6 
24. The position of the rain gauges from which the records 
at my disposal have been compiled, is badly selected for the 
determination of the phenomena of this nature, and there is 
reason to believe that the observations contained in the tables 
show but imperfectly tlie real extent of change which has taken 
place in the more remote districts. And, as the presence of 
forests considerably modifies the distribution of rainfall and 
economises the water supply after it has fallen, and as agri- 
culture is principally dependent on these conditions, the pre- 
sence of a certain amount of forests in a country is not only 
necessary, but of the first importance for this, and those 
climatic considerations which naturally follow. 
2-5. Turning now to the tables in which are recorded the 
total Rainfall and number of rainy days throughout the year, and 
dividing the twelve years, over which the observations .extend, 
into periods of four years each for the sake of comparison, 
by setting off the total Rainfall and number of days on which 
rain fell against each period, any change which may have taken 
place since the first period will become apparent ; and, if 
apparent, it will be difficult to find any cause to which it can be 
ascribed other than to some local disturbance, or influence, 
which has been at work ; and equally difficult to find any such 
disturbing agent except that which lias been effected by man 
in the removal of the natural covering of the surface of the 
island, which is known to have taken place within the period in 
question, viz . : — 
First period ... 
Second period, 
Third period , . . 
m 
c3 
o 
H 
1869 
1.870 
1871 
1872 
1873 
1874 
1875 
1876 
1877 
1878 
1879 
1880 
Total Rainfall. 
>— i ^ 
r r; s3 
o 
^ cd 
C3 § 
^ H u, 
£-5 & 
p a .5 o 
§ EftH 
1ST nmber of days al 
which rain fell at| 
Criminal Prison. | 
• 
Inches. 
398.62 
Days. 
901 
Days. 
787 
• 
356.52 
829 
689 
' ' ■ \ _ ,* , 1 
42.10 
72 
98 
Decrease over first period. 
369.32 
820 
621 
9 
65 
Continued decrease over 
second period. 
29.30 
81 
163 
Total decrease over first 
period. 
26. The above records [shew a steady decrease in the 
number of rainy or showery days, and what is further remark-' 
able is that, although the Rainfall of the last period shows a . 
slight increase (12*80) over the second, the number of rainy 
