The receiver should be supported by another tin of somewhat smaller diameter, and deep enough to contain a bottle which will hold 3 or 4 inches of rainfall.  In order to prevent water entering this compartment, tie a strip of rubber (cut out of an old cycle air tube) or other material round the receiver, and projecting half an inch beyond the bottom (Fig. 152).

All tinned iron surfaces should be given a couple of thin coats or paint.

The standard distance between the rain gauge and the ground is one foot.  The amount caught decreases with increase of elevation, owing to the greater effect of the wind.  The top of the gauge must be perfectly level, so that it may offer the same catchment area to rain from whatever direction it may come.

[Illustration:  Fig. 153.—­Self-measuring gauge.]

Another Arrangement.—­To simplify measurement, the receiver and tube may be arranged as shown in Fig. 153.  In this case the water is delivered directly into the measure, and the rainfall may be read at a glance.  On the top of the support is a small platform for the receiver, its centre directly over the tube.  The graduations, first made on a rod as already described, may be transferred, by means of a fine camel’s hair brush and white paint, to the tube itself.  To draw off the water after taking a reading, a hole should be burnt with a hot wire through the bottom cork.  This hole is plugged with a piece of slightly tapered brass rod, pushed in till its top is flush with the upper surface of the cork.

If the tube has small capacity, provision should be made for catching the overflow by inserting through the cork a small tube reaching to a convenient height-say the 1-inch mark.  The bottom of the tube projects into a closed storage vessel.  Note that the tube must be in position before the graduation is determined, otherwise the readings will exaggerate the rainfall.

[Illustration:  Fig. 154.—­Gauge in case.]

Protection against the Weather.—­A rain-gauge of this kind requires protection against frost, as the freezing of the water would burst the tube.  It will be sufficient to hinge to the front of the support a piece of wood half an inch thicker than the diameter of the tube, grooved out so as to fit the tube when shut round it (Fig 154).

XXVIII.  WIND VANES WITH DIALS.

It is difficult to tell from a distance in which direction the arrow of a wind vane points when the arrow lies obliquely to the spectator, or points directly towards or away from him.  In the case of a vane set up in some position where it will be plainly visible from the house, this difficulty is overcome by making the wind vane operate an arrow moving round a vertical dial set square to the point of observation.  Figs. 155 to 157 are sketches and diagrams of an apparatus which does the work very satisfactorily.  The vane is attached to the upper end of a long rod, revolving