148 E. T. Quayle: 



of thunderstorms giving a total fall of 3i inches in about 19 hours, 

 with its heaviest downpours during the last hour over the head 

 streams. The great flood of September, 1916, was due to 260 points 

 falling in two days. 



For the whole 32 years, 1890-1921, the number of rainfalls with 

 flood possibilities (over 50 points for one day or 75 for two) were 

 as follow: — January 9, February 13, IMarch 20, April 20, May 27, June 

 43, July 28, August 21, September 25, October 15, November 14, De- 

 cember 15. These are very different from the actual numbers of 

 floods occurring these months, except in August and September. 



Flood Volumes Increasing. 



Owing mainly to the variability in the dates on which the winter 

 Tains may be said to have begun, and the soil to become well mois- 

 tened, the only months which show any approach to consistency in 

 the relation between the rainfalls and resulting floods are July, 

 August, and September. Taking an average of all the rains with flood 

 possibilities, we get the following results for the earlier and later 

 periods: — 





July. 



Auffust. 





Septeiij 



her. 





6 



S5 



11 



^1 



11 



d 



11 



11 



6 



« — 



<2i 



1 i 



1890-1909 - - - 

 1910-1921 - - - 



13 



10 



103 

 93 



1444 

 1641 



11 

 9 



116 



no 



2135 



2092 



10 

 14 



114 

 149 



1766 

 3564 



So far as these results go, they show an improved run-off during 

 the later period. 



Flood Rains and Flood Volumes. 



As already remarked, the run-off after any particular rainfall 

 varies enormously and can only be guessed at with any probability of 

 success when other conditions are known. For example, it is only the 

 very exceptional rainfall which will cause serious flooding of the river 

 during the six summer months, November to April. November can only 

 show three floods in 32 years, a rainfall of 194 points, giving 3600 eft. 

 in 1893, a. rainfall of 277 points, giving 3850 eft. in 1903, and a rain- 

 fall of 166 points, giving 1900 eft. in 1906. In December there have 

 been no floods. In January only one, a 411 points rain in one day, 

 in 1897, giving a flood of 4100 eft., in February only one, a three-day 

 fall of 430 points giving a flood of 3850 eft. in 1911; in March only 

 cne, a three-day fall of 552 points, giving a flood of 4000 eft. in 1910; 

 in April only one, a fall of 313 points in 1890, giving ai flood of 3500 

 <".ft. The comparative rareity of floods in May and October is prob- 

 ably also due to the liability of the soil to be dry in those months. 

 In the case of the former, because the dry season is usually scarcely 

 over; in the case of the latter, because growing vegetation is making 



