taken in darkness so that the preceding notes taken in daylight will 

 not be appropriate for use in coding clouds for the synoptic message. 

 However, it is intended that the observer will take every advantage of 

 periods of daylight or visibility of the moon and stars that is practica- 

 ble without undue hardship in extending his hoiu's of duty or depriving 

 him of needed rest. The cooperation of other employees in the office 

 should be secured in making cloud notes durmg periods when the 

 observer cannot be on watch. 



It is also important that doubtful cloud observations be omitted 

 from the coded messages. When there is serious doubt about the 

 direction of movement, it should be coded as "unknown." Wlien 

 there is real doubt as to the form of scattered clouds because they are 

 seen only near the horizon, or for other reasons the clouds cannot be 

 accurately observed at or prior to the observation, no cloud word 

 should be sent. However, doubt should exist only after every effort 

 has been made to ascertain the facts. For example, a definite cloud 

 bank on the horizon should be reported and if it is observed to be ad- 

 vancing slowly its direction of movement should be given. It is im- 

 portant to note that a bank of what appears to be cirrostratus on 

 the horizon, definitely advancing, is coded H5 when the high clouds 

 do not extend 45° above the horizon. This may be the first indication 

 in the sky of the approach of a disturbance. 



It is especially desirable that care be exercised in observations of 

 high and intermediate clouds, particularly altostratus and altocumulus 

 clouds, because accurate data concerning clouds at these levels are 

 of very great value to the forecasters. Preconceived ideas as to the 

 probable direction of upper and intermediate clouds based on average 

 conditions or past experience should not influence the observer. 

 Intermediate and high clouds may be visible at one station when the 

 sky is completely covered by low clouds at surrounding stations. In 

 such cases, isolated observations of intermediate and liigh clouds 

 become of critical importance. 



Special instructions are necessary for observing and coding clouds 

 at mountain stations where they may exist at levels below the ob- 

 server's station. 



18 



