ARCTOWSEI, rOLCANIC DUST VEILS 



165 



peratiire curves simph' to show how cautious one has to be in the study 

 of the cause of local climatic variations. 



In the following diagram (Fig. 5), I reproduce the curves of the con- 

 secutive means of temperature for Para/^ Cayenne^^ and the West Indian 



Fig. 5. — Comparison of the consecutive temperature curves of Cayenne and Pard with 



those of the West Indies 



stations: Port-au-Prince/"^ St. Croix Christianssted/^ St. Lucia and 

 Barbados/^ together with the curve of Fort-de-France. 



The curves of St. Lucia and Barbados are so very different from the 

 curve of Fort-de-France that one would be inclined to admit some errors 



=1 MeteorologiscLe Zeitschrift, vols. 23, 24, 28, 31, pp. 517, 431, 215, 139. 



s^Annales Bur. Centr. Meteorol. de France. 



^Annales Bur. Centr. Meteorol. de France. 



" Meteorologisk Aarboj? for 1800. . .1909. Udgivet af det danske meteor. Institut. 



^Received in manuscript from the Meteorological Office in London. 



