THE GREAT EARTHQUAKE OF S. E. JAPAN 157 



the Clotliirig Department of the Army in Honzyo (No. 6 in PI. XVIII). 

 This is a piece of quite open ground of 60,000 square metres, and 

 nearby, only separated by a street, there is an extensive garden of a 

 wealthy banker, Mr. Yasuda, covering an area of about 40,000 square 

 metres. Thus we had here an area of 100,000 square metres, or about 

 250 acres, lying just on the eastern bank of the Kiver Rumida. People 

 assembled here from all quarters to save their lives and property. 

 What a sad fate it was for the people seeking refuge, not knowing 

 in the least the calamity impending over them ! The whole space was 

 so thickly packed, according to eye-witnesses, with men, women and 

 children and their belongings that they found themselves almost unable 

 to move. 



" At 4 o'clock on September 1, fire approached from three sides, 

 leaving only the side next the river. Suffocating fumes and horrible 

 fires threatened the unhappy people, sparks falling over them in 

 showers. All at once the peo[)le heard some unearthly sound approach- 

 ing them, the heavens darkened, and they were terror-stricken to 

 find that a furious tornado was sweeping toward them lifting or 

 setting in flames everything before it. When it had passed over 

 the ground, what was left behind ? The charred remains of 35,000 

 human beings ! " 



It is said that the exact number of those who perished was 

 38,015 ; only about 2000, who were mostly seated in the southern 

 corner of the ground, were left alive, but the majority of the people 

 were terribly burnt On the other hand, there were many corpses 

 which had quite peaceful countenances with hardly any effect of heat 

 perceptible on their skin or clothes ; their death seems to have been 

 due to the poisonous action of carbon-monoxide, which is decomposable 

 from dioxide under temperatures higher than 850°. 



According to Prof. Terada, the burning area of the city was suc- 

 cessively visited by a few score of tornadoes, the majority of them 

 having taken place during the four hours from 3 p.m. to 7 p.m., 

 Sept. 1. Most of these tornadoes were mere whirlwinds produced at corners 

 of acute-angled areas bounded by firefront, yet there were nevertheless 

 a few others which showed characteristics of real tornadoes such as 

 that which has just been alluded to. 



It is said that the progress of the fire was more rapid in 

 Yokohama than in Tokyo; 65% of the former city was burnt in 12 

 hours, while in tlie latter it took 18 hours to burn ^^.°/o of the houses 



