drawn for the 3,000 m. level will often show a ridge of high pressure over 

 the surface location of the cyclone, extending sometimes 1,000 km. 

 westward. Pilot balloon soundings have shown this N.E. trade wind 

 current reaching even the Gulf of Bengal, when a typhoon was nearing 

 the Island of Luzon, mo\'ing westward. However, this upper level map 

 will show the real track along which the typhoon will move for the next 

 twelve hours. We do not think it possible to forecast its motion for the 

 next twenty-four hours. 



We have chosen the 3,000 m. level because of special seismological 

 registrations obtained during typhoon weather. As far back as 1924 

 we correlated the presence of " group-microseisms " showing a constant 

 proper period of 4 to 6 seconds, independent of the amplitude, with a 

 t5^phoon over our neighbouring seas. These microseisms were never 

 registered when high seas, raised by the strong anticyclonic winter 

 monsoon gales, were breaking over the rugged Chusan Islands some 80 

 miles from Zi-Ka-Wei. But as soon as a tropical storm was located 

 even at 1,000 km., these special tremors were clearly obtained. 



This period of 4 to 6 seconds, for a single vibration, was registered by 

 means of seismographs of very different proper periods, although always 

 longer than 5 s. We considered this period of vibration as being, approxi- 

 mately, the period of vibration of the column of the' cyclone, progressing 

 over the ocean and oscillating under the pressure of the Pacific air mass 

 aloft. The fierce rain and wind squalls were also found to have a decided 

 period between 4 and 6 s. putting out of action, by resonance, any high 

 structure, vibrating with proper period of 4 to 6 seconds. 



If a typhoon on a westerly track appears to be recurving northward 

 or northeastward the trend of the 3,000 m. level isobars will show whether 

 the Pacific air mass is withdrawing eastward (recurve to the north-east) 

 or remains stationary (recurve to the north). 



Another kind of puzzling but very efficient correlation has been in 

 use at Zi-Ka-Wei since 1938 for ascertaining the future motion of a 

 typhoon already located by the usual means. 



If radar pulses are sent out on the meah'critical frequency for the E 

 layer reflexions (around 6 megacycles at Zi-Ka-Wei) a very useful cor- 

 relation has been found which helps greatly in deducing whether the 

 Pacific air mass, driving the typhoon, will continue westward or whether 

 it will remain stationary or recede eastward. 



If while the typhoon is over the Eastern Sea we obtain in daytime 

 after sunrise and before sunset echoes, from the E layer, these are strictly, 

 although mysteriously, correlated with the westward advance of the 

 Pacific air mass. 



On the other hand, if, while a typhoon is located over the Eastern Sea, 

 we get reflections from the F layer, this means a coming invasion of 

 Siberian air (or the slowing up of Siberian air if already present over 

 our region) and that the typhoon will not move westward in the next 

 twenty-four hours, but will recurve around the border of the Pacific air 

 mass, over the Eastern Sea. The intensity of the echoes seem to show 

 the future intensity of the Siberian air mass. 



Of course, this correlation is good only for the Shanghai region, and 

 for such a method of forecasting to apply the typhoon should not be more 

 than 300 or 400 miles in any direction. 



