336 
TERRESTRIAL AND SOLAR RELATIONS 
vapor pressures, and the precipitation synchronize with the 
solar changes in their annual variations. The evidence at 
present enables us to affirm that both are true, and that the 
synchronism exists, though in a very complex form, because the 
prevailing local conditions depend primarily upon the general 
circulation, and therefore only indirectly upon the solar varia 
tions. It is not possible in this place to do more than summarize 
the general principles that have been established in a research 
extending over twenty years, and embracing the available solar 
and terrestrial data. The first task is to procure homogeneous 
material of the several observed quantities, extending over a 
long series of years, sun-spot frequencies, solar-prominence 
frequencies, amplitudes of the terrestrial magnetic field, baro 
metric pressures in all parts of the world, temperatures, and 
vapor pressures in all countries, precipitation in many districts, 
direct observations of the solar radiation in calories per square 
centimeter per minute. Unfortunately the difficulties of secur 
ing such homogeneous data of any of these elements is greatly 
complicated by the irregular and inconsistent methods that 
have been employed by meteorologists. In consequence of the 
necessity of substituting a few selected hours of observing for 
the twenty-four hours of each day, it is necessary to reduce the 
means from selected hours to the mean of twenty-four hours, which 
involves a long, special research for each country. The selected 
hours are different in different countries; the series are broken 
by changes in the selected hours in consequence of some admin 
istrative requirement; the corrections change from place to 
place when the same hour is made the basis of the work, as 
where the 75th meridian of the United States is made the hour 
of observing, which involves a range of three hours locally 
between the Atlantic and the Pacific States; or where the 
Greenwich noon is the basis of simultaneous world observations, 
involving variations up to twelve hours in local conditions; the 
altitudes and locations of the instruments in great cities have 
been not infrequently changed, and the instrumental equipment 
and the methods of computing have never been uniform for 
the long series. It is necessary to overcome these obstacles by