40 SCIENCE 
being less euphonious than the other. But it 
may be too late to stop the illogical use of py 
for the millimicron (mp) in the domain of 
opties. | 
There will undoubtedly be a tendency, as 
reckoning by powers of 1,000 comes more into 
use, for work in each particular line to be 
always expressed in the same derived unit. 
Here the advantage of a,common unit more 
than compensates for the fact that in some 
particular cases the unit is not quite the most 
convenient. For instance, workers with ther- 
moelements have generally found it advan- 
tageous to work in microvolts, and to keep to 
this unit even when the number of microvolts 
is over 10,000, that is, more than 10 millivolts. 
In case of doubt between two units, it is 
probably better to use the smaller. For this 
diminishes the use of fractions, and also gives 
records more likely to be concordant with fu- 
ture work, since the increase of accuracy as 
time goes on increases the advantage of the 
lower unit. 
In no class of quantities is more to be 
gained by reckoning by powers of 1,000 than 
with coefficients of expansion, and temperature 
and pressure coefficients generally. If these 
quantities were always tabulated in thou- 
sandths or millionths, instead of with a vari- 
able number of zeros, according to the fancy 
or convenience of the tabulator, a very much 
larger number of them would actually lie in 
the memory of the average working experi- 
menter than are now to be found there. Yet 
these quantities, and some others, being pure 
numbers, have no special name, and therefore 
nothing to which the prefixes, milli-, micro-, 
ete., can be attached. They may legitimately 
be designated as “ parts per mille,” “ parts per 
million,” ete., but these expressions are rather 
awkward, particularly when the whole expres- 
sion is “parts per mille per degree,” or some- 
thing like that. It would be convenient to 
use the fractional prefixes alone as nouns in 
such eases, milli meaning one part per thou- 
sand, and micro, one part per million. There 
is certainly considerable reason to wish that 
some leader, or committee, having sufficient 
LN. S. Von. XXXV. No. 888 
authority, would authorize the use in this way 
of these terms (or something better). They 
have these advantages: They are brief; they 
would harmonize with the terms used for other 
physical quantities; they would tend to extend 
the use of powers of 1,000. For instance, at 
present, most observers, working to an accu- 
racy of (say) 10 parts per million, would pre- 
fer to state it as one part per 100,000, while 
with the word “micro” in use the almost 
universal expression would be 10 micros. And 
the use of powers of 1,000 is quite as desirable 
in stating errors, etc., as in most other cases. 
The use of fractional or multiple prefixes 
also sounds a little strange in those cases 
where, in order to adhere strictly to the O.G.S. 
system, the centimeter is used as the unit of 
all linear measurements. The real difficulty 
here, however, does not lie in the prefixes, but 
in the fact that two different fundamental 
units, the meter and centimeter, are in use, 
and that most physicists are probably more 
used to measuring small lengths in millimeters 
and microns. This difficulty would not be in- 
creased by the use of the term millicentimeter 
and microcentimeter, which are of course the 
logical terms to use if the centimeter is to 
become the practical unit of all lengths. It 
also seems logical to use the centimeter only 
where such other C.G.S. units as the absolute 
electrostatic and electromagnetic units would 
be used, and to use the millimeter and micron 
in eases corresponding to those where the 
ampere, ohm and volt would be considered 
appropriate. 
In any case, it may be well to repeat, the 
main and essential advantage of the newer 
system that is coming into use is in the re- 
striction of notation to powers of ten which 
are also powers of 1,000. And this restriction 
can profitably be adopted whatever may be 
thought or done regarding the other points 
mentioned in this paper. 
W. P. WHITE 
GEOPHYSICAL LABORATORY, 
CARNEGIE INSTITUTION OF WASHINGTON, 
WASHINGTON, D. C., 
November 1, 1911 
