182 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1951 



seems probable that the effectiveness of this process depends on the 

 form of the snow crystals; the dendritic (feathery) type would ap- 

 pear to be a much better source of splinters than the columnar and 

 tabular forms. It has been stated that natural snow often appears to 

 fall from below the —10° C. level, thus suggesting that something like 

 the splintering process is active. A more careful study of this might 

 yield some useful clues, particularly if samples of the snow crystals 

 could be obtained. If splintering does not occur, continuous seeding 

 would be required. 



SEEDING OF PRECIPITATING CLOUDS 



It is next in order to consider the possibility that seeding will alter 

 the timing or the total amount of precipitation that falls from a 

 cloud which is about to release precipitation from natural causes. 

 It is almost impossible to answer this by means of individual seeding 

 experiments. An objective answer might be obtained by seeding all 

 potential rain clouds over a well-defined area for a long period of 

 time (a year or more) and then comparing the precipitation in the 

 area to that in similar control areas. Even here, there are pitfalls 

 and the experiment should be very carefully designed in advance by 

 a competent statistician-meteorologist team. 



Evidence has been presented above that the activation temperatures 

 of the artificially introduced ice nuclei are higher than those of most 

 of the natural ice nuclei. If an active cloud, which is about to release 

 precipitation by natural processes, is seeded, the onset of precipitation 

 may be advanced by the time interval which would be required for 

 the cloud to grow from the temperature level corresponding to the 

 activation of the artificial seeds to that of the natural ice nuclei. The 

 vertical distance between these levels might be of the order of 2 km. 

 The corresponding time interval might range from say 20 minutes 

 for an active cumulus to a few hours for frontal cloud systems. This 

 might provide a means for increasing the precipitation in a selected 

 area from a given cloud even though the total precipitation from the 

 cloud were not changed. The net increase of precipitation in a given 

 area from a large number of clouds would be small because only a 

 few of them would be in the proper stage of development as they 

 approached the area. It is possible that the seeding would decrease 

 the precipitation from the cloud by causing the precipitation elements 

 to form at a lower level, thus decreasing the total water condensed in 

 the cloud. It may be concluded that although the time of precipita- 

 tion may be advanced it would be difficult to obtain any practical ad- 

 vantage from this prospect except under very special circumstances. 



Because of inadequate knowledge of natural precipitation processes 

 it is not possible to make an unequivocal statement as to whether or 



