352 Illustrations of the Vai'iate Difference Correlation Method 
and Coffee (c. + '05 + '25) ; they have insignificant negative correlations with 
Shipping (< — -03 + ■25) and Revenue (< — "15 + -24). 
Savings are thus — apart from continual time change — no very satisfactory 
measure of general prosperity, and a fluctuating increase is usually accompanied 
by a reduction of luxuries. 
(i) Coal Index. The importation of coal has little relation to any factor of 
prosperity besides Shipping (c. + '58 + "IT). With Railways the correlation is 
not quite double the probable error and the value, even at the sixth difference, 
appears still falling. The correlation with Revenue only just exceeds the probable 
error (+ "270 + •232). With International Commerce {+'111 + '243), Stamp Duties 
(+•052 + ^250), Savings (+ •196 + ^241) and Coffee (+ -152 + -245) the correlations 
are less than their probable eriors, small and in some cases still falling. With 
the Postal Index, the correlation is negative, insignificant and falling. Alone in 
the case of the Tobacco Index does the correlation appear to be nearly as 
significant as in that of Shipping, but it is negative and increasing* (—•514 + •184), 
while in the case of Shipping it was steady. It is singular to find that Coal, the 
increased import of which should mark increased industrial activity, is, beyond 
the naturally influenced Shipping, alone effectively associated with the con- 
sumption of Tobacco. 
(j) Tobacco Index. This is of considerable interest as marking the association 
of indices of trade prosperity with the consumption of a luxury. With four 
exceptions Tobacco is negatively correlated, although often insignificantly, with 
the other indices. Revenue (+ '115 ± ^247), International Commerce (+ "015 + ^250), 
and Post (+ "108 + '247) are all positive, insignificant, and in the first two cases 
still falling. The correlation with Coffee is positive and might, perhaps, be 
significant (+ '326 ± ^224), but it appears to be still falling. With Coal and 
Savings there are probably significant negative correlations (— •514 + ^184, and 
- -431 +^204 respectively); with Railways (- ^243 + -236), Shipping (- -271 + ^229) 
and Stamp Duties (— •I 29 + ^246) there are insignificant negative correlations, but 
they tend to confirm each other in sign. Thus we see that the consumption of 
tobacco can hardly be considered as a measure of general prosperity ; it appears 
to be greatest when trade conditions are unfavourable, and in particular when 
savings are least and manufacturing conditions as measured by the importation of 
coal are slack. The result suggests the pipe of the unemployed at the street 
corner, rather than the increased expenditure of the fully occupied artisan. 
{k) Coffee Index. This is another luxury and the results are very similar. 
There appears a significant correlation with Revenue (+ '400 + 210), which might 
easily be explained, and there is a falling but possibly significant correlation 
with Tobacco (+ "326 + ^224). With all other indices the relationships are 
* It is, perhaps, hard to beUeve that so much smuggliug could be carried on in eoUiers, that it would 
seriously affect the profits of the tobacco monopoly ! 
