THE PHYSICS OF ICE CLOUDS AND MIXED CLOUDS 
By F. H. LUDLAM 
Imperial College of Science and Technology and Meteorological Office, London 
THE APPEARANCE OF ICE CLOUDS 
It has long been recognised that clouds of ice particles 
have a characteristically fibrous appearance very differ- 
ent from that of droplet clouds, which look “solid” and 
have more sharply defined edges, even when dissipating. 
Small, isolated trails of snow or soft hail, known as 
Fallstreifen (fall-streaks) or virga, which fall from 
medium-level clouds, also possess this fibrous structure, 
and when they become separated from their parent 
clouds and are seen in a favourable light exactly re- 
semble forms of true cirrus. A fibrous structure on a 
coarser scale can also be seen in rain beneath shower 
clouds. 
It is therefore natural to interpret the fibrous appear- 
ance as the result of the high fallmg-speed of unusually 
large particles falling in trails from small condensation- 
regions, and to consider typical isolated cirrus clouds 
as a form of precipitation. Wegener [21] regarded cirrus 
uncinus as Fallstreifen, and attributed their development 
to the absence of efficient ice nuclei, those available 
acting only at a high supersaturation so that the ice 
particles then formed grow quickly and acquire a large 
falling-speed. There have been other attempts to explain 
the form of these clouds, but careful observation sup- 
ports the Fallstreifen theory, and in recent years investi- 
gations concerning the properties of ice-particle nuclei 
have completely explained why ice clouds so often 
contain much larger particles than droplet clouds, and 
have led to a better understanding of the processes 
attending the formation and decay of both pure ice- 
clouds and mixed clouds (containing both ice and liquid 
water). 
THE BEHAVIOUR OF ICE NUCLEI 
For some time it has been thought that the nuclei 
effective in producing ice crystals are probably solid, 
water-insoluble particles, distinct from the hygroscopic 
nuclei which readily allow the formation of droplets. 
Wegener [21] and Findeisen [5] suggested that quartz 
dusts, carried up from the ground, might be the atmos- 
pheric ice nuclei. The prevalence of clouds of liquid 
water at temperatures down to about —10C was as- 
scribed to the inefficiency of the ice nuclei or to their 
absence. The nuclei were called sublimation nuclei, and 
it was generally believed that the ice crystals were 
formed by the direct transition of water vapour into 
the solid state. 
This conception has been altered by recent researches. 
The first important contributions to the subject were 
by Krastanow [10, 11], who outlined the theory of 
condensation according to the principles of statistical 
physies, and showed that the direct deposition of vapour 
into ice upon such solid nuclei as occur in the atmos- 
phere is to be expected very rarely and only at tem- 
peratures approaching —70C.1 He suggested that ice 
particles are usually produced secondarily by the thermo- 
dynamically easier path through the freezing of droplets 
containing these nuclei, and claimed that such mfected 
droplets exist because condensation can also proceed 
upon solid, insoluble (but ‘‘wettable’’) particles as well 
as upon hygroscopic nuclei. Accordingly, it appeared that 
crystals usually arise as a result of the freezing of super- 
cooled drops (and perhaps also by the freezing of the 
microscopic film of adsorbed water which grows on 
solid “wettable” particles at vapour pressures approach- 
ing the saturated vapour pressure over liquid water, 
followed by direct condensation from the vapour). For 
any such particle there are successively lower ranges of 
temperature in which it may act as a nucleus for a 
supercooled drop, for a supercooled drop which freezes, 
and for an ice erystal formed by direct condensation. 
Krastanow was concerned primarily with estimating 
these ranges, but his work contamed the important 
implication that atmospheric crystal nuclei would be- 
come active only at approximate saturation over liquid 
water—formerly it was believed that they would allow 
direct condensation at small ice-supersaturations. At 
the temperatures of the cirrus levels, however, water- 
saturation corresponds to relative humidities of 150 per 
cent or more over ice. 
Some support for this view was provided by the 
expansion-chamber experiments of Regener [16] in 
which droplet fogs were formed at temperatures down 
to at least —50C, crystals also appearing when high 
expansion ratios were used. Weickmann [22], observing 
the condensation products upon nuclei supported on a 
metal face chilled to about —40C, gave a complete 
confirmation, finding that no crystals could be produced 
at low ice-supersaturations, that a few formed under 
prolonged high ice-supersaturations, and that relatively 
abundant formation occurred in the neighbourhood of 
water-saturation. He also showed that particles of the 
common minerals (including quartz) are not especially 
active as ice nuclei, but concluded that atmospheric 
ice nuclei are small insoluble solids about 1 » or less in 
size, which, from their mode of action, he calls freezing 
nuclet. 
It is interesting that by careful search particles have 
now been found which appear to act as true sublimation 
nuclei (allowing the formation of ice crystals direct 
from the vapour at temperatures below —10C and at 
1. A more detailed account of these references and some 
subsequent ones has already been given by the present writer 
(12). 
192 
