66 



C 

 D 



E 



G 



H 



SCIENCE PROGRESS 



So, too, curves A and B are inverse 

 for the same years — B here agreeing 

 with the March i curve ; and the 

 consequent indication is that by 

 about mid-April the three years 

 1 914- 1 6 were relatively much as they 

 had been before March. 



We do not propose to describe 

 in detail all the many parallelisms 

 and fewer inversions which any 

 reader may trace for himself on this 

 chart ; but we must call special 

 attention to the striking phenomena 

 of 191 7 and 1 91 6 — pre-eminently of 

 191 7. The effect of the abnormally 

 cold winter of 191 6-1 7 is strikingly 

 indicated by the curve on Chart I, 

 and by curves A,B,C,D,E ; but F 

 shows that before mid- June 191 7 

 had recovered, and was rather more 

 forward than 191 6 in June, while 

 G and H show that the year which 

 had begun so much in arrear finished 

 as only slightly less forward a year 

 than the premier summer-years 19 14 

 and 191 8. 



On the other hand, 191 6, which 

 had been ahead of 191 5 before March 

 and in the first half of April, and 

 had just managed to keep ahead of 

 it until the end of May, fell off in 

 June and July and — as we have seen 

 —was utterly beaten afterwards by 

 191 7, which had started under so 

 severe a handicap. We wish to draw 

 special attention to the fact that by 

 taking averages of seasonal groups, 

 as we have done, it is thus possible 

 to trace graphically the loss or gain 

 of each year, month by month, 

 relatively to other years, 



A very little inspection of the 

 eight curves on Chart II should 

 suggest their combination into a 

 smaller number ; and on Chart III 

 are exhibited the results of such 

 combinations — viz. of A,B,C (March 



Chart 2. 



