OCTOBER 29, 1897. ] 
face at 180°C. The importance of the ob- 
Served increase in compressibility and in 
volume contraction in the lapse of time was 
especially dwelt upon. In the latter paper 
he said that elastic tubing punctured with 
5,000 to 10,000 extremely small holes was 
subjected to external and internal pressure. 
The mean diameter of the pores in any 
given case is computed from the pressure- 
difference just necessary to cause a flow of 
gas against the capillary reaction of the 
wet porous septum. The author points out 
various applications of such tubes. Among 
other points, the cyclic character of the 
flows of liquid through the pores, when 
pressure increases and decreases consecu- 
tively, is exhibited. 
Professor Frederick Bedell read a paper 
written by himself, Professor R. E. Chand- 
ler and Mr. R. H. Sherwood, Jr., on the 
predetermination of transformer regula- 
tion. To determine the ‘regulation by this 
method a wattmeter, ammeter and volt- 
meter are located in the primary circuit. 
One set of readings is taken with the sec- 
ondary short-circuited by a stout copper 
wire, the primary voltage being adjusted 
until the normal full-load current or any de- 
sired fraction of it flows in the transformer. 
No other data for obtaining a complete reg- 
ulation curveis required except the one set 
of readings above mentioned and the mag- 
netizing current. If a wattmeter reading 
is taken when the magnetizing current is 
measured, the data is sufficient to plot a 
complete efficiency curve as well as a curve 
for the regulation of the transformer. The 
two sets of measurements then consist of 
the reading of a wattmeter, voltmeter and 
ammeter, first on short-circuit with normal 
current, and second on open-circuit at nor- 
mal voltage. The wattmeter reading in 
the first case gives the copper losses; in 
the second ease, the core losses. Itis com- 
monly convenient to use the high-potential 
coil as primary in the short-circuit meas- 
SCIENCE. 
653 
urements, and the low-potential coil as pri- 
mary in the open-circuit measurements. A. 
high-potential supply is not then needed, 
and as no power is required except to sup- 
ply the losses, the complete test of a trans- 
former may be made with an incandescent 
lighting circuit for the source of supply, a 
50 light transformer being tested from one 
16 c. p. lamp socket. The total drop is 
found by laying off in the proper manner 
the inductive drop, the magnetic leakage 
drop and the drop due to ohmic resistance. 
The method is theoretically an almost ex- 
act one. Practically it is an exact method 
and less likely to error than the ordinary 
method of determining the regulation of a 
transformer by loading it. The results 
given in this paper (given in full in The 
Electrical World), from a long series of tests 
on seven transformers of various makes, 
shows the reliability of the method, the 
secondary voltage at full load determined 
by it varying usually less than one or 
two-tenths of a volt from the voltage as 
found by measurement on the transformer 
when actually loaded. An approximate 
method (by Kapp) used by one of 
our large electrical companies gives less 
accurate results. The reader is referred to 
the data given in the full paper. 
An electrical thermostat, by Dr. W. R. 
Whitney, consisted of mercury in a U tube. 
One arm of the tube contained ether, and 
the other air, at reduced pressure. One 
platinum wire was fused into the bottom 
of the tube and another above the mercury 
in the leg of the tube containing the ether. 
On the rise of the temperature the ether 
gas expanded much more rapidly than the 
air and forced down the mercury below 
the ether; a control electrical circuit 
through the platinum wires and the mer- 
cury is thus broken, and through it the 
heating circuit. Professor W. O. Atwater 
and E. B. Rosa described their apparatus 
for testing the law of conservation of energy 
