94 



ANNALS NEW YORK ACADEMY OF SCIENCES 



grams, representing the conditions for 1906, February, li)06,-January, 

 1907, April, 1906,-March, 1907, and June, 1906,-May, 1907 (Figs. 46- 

 49), show that the transformation began with a slight rotary movement, 

 followed by a displacement of the pleionian center from the north to- 

 wards the south, and then by a displacement of the pleion towards the 

 west and of the antipleion towards the east. 



1907-1908. The pleion of 1908 (Fig. ?0) is so similar to that of 

 1907 that one would think that nothing extraordinary happened during 

 the year and that there was simply a shifting of the pleion towards the 

 northeast. In reality, the consecutive maps show that the pleion and 

 antipleion moved first around, with the hands of a clock, so that for 

 February, 1907,-Januar3r, 1908, the qiiasinoraial line had already ex- 

 tended from North Dakota towards Tennessee and Virginia, as is shown 

 on the following diagram (Fig. 50), then the antipleion extended farther 

 south (April, 1907,-March, 1908), and from then on it was driven away 



progressively in a northeasterly 

 direction. The lines of the dia- 

 gram show the successive posi- 

 tions of the quasinormal line. 

 The 1908 western antipleion is, 

 in June, 1907,-May, 1908, al- 

 ready on the plateaux and from 

 then on its negativity increases 

 progressively. 



1908-1909. Finally, the last 

 twelve consecutive maps sliow 

 that the axis of the pleion of 1908 

 first turned slightly to a north- 

 south direction, then moved eastward, then back again to the west. For 

 November, 1908,-October, 1909, there were two pleionian centers, one 

 in North Dakota, the other in Texas. Tlie advance of the antipleion 

 towards the northeast is seen only on the last consecutive map. 



So far, the consecutive maps have served only to explain the transfor- 

 mation of the departure map of one year into tliat of the following year. 

 One might be satisfied with the results obtained, the consecutive maps 

 having served their purpose in a satisfactory way. 



The principal conclusion gained is that the method used could be 

 applied to seasonal forecasting ; but it is evident that, for such a purpose, 

 it would be necessary to draw the consecutive maps as correctly as possi- 

 ble, by calculating the means for all the individual stations. 



It is also evident that the same would have to be done for the rainfall 



Fig. 50. — Displacements of the quasi- 

 normal line on the consecutive de- 

 parture maps of February, lS07,-Jan- 

 uary, 1908, till July, 1907,-June, 1908 



