1218 
combined in series in a circle. Because of the frequency 
of measurement, no ambiguity will arise in using the 
same contacts for each 360 degrees of motion of the 
pressure arm. Probable error in pressure measurement 
is 0.5 mb. 
For humidity indications, a standard hair hygrograph 
has been modified by replacing the clock and chart 
drum with the coding device. For the range of from 
10 to 100 per cent relative humidity only fifty contacts 
are used. Although humidity changes in increments of 
2 per cent can be recorded, the accuracy of the measure- 
ments is about -+-7 per cent for temperatures above 
freezing, due principally to drift errors of the hair. 
It should be noted that for all these measurements lag 
constants are not as critical as they are in radiosonde 
observations. 
A tipping-bucket type of rain gage is employed to 
determine the accumulated rainfall between transmis- 
sion times. Accumulated rainfall up to 3 m. can be 
measured in steps of 0.03 im. Standard rotating-cup 
anemometers and wind vanes have been adopted for 
surface-wind measurements. Over the range of from 2 
to 150 mph, values are transmitted in increments of 
1.5 mph and wind direction in increments of 3.6 degrees. 
As soon as the equipment is available, measurements 
of the visibility and of the amount of light from the 
northern sector of the sky will also be made. 
General Remarks. To utilize the full potential value 
of an automatic weather station, the installation should 
be operable unattended for as long a time as possible. 
The design goal of this equipment is from six months 
to one year. As a matter of fact, this equipment has 
operated unattended for a little over six months. This 
achievement, however, should be taken merely as an 
indication of the possibilities of the newly developed 
automatic weather station. 
One serious problem has always been that of an 
adequate power supply—a reliable source of power 
that can operate automatically, imtermittently, and 
under a wide variety of climatic conditions for a long 
period of time. To date, the best results are obtained 
with gasoline-engine-driven generators, especially where 
large amounts of power are required. Wind-driven 
power plants offer promise for locations where the 
average winds are high enough to operate the wind 
generators efficiently and maintain the storage bat- 
teries fully charged. Also, for systems with modest 
over-all power requirements of several hundred watts, 
the use of banks of Edison primary cells in underground 
vaults is possible. 
For locations where icing is prevalent, serious prob- 
lems of instrument design arise, particularly for wind 
measurements. Most rotating-cup anemometers and 
wind vanes can become totally inoperative and perma- 
nently damaged when subjected to severe icing. A 
specially designed heated Pitot-tube anemometer and 
wind vane have been designed and are satisfactory, 
except that excessive continuous power is required 
to keep them deiced. In addition, the Pitot-tube ane- 
mometers are rather insensitive to low wind speeds. 
A properly designed instrument shelter is required 
METEOROLOGICAL INSTRUMENTS 
for the arctic use of unattended weather stations. Rime 
ice and wind-blown snow can cause havoc with instru- 
ments exposed in present standard shelters. 
To expand the usefulness of unattended stations, 
additional instrument research is necessary. There is 
a need for a reliable humidity indicator for the tempera- 
ture ranges encountered. A pressure-tendency indi- 
cator, a show gage, a cloud-base indicator, and a 
visibility meter that are readily adaptable to auto- 
matic weather station techniques would collectively 
expand the importance of the unattended weather sta- 
tions manyfold. 
CEILOMETERS—PULSED LIGHT AND RADAR 
Introduction. The general requirement for the meas- 
urement of the base of clouds during the daytime as 
well as at night has been satisfied with the develop- 
ment of the elegant ceilometer by Laufer and Foskett 
[21]. The problem of the overwhelming background 
light scattered by a cloud during daytime observations, 
with the standard light projector, was solved by util- 
izing a modulated light beam in the projector and a 
specially designed photoelectric pickup device in the 
telescope. The projector and telescope were separated 
by a 1000-ft base line. These ceilometers are in wide- 
spread use by the United States Weather Bureau, 
Navy, and Air Force. 
There are, however, two other important require- 
ments in meteorology for cloud height measurements. 
One is a method for measuring the base of clouds from 
a single station. The second is a method for measuring 
the tops as well as the bases of clouds from the ground. 
In this section, a brief description will be given of two 
pieces of experimental equipment that show great prom- 
ise of meeting these additional requirements. One is a 
pulsed-light cloud-base indicator and the other a micro- 
wave radar set capable of depicting cross sections of 
cloud systems. 
Detection of Clouds by Pulsed Light [17, 26]. The 
principle of operation of this British [17] device is 
briefly this. Light pulses of about one-microsecond 
duration are produced by a high voltage spark placed 
at the focus of a good quality 36-in. Cassegrain mirror. 
The light pulse reflected from a cloud deck is then re- 
ceived by a second mirror which focuses the light on 
a photocell. The output of the cell is placed on a 
cathode-ray tube, using the usual radar techniques of 
display. 
The electric spark produced between aluminum elec- 
trodes has a peak power of about 11 megawatts. It is 
estimated that a flux of about four million lumens is 
placed on a cloud deck. With the first experimental 
model, clouds up to 18,000 ft were detected in daylight. 
It is expected that this device, designed at the Tele- 
communications Research Hstablishment in England, 
will be modified to incorporate a means for recording on 
paper the height of the base of clouds. 
Radar Cloud-Base and Cloud-Top Indicator. It was 
concluded from theoretical studies [84] that with a 
radar set operating on a wave length of about one 
centimeter and with sufficient, though reasonable, peak 
