I10 POWER DISTRIBUITION FOR EI,ECTRIC RATLR(’)AI)S.' 
generation is raised and that of distribution lowered. In 
substations of the third class, the original cost of genera- 
tion is low, but the transmission of power to, and the 
maintenance of, the substations has to be balanced against 
the cost of distribution from the main station direct, and 
the cost of generating at separate stations. 
This balancing of costs involves very nice discrimina- 
tions and deals with somewhat uncertain factors. The 
true cost of electric power itself is not easy to estimate, 
and actual data from existing stations are often rendered 
valueless by disingenuous bookkeeping. ‘There is a great 
difference between the cost of power computed from fuel 
and labor alone, as is often done by those who like to de- 
ceive themselves, and the cost with all the items of inter- 
est, repairs and depreciation relentlessly footed up. It is 
not unusual to find the item of depreciation deliberately 
neglected in computing the cost of power and in other 
estimates. Street railways have been particularly prone 
to this sort of financial juggling—it is so convenient to in- 
crease the capital account for ‘‘ improvements’’ instead of 
withholding dividends really unearned or shouldering a 
genuine deficit. Without discussing this question of 
financial morality, we cannot too forcibly remind the en- 
gineer not to deceive himself and the manager that if the 
present trend of legislation and ‘‘labor reform’’ continues 
there is likely to come a dreadful day of reckoning in which 
a sinking fund will be sorely needed. 
To determine the conditions of economy that govern 
the establishment of substations, it is first necessary to 
know the probable cost of electric power in stations of dif- 
ferent sizes and kinds. This is not easy to estimate in 
general, but can be gotten at with fair accuracy for any 
given set of conditions as to cost of plant, coal, labor and 
so forth. In small plants the labor item is disproportion- 
ately large and the general eficiency less than in large 
ones. On the other hand in plants of 1000 k.w. output 
and over the labor item remains proportionately nearly the 
same as the plant increases in size, and the efficiency rises