AN TI- C YCL ONIC STORMS. 186 
There can be no descending currents in one place without ascending currents 
in another; therefore, in every atmospheric disturbance, there must be simulta- 
neous cyclonic and anticyclonic winds. Such disturbances originate either in an 
unusual cooling and condensation, or in an unusual heating and expansion of 
air. In the former case the inflow, in the upper regions of the atmosphere, will 
produce an increased pressure. In the latter, the outflow will produce a diminu- 
tion of pressure. In the restoration of equilibrium, currents of warm air are often 
brought into contact with colder currents. If the currents are both saturated 
with moisture, or if they contain more vapor than can be retained under the 
temperature of the mixed currents, precipitation takes place, in the form of rain, 
hail, or snow. This precipitation reduces the weight of the atmospheric column 
and the barometer falls. Accordingly, there isaconstantly increasing tendency to 
cyclonism about storm centres, and there has been a very prevalent disposition 
to look upon all storms as of cyclonic origin. 
A little reflection, however, will show that the initial mixture of currents 
may be due to either of the causes above mentioned ; either to the flow of warmer 
air into a cold depression at the top of the atmosphere or to a flow of cold 
air, at the earth's surface, towards a region of low barometric pressure. In the 
former case, the initial superficial currents are determined by a downward press- 
ure and they are, therefore, anticyclonic ; in the latter they are determined by an 
upward pressure and are cyclonic. 
A careful study of the weather-maps shows that the heaviest rains and snows 
occur in advance of the centres of low barometric pressure, or in the rear of the 
centres of high barometric pressure. If storms began in the cyclonic currents, 
the reverse should be true ; the greatest effect following the low centre and pre- 
ceding the high centre. 
The frequent failures of forecasts, during the past winter, seem to have been 
mainly due to a misinterpretation or a misconception of these facts, to which the 
writer first called attention in 1871 (Proc. Amer. Phil. Soc. XII, 40). They 
were subsequently embodied in the " Suggestions " of the Signal Service Bureau, 
and the officers of the Bureau communicated to the public journals some remark- 
able evidences of anticyclonism in storms of great magnitude. 
Loomis's subsequent discussions of the Signal Service observations have 
furnished abundant additional evidence of a like character, and have shown the 
great frequency of anticyclonism at the beginning, and during the continuance of 
showers and of storms of all kinds. 
The limit between anticyclonic and cyclonic tendencies, may be approxi- 
mately assumed to be midway between the centres of high and low barometric 
pressure. All cloudiness or precipitation between the limit and the high centre, re- 
presents anticyclonic influence; aU between the limit and the low centre represents 
cyclonic influence. Local cyclonism sets in soon after precipitation begins, and 
the anticyclonic influence is thus partially hidden ; but a critical examination of 
the weather maps will show that the prevaihng currents of the region often con- 
tinue to be anticyclonic until the rain or snow is nearly, or quite over. The 
