JUNE, 1903. MONTHLY WEATHER RXWEW. 271
period of death, and in part to the action of acids on antho-
cyanin as described below. Many of the destructive changes
which take place in the chlbrophyll are oxidation processes,
the same as occur in the cells of highly coloredvariegated
plants, and physiologically they are not very different from
the changes occurring in calathea, caladium, codiaeum, etc.
The approach of niaturity in the leaf, and the coming on of
cool weather in autumn, stimulate the production of osidiz-
ing ferments, and the action of these and the acids of the cell
sap upon the chromogen, or color contents of the leaves, es-
pecially the chlorophyll and anthocyanin, causes many of the
brilliant colors of autumn foliage. There is a popular belief
that these colors are clue to. cold weather or frosts; but while
frosts, if they are light, liesten the solution ancl destruction
of the chlorophyll, they can not be looked upon as more tliau
hastening changes, which would occur in time without them.
Even in the Tropics, some foliage before it matures becomes
highly colored, and on the Japanese maples the writer has ob-
served beautiful autumnal colorations in July in the region of
Washington.
In practically all deciduous trees, bushes, etc., before the
maturing and falling of the leaves, a11 of the valuable food
materials, such as sugars, albuininoicls, etc., pass froin the
leaves through the vascular bundles into the twigs and
branches, so that they tire not lost to the plant. Wheii the
leaves finally fall they are therefore nothing but mere skele-
tons, containing waste materials. I n the passage, especially
of albnminoid matters from the leaves to the stems, i t is nec-
essary that the materials be protected from the strong action
of light, ancl i t is believed that part of the coloration of ma-
turing leaves serves the purpose.
A coloring material, or chromogen, known as anthocyanin,
is always present in such cases, and develops beautiful reds
when the cell sap is acid, blue when no acids are present, and
violet when there is only slight acidity. This, in connection
with the disorganizing chlorophyll, causes the various mixtures
of yellow, brown, violet, red, orange, etc., of autumnal colora-
tion as described above. I n very young leaves of many plauts,
such as di1n)ithits gln)zdirlosn, Jirylutis regin, Vifis, C‘issiis, and
many other plants, this same anthocyanin is developed as a
protection to the albuminoid materials traveling to the youug
cells. Such protective colorations have to be dibtinctly sepa-
rated froin variegations. I n evergreen leaves, during the
winter, the chlorophyll granules are protected by the develop-
ment of anthocyanin, forming a brownish or reddish tinge in
the cell sap.
While, as stated above, these protective and in some cases
transitory colorations shoulcl be clearly clistinguishecl froin
variegation, i t is an interesting fact that they develop when the
conditions for active nutritic in are unfavorable, and may, in many
cases, be produced in maturing leaves by starving the plants or
permitting them to become sufficiently dry to check growth.
This is especially prominent in many conifers.
THE WEATHER BUREAU SEISMOGRAPH.
By Prof. C. F. MARVIN, dated July 1, 1903.
It has always been the policy of the Weather Bureau to
require its observers to take careful note of earthquake phe-
nomena of sufficient intensity to be felt at stations, but no
specific effort has been macle to provide generally the instru-
mental means by which such phenomena could be automatic-
ally recorded and measured. The Central Office at Washing-
ton has, however, maintained a simple form of seismograph in
operation ever since December, 1892, ancl recently has greatly
improved its equipment by the installation of one of the large
horizontal pendulums made by J. Sr A. Bosch, of Ytrassburg,
and designed after the models described by Omori.’
1 Publications o f the Earthquake Investigation Committee iu Foreign
Languages. No. 5. Tokio, 1YU%
The older form of the Weather Bureau seismograph was
described by the writer in the MONTHLY WEATHER REVIEW for
July, 1895, Vol. SSIII, p. 250.
The new instrument is of a very superior type and gives an
accurate record of the movement of the earth at the pendulum
in the horizontal plane. At the present time but one of the
two pendulums constituting the set has been installed and
this produces a record of the north and south component of
horizontal motion.
The iuechanical principles involved in the construction of a
seismograph of this type were first developed and applied to
the ineasureinent of earthquakes in t,he latter part of 1880 by
James A. Ewing, then Professor of Mechanical Eiigineering at
the University of Tokio, but now Professor of Mechanism and
Applied Mechanics at tlie University of Cambridge, England.
Numerous nioclifications have since been incorporated in the
instriiineiit by Gray, Oinori, ancl others, and in its present
form i t is well adapted to measure and record all kinds of
earthquakes, except, perhaps, the most destructive, and is
especially suited to register the feeble, unfelt earthquakes,
which frequently occur in a11 parts of the morlcl.
The instrunleiit as set up to photograph is shown in fig. 1.
As actually iiistalled in a small basement room of the Weather
Bureau. the separate castillgs are secured to thick blocks of
stone cemented firmly into the concrete floor of the building
aiicl projecting but a few inches above the floor level. The
heavy casting A forms the support for the so-called horizontal
pendulum B C I). is a massive lead weight, rigidly attached
to the c o n i d tubular rod H, the end of which, at 11, terminates
in a hardened steel plug, hollowed out cnp-wise and highly
polished. At this point the pendulum is supportetl upon a
sharp, conicd pointed stud of hardened steel fised to the cast-
ing A. The remaining support for the pendulum consists of a
pair of steel wires, faintly seeii at w w in the picture. At the
weight end these are attached to eyes with a knife-formecl inside
edge and there engage two studs that project laterally from
the mass P. At D the wires are united to a stirrup. which at
the point opposite the wires is provided with a bit of hardened
steel formed with a cup-shaped recess and highly polished.
Here the stirrup is supported on a sharp, hardened steel cone
attached to the carrier E forming the summit of the casting A.
The carrier E is provided with several acljusting Rcrews; thus,
n serves to raise or lower the weight C‘ and thus adjust i t to a
horizoiital position, while b causes the point at 1) to move away
from or nearer to the top of the column, and. finally, R pair of
screws, one of which is seen at r, gires I) a lateral motion in
the horizontal plane. In short, the pendulum h’ C 1) is sup-
ported a t H and I) 011 sharp steel points and swings, there-
fore, with great freecloin of motion. If the points 11 and 1)
are rigorously in a vertical line, the pendulum is in neutral
equilibrium and the mass C‘ will then remain a t rest in any
position. For practical work, however, a small degree of sta-
bility must be iinparted to the mass (‘, otherwise minute changes
of temperature and other influences which it is impossible to
control will cause the mass (J to wander about from one position
of rest to another. The desired degree of stability is given to
the pendululn by means of t,he screw b and the azimuth of the
point of rest is adjusted by the screws f-. The degree of sta-
bility is determined by noting the time of vibration of the mass
C:, which can he adjusted to swing as slowly as one complete
vibration in thirty or forty secoiicls. A period of twenty-five
to thirty seconds seems to contribute a sufficient stability for
practical work.
The whole object sought in this construction is to secure a
“steady mass,” as i t is called; that is, a inass that shall remain
quite at rest during an earthquake, notwithstanding that the
earth and the supports for the mass are undergoing apprecia-
ble vibratory displacements. The kinetic property of bodies
utilized in this connection is that which gives rise to the so-
272 MONTHLY WEATHER REVIEW. JUNE, 1903
FIG. 1.
called n.r& ?f imfantam~oris I.ofr/tifJ,l. Tvhenever aiiy force is
applied in a one-sided fashion to iiiove a boclg all points of
the body mill not move in tlie same manner, in fact, one p i n t ,
or a line of points, will actually remain sensibly a t rest for
small displacements. This point, in mechanics, is called the
a.ris (f inaturitaworis rotntioii, or tlir cwtw c f l w c u s s i o i t .
In the case of the pendulum R, r f , P nearly all the mass is
concentrated at C and the result is tliat the axis of instanta-
neous rotation, or, as we shall call it, the steady line, is, at a
point very near the center of t8he mass C'. Consequently
whenever the support d is displrtcecl horizontally with a vibra-
tory motion, as in t,he case of an earthquake, the steady line
of the mass C will remain at rest for all movements transverse
to the plane B, C. I). Motion directly in the line of the strut
is communicated, of course, to C, but the registering meclian-
ism is so disposed that such motions produce no record what-
ever. Although the inass (,' is very largely displaced when-
ever the support d is tilted even in the slightest degree in a
direction perpendicular to the plane B, C!, D, results have
nevertheless shown that tilting is not an appreciable feature
in earthclualie motion, except near the origin of the disturb-
ance. or possibly in the case of large wayes. We find then
that the steady line of the inass Creninins relatively stationary
during an earthquake disturlmnce, and the manner of record-
ing tlie movement of the earth with respect to this point is
shown more in detail in fig. 2.
The magnifying and recording lever shown at L is made of
very thin sheet aluminum bent into an inverted trough-shaped
section to secure stiffness, and is provided with a conical
pointed, hardened steel, axis, d, which is centered and carried
in the stirrup F which, in turn, is adjustably but firmly attached
to the heavy casting C. The short arm of the lever L is slotted
and engages the slender staff f in the manner shown. The
stafi.f is made of hardened steel, with conical pivot points
centered in the stirrup F', which is securely attached to the
mass C in such a position that the center of the staff f lies in
the prolongation of the steady line of the mass C.
The record is traced on a sheet of smoked paper wrapped
around the large cylinder R, fig 1. In order that the friction
may be reduced as far 8s practicable at the tracing point, the
JUNE, 1903.
FIG. 2,
273
coating of soot is made relatively thin and a paper with a highly
glazed surface is employed. Much depends also upon the
tracing point which is a mere bit of steel, s, pivoted to the lever
in the manner shown a t J. The degree of magnification can
be adjusted to suit by shifting the carrier or stirrup F so as
to increase or shorten the distance between the staff .f a i d the
pivot d. Provision is macle for a magnification of from 5 to 15
times. A 10-fold dagnification seems to give about the best
results for feeble earthquakes.
Every precaution must be taken in the coiistruction and
mounting of the lever I; to satisfy tlie following conditions:
1.-To reduce the mass of the lever to the mininium with-
out serious loss of stiflness and rigidity.
11.-To eliminate friction a t the pivots ancl stylus to the least
degree without perceptible shake or lost motion in the pivot
points.
111.-To proportion the two arms of the lever SO that its asis
of instantaneous rotation shall fall a t or near the point where
the forked end engages the staff -f This requirenient is of
considerable importance if the lever L has milch iiiass, but
can generally be ignored by keeping the weight clown to the
lowest possible limit.
1V.-To eliminate every trace of looseness in the pivots and
in the forked connection, but at the sanie time avoid friction.
The record cylinder R, fig. 1, is driven at the rate of one revolu-
tion per hour, and the asis at one end is cut with a steep screw-
thread which shifts the cylinder endwise as i t revolves. The
stylus therefore traces a helical line on the drum, thus sepa-
rating the successive portions of the record.
The mechanism at ilI is an electric time marker, the magnet
of whichis in connection with a circuit closer actuated by a
high grade clock. The circuit is closed momentarily once each
minute, and causes a finger to mark a time stroke each minute
on the trace.
We hardly need to esplain that when an earthquake occurs
all the parts of the instrument partake of the motion of the
earth except the steady line of the mass C'. This remains rela-
tively stationary in space for horizontal displacements perpen-
dicular to R (-'. The lever L, at the point where i t engages
the staff .f, likewise remains a t rest, hence i t follows that the
stylus will trace on the drum a magnified record of the lateral
displacements executed by the pivot cl; that is by the ground
supporting the instrument.
An example of a record of feeble earthquake is given in the
MONTHLY WEATHER REVIEW for March, 1903, on page 136. Since
that record was macle the sensitiveness of the pendulum has
been increased a little, that is, the points B ancl P have been
brought still closer to the vertical line. The time of a com-
plete vibration of the pendulum is nom about twenty-five
seconds. A simple pendulum that would vibrate in the same
time would require to be nearly li93 feet long.
NOTES UPON THE THEUHP AND USE OF THE PENDULUM.
The equation for the time of vibration of a horizontal pen-
dulum is found as follows:
Let I equal distance from the axis of suspension to the center
of oscillation of the pe~iduluin considered as a siinple penclu-
lum. I n other worcls let I be the radius of gyration of the
mass wit,li respect to tlie axis of suspension. Now as a simple
pendulum tlie time of vibration will be:
t =2 7 -
where y is the acceleration due to gravity or whatever forcc 4:
I
I
F~ci. 3,
causes the pendulum to oscillate. I n the case of a horizontal
peiicluluni only a very small component, .f, fig. 3, of gravity
acts to produce the oscillation. hence the time of oscillation
will be reduced iii proportion. Thus we have *f = 9 sin i, and
the time of a complete vibration is
I .-
2 74 MONTEfLY WEATHER REVIEW. JUNE, 1903
from which
4 2 1 sin i =
~
9 T'
which gives the angular inclination to the vertical of the
asis of the pendulum corresponding to a given time of vibra-
tion T! I f k is the linear distance between the two pivot
point bearings of the pendulum, then the linear displacement
of the upper point from the vertical line through the lower
point (that is, D 1)' = n, fig. 3) will be
. . d x ' h l n = h s1n 1. =
--.-e
y T'
It is of great importance to investigate the effects of very
small changes in the level on such a penclulum as that nom
under consideration. It is plain that such influences as varia-
tions in temperature, either in the iron column -4. fig. 1, or
the stone base, or any slight tiltings of the ground will pro-
duce changes in the position of the point D with respect to B,
the result of which may cause the pendulum to wander from
one position of rest to another whenever such changes occur.
Now, while the point I) is very nearly vertically above the
point H yet in general the distance I ) I)' = n is a magnitude
of large order relative to the iiiinute effects of temperature
changes or the tiltings of the foundations clue to ordinary in-
fluences. Th~is, for example, in the Weather Bureau pendu-
lum we have the following dimensions: 1 = 753 mm., h = 958
mm., If7= 11.4 kilos. When T= 25 seconds, n = 4.6.17 mm.
Whereas the displacements due to the influences mentioned
are but a few hundredths of a millimeter in magnitude.
Let fig. 4 represent a cliagramatic view of the pendulum and
magnifying lever, as seen from the top. The point at I)' is
the end projection of the vertical line through the bottom
pivot point R of the pendulum. C is the heavy mass and D
is the top pivot.
I -r- , /I
I -'.
r--x--h Y ;---,.I --------- 1 ____--_____ J I
__---
FIG. 4.
Let t i = the ratio of the long to the short ann of the mag-
nifying lever. I f now we consider the pivot of the lever sta-
tionary and the pivotf, fig. 2, to move, then the tracing point will
give a record of the clisplacement o f f magnified 11 times. 80
likewise, as happens during an earthquake, i f C! and the pivot
fremain at rest and the pivot d, fig. 2, moves with the earth, the
magnification will still be 71 times, since tlie record sheet is
itself displaced an amount eqiial to that of the pivot.
I f now any of the influences mentioned causes a siuall dis-
placement of the point I) esactly in the direction of the line
I)' C', the only effect will be to alter the t,ime of oscillation of
the pendulum. A clisplacement in any other directiou, how-
eves, as, for example, to F will not only modify the period, but
will cause the pendulum to swing to a new poAition of rest in
the azimuth I)' (' I . Inasmuch, however, as the distance 1, D'
is relatively large compared with the minute displacements we
have in mind, the alteration they may produce in the period
is of no consequence whatever and we are therefore nom con-
cerned only with the component of displacements perpenclicu-
lar to the line I)' C'. We virtually assuiue that the clisplace-
ment is tangent to the circle through 1) with center at, 1)'.
Let J D be a small lateral clisplacement of the point 1) a d
let u be the corresponding angular change of the pendulum
J 1) in azimuth; then tan u = and i f d is the clisplacement on
A D the record sheet then d = vi 1 tan u = -- )1 1 from which (I.
That is, the value of 3 B in linear iiieasuse is
I f p is the corresponding angular tilting of the ground its
measure is
(1 )
Since for the sinall angular magnitudes now uncles consiclera-
tion, sin p is proportional to q we may write-
I n which y" = the angular displacement t.SlJreSSeC1 in seconds
of arc.
I t is seen from the laht equation that the seiihitireness to
angular tilting depends entirely upon the time of vibration of
the pendulum, so that two horizontal pendulums of the same
time of vibration, but otherwise of different proportions are,
nevertlielebs, equally sensitive to angular tilting of the verti-
cal axis.
It iH plain also that i f the motion during an earthquake con-
sists of tiltings rather than horizontal displacements the evalu-
ation of the record must lie macle by means of equation (2).
With 1 2 = 10, T = 25 seconds, and d meabhred in millimeters.
y." = 0.1329 t l , (3)
(4) ana J L, = 0 .0 ~i ' i (I.
Equations 3 and 4 give the numerical values of y'' and J LJ
for the Weather Bureau penclulum when the magnification is
10 and the period of vibration is twenty-five seconds. The
extreme sensitiveness to tilting is eshibitecl in several ways.
The weight of the observer almost anywhere on the floor of
the small room in which the instrument is installed suffices
to tilt the pendulum enough to show on the record, a large
displacement is produced by standing at one side of the ped-
estal. It has been noticed also that the weight of an ice
wagon which stops daily to deliver ice at a basement entrance
to the building causes a definite displacement of the trace of
about one niillimeter which disappears when the wagon drives
away. There are no vibrations or oscillations registered, only
a distinct elastic bencliug of the ground due to the load.
This motion, moreover, is c.oiumunicaterl through the foun-
dation walls of the builcling. The distance of the wagon
from the seismograph is about 2U feet; the asphalted drive
and the basement floor are on the same level. The subsoil is
a hard clay.
It may be added that tlie road is a private driveway back
of the building and is rarely used SO that the eEects and dis-
turbances of the record due to its proximity are not regarded
as interfering in the least with the validity of earthquake
records.
Aside from transitory tiltings of the ground of the kind
just cliscussed others of a more gradual character are also ob-
wrvecl. If the pendulum were to remain absolutely stationary
during twenty-four hours tlie record sheet would contain
twenty-four straight parallel lines quite accurately spaced
three millimeters apart.
The qmcings are never exact, lmt are soinetimes quite uni-
form. Generally, however, there is a distinct and progressive
widening os narrowing of the spacings acrom the sheet, show-
ing that a slow progressive tilting of the column or the ground
has been in progress during the twenty-four hours in question.
JUNE, 1903. MONTHLY WEATHEX REVIEW. 276
While some of these displacements must be attributed to
temperature changes and effects entirely within the instru-
ment, yet slow tiltings of the ground also occur, due to a
variety of causes. The seismograph, as now installed, answers
every purpose for the registration of distinctively earthquake
movements, but the slow tiltings referred to can not be
studied satisfactorily in the present location of the apparatus
which for such purposes should be isolated as far as practicable.
OBSERVATIONS OF SOLAR RADIATION WITH THE
AN(3STROM PYR-LIOMETER, AT PROVIDENCE, R. 1.l
By Mr. HARVEY N. Davrs, dated Marrh 9.1903.
During the fall of 1901 arrangements were macle by Prof-
Cleveland Abbe, on behalf of the United States Weather
Bureau, and Prof. Carl Barus, of Brown University, for mak-
ing a series of observations upon the amount of solar radiation
received from day to day a t the surface of the earth. AJI
hgstrom electric compensation pyrheliometer, No. 28, and a
Weston milliamperemeter, No. 4315, were accordingly sent to
Providence, E. I., and the work placed in my hands.
As an observing station we finally decided upon a room in
the third story of a house situated in one of the highest parts
of the city. The galvanometer, resistances, and batteries were
permanently fastened to the wall just inside a southern wiiidow,
while the sloping roof outside offered a convenient and ex-
posed support f o r the tripod and pyrheliometer. When in
position the observing tube was about 18s feet above sea level.
As is already well known,’ the instrument consists essentially
of two thin narrow strips of blackened platinum so mounted
as to be exposable to the sun’s radiation. While one is thus
exposed the other is shielded and heated to the same tempera-
ture by the passage of an electric current of known intensity
(usually .2 to .4 amperes), the ammeter and a variable resist-
ance being included in the circuit. The desired equality of
temperature is recognized by means of a secondary thermo-
electric circuit, including a very sensitive galvanometer of the
D’Arsonval type, and a constanten-copper thermal element,
whose junctions are very close to, but electrically insulated from,
the centers of the two strips. At first the instrument was used
with its electrical connections just as they were packed, but a
considerable shifting of the zero point of the galvanometer
soon appeared, and seemed to be due to a bet in the torsion
suspension, caused by the estreiue deflections to which SO sen-
sitive an instrument is liable, before the current strength can
be ~roperly regulated. On this account I was led to introduce
a platinum key into the galvanometer circuit and to use a zero
method, adjusting the current in the main
circuit until no throw was observable when
the key was closed. This key was almost
immediately replaced by a mercury com-
mutator, symmetrical with respect to the
galvanometer ancl iiyrheliometer tube, to
avoid any spurious thermal E. &I. F. in the
circuit. It was also found convenient to
modify the connections of the main circuit
for various reasons, until i t assumed a form schematically represented in fig. 1. P is Frit. 1.
the tube, C: the commutator, and (3 the galvanometer of the
thermo-couple circuit. In the main circuit r is the variable
1 A similar report by Mr. H. H. Xiriball will follom.-ED.
2 See Prof. C. F. Marvin: ‘* The measurement of sunshine and the pre-
liminary examination of Angqtriirn’s pyrheliometer, ‘’ MONTHLY WEATHER
REVIEW, October, 1901.
See also Knut .hg.strijm, Intensit6 de la Radiation Aolaire-Reclierches
faites A T6n6riffe, 1895 et 1836.
See also K. Angstriim, Nova Acta Upsal, 1893: The Physical Review,
1, p. 365, 1893: Wied. Ann. 67, p. 636. 1899; Astrophysical Journal. 9, p. 334,
1899, and Annalen der Physik uncl Chemie, Neue Folge, Baud 67 [ 18991,
p. 633.
Uysal, 1900.
37-3
resistance (which could be made Q) ) and M the ammeter sup-
plied with the apparatus. R is a resistance box of considera-
ble size, which was used, partly to cut down the current on
cloudy days, and partly to keep two Daniels’ cells (B) in con-
dition when the apparatus was not in use. T is a mercury
three-way key.
*Early in December the behavior of the observing tube be-
came very irregular, its resistance often becoming infinite for
no apparent reason whatever. It was, therefore, returnect to
Washington and its contacts thoroughly examined, and, al-
though no trouble could be found, the bad contact was in
some way improved, for it functioned properly during the
rest of the year.
During the summer of 1902 the writer was obliged to give
up the work on account of his absence from the country, and
Mr. Robinson Pierce, jr., also of Brown University, very kindly
undertook it. The apparatus was moved to a similar situation
at his home. a quarter of a mile away on the same ridge, the
altitude of the tube being 163 feet. Here observations were
macle during July, August, and a part of September. Others
were made later in September at the first place. Besides
making these observations, Mr. Pierce has also carried through
a considerable number of the calculations, whose results appear
in the accompanying tables.
The method of observing was as follows: The tube was first
set up and oriented, both strips being esposed to the sun, and
the standard time, the neutral reading of the galvanometer
(the key being open) and the temperature inside the tube were
recorded. The “throw ” when the key was closed was also
observed, and both strips were exposed until this was a mini-
~n u ~n , and usually very small. Two current determinations
mere then made, the first with the left-hand strip in the cir-
cuit ani1 the second c L switch right; ” the tiibe’s orientation
was corrected; two more cletermiiiations were made, the first
‘’ switch right ” and the second “ switch left,” ancl then the
time, teniperature, and zero point were again observed and
rec*orcled. The mean of four such current determinations was
taken as the i of the set, corresponding to the mean tinie and
the mean temperature. The total time necessary to complete
a set was from four to eight minutes. The state of the sky
was also recorded.
The sources of error to which such work is subject are very
many. In the fir& place, a brisk breeze, i f it were from the
right direction ancl a bit gusty, was sometimes enough to
cause a throw of 2 or 3 centimeters in a male distance of
some 50 centimeters, and the resulting error in the determi-
nation of radiation is 5 or 10 per cent. It is almost always
possible, however, to take readings between times when the wind
is gusty; when it is steady, the effect upon the mean 1 should
be zero, so that this trouble is not particularly formidable if
one does not care for accuracy within say 2 or 3 per cent. A
further difficulty is caused, on all but the best days, by varia-
tions in the amount of heat absorbed by mists or clouds in
the path of the Run’s rays. I f the cloud layers are a t all
thick, the resulting fluctuations in the radiation received are
so considerable a i d so rapid that anything but the roughest
kind of an approximation is a t once impossible and meaning-
less. The presence of either of these difficulties is indicated
in the nccompanying tables by the words “ readings variable,”
and when a full set of four determinations could not be ob-
tained the resulting radiation number is marked with ( ? ).
INTERTICtATION OF INSTRUMENTAL ERRORS.
Besides these meteorological troubles there were also in-
strumental ones to be reckoned with. The most obvious of
these was a scale error in the ammeter, the pointer of which
quite evidently read some 0.008 amperes too low when the in-
strument was first received. It was accordingly connected in
series with a Thomson current balance, No. 131, a variable re-