286 Scientific Proceedings, Royal Dublin Society. 
virtually becomes the higher of the two siphons, and now it is 
the siphon supplying bottle 2 which operates. The second tenth 
of an inch is thus stored in bottle 2. If the rain continues after 
this, it remains stored in the reservoir, which thus receives the 
third tenth. Finally this may overflow by a tubulure attached at 
the upper end of ¢ into a large bottle placed on the shelf beneath 
(not shown in the figure). 
It is apparent that the bottle 2 may be at once arranged to 
receive a second or a third charge by simply using one sufficiently 
capacious, and bringing its siphon but a little way downwards 
within its tubulure—just so much that the last charge received 
will throw the siphon out of action by opposing to its action a hydro- 
static head greater than the difference in level between the bend 
of the siphon and the overflow tubulure on ¢. This sufficiently 
explains the manner in which this gauge may be made to deliver 
its catch into a succession of bottles in any aliquot parts of the 
content of the reservoir a. ‘To suit the possibilities of this climate, 
a total storage capacity for about one inch rainfall is sufficient. 
This would, perhaps, be best received as follows—two bottles and 
siphons being used, the first bottle receiving one-tenth, the second 
three-tenths, the remainder being caught in the reservoir and over- 
flow. <A rainfall of from 0:1 to 0-2 in. is in this case subdivided 
between 1 and reservoir, from 0:2 to 0:4 in. between bottles 1 and 
2 and reservoir; from 0:4 to 9°5 in. between 1, 2, and reservoir. 
There will but seldom be a rainfall exceeding this between daily 
periods of attendance on the apparatus ; but when this does occur 
the subdivision is between 1 and 2 and reservoir + overflow. A 
threefold subdivision will be sufficient. Ifno more than a twofold 
one is sought, bottle 2 may be made so capacious as to receive any 
quantity over the first tenth. Again, if it is desired to bottle off 
the first twentieth, the reservoir must be made of the suitable 
dimensions. Obviously any desired number of siphons may be 
used. 
Returning now to the construction of the siphons, it is requisite 
to say a word explanatory of the use of the indiarubber continua- 
tions of the shorter limb. ‘Those who have used cup-of-Tantalus 
arrangements will be familiar with the difficulty of effecting the 
sudden and complete cessation of the siphon when the liquid in 
the cup sinks sufficiently to uncover the mouth of the siphon 
