1168 
Morning Sé¢ from the south is usually a prognostic of 
a marked rise in temperature, for it is formed by the 
mixture of warm, moist air at a low height with colder, 
surface air, and the mixing layer will soon work through 
to the ground. This is true especially if there is suffi- 
cient insolation to warm the surface layer to a temper- 
ature high enough to engage it in turbulent or convec- 
tive interchange with the warm layer above. Insolation 
thus favors daytime arrivals of warm fronts at the 
surface. Unless the advection is raising the dew point 
rapidly, the mixing through a deeper layer and the rise 
in temperature tend to reduce the cloudiness by mid- 
or late morning, changing it from Si to Sc or even Cu. 
Altocumulus clouds observed in the evening are more 
likely to last all night, and to increase, growing upward 
and downward, when the surface wind is south than 
when the wind is northwesterly. In the first case it is 
usually the southerly wind, with its transport of heat 
and moisture, that is responsible for their formation, 
while in the second, such clouds may be merely the 
leftovers of flattened tops of convectional clouds formed 
during the daytime convection from the surface. Radi- 
ational cooling from the top of either type, however, 
favors growth and continuance. 
The occurrence of sharp-topped Cu or Sc at sunset, 
when the wind is westerly or northwesterly, is a prog- 
nostic of a marked fall in temperature, for the sharp- 
ness, due to strong convection, indicates the steep 
lapse rate which is characteristic of a fresh cold air 
mass. At night radiational cooling as well as advection 
strengthens cold air masses and, indeed, favors 
nocturnal arrivals of cold fronts. 
Interpretation of Observed Trends in Cloud Trans- 
formations. An increase in cloudiness will flatten the 
diurnal variation of temperature. 
A wind-shift line at the earth’s surface is usually 
marked by some forced ascent of air; thus the appear- 
ance and approach of cloud forms due to forced ascent 
indicate coming change, in anywhere from a few minutes 
to a few hours, depending on the distance and speed. 
This change may often be accompanied by rain. If 
there is a well-defined cloud base, the height of which 
is estimated, successive measurements of the angular 
altitude of the base as it approaches will give a reason- 
ably accurate estimate of the time of arrival of the wind 
shift. 
On a dull day we expect clearing if it becomes lighter, 
and vice versa. It is chiefly a change in light which 
attracts attention. A pall of forest-fire smoke which 
darkens the sky often brings out the umbrellas. Indeed, 
a solar eclipse has brought in washing off the line! 
A meteorologist can intelligently make up any num- 
ber of qualitative rules for short-term forecasting at 
a single station [ef. 14; 45; 47, p. 213; 58; 101]. The 
foregoing remarks must suffice as samples. 
The expectancy of any rain following the first appear- 
ance of various cloud forms was worked out by Clayton 
[23], apparently in more detail than by anyone else. 
Sweetland [98] added St nimbiformis (= Ns) and Palmer 
[75], halos. 
Clayton’s analysis of cloud sequences by layers be- 
CLOUDS, FOG, AND AIRCRAFT ICING 
fore 135 rainstorms [23] showed that the pre-rain se- 
quence of Cz or Cs to As to Nb or Cb occurred in 84 per 
cent of the cases. The average expectancy of rain at 
Blue Hill, including cases when rain was falling at the 
zero hour, was 43 per cent in twenty-four hours; but 
when, as in the cases in Table I, precipitation was not 
‘“Tasie I. ExpEcrancy oF Rain at Buue Hitt, Mitton, Mass., 
Fottowine Various CLoup Forms 
Per cone igs cases Most freuen 
n ai Vi 
Genus Cases falisssqainn || pcs 
4 hr (hr) 
Ci 159 33 26 
Ce 137 36 22 
Cs 160 44 13 
Ac 84 45 8 
As 142 68 6 
Ns* 70 73 4t 
Halos 569 <57 16f 
* St nimbiformis. 
{ Cases when precipitation began at an unknown hour 
during the night have been omitted; intervals are therefore 
somewhat weighted toward daytime conditions. Since rain 
occurs more readily at night, the intervals would probably 
have been somewhat less without these exclusions. 
t Average interval. 
already im progress, the average expectancy was 36 per 
cent. Only in the case of Cs with halos, As, and 
Nb(=Ns) is the expectancy substantially above average. 
Halos Before Rain. The Zui Indians say, regarding 
halos, ‘When the sun is in his house, it will ram soon.” 
Several tabulations of the occurrence of rain after halo- 
producing C's are summarized in Table II. 
In most of the studies, solar and lunar halos are 
separated, but the differences are mostly small—fre- 
quency of precipitation is 1 per cent greater after lunar 
halos at Wauseon and 5 per cent greater at State 
College, but 9 per cent less at Blue Hill and 10 per 
cent less in London. 
Differences between summer and winter are given 
by Mikesell [57], Neuberger [73], Palmer [75], and 
Russell [83], the precipitation being decidedly less in 
summer: Wauseon, 54 per cent vs. 62 per cent, and 
State College, 42 per cent vs. 78 per cent at twenty- 
four hours; and Blue Hill, 61 per cent vs. 70 per cent, 
at thirty-six hours. In London, the frequency in Decem- 
ber is as high as 83 per cent (vs. 68 per cent for the year) 
at twenty-four hours. 
Mikesell [57] found that the prognostic value of 
halos at Wauseon varied markedly with the pressure 
and pressure tendency (see Table III). The marked 
concentration of halos with high but falling pressure is 
in line with the typical position of C's on the baclx slope 
of a high. 
A few other points from studies of halos and Cs 
follow. Martin [65, 66] reports that at Fort Worth a 
halo with an east wind is followed by precipitation 
within thirty-six hours in 87 per cent of the cases, and 
within forty-eight hours in 95 per cent of the cases. At 
Columbus, when the wind at the time of the halo or 
soon after is from the southwest, the precipitation pros- 
pect in forty-eight hours is 88 per cent, according to 
