112 SURFACE CONDENSERS. 
corresponding change in pitch is unnecessary; also that better results will be obtained 
by leaving out all the baffles except such as are necessary to protect the cooler. The 
straight and uniform drilling with uniform pitch can be offset, as necessary, by scientific 
arrangement of passages and spacing between the tubes. All this will be developed in 
marine work the same as it has in stationary work, provided that advantage is taken 
of the stationary experience. 
As regards the tables of performance given in Plates 45 and 46, I do not consider 
these figures of value as giving data to make comparisons with other tests as usually 
reported; data are not given by which various essential figures can be checked. The 
tables are evidently prepared more with an idea of getting a general comparison rather 
than engineering or scientific discussion or study. 
In the first place the number of vacuum gauges on a condenser is of interest and 
necessary to the builders who study the design, but the average vacuum in the condenser 
is of little importance so far as the guarantee is concerned. The only vacuum that is 
really desired is the one which is not stated—that is, the vacuum at the exhaust or inlet 
flange of the condenser. If the turbine exhaust flange bolts direct to the condenser, 
that figure is given, but is somewhat different from the figures assumed in calculating 
the performance. However, great care has to be exercised in placing gauges in different 
sections of the condenser to insure getting accurate readings not influenced by eddies 
or velocity head, and while the condenser performance might show better by taking the 
average vacuum when there is a large loss through the condenser, that does not interest 
the buyer or user of the condenser. 
In the next place, tests cannot be figured by taking the average vacuum for a long 
run and then correcting this figure to the corresponding temperature, which varies on 
a different curve entirely from the vacuum curve. The heat contents of the steam en- 
tering the condenser are stated. These look rather high to us, but there is no means for 
checking them. Also an average steam consumption is assumed throughout the run 
which cannot be possible provided a uniform quantity of circulating water is used, as 
the different observations show as much as 2% degrees difference in the temperature 
rise in the circulating water. 
There is no statement given as to the method of the plus and minus corrections 
shown on Plate 45, and this essentially modifies even the average vacuum taken for 
figuring the tests. The geometric mean is also taken for determining the performance 
of the condenser. Ordinarily the arithmetric mean is taken for the stationary tests with 
which comparisons will be made, and the arithmetric mean is probably the safest for 
purposes of comparison under ordinary conditions of operation. In this case the range 
in temperature is considerable, and the discharge water is brought very close to the 
temperature corresponding to the vacuum. This is not the place to discuss the relative 
value of the arithmetric and geometric means, both of which are based on conditions 
which do not exist in practice. 
For example, the geometric or logarithmic mean, while correct for an ideally perfect 
condenser, assumes that the steam temperature and heat transfer are constant and uni- 
form throughout the condenser and that the heat transmitted at all times is directly 
proportional to the temperature difference between the steam and circulating water. 
These conditions are not met in practice as we all know. The logarithmic mean becomes 
of less and less value as the temperature of the discharge water approaches that of the 
steam. For example, if we assume the case where the discharge circulating water is 
