126 
between the means of the two sets of 70 each made under a 
difference of temperature of 70° It is obvious, therefore, 
that Mr. Spence’s experiments, though evidently made with 
great care, afford no certain evidence that any sensible 
change takes place in the strength of cast iron when its 
temperature is reduced from 70° to zero of Fahrenheit’s scale. 
As showing the little reliance to be placed, in certain 
cases, on results derived from short series of experiments, I 
may mention that in Mr. Spence's experiments, notwith- 
standing the very great diversity in the breaking weights 
of the bars used, and the care taken to mix them as much 
as possible before testing, there is in one case a run of 
eleven consecutive experiments in all of which the breaking 
weights are below the general average; while in another 
there is a run of eight in which the breaking weights are 
all above the average. Similar runs of six and five each 
occur several times. Facts like these will show to those 
who have little experience in the application of the theory 
of errors how necessary it is, in some inquiries at least, to 
multiply experiments as much as possible before proceeding 
to deduce results and draw conclusions. Taking all the 
experiments on the effect of cold on iron which have yet 
been brought before the Society, they can only be regarded 
as indicating that if any effect at all is produced, it is more 
apparent on iron of good quality than on inferior iron, but 
that its amount is so small as to be wholly inadequate to 
account for the railway and other accidents which have 
been attributed to it. 
Mi\ Brockbank stated that at the time he entered upon 
the experiments communicated to the Society, he had no 
knowledge of those made by Mr. Knut Styffe of Stockholm, 
and C. P. Sandberg, A.I.C.E./of London, as detailed in the 
English translation of Mr. Styffe’s work on the Strength of 
Iron and Steel. He was however pleased to find that the 
researches of these gentlemen confirmed the conclusions 
drawn from his own experiments ; and he especially pointed 
out that in Mr. Sandberg’s experiments on the Strength of 
Rails, the objection raised as to the hardness of the ground 
