1164 
physical constitution of clouds is that proposed by 
Bergeron [2] on the basis of his investigation of colloidal 
instability of clouds. His classification is made in tabu- 
lar form as follows: 
1 
Group} Cloudelement | Characteristic | Cumuliform | Strife 
I | Complete Ice needles | Ce with Cs neb, 
erystals halo 
II | Complete Dry powder | Ci unc Cs fil, Ce 
crystals snow without 
and skele- halo 
tons 
III | Skeletons Snowflakes Cb (winter)| As 
and fog and/or 
droplets rime- 
graupel or 
frost- 
graupel 
IV | Fog droplets | (Dry fog) Cu St, Sc, Ac 
V | Drizzle drop-) Wet fog, Cu St, Sc, Ac 
lets drizzle 
VI | Raindrops Rain or hail | Cb (sum- Ns 
mer) 
According to Bergeron, Groups I and IV (only com- 
plete crystals or only droplets) are colloidally stable, 
Groups Ii and V exhibit slight colloidal instability, and 
Groups III and VI are wholly unstable. 
Assessment of Adequacies and Inadequacies 
The international classification of clouds has behind 
it nearly a century and a half of development, and a 
half a century of experience has been gained with it in 
essentially its present form. Many great meteorologists 
have devoted their attention to perfecting it. It would 
seem to represent a high degree of perfection and re- 
finement. However, there are serious criticisms that 
can be leveled against it. Intended to permit accurate 
classification of any observed cloud form, its genera are 
not mutually exclusive and many cloud forms fit none 
of the genera defined. Intended to bring about uniform 
usage, observers still fail by a wide margin to agree on 
the classification of common forms of cloud. 
Two examples will serve to illustrate this unsatisfac- 
tory status. In 1942 the author showed a set of color 
pictures of clouds before a group of meteorologists all 
of whom had had at least three years of professional 
training in synoptic meteorology. The slides were se- 
lected to illustrate common states of the sky and no 
attempt was made to select either the most typical 
appearances of the cloud genera or the most ambiguous. 
Every person in the group was asked to write down, as 
each slide was shown, the genus of the principal cloud 
form illustrated. When the results were tabulated it was 
found that only 56 per cent of the group, on the average, 
agreed on the principal genus, 27 per cent preferring a 
second genus and the rest of the group scattering their 
selections among as many as six additional genera. 
It may be argued that the group test showed only 
that the training in cloud recognition of the subjects 
was insufficient and nonuniform, and indeed this may 
have been the case. Nevertheless, if the classification is 
so ineffectual in the hands of meteorologists, regardless 
of its intrinsic quality it fails in an important part of 
CLOUDS, FOG, AND AIRCRAFT ICING 
its intended purpose. It must be so fitted to the de- 
mands that meteorologists will find it easy and natural 
to use accurately. 
The second example is the result of a review the 
author made of more than thirty books appearing be- 
tween 1938 and 1942, comprising both standard me- 
teorological texts and training materials on meteorology 
intended for aviators. Nearly half of the books take.no 
notice of the 1932 International Cloud Atlas, but retain 
the terms and definitions of the earlier editions. A size- 
able number mention cirrus, stratus, cumulus, and 
nimbus as the four fundamental cloud forms—a notion 
that has no basis in the international classification or its 
antecedents. A few reduce the number of basi¢ forms to 
two, cumuliform and stratiform, disregarding the oc- 
currence of cirrus which conforms to neither. Another 
book explains that ‘nimbus” means “head,” and that 
therefore a nimbostratus is a stratus with a head on it! 
Nor are these examples of divergences by any means all 
drawn from nonmeteorologist authors. 
The day is past when meteorological terms were prac- 
tically the private property of professional meteorolo- 
gists. The airmen who use meteorological terms now out- 
number the meteorologists, and their use of the word 
“stratus” to describe any low cloud that forms an ex- 
tensive low ceiling, or “cumulus” to denote the most 
spectacular thunderhead, is an example of word usage 
that we cannot ignore. 
At this point let us review the nature of a classifica- 
tion as such. A classification as conceived by Linnaeus 
corresponds to the geometric concept of a partition; that 
is, a whole that is divided into a finite number of mutu- 
ally exclusive parts, the sum of all the parts constituting 
the whole. Application of this concept to the classifica- 
tion of clouds would require that every conceivable 
form of cloud be allocable to one and only one family; 
within that family to one and only one genus, species, 
etc. The problem of defining the classes is that of draw- 
ing the lines between them. 
An attempt to subject the international classifica- 
tion to rigorous tests of logic meets with difficulty. For 
example, the definition of cirrus gives only three char- 
acteristics as belonging unequivocally to this genus: 
that they are detached clouds, have a delicate and 
fibrous appearance, and are without shading. No other 
genus is open to detached clouds of delicate and fibrous 
appearance. What is to be said, then, of a detached, 
delicate, fibrous cloud that is thick enough to show 
shading? Plate 6 of the International Atlas shows just 
such a cloud and the explanatory text even calls atten- 
tion to the shading. Yet it is classified as cirrus. The 
example would be trivial if it were unique; it is the 
multiplicity of such ambiguities that makes the classi- 
fication open to grave criticism. No matter how strictly 
the definitions are made mutually exclusive, the human 
element of observation necessarily blurs the boundary 
between them in actual use. No effort should be spared 
to eliminate indefiniteness from the definitions them- 
selves. 
Brooks [6], taking cognizance of the same problem, 
questions whether it is not too much to expect that the 
definitions of the genera could be made completely in- 
