Jotyv—A Fractionating Rain-Gauge. 287 
within. The siphon leaves off by sucking in a mixture of air and 
water, and if a continuous trickle is entering the cup, one of two 
things will then happen—the siphon will start prematurely into 
action owing to the downward pull of the short beads of water in 
the longer limb, or it will fall into continuous action, such a quan- 
tity of water passing over mingled with air as will just equal the 
inflow. This difficulty recently presented itself to me in the con- 
struction of an apparatus involving the use of an intermittent 
siphon, and I then found that the method adopted in this gauge 
completely surmounted the difficulty. A moment’s consideration 
_ will,in fact, show that the bell-glass, being full of water till the 
last moment, will stretch by a little the thin rubber tube supporting 
it. When, therefore, the mouth of the bell-glass is finally 
uncovered, and its contents thereby discharged, the rubber 
contracts, effecting the complete disengagement of the siphon for 
an interval which, even with a fairly rapid inflow, is sufficient to 
put the siphon entirely out of action, and leave it filled with air. 
Any form of spring suspension will evidently effect this end, 
provided it permit of some appreciable lengthening of the shorter 
limb of the siphon due to the weight of water within it. 
As regards the dimensions which I have adopted in the case of 
the apparatus which is before you, and which is intended for trial 
only, I have made the catchment funnel 16 inches in diameter. 
The area of this is just 201 square inches. A rainfall of ~5 in. 
gives, in this case, a volume of 20°1 cubic inches. The first bottle 
just possesses this volume when filled to a depth of about 6 inches. 
The second bottle possesses an equal capacity, but, of course, may 
be interchanged for one of greater capacity, so as to effect a 
different subdivision of the rainfall. The reservoir is an oblong 
tinned-iron box about 5 by 2 inches in plan, and 2 inches deep. 
This permits of sufficient room for the thistle-shaped bells which 
are cut from thistle-funnels, and each about 1 inch in diameter. 
The reservoir would be better made in glass. 
Finally, I may point out that, while preserving two siphons 
only in the reservoir, any sub-division may be effected by arranging 
that the bottles first receiving a charge from the reservoir shall be 
also fitted with siphons on the foregoing principles, so that, when 
just filled, they transfer their contents to yet other bottles placed 
beneath. ‘The latter siphons, being thrown out of action by the 
SCIENT. PROC. R.D.S., VOL. IX., PART III. a 
