86 REPORT—1885. 
of the original purposes as to the uses to which the records were to be 
applied. 
PPS. The fact is that funds have been expended too exclusively upon 
material appliances, and upon agency for working them: the statesman 
can understand that his country gets a tangible return when observatory 
buildings, instruments, operators, records, and reports appear before him 
as a result of the grants that he makes; but it is for the man of science, 
the original adviser, to make him understand that these are very delusive 
results unless supplemented by appropriate measurement, computation, 
and discussion. 
4, And this is the more important inasmuch as the cost of utilising 
the records, even up to the point suggested by Sabine’s examples, will 
at least equal the amount that has been expended in their production. 
It is indispensable that inexpensive measuring, copying, and computing 
power should be used, under skilled direction, on a large scale; and here 
it is that the main part of the cost arises. It would be simple waste of 
superior energy to set a cultivated physicist to the appalling task of per- 
forming the simple but multitudinous series of operations that are involved 
in any adequate treatment of the observations ; and it is to the insufficiency 
of suitable agency in the working power of existing observatories that is 
probably to be attributed the fact that so little has yet been done in the 
way of independent reduction and discussion of the records of the auto- 
matic magnetic instruments. That the work before us is laborious and 
costly is, however, no argument against the undertaking of it if we have 
reason to believe that an adequate return will be obtained; and a more 
costly process is to be preferred to a less costly one if the quality of the 
results that are the outcome of it is higher in a corresponding degree. 
5. I cannot but think that the wonderful progress made during the 
last century in the experimental sciences is apt to make us unduly im- 
patient of the necessarily slower progress of the observational sciences. 
tf astronomy had, during the progress of observation, to have its period 
of phenomenal generalisation—its Ptolemy, its Copernicus, its Kepler— 
before light as to the mode of physical causation dawned upon its 
Newton, is it much to be wondered at that a much more complicated 
science, as terrestrial magnetism undoubtedly is, should have to pass 
through its period of discovery of general phenomenal relations—relations 
which the physical theory will ultimately have to explain—before the 
conditions essential to the conception of a general theory can be laid 
down? 
6. It will be seen that whilst I have no faith in the flights of genius 
that would look at the crude facts as nature presents them to us, and 
from such complex data devise a theory to unravel the complexity, I have 
the greatest confidence in appropriate methods of analysation as leading 
to relatively simple phenomenal generalisations, and thence, inevitably in 
the long run, to the desired physical theory. The first step to be taken 
should, I think, be to collect together all accessible results that have 
already been worked out and published of the nature of— 
(1) The regular solar-diurnal variations ; 
(2) The disturbance variations—dinrnal, annual, and secular; and 
(8) The lunar-diurnal variations ; 
and to convert the expression of them for each of the elements, declination, 
horizontal force, and vertical force, as far as available, into metre-gramme- 
second or ©.G.S. units of force. If not already done, the averages of 
