340 VOLCANIC ERUPTIONS ON MARTINIQUE AND ST. VINCENT. 



facts presented b}^ me in the July nunil)er of this magazine/' which 

 sustains conclusions then reached. I refer to the narrative of George 

 Keunan pul)lished in the Outlook for August 16, the preliminary report 

 made by Tempest Anderson and J. S. Flett to the Roj^al Societ}^ of Lon- 

 don, and the preliminary report made by E. O. Ilovey to the American 

 Museum of Natural Histor3\ Although f rankh^ confessing that I am 

 not an unbiased judge of the printed testimon}', yet it seems fair to 

 claim that the evidence presented is conclusive as to the important 

 part taken by steam and hot dust in the sudden destruction of the peo- 

 ple of St. Vincent on May 7 and of the inhabitants of St. Pierre on 

 May 8. 



Dovjmnard volc<inu'hlaf<ts. — Intimately associated with the destruc- 

 tion of St. Pierre is the direction taken by the blast of dust-charged 

 steam, with its volleys of stones, which swept over the city. The 

 hypothesis that St. Pierre was destro3xd b}'^ an eruption from the 

 "Riviere Blanche subcrater"" being rejected, and the further sugges- 

 tion, based on the earlier reports in reference to the opening of a fis- 

 sure in the side of the mountain, not finding support in later evidence, 

 the way is cleared for a better understanding of the true cause of the 

 direction 'taken by the down blast that came, as now seems definitely 

 proven, from the Etang Sec, which is essentiall\" a suuunit crater. 



To 'understand the nature of the A'olcanic blast which destroyed St. 

 Pierre, one needs to visit the region swept by a similar eruption on 

 St. Vincent. The volcanoes on these two islands have not only shown 

 a direct relationship in the times of their eruptions, but the surface 

 phenomena exhibited by one is the counterpart of what took place in 

 the case of the other. Happily on St. Vincent, however, there was 

 no densely populated city within the radius of greatest destruction. 



On St. Vincent the region throughout which the previously luxu- 

 riant vegetation, plantations, etc., were swe})t away or l)urie(l beneath 

 hot dust and stones, encircles the mountain. The direction in which 

 trees wei'e swept down, and on the peripheiy of the devasted area the 

 erosion of the l)ark of trees still standing, on the side facing the vol- 

 cano, as well as luut-h other evidence, show in a most conclusive man- 

 ner that a blast charged with dust and stones .swept down the sloj)es 

 of Ija Soufriere in all directions. The infiuenc(^ of hills at a distiince 

 of some -i or .5 miles from the volcano in shielding the veg(*tation on 

 their slopes facing away from it shows that the topograph}' of the 



«It was impracticable for me to read the proof of the article referred to, and in the 

 titles of some of the illustrations, especially, there are serious errors. In thc^ title ot 

 the plate opposite page 278, "Georgetown" slionld besuljstituted fi»r " Ivingstown;" 

 the title of the ])late opposite jiage 282 should he " Valley of Wallihu Kivcr deeply 

 filled with hot debris;" on page 284 the title ol tlie ilhistration should " Sunnnit of 

 INIorne d' Orange, St. Pierre." The map on page 282 fails to show tlu; area at the 

 north end of St. Vincent, as indicated on the original, wliieli was not devastated, and 

 is nuu'h generalized in nther ways. — I. ('. R. 



