480 
On Scientific Method . 
[Odtober, 
observation beyond certain limits. Our powers of hearing 
are not delicate enough to perceive vibrations exceeding 
somewhere about 38,000 per second ; hence if there be sound- 
producing vibrations quicker than these no amount of observa- 
tion will enable us to detedt them. Not only is this method un- 
able to pass beyond somewhat narrow limits, — it is also liable 
to lead us to untrue conclusions, unless it be very carefully 
used. Do we not often hear it said — “ See how the buildings of 
our great ancestors have lasted during the centuries, — strong 
and firm these old temples, or walls, or roads remain to this 
day, while the structure which we have raised to-day by to- 
morrow begins to decay”? Observation tells us that the 
older buildings remain ; observation tells us that the newer 
quickly disappear ; but observation does not tell us that it is 
only the great buildings of antiquity which remain — the 
buildings which, from their purpose and design, we should 
expedt to be very strongly put together ; and that the ordi- 
nary houses and the common buildings have all long ago 
utterly disappeared. The conclusion drawn from observation 
alone — viz., that our ancestors built more strongly than we 
do — is therefore a conclusion which is not proved by the evi- 
dence adduced. 
Again, the mind of the observer may be so overcast with 
prejudice or fancy, or may be so dim and dull as either not 
to receive aright the image of outward things, or to transpose 
that image so that it becomes a caricature, not a truthful pic- 
ture. He who when shown, in the old heathen temple, the 
pidture of all those who had been saved from shipwreck after 
paying their vows, and asked to believe now in the power of 
the gods, replied “ But where are they who paid their vows 
and were not saved from shipwreck?” was a man whom we 
have often need to copy. 
By observation alone we cannot tell the exact conditions 
regulating the occurrence of any phenomenon ; these 
conditions can be determined only by experiment. But where 
an observation has been made, and we are wishful to deter- 
mine the exadt conditions which regulate the occurrence 
of the observed fadt, we shall find that it is very diffi- 
cult to fix on these conditions. Every fadt is so closely 
related to so many others that it becomes very hard to strike 
out those conditions which are really non-essential. Nay, it 
often happens that what at first sight appears to be the most 
important condition of a phenomenon proves, after experi- 
ment, to have little or no influence on this phenomenon 
while other overlooked circumstances are the true governing 
causes. Thus if we let fall a piece of lead and a. sheet of 
