ON MAGNETICAL AND METEOROLOGICAL OBSERVATIONS. 4] 
The pressure of the whole atmosphere at Toronto is therefore, as was to 
be expected, superior to that of the Pacific and inferior to that of the Atlan- 
tic in the same latitude. The pressure of vapour at this point, however, is 
considerably less than under the same latitude in either of the two oceans. 
This circumstance is explained, partly, by the hygrometrical observations 
giving the humidity of the air at Toronto = 0°78, and on the ocean at 40° 
lat. = 0°84, and on the ocean at 45° = 0°85; partly by the mean temperature 
of the continent at Toronto having been found to be considerably lower than 
under the same parallel at sea. 
You will pardon these superficial reflections, considering how difficult it is 
to avoid the temptation of making a preliminary use of a treasure like the 
Toronto observations, even when hoping to devote one’s leisure to its further 
study and development. 
In conclusion, I avail myself of your kind permission to submit two pro- 
posals relative to future magnetical and meteorological labours. 
1. The determinations of mean magnetic values for the year 1829 have not 
yet been completely applied to the deduction of the Gaussian constants of 
terrestrial magnetism for the same year. A comparison of the magnetical 
maps, representing the empirical results on the one hand and those calculated 
by the Gaussian constants on the other, is still far from presenting a perfect 
agreement. For the above year there is still wanting, therefore,— 
_ 1. Those values of the constants which best correspond to the existing 
observations; and 
2. The probable errors of each of those first bases of the theory. 
This deficiency appears to me a material one as regards science. The 
English and Russian observations combined will afford the most probable 
values of these constants for the year 1845, and it is consequently most de- 
sirable to learn, by a comparison with equally probable values of the same 
for the year 1829, their annual variation. Even should these values for 1829 
be somewhat less exact than the later ones, this circumstance is not of mate- 
rial importance, if the amount of their probable error is ascertained. Now 
both these requisites, the best possible determination of the constants for 1829 
and the calculation of their probable error, can be effected in the following 
manner :— 
1. By forming from each magnetic element observed in 1829 an equation 
containing on the one hand a known numeric value (i. e. the difference 
between the observed value of this element and the same calculated 
theoretically with the assumed preliminary amounts of the constants), 
and on the other certain given multiples of the corrections to be ap- 
plied singly to each of the 24 Gaussian constants ; and 
2. By resolving, according to the method of least squares, the linear equa- 
tions with 24 unknown elements, obtained in this manner (amounting 
in all to about 1000). 
By this means we shall obtain not only the most probable corrected values of 
the above constants, but likewise the probable error of each of them. 
In Schumacher’s Astronomical Notices (Astron. Nachrichten, Nos. 4.50, 452 
and 4.54:), the commencement of this undertaking has been published as exe- 
cuted at my request with exemplary zeal by a scientific young friend of mine. 
At my own suggestion, however, the prosecution and conclusion of this work 
has been deferred to a period when it might be performed without a ruinous 
sacrifice of time and trouble on the part of the individual engaged in it. It 
became evident indeed that a calculation of this extent would necessitate an 
entire devotion to the task of one or two years, for which the pecuniary assist- 
ance of some government would be indispensable. If you should be of the 
