accordingly the distance along it from a G-kick to the following 

 echo-kick is a measure of the "delay of the echo." 



The delay of the echo is an indication of the altitude of the mirror 

 where it was reflected — -the layer E or F, as the case may be. Signals 

 of relatively low frequency being reflected from E while those of 

 medium frequency are echoed at F, one adjusts the frequency according 



|G F| iG F| 



I * L^ 



-^►t 



/Vv^^AAAV^A^AAAA^^^^v^A^/\/VvVv^vv\^AA/^AA' 



lil- 



1110 CYCLES PER SECOND 



Fig. 2 — Echoes. G, original signal; E\, F\, echoes returning after a single re- 

 flection from E and F respectively; Fi- • • F5, echoes which have suffered two to five 

 reflections at 7^-layer. (Appleton and Builder.) 



to the layer which one wishes to locate. If there should be not two 

 but several layers of the ionosphere, each having a greater TV-value 

 than the one beneath it, one would locate them all with appropriate 

 frequencies. If there is a continuously-rising distribution of N with 

 height in the ionosphere, one may plumb it by varying the frequency 

 continuously. Now we are at the principle of the echo-method; but 

 before it is used, there are many details to clear up. 



First as to the "signals," a term which (it must have been noticed) 

 replaced the term "waves" in the foregoing paragraphs. These 

 signals are wave-trains indeed, but not the long continuous uniform 

 trains tacitly assumed in the description of the other two methods. 

 Those methods are adapted to trains of indefinite length, but not the 

 echo-method, which for an obvious reason requires wave-trains of 

 limited length — and the more limited, the better. 



