28 BULLETIN 1300, U. S. DEPARTMENT OF AGRICULTURE 
other hand, winter slaughter and summer weight are determined 
only 27 per cent and 24 per cent, respectively, by preceding crop and 
37 per cent and 31 per cent by preceding prices. They must be 
determined largely by factors in the hog situation other than the corn 
crop. Both summer and winter prices are determined relatively 
little by the preceding corn crops (22 per cent and 24 per cent, respec- 
tively), but are at least half determined by corn prices (50 per cent 
and 55 per cent). A direct influence of corn price on hog prices is 
again suggested, but also the influence of fluctuations in the general 
price level. 
THE HOG VARIABLES 
The correlations among the hog variables are given in Tables 4, 
5, and 6. The more important ones are shown graphically in 
Figures 18 to 25, which are made according to the same plan as 
Figures 16 and 17. Each figure shows the deviations which occurred 
in the various quantities relative to a unit deviation of a given one 
of them. 
Inspection of these figures shows at once that the various factors of 
the hog situation are closely bound together. It is a considerable 
problem, however, to disentangle the network of correlations in such 
way as to give a consistent interpretation of the causal relations. 
Pork production, of course, is simply the product of slaughter and 
dressed weight, the latter of which may be assumed to be a nearly 
constant proportion of the live weight. In the case of winter pork 
we find 92 per cent determination by slaughter, 7 per cent by live 
weight, with 9 per cent to be subtracted because of the negative 
correlation between the two factors. The sum is 90 per cent, leaving 
a spurious 10 per cent. In the case of summer pork, there is again 92 
per cent determination by slaughter, but 4 per cent determination by 
weight with 1 per cent to be subtracted because of negative correla- 
tion — 95 per cent apparent determination, leaving only 5 per cent 
spurious. In both cases pork is almost wholly determined by 
slaughter, the percentage fluctuations in weight about its trend being 
too small to be of much effect (compare Table 2). Live weight is 
thus of little interest except so far as it gives an indication of the hog 
situation. 
In general, the correlations involving pork production are closely 
similar to the corresponding ones relating to slaughter, with, however, 
slight differences which follow live weight. In a few cases involving 
summer pork there is an apparent discrepancy due to the different 
periods of time involved. The data for summer slaughter were used, 
beginning with 1871; while data on summer weight, and hence pork, 
were available only beginning with 1889. 
Total western slaughter is simply the sum of the western summer 
and the western winter slaughter. The correlations indicate that its 
fluctuations are determined 35 per cent by the former, 24 per cent by 
the latter, and 37 per cent jointly, giving a total of 96 per cent. The 
remaining 4 per cent is spurious. 
As already noted, the farm price was found for the very middle of 
the winter season (January 1). Accordingly it shows a correlation of 
+ 0.90 with the price of the western winter pack. The high correla- 
tion ( + 0.77) with the preceding western summer price, with which 
winter price shows a smaller correlation ( + 0.69), indicates that farm 
