Full text of "Memoirs of the Science Department, University of Tokio, Japan"

See other formats

Ree ye in he phe ee nae ht re
Lee een Pe Ont enge ar ma man anne m
2 で で だ TI

en hy nee ere mine meen ea
Serre,

phere cee geen
で だ rr お は な て ウン
DE な CIRCCRANOC た COOEC が 1

”
Geers Da where ee
ah に トト ゃ すか こい "の に ピー か

Pe に 5
ER a Ban het A le stom ey harsene ts wet be beeen
or ET ee も TI
hatha いで いこ いで と いき や か いけれ 人
ee ergs ne ner 00

に と しい
+

nee rte edelasraghris ay Uenrprae bgp dienen tin ich? た
oh で SE
CD rin 5 の いい
i a SEI pete el hehe mis babe Dees hae te Svouey pe hr anna pay aoe ewer meinte Gee area
も も いい セー で いとこ こい に いこ の た た ャ いな pt のか のど わな い マ セン いな に も 70 いた に し や たら Abtei こい
(お に に に も が ル ス ス し 42 CC に に に に に だ に 3 (こい とい に ee 1 に
here つぐ に YR て し ト 1: eee
で と ーー て See rs Kun PO ry
CC N ET En et ACGT を 7 KULT er かう tt

や コート い めい
ミー て まい TTA)
ee om he mm

rg REN ER te a ram hy Apres a rar いら arten ren tones

oreHtv1v ve Bogs 1 ーー ペー pharmacies

2 suber nett En BEL el お Rm
Te riet

ta Nenn he HEHE AN a ee ne
Ah tc or et de 1 の いこ こい"! ment
ee pean ano OKXKC で の さだ

に 2 は に て らら て ャ ーー ここ へ an

en
oe か し

nals ee eh mantener ae oh
to 人 MER < いい phe ran
r
m ann.

で ンー パャ ーー ムッ れい いさ
Mey Pune ee
PC だ だ
tt

eb mean nahe foe arte mar
ID De FI

“
ee er runde thes ”

ne bea
ve. white ity he ane we bs Mara os Phage 6
TH Meek ala テ
ee ara babe la ao be f Bains Picele tasers meee itt
Pe 入ら な サウ セ バ :0m で に 42440er3 な
これ アト マサ を さい れいこ いう け Re い けら に われ 和 せ じ や て ヤレ 19 い ef114202 人
すす 4 た で に ECO て Oo は CT
Denen alba han Hate Make mehrere I
eeyyrurrr が ROOTCUCKC21
inks oy ar prarg Benen ys 4
SC

For mich
oh ce

か も に いこ Silent m ie wire
rtBCEHI に お CI ・
jene haare hate Se

ー い デー
CC

いて
ECGEE や RDC だ に の な ここ 1

earners rs
ee era に
Mehmet es vue sabe mhe fate satbaese ~elshs bere dece Ape sabe pen
enor ernennen mr reader be les bush
De m ae nn ebb his ewan eA ion ir amie meprn mie ea

ED いっ

each hem be i \
i = た いい 6 rhaty wyert perma Beh On PI LL SL toda celoced Da Salt in ots DR wann EN pictrtaly) Spamnagr ity rede
f PD ee ee lable een artieet tiepearabrsmepiers en resists
KEOCCS だ ID

RL EA nes nn EE う

Ben ‚Tat kebehtan tt oooh tence here BE be latest tte akties ET

CRONPTCI2R の CEC
letehepene

と も か か トド
er e0 は は テイ ane タロ ン ューーー
se He Ra ew on a eee oot

nern me! Dt Pate 9 Mew
sm eae ms mons Mahn bee be mn Aue ore ep ve mt meet
SNe votre ios inee more nes RATT

は
ne are ee Se
abi Een

PhD

a shes eke Me mele it mann
が で で 1

HRC
TU a wee ee
19 さい いな か
ーー ドニ ドー キャ ww ドラ
Pichu es amet

EYE
ah an re ie

mm mann

nile mips Stehen eaepe nett aoe
leith ro am 1 Dr man a gaged se mann trees
も と 人 の た がい か や で で

Sper men
beng He eae

セツ 5 な

Tein’
ee
~ eer
erent
nt a ner
KC

CO70DCOST の TCP
lm < や ea に トド い に ド セー トト ドー トー ーー
Meter m ne en ann er

N ine

CC な
PT
で コー

Sere a
な で か 1

MC だ 2
pine

rna パロ に
De a ee

nur ehe Lei
rich a ee 1
} DR ee el RE
ite ne ee Haren eigenen sahen eh
ORPTTTT PPRTKK Riser) seh cen Mira
RF は DVC で た は

7 だ だ だ
range rennt he min mm sn

Lee
Pome pat Nee we

ako te Sele a ene aan

Py で で PO

porns rte srt te
COCO が NE

KDE
eh tase Sh ares
h singh FONE et ott rar peas EEE EN
ate psa eres ee eg aa ed a A BM
cH CRER ET or.
It ド すれ だ ア トト で の ies oe repre fare geet om Satie se
ee ar PCE1 は で まで で が ね 2 だ と]
CECPCESH が OCRET が CC に TNGPKECOSYE
せわ で で と の 0 よ 1
ROCK
* ーー ニーー・ a
CLES
eh ca ee metg toe
ern abo winds nn eo one
ihe ened ener vee
ed ann me nein.

ど

CD
COCCO
に ヤー コム さか か

CTO
や た た か が か

か
ee ehe eh
TEILE RT
a ee ee repealed re
a al eae m see ern nn ein he m en © thet
ten ee a eran! er ne Von
Seles Denen ed Ho Bm FR da Ont ane ate inate Ht anrran tent m Aree Pe te.
た れれ 電 ヤ WW mn の td dp に
セテ ン トリ ろだ ちや シバ ロン の 0 4 Anand ni vices arena 6 ay
イイ ルー ジッ て 作 で hr か の を も が の 06 の の 人 の 介 < 9 アデ ロマ ッ つと マ
な ott “a pele ep eae tay te ee em area.
6 fhe? N a EI) の とら
PC CGC に ICE の ウン
CO お に CROSS

PT Bint Ao eee
7 の は か セ ヤ ポ の BER Dee: Er が
oe に Pt ART,

Pe EEE

エバ アト EI LL etl tlle hoteles th tice Pe pal op othe
Sas A nen wen

アーレ" マー ハー アー リア
さか の から と どい いさ an
の と だ が れれ れい の か
ee re har mt E38
a me che

pepe oven

01 アドバ で で
stot = are MIM eT catemiepere phron pemare es ab berets
een na war niederen
も の ドロ > を さっ
me こと か の は めか の いか か ば か で か われ かれ の h が せ と アコ か か CTCET で 7 の の が et も の の 人 て
いい ゃ ー ャ わな わか か ね われる どか とか なか いて の れれ 人 作 結 ながれ の か の と うら が 1 いか か に
こし で v ん 5 ね の わり oi kb dyad shy gypsy We ee a ee en
Sore Ve Oe er ee ange i mg Amer ser
Wolds ha analy En か ip ematecesteph neers
elder eePryr lyri slelersseii bet tetal Semper m hen a ED een
2 ドル な し ACT て て て が に つう こう 7 が

ST atten pe eet ne ponies
と VT ee eh en penitent oe
RT antl apr een poet th eet er
ele ee が よっ アパ
Ind いせ Cf

CR だ CC と CC 0
ety ce banal ea

Ada A TER dl tacts
nee nme ern Tener
“pepe oo amis mae Rem nr en stm need

と ュ ュー ニー パャ ーー を で デリ py ee re ley
° ’ に 2 か で パイ 1 +

eee ne Ae GE
DOCPPECCACCCCOC
PF phi e ep
een ee
ed me

ELDER
"rernnnenr

fer
rer ide
hy.

Serer iy
the Rene ee
の ct に だ だし た ず コ

f . DOG. at : の tem
‘ ・ - REEL
BD erat gteete hale gen
reer ETT POLL LOLCat
EE Tee ened ree rae ee eae ign
REDET DOLL UN
ER EA ate o barkeit
ee MIT DS hee
Sake wee iae ine
er en teens mie shore Meme
NN En AN ee
MCNCOCDCOZI
um
NEN BERGER
ane wah
he
v +
PA Ae aa ne
Coke pele mer
POLE en be
ee ¢

MCOOO
ERBEN
Parone

か
の

a
ct
Ota

a
Ge eee ep
io eerie

ャ ャ ーー・ plete neces
Ya 0
oe webs he weap wren
EPCPCCCODRCOL2
ィ ー 1

vorn ip yon bei, bb ae,

*
.

N ti

© eS >
<《 Can, © OA [UELI EL €
« 人 < #4 (Oe A Gad
<《 KE EH & ーー の CM aC (OEE
« @ CRE K « ! 4ı , \ a«c_c い Ca [4
6 amade Ode € GE ORC Kida
T NPC 4 COQ (A € TER KAT | {CC CC @
im Cragg Gage a GMX CIC GCE |
C OO Ga, AREA OF OME Luc ath OC QE €
.. OC te Ge Cg _< oy - KO OBE F
« CMa dt La ALT 07 MC CEE ©
< qr BR QC % GE 《 AOE CM CU ©
a c «fe < ada (6 Ga ( て € < € a
«U TU WEE ANGE CXQE C AEC «
a Wet ir ( と て る ER ( GtEaR «4 ce
€ CON Wi © RT ECE (< CAE ' (
@ 4S er 5 eG K《 <Z< ( CAR er «
@ Oy Ce る の 70 € a
@ cc < - WW € € € 4
| ur ii @ 、 GAG! vA
re. Ci (EX , CI 4
Pt ( a ° > 2“
7 Ce <7 C Ga’ ‘
< で Cr am <
At ( ie 人 '
TE u URL 1 OC @ > SS
Ca oe “x (La Cl EX TE
Cm « £4 TEE OH < で e(@ ( € Co
OM OO TE EAU CH OM
CU ae Ca aC CEE
Hie Qe CC Ga ぐ <7 CCK
Ne a < as 人 す - vd rs CC て — cs
€ CC
人 GQ El CQ (Qa a
5G < 《 me. G CK aC CGS |
(< Ge ai (( Ce: « も ((。 < «
2: “CC a . — 4
/ CK | ( = の
Or を <<a Ta “ag Wire
(人 «€
4 CMON €
7 Gr wm
\ k( し る _ < く @&
(on CAME &
C@. CLEC
¢ ¢ Ce CCe
(K (( 人 <
GC fc cc (6
と ae rar d
で る 6 6 Ue 4
- @ Ca @
tos oe る
@ < aia <C(
< ac
7 々 て Cc a
7 で ka @
4 CO _€ a &@
aC .. (CE 作

A Ei ( OTE AE OC CS ELC COE
Au ce トー ar LT CCT KH Ca CC OL ECS Kae
Ce く ぐ Cue [UT CMC COE Ka «OG EEE QO
(a Ge ぐ ぐ egasgk (CC €: SG Ce Ge OL CE < 2
「 Gama ag 人 6 6 ぐ く CEG (店 る ae ces ?
Gate’ て aqua AR CC ER (( 電 て COE COE
Ci 《< rw ag (Ua «is « Cod C COC CG
(LOD Ca X mcg, Cr [TO a EE
ACE Ce CO! Mi の 4q Ca ER COC CE OTC EC
OC て Cay dA CME CE CCC CO GUC EEE
(Nr eg gem IT TAU —— Sb aia &| NT CO EK
(@: a ra ) + qe @ «(t Lil < € IS < ce
COMING Cee gan KC CCUM OC Gare Cam cece eqcq
CPO Quel ae Aa (€ 6% «ce CCC GS
GG EL a « (q ( GES @ CGR ay die
[X En iq ac. a © de < OK 電 Owen a? ¢ Cnc ( AL Cen” & (CE .
ee ae gue «4 1a aa a A ⑯ 所 《q < GAG
CE EC EC g4 て IE CC € €4Gl (Ee
Er «Ke ca atti € (Oma ECT ace a U CUO a © ぐ <6
も ea ea, a € CR KU RACE ずる
‘CNG 。 © cad ali ( td cat @ ¢ ct (Cc va ( W (OGG
LEE ae COE ai LEE TE ぐ の <
を < 6.4 (RT (al (« Cay SOT K UT AE AE (6 《
CET REEL IK (id? KAT (MCE
TEE CC ac dd COME, COC CCC 48 COC AG
CC f(( acme Gl (fo: @euw (CC TUR UCT 7Wee & a
CE «UI or. て ぐ ぐ ぐ て COE CC CGC <e<
(Ce Ku COT GR OKC CONT C40 ーー cg Cet
く 7 OO Oi Oa て <@( て UTAH ui A ar CG
QC OEE CU 1 CEO OE MO, OO @ と
CC CEM ak CHO CC & or « QO CC
ca て る < LER aX GT AR Ai EC EUER
Ua で る WEG KT Yar (ck gam we ず (all
KK LAUFE Qa qa as (€ ur る € し 6
Ge ems a LPM ( OVER @ fe a
wee る &@ LOCKT (C ( OWE « re 4 = wae @
of a au: KALTE“ 4 u CC 電話 ( で (6
| ET -《 KIT (て C (a Bar (で 【6
Cie © Gin ¢ <(CCCC く て Cane CCM RE COE
< OF i COE (で < Far MK x Me S cd
ne ‘9 Gir i Car CL CIE K Santa € CC
it CC Gay LOL Ka CU wa COR Een で
MOR < Ad ur KU OTM CCE ma (Ca aM RU <
CGE CE (UT (GrGCC CE RE TE GT AC (4
I I て @ quan a < で reKC Ce TEA GT AC 44
(Qe eg war Ex < で Ce 4G CC GE (で KU
SC Ce reg €G 〈 で WW《( CE RE. OF ET. CU
f € ı Ca dm yf s We qh ac ag ACC ((é «
LEE <i@ eax C CE AK «a ar Es aa alae C4
KOE て <em Cae CE € CO AC Gar ac ぐ ((
(MT Cam Kae ( CORK Cit GE ES RR 7 GC
(Sa Se 4 Ce "(SG mn CC RER au Ta
CO |]. ane « ET i ree SS Ge € GK “at age
UT Cie ee KE (CEE El CU KC
I. (Cu «cc va る i AT TEE GA (ie a EI
で 人 を て (df 7 ‘ae © ET AM AT (iC a Cx Cie ( 《 AG
<Y (町 eg gud ara (a a AU CCGE | « CG
Ca @& « ン at て CAE 。 @C€@act Gia EG

MEMOIRS

OF THE

SCIENCE DEPARTMENT,
TORIO DATSAKU.

(University of Tokio.)

REPORT

ON THE

METEOROLOGY OF TOKIO.

FOR THE
Yuar 2540 (1880)
BY

T. C. MENDENHALL, Pu. D.

PROFESSOR OF EXPERIMENTAL Puysics IN TOkrO Daicaxu,

PUBLISHED BY TOKIO DAIGAKU.
TOKIO:
2541 (1881.)

Tokio. Tapan/
MEMOIRS

SCIENCE DEPARTMENT,
TOKIO DAIGAKU.

(University of Tokio.)

JHNNKSs 7 に

REPORT

ON THE

METEOROLOGY OF TOKIO.

FOR THE
Year 2540 (1880)
BY

T. C. MENDENHALL, Pu. D.

Proressor OF EXPERIMENTAL Physics IN Tokio DaiGaku,

PUBLISHED BY TOKIO DAIGAKU.
TOKIO:
2541 (1881.)

SEP 13 1993
LIBRARIES

YABLE OF ‘CONTENTS.

4 NORTH DIST SUIT EEE Re Sra Its RIE aS TIC Im cra cP roma)
SETI en NR uta heel
hE LVEHOON On OCTOBER LO ANDAs =. nenne dessa een ts 23

LEANED MET VTNDUTY Upper yar as aes ee er 39

—=t-—

ELSI inva OPSER WAMIONS Minor N ala AT |
HVCRNHEAELC 電 43 お REED
EZKGXEHUHNCDRWBIIZNUHN(O2AUMOA 寺 お 58

METEOROLOGICAL OBSERVATIONS ON FUJTINOYAMA.… せ ……………….…………………… 68

BIE EDEL EM ONIONS ORM LO RL OM srt een canis pice Rena ae era fal

REPORT ON THE METEOROLOGY OF TOKIO.
For tue YEAR 2540 (1880)

T. C. MENDENHALL.

PRELIMINARY.

Tn the report for the year 2539 (1879) the first of this series, the location
of the observatory, in which the observations have been made, is given and the
various instruments used are enumerated and described. It is, therefore, only
necessary to state that no change has taken place in regard to either of these.
As stated in that report, while the location of the observatory was satisfactory,
the instrumental equipment was by no means so and it was at that time confi-
dently expected that before the end of the year several important and much
needed additions would be made. Unforseen difficulties have prevented this,
however, and the observations of the year have been made with the instruments
previously in use. Several very desirable additions and improvements were
recommended in the conclusion of the report of last year and it is to be regretted
that, up to the present time, it has not been found possible to make them. It
is now proposed, however, to equip the observatory, during the next year, with
an entirely new and complete set of apparatus, such as is in use in the first class
stations of the Signal Service of the United States. Such an addition to the
instruments already in use would add greatly to the value of the observations
and it is sincerely hoped that nothing may interfere with its accomplishment.

During the year several changes have been made in the observers employed.
When the hourly observations began in March, Mr Nobutani was added to the
force and he has remained permauently attached to the observatory, taking the
place of Mr Wuyeno. Mr Namba continued until August when, unfortunately
for the interests of the observatory, he was ordered to France by the Education-
al Department, for the purpose of continuing his studies in Paris. Mr Namba
was a careful and trustworthy observer and it is to be regretted that the obser-
vatory was so soon deprived of his services In his place Mr Kiriyama was
appointed regularly and Mr Tsubota has also been employed during the months
in which hourly observations have been maintained. Although, considering the
interests of all parties concerned, the changes thus far made haye been, on the

whole, desirable, it must be admitted that frequent changes, implying generally
the introduction of inexperienced persons, cannot fail to be detrimental to the
accuracy and value of the work and it is hoped they may be rare in the future.

The hours of observation for the year 1880 were the same as during the
year 1879. With the beginning of the year 1881, however, the hours were
changed and the number of daily observations increased to six. They are at present
as follows; 6 a.m. 10 a.m. 2 p.m. 6 p.m. 9-27 p.m. and 11 p.m. The change
from 10 p. m. to 9-27 p. m. was rendered necessary in order to make that obser-
vation agree with the “international hour” which has been changed in accordance
with the desire of the chief of the U. S. Signal Service.

In the preparation of the report for this year the aim has been to arrange
it that comparison between it and that for the previous year would be easy.
The tables are, therefore, in general, arranged in the same manner and corres-
ponding tables are indicated by the same letter of the alphabet. The charts are
also given numbers corresponding to those of last year and comparison of these
graphical representations of the meteorological phenomena for the two years
will be easily made. Finally a series of charts is given, showing the means of a
portion of the observations for the two years and a comparison of these with
those for the present year is interesting as indicating a considerable degree of
constancy on the part of several of the meteorological elements, that is, so far
as can be established from the results of two years’ observations. Hourly obser-
vations were maintained during the months of March, June, September and
December. These four months, it is believed, afford a good representation of
the varying meteorological conditions of the year, but it is intended to continue
the hourly observations during several months of the year 1881.

During the first week in Augustran expedition to the summit of Fujinoyama
was made by the special students in the Department of Physics, accompanied
by Professor W. 8. Chaplin of the Department of Civil Engineering, and the
writer. The special object of the expedition was the determination of the force
of gravity on the summit of the mountain. At the same time a series of mete-
orological observations was made during the stay of the party upon the mountain.
The value of these has been greatly enhanced by the addition of a series of
barometrical observations made at the same time by Messrs Nakamura and
Wada of the Surveying Department who carried out a series of simultaneous
observations on the summit of the mountain and at the sea level near the base.
I am greatly indebted to Messrs Nakamura and Wada for their kindness in fur-
nishing these observations to be made use of in this report in any desirable
manner. Taken all together, these meteorological observations are, of course,
of considerable value as affording a means of computing the height of the moun-
tain. To that end Prof. Chaplin has kindly consented to discuss them in con-
nection with others, bearing upon the same point, which he has been able to
secure and the paper in which he has contributed the result of this discussion
will be found to be one of the most interesting and valuable portions of this report.

In the report of last year reference was made to the proposed undertaking
of a series of measurements of earth temperature, by means of the thermo-electric
method. The labor of determining, experimentally, the best method for accom-
plishing this end was undertaken by Mr Hiraoka, Lecturer in the Department of
Physics, and he has worked with great industry to overcome the difficulties
which seemed to be in the way. Unfortunately complete success in that direc-
tion cannot yet be recorded. The chief and, apparently, only difficulty seems
to be to secure sufficiently perfect insulation of the wires leading to the thermo-
electric couples. 1t will be remembered that the ordinary means of insulation
are not procurable in this country, or, if at all not without great difficulty.
Various devices have been resorted to and in some instances success seemed
assured, very satisfactory tests being obtained; but after a few days the insula-
tion has broken down and temperature observations, therefore, rendered impossible.
Success is by no means dispaired of, however, and efforts to secure it will still
continue.

Reference was also made in the report of last year to a series of experiments
about to be undertaken for the determination of the velocity of sound. It was
proposed to take advaritage of the signal gun, fired at 12 m. each day from
Tenshudai and of the telegraph line connecting the observatory with the
University. By the aid of the latter the time of the arrival of the sound at
each of these points is recorded upon a chronograph in the physical laboratory
at the University. This plan has been successfully carried out and observations
to the number of more than two hundred have been already secured, giving the
velocity of the sound wave under widely varying meteorological conditions. In
addition to these, made, one each day, by means of the noon signal gun, a large
number have recently been made, by means of successive firings upon the same
day and, as nearly as possible, under the same meteorological conditions. These
will greatly increase the value of the regular series and in the end, a result of
considerable accuracy ought to be obtained from the reduction of thes observations.
It has been thought best, however, to publish these results in a separate memoir,
hence they do not appear in this report, as might be expected from what was
said concerning them in the last.

A suggestion was also made concerning the use of some form of seismometer
or seismograph in the meteorological observatory. Early in the year 1880 interest
in seismology, quickened, doubtless, by the unusually violent earthquake of
February 22, became so great that a society was formed in ‘Tokio known as the
“Seismological Society of Japan.” ‘This society is in good working condition.
It has already published volumes of its;proceedings and}it is looking well after
the interests of seismology in this country. It is desirable, therefore, that all
observers co-operate with the society and it would be extremely useful to have
one or more seismometers or seismographs, of a form which shall be approved by
this society, placed in the meteorological observatory. For instruments which
are not self-recording, or for those requiring frequent observation or attention,

this location would be especially favorable, owing to the continued presence of
one or more observers.

The frequent occurrence of extensive and disastrous conflagrations in the
City of Tokio has attracted the attention of all who have resided for any length
of time in the capital. The immense destruction of property in this way within
a year or two has excited great interest in all investigations pertaining to the
origin of fires and possible means of preventing their spread. It has long been
known that disastrous fires were most frequent, in general, in certain months of
the year and it is easy to see that their occurrence is intimately related to the
velocity and direction of the wind. Mr. Yamagawa, Adjunct Professor of Phy-
sics in the University, has recently devoted much time to an investigation of the
origin and course of these fires, and to their classification in reference to
atmospheric movement. The results of this investigation cannot fail to possess
great interest and I am indebted to Professor Yamagawa for the preparation of
some of the most important of these results, which will be found, with charts
illustrating the conclusions reached, at the end of this volume.

BAROMETRICAL OBSERVATIONS.

Table A exhibits the readings of the barometer at all of the regular obser-
vations during the year, after having been reduced to the freezing point and to
the level of the sea. The actual fluctuations in the barometer during the year
are graphically exhibited on the six pages of Chart No I. A resumé of the
observations contained in table A will be found in table B. In this will be
found the means for each hour of observation for each month in the year, as
well as general means, maxima and minima and range. Concerning the first it
will be seen that the mean for the 2 p. m. readings is tlıe least, and that, with
few exceptions, the mean of the 7 a. m. is greater than those of 10 p. m. In the
means for the year, that for the reading at 7 a. m. is .005 inches higher than
that for the readings at 10 p.m. But a glance at either the table or chart,
given under the head of “hourly observations,” representing the means for each
hour of the day, during the four months in which hourly observations were made,
will show, of course, that a maximum does not occur at 7 a. m. but about three
hours later, and that 2 p.m. is not, in general, a minimum point but that
10 p. m. represents very closely a second maximum. It is believed, therefore,
that the hours of observation established for the year 1881 will furnish a much
more truthful record of the barometrical changes.

A brief comparison of the principal facts found in table B and those from
the corresponding table in the report for 1879 will be of interest.

1880 1879
Mean height of Barometer for the year.. 29.958 inches 29.952 inches
I En aD es acre Raa ERE BERR ENTER SOSSE) wur BO
EGO 2ER ae AOS wes:
est Se creas ere Se Me SEIN FERNAB ME dos 1 1.426 ,,
Maximum Monthly Range.… せ ……………………… a er als,
Minimum ,, 内 で OF ot, BEN) ge

The maximum for this year occurred at 7a. m. on February 26, while
that for last year occurred in April. The minimum for this year was recorded
at 7 a. m. on March 20 while that of last year was in February.

In regard to the maxima and minima given above, it must be remembered
that they are derived from the series of regular observations. On occasions of
extraordinary and rapid fluctuations in the barometer this series will often fail
to contain the extreme points reached. Without doubt, the lowest point reached
by the barometer during the year was touched at some time during the Typhoon
of October 4th. At 2 O’clock a. m. of that day a height of 28.735 inches was
observed. If this be compared with the maximum given above it greatly increases
the range for the year, making it 1.804 inches instead of 1.361 as obtained from
the regular series. Further reference to the barometrical fluctuations on that
occasion will be found in the brief account of the typhoon given under “the
wind.”

The maximum monthly range for the year occurred in March while the
greatest for last year was in February. The minimum fluctuation both for this
year and for last was in July.

In Chart No 2 are exhibited, in the first diagram the fluctuation in the
monthly barometric mean, and in the second the maximum, minimum and range
for each month in the year. The curve in the first diagram shows three well
marked maximum points, these being for the months of February, April and
October and this will be found to agree with the results of last years’ observations
upon an examination of the corresponding diagram in the report for 1879. In
the latter there is also a slight indication of a fourth maximum in the month of
July, but this is not repeated in the curve for this year, the mean for July being
lower than that for any other month.

TABLE A. SHOWING READINGS OF THE BAROMETER
THROUGHOUT THE YEAR.

7 2

29.742) 29.801
29.993 29.939
29.698 29.563
29.704! 29.643
29.822] 29.795
29.859) 29.767
29.969 29.969
‚30.091 29.982
29.996 29,847
| 29.763 29.784
29.940 29,860
| 29.976 29.968
| 80.056 30.014
80.179 30.079
(30.058 29.835
30.109, 30.073
80.140 30.057
30.250 30.202
30.342 30.287
30.410 30.352
30.370 30.192
29.865 29.917
30.244 30.140
30.364) 30.323
"30.346 30.089
29.739 29.570
30.032) 30.004
30.146| 30.147
30.296 30.143
30.314) 30.285

January

29.868
29,926
29.652
29.690
29.886
29.903
30.105
30.023
29.744
29.886)
29.939
30.041
30.160
30.025,
30.057,
30.151
30.162
30.325
30.331)
30.430
30.023
30.142
30.251
30.380
29.995
29.826
30.046)
30.218
30.174
30.389

| 80.442 30.387

30.399

February

7 2 10

30.368] 30.239] 30.339
80.190| 80.015 29.871)
29.715) 29.715| 29.943
30.016) 30.034) 30.144)
30.294) 30.335) 30.466
30.489} 30.372) 30.316
30.249 30.178 30.136,
30.030 29.977 30.072
30.177) 30.169 30.133
29.905, 29.864 30.049,
30.150, 30.081 30.100,
29.986 29.840) 29.853
29.922) 29.881 30.097
30.229 30.209) 30.296
30.279 30.083) 30.002)
29.760 29.788 30.040
30.130 30.057) 30.176
BR 30.210) 30.320
30.383) 30.341) 30.213,
29.881) 29.821) 29.908
29,904, 29.800) 29.752)
29.914) 29.905) 30.042
30.111 30.099) 30.269
30.312] 30.220) 30.284
30.481 30.483] 30.535
30.539 30.444 30.474
30.384 30.198} 30.014
29.649 29.751] 29.985
30.196, 30.281) 30.355

22 |

30.365
30.241
30.177
30.258
30.337
30.454
30.329
30.236
29.979
30.180

30.103

29.991
29.974
29.741

29.988)

30.044
29.854
30.036

30.027

29.178
30.113
30.357
30.025

29.846 29.927] 30.066

30.091
50.038
29.934
29.654
30.079
29.735
29.976

March

2 10

30.262) 30.266
30.175| 30.207
30.146} 30.243

30.532) 30.416
80.337) 30.342
30.235] 30.291
29.969] 29.876

| 30.049) 30.036,
30.043, 30.064
29.900) 29.965)

| 29.655] 29.886,
29.939 29.983
29.967| 29.839
29.902) 30.031
29.994) 30.044
29.888) 29.574!
29.355] 29.891
30.173} 30.302
30.258] 30.234
29.800) 29.747

29.941] 29.988
30.005) 29.988
29,809) 29.767
29.715) 29.983
29.927] 29.643,
29.881) 30.033

30.216) 30.312)

29.997) 30.115)

29.857] 29.810,

April

7 2 10

29.514) 29.604) 29.843
30,015) 30.004) 30.068
| 30.063} 29.953) 29.946
29.901] 29.783} 29.916
‚ 30.000) 30.066) 30.238)
30.303) 30.280) 30.319
30.316) 30.247) 30.28

30.330) 30.302) 30.35

30.894) 30.364) 30.353
30.234) 30.027) 29.95

29.779] 29.765) 29.922)
30.056, 30.064) 30.134
30.227) 30.281) 30.332
30.417) 30.390) 30.427
30.446) 30.342) 30.238)
a 29.661) 29.649
29.811 29.869] 30.012
30.097} 30.109) 30.234
30.336) 30.309) 30.331

30.345) 30.283) 30.294
30.332 30.232) 30.24

30.194) 30.066 29.923
29.689 29.510) 29.45

29.486) 29.430) 29.52

29.572) 29.637| 29.922
30.099) 30.105) 30.11%
30.120) 30.011) 29.941
29.825) 29.857) 29.974
30.101) 30.085) 30.100
30.032) 29.868) 29.599)

29.072] 29.415] arena | 記 語 | 555。

TABLE A. SHOWING READINGS OF THE BAROMETER
THROUGHOUT THE YEAR.

29.698
30.100
30.214
30.054
29.993
29.925
29.449
129.985
30.038
29.835
29.709
29.993
29.823
29.789
29.906
29.958
29.802
30.033
30.208
30.139
30.044
29.723
29.769
29.863
29.863
29.867
30.093
| 30.090
129.905 29.940
29.927) 29.846
29.801 29.748

29.671
30.180
30.094
29.999
29.943
29.780
29.552
29.995
29.941
29.791
29.779
29.879
29.742
29.736
29.883
29.936
29.785
30.033
30.196
30.055
29.949
29.710
29.720
29.813
29.783
29.868
30.062
29.971

|
29.940
29.665.
29.836
30.009)
29.925
29.642
29,925,
29,877
29.780
29.839
29.950
29.897
29,911
30.139
30.189)
30.115
29.910
29.765
29.786
29.818
29.806,
29.983
30.104
29.988
29.928
29.853

129.608
129.655
29.905

29.956
29.970
30.009
29.532
29.521
29.703)
29.816
29.907
29.923
29.860
29.959
29.991
29.842
29.603
29.677
29.691
29.765
29.935
30.179
30.279
30.167
29.917
29.776
29.892
29.836
29.782
29.745

June.

29.620
29.688
29.896)
29.909)
29.940
29.952
29.291
29.505
29.681
29.837
29.900
29.882
29.828)
29.927
29.918
29.739
29.607
29.646
29.684
29.760
29.999
30.204
30.224

30.098

29.813 :

29.758
29.852
29.773

29.707)

29.683

|
26.647) .

29.653
29.837
29.949
29.963
29.995

29.927

29.342]:
29.645) 29.

29.750
29.867

29.918

29.865
29.901
29.981

29.928

29.627|
29,678)
29.675)

29.736
29.827
30.108

29.786
29.731

29.712

229.642
529.533
229.666

| duly.

| 7
29.697
29.697
29.901
29.963
29.821
29.798
29.658
29.674
29.676
29.743
29.759
29.731
29.777
29.928
30.000

| 29.656
29.619)
129.892
"30.004
129.894
(29.858

29.716
| 29.764
129.797
129.764
29.793
29.920
30.049
30.0231 29.969
29.952) 29.900
29.896] 29.829
29.821 29.769
|29.761| 29.726
| 29.731] 29.583
29.561
29.495
29.639
29.666
29.621
129.704
29.653
29.725

129.702
29.658
199.715
(29.767
| 29.701

129.888

29.893

29.886

129.932
|

10 |
se
29.694

29.830
29.951
129.957
| 29,842,
129.801
29.705
29.717
29.732)
29.817
29.784
29.750
29.872)
30.005
30.031
29.998
29.900
29.839
29.779
29.759
29.707
29.554
29.601)
29.686)
29.687
29.663
29.782
29.682
29.886)
19.905

I
29.384 |

29.868
29.836
29.849
29.925
29.962
29.962
29.994
29.958
29.934
29.858
29.683
29.530
29.570

29.757] 2

29.852
29.934
29,982
29.995
29.957
29.945
29.895
29.821
29.761
29.730
29.786
29.568
29.749
29.689
29.714
30.024
50.000)

29.813
29.781
29.814
| 29.911) 2
29.919
29.924
29.966
29,957) ¢
29.8721:
29.762
29.608 :
29.465
29.595} ¢

29.849) 29.922
29.929) 29.984
29.938] 29.985
29.940) 29.956
29.930) 29.955
29.885) 29.923)
29.818) 29.847
29.761) 29.744
29.711) 29.72
29.688) 29.734
29.699) 2
29.645) 2
29.724 ¢
29.675} 2¢
29.923
30.034

29.932

29.759
30.004
20.927

TABLE A. SHOWING READINGS OF THE BAROMETER
THROUGHOUT THE YEAR.

September

29.954) 29.942) 30.005
30.033) 30.009) 30.066

30.052] 29.975) 29.995
| 29.977] 29.977) 29.988
29.962) 29.938) 29.937
29.952) 29.888) 29.934)

| 30.008 29.970) 30.027
30.053) 29.977] 29.972
29.886) 29.772] 29.807
! 29.840) 29.824] 29.874

30.118} 30.069) 30.112)
, 30.081} 30.005} 30.004
29.951) 29.856) 29.856)
29.701) 39.690) 29.869)
30.014) 30.013) 30.080
‚30.0731 30.018) 30.077)
30.060) 29.993) 30.034
30.001) 29.910) 29.924

29.815] 29.783) 29.786
29.646) 29.819
30.051) 30.112] 30.22]
30.238) 30.154| 30.134
30.031) 29.905} 29.890
30.058} 30.095) 30.161
30.131} 29.980) 29.964
29.971) 29.955) 30.043
30.145) 30.121 30.214

29.886] 29.803} 29.801) 30.102

30.183
29.973

| 30.092) 30.083} 30.067) 29.793} 2¢

29.464

30.297
30.378
30.237
30.029
30.013
30.278
29.990] 29.997 30.097] 30.165

50.188
30.293
30.209
30.098
30.062
30.022
29.966
30.094

30.176
29.949
29.875
30.009
29.812
29.872
30.148
30.105
29.925
29.816

October

29.972] ¢

2 10

30.077 30.073
29.907| 29.943
2) 29.356
;s| 29.958
30.141)
30.378
30.303
30.152] 30.094
29.952] 29.987
30.017| 30.180
30.221 30.277
30.015] 30.035
30.249] 30.303
30.225 30.259,
30.121) 30.112
30.023] 30.059
30.014) 30.058
29.916 29.928
29,956) 30.073
30.053) 30.098
30.101} 30.208
30.053) 30.041)
29.865 29.896
29.846 29.975
29.942) 29,955
29.590 29.706
29.955 30.097
30.104] 30.128
30.016) 30.023
29.748 29.677
29.839 29.993

November

30.053
29.960
29.942
30.068
30.060
30.133
29.975
29.788
29.811
29.940
30.260
30.340
30.174
29.921
30.103
30.258
30.067
30.003
30.111
29.961
29.849
30.015
29.918
29.906
29.988
29.746
29.540
29.821
30.110
29.957

29.979
29.860
29.908
30.000
30.003
30.134
29.917
29.683
29.716
29.951
30.248
30.244
30.108
29.885
30.080
30.110
29.979)
29.891
30.054
29.820
29.799)
29.993
29.771
29.845
29.911
29.453
29.471
29.882
30.049
29.864

30.011
29.917!
30.028
30.041
30.113
30.111
29.854
29.844
29.897
30.158)
30.286
30.206
30.037
30.029
30.164)
30.115
30.028
29.964
30.079
29.852)
29.889
30.043
29.831
29.923
29.899
29.558
29.684
30.016
30.029

29.969

December

30.007
29.950
30.116
29.991
29.943
29.961
30.102
29.905
29.536
29.482
29.658)
30.023
30.228

30.093

29.803) :

29.772|:

30.105

30.0891:
29.849:

29.894
29.919
29.895
29.875
29.831
30.158
30.099
30.113
29.739
30.049
30.283
30.440

29.966) 29.966
29.924] 30.085)
30.046) 30.066
29.867) 29.932
29.856] 29.901
29.968] 30.108
30.012] 30.014
29.719 §
29.371} 2
29.476
29.661
30.049] 30.
30.159| 8
29.932] 29.817
89| 39.789
930.054
i] 30.107
29.965
29.803
9129.97
29.788] 29.957
29.789] 29.924)
29.736] 2

29.880):
30.088
30.088
29.913
29.798
30.030
30.289} 30.448
30.342) 30.293

TABLE B. SHOWING MONTHLY MEANS, MAXIMA, MINIMA AND
RANGE OF THE BAROMETER FOR THE YEAR.

| ‘if
|| | | |
Means | General | | |
Month. of | N | Max. | Min. | Rang
Fach Reading. Ne | |
7 a.m. | 2 p.m. 10 p.m. | | | |
| | _
| N
January | 30.073 | 30.000 30.069 | 30.017 | 30.442 | 29.563 879
| |

Fehrifäry 30.135 | 30.081 | 30.144 | 30.120 | 30.539 | 2045 | .80

March 30.043 | 29.978 | 30.012 | 30.011 | 30.454 29.178 | 1.276

April | 30.067 | 30.017 | 30.055 | 30.046 | 30.446 | 29.430 | 1.016

May 29.925 29.882 29.917 | 29.908 30.215 | 29.449 | „766
| N | |
1 | |
| | |
| | | =
June 29.817 | 29.811 2.845 | 29.83) 80.279 | 29.291 988
|
| | |
| | |
July 29.792 29.756 | 29.797 29.782 | 80.049 | 29.495 | dad

Argust 20.841 29.807 | 29.829 29.826 | 80.08! | 29.401 | 633
| I
September 20.991 29.947 | 29.992 20.978 20.238 29,616 || 502
4
| |
October 30.018 | 29.996 | 30.012 30.029 30.378 2.356 | 1.022
| I |
|
ー| 一 一 - 一 92.21 (ee
November 20.902 20.990 | 20.986 20.006 30.310 29.453 | 887
December 20.06 | 29.903 20.074 20.947 30.418 29.371 | 1.077
|
% 9,177 9.09% | 90.1479 on O72 aon am 5) =o (| の
ear 29.977 20.025 29.913 20.958 30,559 29,178 1.361

I
」
ーーーーーーーーーーーーーーーーーーーーーー- 一 一 ーーーーーーーー 一 - 一 ーーーー- 一 ーーーーーーーー- 一 ーーーーーーーーーーーーーーーーーー

A)
グ

— |
Co \

“ . x
Im © )
ご

x (
in

>
5)
m

ear

during the Y
N TREE

>
te 2
ORF
|
|
sl

‘Chart N°1

Showin o Fluctuations tn che Barometer

| seep

7 04 a 一 #8

Chart N°; yg eee 2 の as
ma 三 年 hk be

ーー FC
cy 2 > # - で 7 ” = BA E7 し っ
Showing Fluctuations tn the Barometer iF re ww =
。 . - A > RS = | ー =

during the Year Z5tO7§ SY) re Wy —+ I 1

< ノ
|
| 1.27 > > + 5 ö 7 5 9 70 アカ 22 i3 t 15 | 16

Merch. xs |ı

April.» &

hoy

aoe —— FE —— wor =
ご | / | = 17
1 ア
3 y; 8
n / | ~ 6
| / | {
N 3
| \ や ) ~ 11
| / { I |
) |/ こ
| I (
I} 5 1 テー
| 90 } aa / N
i( ( | | |
\ we
o = | N \ に
va ™N \ \ ” \
N ) >) N
{ |
x x by x N い |
‘ | ( 1
en ik | > j I
“ EN N ) | a ー
| > ge mei {
アア CN ) © |/ \ 」
| er: ( 1 1
‘ > ” 【 ~ \ |
| } )
| f ( 本 |
t ) や N ~ \ | = |
} ) ) = |_| |
{| | \ | |
| Q a | x 1
グ tee ) |
( \ Te \ |
wl | 1 sl | x RS x |
1
\ ト | ae
~ an | i FN ws | \
ee ee ee | ite aS | EN — ニーー エ ーー )- ー | mo | eee HORE ) ined ーー 4 \

Chart N®1

. 2 ma) ; 2
Showing Fluctuations tn the Barometer

6 ff
durino the Year 2540/18 80)
ノ

a Bra

Ye =
針 af
+ +

pe 4 月

EEE
Chart N1 * PR ei

Y ンー ’ . 7 Ze Peas
Showing Fluctuahons tn the Barometer # 中
during the Year 2540/1589) FR
2

ag a |

Sep?. = き

ーー を ーー

7 の 7 了 2 18 73 15 7 の

~
m
uo
2°

ur
a
N
&

Oct.

10 mn Sn on 4 Se A EEE eS 2 32 RER SS ees Ee ae

|
| _-
Ba し a ARTE he Ee es ee Ean eee ame |

『 0. wi — の 7
Chart N 7 ー re er 5
od = f- fn
m 4 ーー - ms
cy : 7 = 7 アリ 7 ot 15 m ウ
SROWLAO F 7 uations tn dhe baromeler Aa d な vg
7 >

/ m pe =
ina the Year 251075 80) IF 1 J 1

oh ES Ta EM Rs OS CES BS ee CA ee ai

i rad 2 ) の is 5 シン 10 /I 12 13 / テ /5 7 の |
| |
or.
| トー 4
| N
| N ~ se \ ーJ
m we a> ンー キミ aes ES 0
| N ee = ナー SE
IE SN N
or
| |
FE |
iy i |
| sg J EU re aa
ay
SQ I に -
= | 17 1S 19 20 27 23 23 24 25 26 27 28 の 3
~ ol
に \
„S 4
— \ „|
> |
i i ーー と) ie
i hod 】 tin 4
| er に
al ee \ Fae, | 2 >
| = \
I | \
ひ ‘ /
ま cr |
rh
I
ee ここ ドー ニテ ーー トー SSS ES Se SETS a ail
| |
た っ に =
2 > 3 t 5 の 7 5 の 10) I 12 14 5 77 |
I +} E73
| |
cS ar zei =
€. N
| ーー ーー By ザイ ae GE: x 7
ーーー 2 5 x \ /
5 = \ > 7
| — |
ir |
El ;
ry | 『 — ー ー ーーーーーーーー デ ーーーーーーーーーーーーーーーーー — me = + ge ver |
ま
P |
Q Kira | 7 の 1 る の | 27 122.122 | そま | 25 | 26 | 27 | 28 7 30 | 31
bw | |
ここ | ;
eo |
~
S| | :,
‘nn | /
| ューーーーーーーーーーーーーー 5 SSS SSS BEST fae © ン -
al 4 き
1
rl
|
|

nu > Zn

Chorrt N°2

; : fe. = Eo
Shou tng Ok は a, i a
6 ; 1% 第 s+ = F G
?" Mean height of the Barometer Fag oe 9 a ee i eg a
- り っ r +r. = : TP に
for cach month, & \ # 7 = |
ond Maxi の だ = = 同 度 月 年 % |
2" Maxim um Minimum and Kerr, wer |
Range for each month, re ae & a ie =
に 、 : L w] = AR ip
during the Year 2540//880) FR £ 月 = a
ー 3 |= EEE Ae eee RT PES ST Bal Moe Pa Is +
] 3 S N ミ 、 Sra, oS = => ale
ASASASASRZIHRSHS Aa sla Sly Sia &
r
ar seni nes
wun ーー
7
; a ed ee ee
= = = Br, I - ヒ ™ Riis alla el
ミ 3 一 まで > ミ > な っ 一. 、S d
ASSAS ASIANS AS AS ASA S ASA SASS
%
の
r
3Dın — ー = se = ーー ad
N
6
X A
*.2

29in

TEMPERATURE.

Table C contains all the regular readings of the standard thermometer
during the year, and from this table Chart No 3 is constructed, showing the
fluctuations in temperature throughout the year. In table D will be found the
maximum and minimum temperatures for every day in the year, the mean of the
readings of the standard thermometer and also the mean of the maximum and
minimum temperatures for each day.

The following is a comparison of extreme temperatures in the year 1880
with those of the previous year.

A minimum of 32° or under was observed as follows ;

1880 1879
In January on............29 days.…………27 days
SAPUGUNUANY Aussen 5 TaseccsvexcalO’, 55
green ee A
» November... iL en. N
LE) EE つこ Daher
Total for the year 72 5, zn 507,
A maximum temperature of 90° or over was observed as follows ;
1880 1879
ALT, DI GP een csecectucasssceters (OH as 7 days
リン CR SHEE aah res の II
Total for the year 2 4, zn 由記

It will thus be seen that the temperature during the year has been, on the
whole, lower than during the year previous. In the general means the difference
amounts to 1°.2, that for this year being 57°.3 and that for 1879 58°.5.

Table F contains monthly means, maxima, minima etc., collected from the
previous tables. If it be compared with the correspending table for the previous
year it will be seen that for several months the means for the two years agree
very closely, the principal difference occurring during the last four months.
During the months of September, October and November the temperature was
higher than during the same time in the previous year but in December a sudden
fall occurs. In 1879 the difference between the November and December means
was 2°.8, in 1880 it was 11°.5. The minimum temperature for 1879 occurred in
January; that for 1880 occurred in December. A low temperature continued to
prevail in January 1881 and the winter of 1880-81 must be regarded as one of
unusual severity. It must also be noticed that lower temperatures prevailed
during July and August ofthis year than during the same months of the year 1879.

The highest temperature reached during 1880 was 90°.5 on August 21st
and only on one other day was as high a temperature as 90° recorded, whereas
in 1879 the highest point reached was 93° and a record of 90° or higher was
made on 12 days of that year.

The lowest temperature for the year was 22°.3 which was reached on
December 25th and also on December 29th. The lowest point reached in 1879
was 24°.1. The total range for the year 1880 was 68°.2 and for the year 1879
it was 68°.9.

The highest minimum temperature for any month was 65° in August, the
minimum for the same month of last year being 69°.4. The greatest daily
range in the year was 30°.4 on December 22, the maximum daily range for 1879
being 29°.8. The minimum daily range was 1°.4 on March 2, that for 1879
being 3°. During the whole twenty four hours of March 2 the temperature
differed from that of freezing by barely 1° and again on the whole of July 5 it
did not differ from 70° by more than 2°.

In Chart No 4 the curve of monthly means of the standard thermometer is
shown in the first diagram. In the second the maximum, minimum and range
for each month and in the third the maximum and minimum daily range for
each month are exhibited. In their general features these diagrams will be
found, of course, to resemble those of last year very closely.

TABLE C. SHOWING READINGS OF STANDARD THERMOMETER.

April

46.0
40.0

33.0 |

TABLE C. SHOWING READINGS OF STANDARD THERMOMETER.

| 7 |
October. | November. | December.

| 57.0

| |

| 56.0
60.9

| 59.0 | 45

TABLE D. GIVING MAXIMA, MINIMA AND MEAN TEMPERATURES
FOR THE YEAR.

| January | February | March April
Day | 8 | A | etl | ee
|Max.| Min.) M.| m. |Max.|Min.| M. | m. |Max.|Min.) M.| m. |Max. Min| M.| m.
i | |
1 | 51.0/ 335) 41.2] 42.2) 40.9) 30.4| 85.8] 35.6) 43.7 | 28.0 | 36.8) 35.8 | 59.4 |39.0| 49.5
2 | 47-4) 28.0) 38.9) 87.7) 54.5) 80.4 41.8 | 424 1223) 309/821 |31.6 56.0394 | 47.2
3 | 52.9] 28.5 | 43.0| 40.7| 61.5 | 81.8 | 43.4 | 46.6] 41.0] 28.0 | 34.8) 34.5 | 45.7 | 35.0| 39.7
4 | 49.3|33.0] 39.4] 41.1) 43.7 | 52.0 | 38.6 | 37.8} 42.1 | 29.4 | 35.8 | 35.7 | 51.2] 87.8 | 44.3
5 | 47.5| 26.5 | $7.7 37.0) 43.0 | 31.0 35.8 | 37.0| 43.7 30:8 |37.1| 37.2 59.5 | 40.0| 49.3
6 | 45.2|27.5| 35.9 303) 39.3] 27.0 | 84.7 |33.1 | 47.4 |29.0 39.8 | 38.2| 51.2| 40.6| 45.4
7 | 44.8| 23.5 | 34.4/34.1| 44.8] 28.8] 414 | 36.8) 53.5 | 31.5 | 44.4 | 42.5 | 59.8 | 39.0 | 52.0
8 | 46.5 | 25.0| 36.2 35.7| 44.4| 37.5 | 41.6 409| 44.0| 39.8| 41.7 41.9) 64.2] 43.4| 54.7
9 | 44.0] 27.0| 35.3|35.5| 99.8| 35.4 | 37.7 37.6| 52.5 | 36.8 | 45.6 | 44.4) 63.6 | 46.0 | 55.2
10 | 46.2| 27.2| 39.8] 36.7| 51.8| 30.5 | 41.2 | 41.1) 58.0 35.4 | 47.6| 46.7 | 69.3 | 46.1 | 60.0
11 | 48.7/ 26.5 | 39.9] 37.6) 46.9 | 38.0 | 43.9 | 42.4 56.5 | 43.0 | 49.8 | 49.7 | 71.3| 55.6 | 60.3
12 | 47.5/31.8|387|39.6| 48.5 | 40.5 | 45.0 | 44.5] 57.9| 41.5 | 49.7| 49.7 56-7 | 46.6| 49.1
18 | 47.0| 26.8] 35.7 | 36.9) 52.8 | 38.0 | 42.9 45.4) 59.0| 38.0] 49.8 48,5 50.0| 44.0] 47.7
14 | 42.2| 26.5 | 33.4/34.3] 55.5 | 35.8 | 44.0 | 45.6] 63.7 | 42.3 | 49.9 | 53.0] 58.0] 43.6| 51.1
15 | 44.7| 26.8] 35.1] 35.7 | 62.9] 34.6 | 47.2 | 38.7| 63.0 | 35.8| 49.1] 49.4] 58.3 44.0] 51.5
16 | 42.7] 28.9] 34.9]35.8) 63.4] 35.3 | 50.8 | 49.3) 57.6 | 37.8) 48.9 | 47.7 | 53.8 | 50.0 | 52.8
17 | 47.5] 28.5 | 33.6 | 35.5| 43.5 | 39.3 | 40.9 | 41.4) 68.0 | 47.8 58.0 57.9] 55.5) 49.6 | 52.1
18 | 41.5 | 25.0] 83.4 | 33.2) 48.5 | 37.5 | 43.7 | 43.0) 63.0 | 51.0 55.7 | 57.0), 56.8] 43.5 | 50.9
19 | 39.3/ 27.5| $4.4|33.4] 87.8] 34.0| 35.7 | 35.9| 66.8] 53.0] 59.9) 59.9] 60.8) 42.5] 54.8
20 | 42.8| 28.0] 35.7 | 35.4 | 54.4 | 36.5 | 47.3 | 45.4 | 61.6 | 34.9] 49.8 | 48.2] 64.2 | 46.1 | 56.6
21 |37.0|29.3|33.3|33.1) 51.0| 44.0) 46.9 | 47.5 | 53.0| 33.8 | 44.4 | 43.4| 69.0 46.0| 58.8
22 | 46.9] 30.0 |38.2 38.4) 45.8] 37.4 | 41.3 | 41.6) 55.2 | 38.4 |47.9| 46.8 | 67.5] 53.0| 60.0
23 | 49.9|31.4/ 37.8 | 40.6| 44.4 | 31.4 | 37.0 37.9] 55.8| 43.5 | 47.1] 49.6] 68.0| 56.2 61.5
24 | 40.2/26.0| 31.0/ 33.1) 44.8 | 27.3 | 36.4 | 36.0) 47.0 87.0 | 41.1] 42.0] 72.7 | 56.5 | 63.5
25 | 42.0/ 23.0] 82.4/32.5| 42.5 | 29.4 | 35.9 35.9) 57.0 | 30.5 | 44.9 43.7 || 60.0 | 49.0 | 54.4
96 | 47.0| 25.0| 37.1| 36.0] 51.2) 26.4 | 40.6 | 38.8] 63.9 | 38.6 | 52.7 | 51.2] 61.1) 40.2 | 54.3
27 | 47.0| 28.0| 35.1] 37.5] 53.0] 40.5 | 47.2 | 46.7 | 64.5 | 41.0 | 54.0 | 52.7] 65.5 | 51.0] 59.8
46.5 | 27.4 36.9| 64.0| 45.0 | 52.1/ 54.5 | 66.0 | 52.4 | 56.4] 59.2) 73.5] 60.2] 65.6
45.8 | 26.6 36.2] 40.1 | 33.0 | 35.6 | 36.5 | 57.0| 40.0 | 47.4 | 48.5] 61.5 | 53.5 | 57.2
43.0| 27.8 35.4 58.2 | 39.6 | 50.8 | 48.9 | 55.4 48.5 | 52.4
40.4

in.| M. | m. ‘Max.

59.6 |
60.2 |
60.3

2| 59.8) 74.9 |:

614

643 78.3
56.8 79.2
|
58.5 |

63.1

61.6

66.2 |
62.0}

64.5

kJ Min,

August

74.5
74.7
75.0
715
73.9
71.9
73.0
74.0
73.7
72.0
68.5
72.7
67.5
73.0
711
75.6
74.8
70.4
72.8
69.1
71.3
68.5
73.0
71.8
74.8
76.1

69.6
68.7

67.7
65.0
65.0

82.9
817
80.9
79.9
81.9
79.6
80.1
76.7
76.8
78.6
77.1
75.5
78.2
79.6
82.2
80.6
80.6

77.3 176.9

77.2
76.6
78.3
79.0
80.0
82.0
81.0
83.2

75.6
74.6
72.7
74.0
69.2

TABLE D. GIVING MAXIMA, MINIMA AND MEAN TEMPERATURES
FOR THE YEAR.

82.3
81.6
80.3
78.8
80.8
78.6
78.8
78.0
76.8
77.7
75.7
76.0
76.1
78.3
72.9
80.0
79.9

77.6
76.3
778
77.7
79.5
81.1
80.7
82.2
74.2
74.5
73.8
71.9
69.0

TABLE D. GIVING MAXIMA, MINIMA AND MEAN TEMPERATURES
FOR THE YEAR.

Max.

Min.) M. | m.

September

October

Max

.Min.| M. | m.

November

Max.| Min.

December

75.4
78.4
77.6
80.4
73.4
73.1
82.0
85.1
81.1
84.3
84.6
81.0
83.0
82.5
83.1
81.5
77.8
79.0

| 79.9

73.6 | 57.7 | 68.8
76.0 | 57.0 | 65.6
62.5 | 95.0 | 57.9
67.8 | 55.6 | 61.8
69.7 | 56.0 | 60.9
72.6 | 52.9] 63.4

60.9 | ¢

66.2

| 68.2
| 66.9

68.0

57.0} 51.

64.6
65.9
63.1

| 58.9

57.9
51.1
50.3
60.6
65.0

| 62.1
| 60.0

2.1 | 59.1

2 | 63.3

56.0
57.9
53.9
53.9
48.0
50.2
49.5
49.9
53.1
49.2
54.3
54.8
54.6
52.1
47.1
46.6
52.3
49.1
49.3

32.3
34.5
36.0
315
30.5
33.2
26.8
30.0
29.0
26.3
29.1
34.3
35.5
29.0
36.4
29.0
30.2
33.0
30.7

44.7
45.6
44.2
42.6
37.8
41.4
37.4
39.6
40.1
36.3
41.7
44.7
43.9
40.8
40.7
35.4
39.8
39.5
39.6

80.1
83.1

57.0
61.0

47.5
45.7

30.2
26.0

37.7
35.8

| 82.9

60.4

55.1

24.7

38.5

| 88.9

74.1

48.5
57.8

45.3
39.1

32.0
25.1

39.2
32.9

72.0
73.0
76.0
74.0
717

74.0 | 59.2

57.2

‚61.1

60.9

| 53.9
N

| 52.9

57.3

45.3
47.6
48.3
42.5
47.0
47.2
44.1

22.3
22.8
24.0
25.7
22.3
32.3
25.6

34.4
33.6
35.8
33.4
34.2
38.3

TABLE E. SHOWING THERMOMETRIC RANGE FOR EVERY DAY
IN THE YEAR.

| September

| December

| October
| November

| February

mm
=
i
の

an
bo
oa

Hm
Sm
ie

~
bo

L6G

rtG

SUR YY
WO}
SUBOTY

“Tau, "PS
uroa}
suvayy

"wo9al]

*WaAAON | "0190 | "lag

ysndny

oun

Avy

IndY

"IILANONYAHL GAL 70
AONVU NVY VNININ VIVIXYIV ‘SNVAN AIHLNON DNIAOHS “A TIIVL

Tore yy

"urp

9 id +

dio gt
oe

Heart
ul

ah
中

BERN

23
Showing the readıngs of standard

Chart N
thermometer during the year 2540

(1880)

ーー ーー シー

2 = た
Chart N°2 u eS
Chart N > | & = = 年 - に yo,
Showing the readings of standard F = a. G 2G
Ihsrmomeler during the year 2370 『 = = as =
J 2 2
(1880) REIT

Lo RS

a
©

March x0 ID
き

ゲド の ZI 20 27 22 23 24 ee 26 27 28 29 70 31

ー ペ ーー

ー = 一 ーー デーーー ニ ーー デー ©

100

プア 18 19 £0 21 22 28 タタ 25 26 27 28 29 30

April mh

の
>
つ

| gol

Chart N23

Showing the readings of standard
thermometer durng the year 2540
(1880)

wi Bb
ik

(Hae +
\\

nahe
2 ーー
Oy WT SW
SB

ーー

Chart N23
Showing the readings of standard
thermometer during the year25#0

(1880)

1 ne nn nn

INES

Chart
Showing the readings of standard

thermometer during the year 2540

(1880)

Chart N23
Showing the readings of standard

thermometer during the year 2540

= 4

157 yu? な a
y 0, Sra ol *
Chart N f = 4B iD # 5 第
: Zee ah ANE Ep 開
Showing I mon thly means of slandand lher 7 月 降 ial
J . ロロ 中 4 | Ei
momeler r pF a et re.
and max min and ran for cach month CE 2 f =
ran ny Oe A aa 4-
3 mar and min daily range foreach mouth I た |
ー 2 “a
に = = i= "4 vr SS Ih に ヒ AN by ie, r BY = ”
] 、9 ro = 4 Sin we = => Sn T = =.
ISA & ASA A PA SaaS A AS AS eig 5
& いこ シン ーー で
iv
me N
3) ve < “
65 ner

A- AA SATA BAA ASA Ng + AH AS
Jan. Feb. Marh April. May. June. Iuly. Aug. Sepr. Oct.‘ Nov. Dec.

= Serra

10
$0

20)

月 朋 月 - 月 eıfl & 月 A 月 + / で 月 ん ) | )] ly

Jan. Feb. March ont May Sune. Ju ly. Avo xp. Del Voi Nex

: | | | | | |

THE WIND.

Observations upon the velocity and direction of the wind have been continued
by means of the anemograph and Robinson’s Anemometer referred to in the
report for last year. ‘The anemometer was blown from its place and considerably
injured in the typhoon of October 4th. Fortunately it was possible to repair
it but for a few weeks subsequent to that time the results are entirely based on
the records of the anemograph. Experience seems to have demonstrated that
the ordinary lighter forms of anemometer have not sufficient strength to resist
the high winds to which they are so frequently exposed at this point. It would
be extremely desirable if their construction could be modified in some way so
that their power of resistance might be increased without at the same time
diminishing their sensibility. ,

Table G. gives the actual number of miles of movement of the atmosphere,
since the last observation, at each of the three observations for every day in the
year, with totals. From this it will be seen that the total movement of the air
during the year in all directions was 53279.5 miles, being an average of 145.5
miles for each day or a trifle more than 6 miles per hour. Compared with the
total movement for the previous year, this is an increase of nearly four thousand
miles or about eight percent of the whole. During the Months of March, May,
July and September of 1880 the movement of the atmosphere was less than for
the corresponding months of 1879, but during the remainder of the year it was
considerably greater. Table H gives the direction of the wind at every observa-
tion made during the year. When there is no perceptible motion of the air at
the time of making the observation, the word “calm” is inserted.

Table I exhibits the total movement of the atmosphere for every day in the
year with the prevailing direction of the wind during the day and the maximum
and minimum movements for each month, As stated in the previous report it
is very difficult, in many instances, to determine the prevailing direction cor-
rectly, hence it is highly probable that this table will, in certain cases, appear
inconsistant with others.

The maximum movement of the air in a single day occurred on March 20th
and amounted to 565.9 miles or an average of about 23.6 miles per hour for the
whole day. This is less than the maximum movement for 1879 which amounted
to 597.5 miles on December 26th. Although on March 20 the average velocity
during the whole day was higher than on any other day of the year, the real
maximum velocity of the wind was by no means reached on that day. This
occurred on the night of October 3 and 4 at which time the wind reached such
a degree of violence as to entitle it to a place among the most violent “typhoons”
which have visited this region for several years. Some special observations of
velocity of the wind and the barometric height were made during the storm and
they are referred to in another place, diagrams of the most important changes

being given. Then were three days in the year in each of which the total
amount of wind exceeded four hundred miles. They were February 28, March
20 and October 4. If the barometric curves be examined at these dates it will
be seen, of course, that these high winds were preceded or accompanied by a
rapid and considerable fall in the barometer, the minimum barometric height
for the year occurring at 7 a. m. on March 20, the day on which the maximum
number of miles of wind is recorded. The minimum amount of wind recorded
on any single day in the year was 34.7 miles on May 19 and the smallest amount
for any month was 3288.9 miles in the month of September.

Tables K, L, and M, exhibit the winds classified as to direction and velocity.
Table L shows the classification of the winds exceeding twenty miles per hour
in velocity, according to direction, for all the months of the year. From this
it will be seen that on 83 occasions the wind was recorded as reaching a velocity
of twenty miles or more. By comparison with the corresponding tables for the
year 1879 it appears that high winds were more frequent during this year than
the last, in about the ratio of two to one, the number recorded for last year
being 40. A very similar distribution will be observed, however, as to direction
and also as to time. The same marked excess of these winds from the North
and Northwest is shown in both tables. During the year 1SS0, out of a total of
83 high winds, 61 are recorded from the North and Northwest; in 1879 the
number was 30 ont of a total of 40, the ratio being almost exactly the same.
Concerning these two directions, however, the proportion for 1879 is reversed in
1880, by far the greater number being recorded from the North. An examina-
tion of table M, which shows the classification of the total miles of wind as to
time and direction, proves that the largest part of the atmospheric movement is
from the same direction. Out of a total of 53279 miles of wind during the year,
28791 miles, or something more than one half came from the North and North-
west. In 1879 the proportion was slightly less than one half. Both of these
tables, in connection with table K, show the excess, in both velocity and
frequency, of winds from the North and Northwest.

A clearer conception of these facts, however, can be obtained by an
examination of charts numbered from 5 to 10 which have been construct-
ed to represent the distribution of winds as to direction and time, according
to the method described in the last report. Some peculiarities of these diagrams,
to which attention was called in the last report, are again repeated in the
curves for 1880. In chart number 5, which exhibits the prevailing direc-
tions for the months of the year, there is the same appearance of a shifting
of the wind from the North to the South through the East. during the
months of May, June, and July, and the same sudden change from South winds
in August to North winds in September. Indeed, if the general chart for the
Year, number 7, be compared with the corresponding chart for 1879, it will be
seen to resemble it very closely, about the only difference being dne to the fact
that winds from the North exceeded in number those from the Northwest to a

greater degree in 1880 than in 1879. In fact a comparison of all of the wind
charts of 1880 with those of 1879 indicates a considerable degree of permanency
in the annual atmospheric movements. To facilitate this comparison, a series
of charts, numbered 5a to 10a, have heen prepared, showing results corresponding
to the foregoing, their construction having been based on the mean of the results
for 1879 and 1880. ‘The method of constructing the first diagram in chart
number 10 was explained in the last report. It shows the general course of the
wind during the year 1880 and it indicates a resultant movement of the atmos-
phere of 18000 miles from a direction North by about 4° East. Number 10a
shows a similar construction for the mean results of 1879 and 1880. It indicates
a mean annual movement of 14000 miles from a point North by 8° West. The
second diagram in chart number 10 consists of a curve showing the number of
miles of wind for each month of the year. It agrees very well in its general
features, with the corresponding curve for 1879 and in the second diagram of
chart number 10a the mean curve for the two years will be found. This exhibits
a great maximum in March and two other maxima, one in July and the other
in October.

If these mean curves be continued from year to year as observations accu-
mulate, they will, in time, approximate very closely to a stability in form, and
it does not seem to be assuming too much to affirm, upon the evidence of but
two years’ observations, that this form will not differ greatly from that in which
they now appear.

THE TYPHOON OF OCTOBER 3 AND 4.

On the night of October 3 and 4 occurred a wind storm of extraordinary
violence, although, fortunately, of short duration. Indeed, so rapid were the
barometrical changes, that, coming as it did between two of the regular observa-
tions, it may be said to have hardly left an impression upon the regular series.
Although without doubt the barometer reached the lowest point for the year at
that time, in the regular series the minimum is credited to March 20 as is
previously stated. Fortunately Mr. Nobutani, recognizing the unusual violence
of the storm, began making hourly observations of the barometer at 1 o'clock
a. m., afterwards increasing their frequency to every half hour, and continued
them until the following morning when the storm had entirely subsided. The

anemograph furnishes a continuous record of the velocity of the wind and we are
thus able to trace, with tolerable completeness, the most important elements of
the phenomenon. The storm was by no means local as it is known to have
caused much damage along a considerable extent of the coast, and to have
brought about much disaster to shipping in the adjoining seas. In order to
undertake anything like a complete discussion of the phenomenon, it would be
necessary to collect from numerous and widely distributed sources, all of the facts

possible concerning its rise and progress. This task, which is by no means an
easy one, has been undertaken by others who have unusual facilities for accom-
plishing it, and it will be desirable to confine this brief account to the results of
the records made in the observatory, and their description may be borrowed, with
slight modifications, from a discussion of the principal meteorological features of
the storm, given by the writer in the columns of the Japan Weekly Mail, soon
after its occurrence.

Although it can hardly be said that this storm gave any marked indications
of its immediate approach, yet it is important to observe that there was a steady
fall in the barometer from the previous Thursday

September 30th—up to the
time of maximum violence of the wind. Chart number 10c exhibits the baro-
metric curve during the most interesting period; 7. e. from 7 o’clock a. m. on the
3rd of October, to 2 o’clock p. m. of the 4th. Previous to one oclock on the
morning of the 4th but three observations are recorded; at 7 a. m., at 2 p.m,
and at 10 p.m. ‘These indicate a steady decline in the barometer and it is not
likely that any extraordinary fluctuations occurred during this time. After one
o'clock a. m., the observations were made hourly, and during a considerable por-
tion of the time they were half-hourly. It will be seen, however, that a very
important portion of the curve, from 10 o'clock p. m. to 1 oclock a. m., is
doubtful and it is not at all unlikely that, had intermediate observations
been recorded, the fall of the barometer would have appeared much more sudden
than it does. The minimum observed height was 28.735 inches at 2 o’clock a. m.
At 3 o'clock the height was only a trifle greater than this ‘and, from the nature
of the curve before and after the interval from 2 oclock to 3 o’clock as well as
from the velocity of the wind, it seems highly probable that between these hours
a lower point than any observed was reached. The curve is constructed to show
the actual vertical movement of the mercurial column. From the minimum it
rose rapidly until 6 o'clock a. m. at which hour the height was 29.386 inches,
and from that hour the rise coutinued with less rapidity but with great steadi-
ness, until the night of the following Wednesday, when the reading was 30.378
inches. Thus the range of the barometer in three days was 1.643 inches. This
is more than two-tenths of an inch greater than the range for the whole of the
last year and nearly three-tenths of an inch greater than the range from the
regular series of observations for this year. At no time during last year did the
barometer reach so low a point as 29 inches, and the mean height for the year
was 29.952 inches.

Chart number 10d shows the velocity of the wind at different hours extend-
ing over the same interval of time. These velocities are computed from a
continuous record made by the anemograph consisting of a Robinson’s anemometer
with Beckley’s registering apparatus attached. From this curve it will be seen
that, so far as the wind is concerned, up to about 11 o'clock p.m. there were no
indications of the coming storm. At that time a breeze sprang up, which conti-
nued at less than twenty miles per hour until about 1 oclock a.m. when it

suddenly increased in velocity, and at 2 a.m. the record shows a speed of sixty
miles per hour. Unfortunately shortly after 2 o’clock, the clock-work which
keeps the registering portion of the apparatus in motion was stopped, the motion
of the pendulum being undoubtedly arrested by a sudden blast of great violence.
This stoppage was not dircovered until 3 o'clock a.ın., so that between these
hours the record is lost. At 3 o'clock the instrument was put in motion again
and for about fifteen or twenty minutes after that hour, the record shows the
extraordinary velocity of ninety-five miles per hour. From this time the violence
diminished rapidly, a velocity of fifty miles per hour being registered at 4 a.m.,
and at 5 a.m. it had fallen to less than twenty miles per hour. Twice afterwards
as will be seen by the chart, the speed rose to about twenty-five miles per hour,
after which it rapidly declined.

Owing to the interruption in the continuity of these records, it is impossible
to affirm that the maximum velocity of the wind was recorded. In fact there
are reasons for believing that the storm reached its greatest violence somewhat
before 3 o'clock.

It seems quite certain then, that at times during the storm the velocity of
the wind exceeded one hundred miles per hour; and especially must this have
been the case during some of the most violent blasts which were generally of too
short duration to show with their full effect upon the register made. The fact
that the pendulum of the anemograph was stopped between 2 o’clock and 3
o'clock by one of these blasts, and that after 3 o’clock its motion was not inter-
fered with, would indicate that more violent disturbances took place before than
after that hour. A smaller anemometer of Robinson’s model was torn from its
fastenings between 2 o'clock and 3 o'clock, and so completely demolished that no
record even of the work which it had already done could be obtained. This is
much to be regretted, as otherwise a means of verifying the extraordinary velo-
city registered by the anemograph would have existed. Concerning the latter it
should be said that, regarding the smalicr anemometer as a standard, it has been
fonnd upon examination to somewhat over-estimate the velocity of very high
winds, and to under-estimate those of low speed. At the same time it cannot
be positively stated which of the two instruments was in error.

A continuous record of the direction of the wind is kept. Upon examining
this it is found that dnring the whole of the period considered, the direction
varied between North and West. Up to 1 o'clock a. m. of the 4th the wind was
stealily from the North-north-west. From that hour until 5 o'clock a. m. its
fluctuations were confined between North west and West. <A decided change in
direction seems to have taken place between the hours of 2 o'clock and 3 o'clock.

The early part of the storm was accompanied by an unusually heavy fall of
rain. The violence of the wind prevented the reading of the rain-ganges during
the night, but when emptied at 7 a.m. they showed a total of 4.66 inches, nearly

all of which must have fallen during—at most—two or three hours.
It may be interesting to mike some comparisons between the violence of

this storm, and that which was undoubtedly the immediate cause of the destruc-
tion of the Tay Bridge, on the evening of the 28th December, 1879. Unfortunately
it does not appear that any very exact or reliable obsersvations of the velocity
of the wind on that occasion were made; but an approximate measure of it may
be obtained from the testimony of several of the witnesses, who were men of
considerable experience in the observation and estimation of high winds. The
following selections from the Times report of the Board of Trade inquiry, are of
interest in this connection. Captain Scott, R. N.—who was superintendent of a
training-ship stationed in the Tay, testifies that his barometer fell from 29.60
inches at noon to 29.09 inches at 7 oclock 一 that being the lowest point reached.
Also that in the Navy, storms were described by numbers from 1 to 12, 12 being
the maximum. Upon that scale he would describe this storm in the Tay as
from 10 to 11. He had on rare occasions in China and the West Indies rated
storms as high as 12. :

Admiral William Heriot Maitland Dougal, who had resided at the mouth of the
Tay continuously for twenty nine years, stated that his barometer fell from 29.49
inches to 28.80 inches. 'The difference hetween these and the previous barome-
tric heights, is easily explained by the fact that his house was at an altitude of
200 feet above the level of the sea. ° He declared that the gale-was like a typhoon
in violence, and that in all the time during which he had lived on the Tay, he
had never experienced a gale of equal severity. In his opinion the velocity of
the wind was from seventy-five to seventy-eigat miles per hour, and that during
the lulls it would fall to something like thirty.

Charles Clark, who was an amateur observer, gave evidence that 29.00
inches was the minimum point reached, and that he had marked the storm 4 on
a scale of 6; and that he had never yet recorded 5 or 6.

Other witnesses testified in about the same way, all agreeing reasonably
well as to barometric depression and probable velocity. :

On comparing these statements with those already made concerning the
recent typhoon here in Japan, it will be seen that both in barometric range and
in wind velocity, the receut storm considerably exceeded that which was the
occasion of the Tay Bridge disaster. The barometric change was not only greater,
but more sudden in the former than in the latter. Concerning the direct meas-
urement of the pressure of the wind in pounds per square foot, it must be said that the
instruments for doing this are, at present, to a great extent erude,and unreliable.
It is generally assumed that the pressure is proportional to the square of the
velocity. Upon a scale 1optst by the Smithsoniu Institution and by the
United States Signal Service, the velocity of twenty-five miles per hour corresponds
to a pressure of 3 tbs. per square foot. Assuming the correctness of this and
also of the law given above, the pressure per square foot in the Tay storm must
have been nearly 30 Tbs. and in the recent typhoon here it must have been
nearly 50 tbs. It was shown in the tests made upon the material of the Tay
Bridge, that it might have been expected to give way under a wind pressure

considerably less than 40 Ibs. The French and many English engineers have
adopted 55 tbs. per square foot as a standard, and about the same number is used
in America, but it seems doubtful if even that furnishes a sufficient “factor of
safety.”

In conclusion, it may safely be said, especially in view of the damage done
to buildings. shipping, ete., that this was one of the most violent storms experi-
enced here for many years. From facts already known concerning other points
along the coast of Japan, it would seem that, liad an efficient system of observa-

tions, telegrams, and signals existed, timely warning might have been given

of its approach and, possibly, much property and many lives saved.

TABLE G. SHOWING MILES OF WIND RECORDED AT EACH
OBSERVATION IN THE YEAR.

April

4691.4

5814.1

93.6| 37.8

23.5) 56.3

242| 28.6

78| 36.7

| January February | March
| 『 | 2 | Beare). Seo ae
| 90.6] 28.6] 816| 882| 48.5) 823| 55.1] 03.5) 438] 285| 954
| 972| 22.3) 20.3) 33.9) 879| 25.0) 845| 755 63.7 |
| sg45| 568 249 | 29.4| 468| 939] 734| 469| 84.1| 1073| 115.1
| uo} 93a] 243] 5325| oss] 1077] 423] 535] 55.64 1274] 633
| 21.6| 223] 811| 166.5| 1332| 39.6) 529| 802] 52.7] 1118] 1134
| 304] 240] 732) 654) 67.0] 596] 307] 372] 326
| 43} 385] 47.7) 554] 486| 352] 472| 346] 298] 811| 391
| 423| 237] 158] 217] 270] 724] 223] 326] 599] 252] 344
| o47| 129] 152) 922] 716] 649] 204] 285] 29.1
| sı2| 1160] 857) 1007| 95.8] 364] 23.4) 40.0] 564| 294) 9927
| 717| 60.1| 51.7] 255! 424 499| mns| 537| 714] 140} 1304
| 856| 502 503] 78.8| 568| 11.9] 36.6] 45.8] 827] 47.8] 493
| 171| 552) 768] 123.5] 1067] 1124) 89.7) 92.6) 591| 568] 61.7
| 432] 99} 215] 1292] 434) 358 | 30.8] 28.0| 1050| 584| 610
| 29.6] 609} 873| 218| 35.0] 646] 615| 216] 279] 35.9] 366
| 893) 101.8] 74.3] 90.0) 80.5 136.3] 311] 263] 25.8] 608] 446
190| 519] 79.4] 776) 472] 510| 340] 41.1] 73.0] 117.3] 440
72.5| ssd| 754/ 502] 485] 362) 559] ea| 246! 638] ez
85.0| 756| 18.0 | 500| 582) 92] 149| 537) 248| 146| 401
14.0) 30.6) 14.0) 169.1] 126.2) 26.5| 169.1) 181.8| 215.0] 280| 45.3
33.4| 45.1] 59.5 | 468| 338] 484| 854] 752] 846| 299) 501
| 576) 1508 1248| 92.6] 72.1 79.7 | 992| 939| 178) 148] 856
| 51.9] 512] 105.7] 67.3] 109.8} 86.0] 20.5] 28.5} 114.0
| 7a| 528] 221] 23.7] 406] 294 | 184.4| 180.6) 129.1] 568| 268
sia] 928] 226/ sso| sz1| 406| gp4| gas| 448] 547] 1445
248| 244] 995| 488] 650| 194] 38.5) 864| 46.5] 562] 519
103.9] 52.0) 22.1 | 822] 199| 27.6) 806| 505 2) szg| 618
38.5} 49.1] 189| 1553| 125.3] 156.6) 73.6] 68.5] 1092| 122.8] 74.7
164| 417| 11.5] 1156] 707| 643) 547| 35.7] 928| 941] 367
| 1014| 1472| 904] en | … | ann | 741] 1371] 954] 292] 612
1877 ee (ieee West heer | 612] 482] 814) ......
1547.9 |1632.7 [15108 | 2161.9|1909.4 1742.8 | 1716.7 |1801.8 [2013.8 | 1657.6 |1847.8 |

5603.9

140.6
78.4
126.1
85.4
58.1
44.8
43.3
45.2
78.0
121.9
55.0
42.4
40.0
48.4
49.3
24.5
42.6
53.9
64.6
60.2
49.7
29.0
121.4
67.2
191.4
88.0
124.7
36.0
19.1
79.3

TABLE G. SHOWING MILES OF WIND RECORDED AT EACH

OBSERVATION IN THE YEAR.

| August

| May | June | July
Day. | aa
74 -2.|.0).7 | 2 | wo fe7-| 2
4 ee ier i ee a
1 | 295| 369| 329] 576) 459| 488| 468
a | al wıl 459) 404 204| 8.6) 69.0
3 | 1| B11) 802) 127| 236) 380) 209)
4 | 222] 22.0] 273] 369] 309| 341) 103
.5 | 96| s48| 794) 131] 264| 102) 408
6 | 197| 618] 1052) 123] 35.5] 479| 75
7 | 1284| 726) 446! 866) ua) 98.7 | 182
8 | 220| 476| 546) 826| 568| 701) 107
9 | 168| al 470) 248] sma 13.3) 126

sı | 160) 871) 58.7] u... |... Pam |
Sums | 901.5 (11004 1746 11044.3|1123.1 1449.1 10183
Total] . 856065 | 380165

TABLE G. SHOWING MILES OF WIND RECORDED AT EACH
OBSERVATION IN THE YEAR.
| September | October | November December
Day. | | 5
7 | 210 | 2) 2 ] ee
|
1 | 801! 726] 49s| 582] 433] 50.0] 480] 38.5] 932
2 | 679] ez2| 550| 654| 583) 329] sos| 809] 30.0]
3 | 516) 440| soo| ul 445] 133] 260] 135] 140 |
4 | 272] 239| 205) 2730| 1240] 305] 465] 348] 190|
5 | 127] 269| 254| 3.0] 150) 520] 188] 220] 220) 479
6 | 269! 150| ssl 220| 455} 2401 915! 140] 85.5} 1058
7 15.5} 15.2] 208] 51.0] 70.0] 42.0] 47.0] 40.8] 26.5
8 | 208) 309) 218] 50| 540] 600| 240) 845) 66.0]
9 | 233| 45.0) 61.1 43.0) 265) 15.5 | 17.0) 52.0 eo
10 | 17.0) 46.5) 28.7] 385) 57.5) 33.0 | 37.0) 38.5] -55.5 |
11 | 143] 23.0] 56.7) 29.0] 325] 230) 470| 48.7] 340]
12 | 45.6] 289| 273] 195| 550| 460] 778| 678) 457)
3 | 91{ 273! 3753| 165] 347) 230) ees| 445] ssl
14 lm 4827| 733| 260| 37.5| 22.5) 1408| 982 93.0 |
15 | 320| 577] 910) 100] sl 754 968! 1070| 1855]
16 | 85.9) 782] 382] 24.0 16.0| 3.0) 42.2) 35.1 202 |
17 | 294] 404] 35.0] 23.0] 505] 265) 3832| 450| 191
18 | 22.0| 26.4] 39.0 | 63.5| 57.2 918 43.1| 87.4| 28.9
19 47| 155| 405| 715| 60.0) 237) 990) 383) 237
20 | 15.5] 22.8] 266) 42.0) 838 15.0 | 36.4) 12.4] 2041
21 | 10.5) 1083| 185] 59.0] 550| 65.5] 36.6] 508| 392
go | 147| 196] 152] 773] 730] 545] 499| 869] 296
23 | so5| 556] sz| 810] 300 734) 405| 208| 13.8|
24 | 1028] 467| 35.0] 85.5] 808] 590] 958] 852| 172
25 | 49.5| 488| 35.7] 588| 87.5] 137] 43.5] 340] 163
26 447| 24.7) 68] 33.9] 90.0] 103.0] 928.0] 55.5) 122.9
27 | 432] 679| 314) gls| 86.0) 418} 99.6] 524] 1115
og | 30.5} 279] 114{ 19.0] 250! 45] 133.7] 98.8) 75.7]
29 | 4s| 352| 35.0) 2835| 31.0' 100] 716) 418] 193]
30 | 8s7| 654] 59.3] 240] 238] 215] 305] 26.0] 284} 11
Tag Re Ne I: 1408| 1400) 62.5] u... | u... .|
Sums |1003.6 |1159.2 /1126.1 | 1614.5 1610.4 1084.6 | 1541.6 |1400:6 1257.1 | 1502.7 |1952.7 [17899
Total 3288.9 4309.5 4199.3 5245.3

TABLE H. SHOWING DIRECTION OF THE WIND AT EACH
OBSERV-ATION IN THE YEAR.

3 SE N a

4 | N | N | NE| NE] N | VE| NW

5iwi|s|N]iNine| FE] N|N] E] N [NW] E

elswls lswlslslsx | N | sm| ElIX.IN| KN

7 | sulnisisisiwisimls/iw|s|s

8 | Nw SE SW: | N|N|w|N INWI8SW| SER| 8
9/wi|si|nw] N|N/| NIN] E| EINW| E | SE

olswIisis/isisielelsiw/is|s|s

u |Nw|Nw| N| N| N/N | N | SE/ SE| SW|NW| N

12 | N |INBINWI N]|N|NI/ NN] N | SE] N | NE/ N
13 abe nw] N | N| N INW|I N| s | S | NW] NE| NE

IHN | MINI SEI Ss tw ea | N.I | mE

Bi NI NIN s | s |Nw| s |sw] N | NE| N

16 | N | N N | N INW| N | Nw] N]| E/N

7 |N|x N | ni ni wf]|eE]Ne| E| N

is | N | N ElzEIn/sn/iEeins|se|e

19 |xw| NE N|N]E]sE| S| E 8

wo IN | E N | | win|niw 8

21 |N|N al ot wee) Bl 8

2 iN] N N | nN |nw] se] se] s | s | 8
23 | se N N | N IN INWIR| 8 | s |sw

| 24 | N | NE | NE| N INWINW| N | sw] E| N
| 25 Inw|nelse|l N Insel ve Inw| s | s | Nw{|Nw| Nw

NE | NE} W 8 S | Nw] 8 8

N
27 1N|slslNlsslswlwlslslwlsls
28 |W| sl|NINwINw| ls|NlNwlsw| N | NE

NE | NE

TABLE H. SHOWING DIRECTION OF THE WIND AT EACH
OBSERVATION IN THE YEAR.

May

| Calm |

Calm |

June

bho

Calm

E
Calm
NE
Calm

Calm |

July August
9 2 10 7 2 10

| NE

Calm

| SW
| NE

SE
NW
| Calm
ae

の の

の

レン

TABLE H. SHOWING DIRECTION OF THE WIND AT EACH
OBSERVATION IN THE YEAR.

ーー

Calm
NW
NW
SE | N NW
Calm| 1 | N

Calm
SE

|
Calm |

TABLE I SHOWING MILES OF WIND AND PREVAILING DIRECTION
FOR EVERY DAY IN THE YEAR.

January February March

Miles I DE Miles PSDs Miles Rum

150.2 | 119.0 N 162.4

69.8 | 96.8 7 223.7

166.2 | 170.1 154.4
59.3 | 284.0 151.4
75.0 339. 185.8
100.5
111.6
121.8
78.0
119.8
206.4
115.1
131.4
163.8
111.0
83.2
148.1
149.8
142.9
565.9
195.2
90.4
163.0
494.1
150.5
116.4
163.3
251.3
183.2
306.6
191.3
565.9
78.0

TABLE I SHOWING MILES OF WIND AND PREVAILING DIRECTION
FOR EVERY DAY IN THE YEAR.

| |

May June | り uly August

Day. Ti Te T Fer

Miles | P.D. Miles P.D. | Miles P.D. | Miles Pos

レイ
ドコ
つ
~
の

1 | 993 | N | 183 | NE | 173
21 1322 N 69.4 NE 176.7 N ye GRO S
3 | 1224 S 74.3 E
4

| 8.6 | NE 103.7 8

71.5 SE | 1019 N | 952 S 102.5 S

eh! ae «Sd. 2 Soe soe, | Ne | 778 S
6 | wer} s | 957 s | | s | sol s
7 | sse| oN | gg9 s | aoz| ne | wı| se
9 | m2| s | 159.5 w | 37 SE | 1673 SE
9 | 872 | s 95.2 E | 762 SE | 1511| SR
io} 143] N | 969 s | uss | NE | 29] 8
ll | 166.2 | N 58.6 sw. | -815 | se | 1002 | NW
12 | 1313 | s | 66 E | 99.8 S 46.3 S
13 799 | Nw | u E-| wz| m | 1088 S
14 | 807 | SE | 903 E | 105 | NE | 192 8
15 | 1794 N | 1160 8 | 1250| NE | 198 S
16 556 | -SW | 294.5 SE 94.8 E 259.3 S
7 | 1906] we | 1982 sw | 676 8 127.9 8
i} on8 | s I sn NE 85.5 8 | 952 E
19 34.7 | NW | 167 E | 83.5 SE | 127.0 NE

20 | 1390 | NE | 187 NE 153.9 NE | 1247 sw

2 93 | S 181.6 N 184.8 NE | 102.7 SE
23 i646 | s | io| ne | 202 woul zu S
2% | mal s | ul gs | ol xs | mil s
vg | 148 | SE (0.9 N .| 125.1 sE | 9i.9 SW
25 | 72.6 S 144.6 N | 1080 NE 266.5 S
26 | “A| NE | 2017 N 120.4 E | 1984 8

30 86.8 8 47.6 NE | 217.5 S | 91.1 NI
引 1118 Bi aoe | 1016 sw | 1178 I
Max.) 256 | N | 2999| sg 207 | N | 265 S

Min} 97 | NW | 476 | NE 407 | NE 46.3 R

TABLE I SHOWING MILES OF WIND AND PREVAILING DIRECTION
FOR EVERY DAY IN THE YEAR.

|
September October | November December

Miles P.D. | Miles P.D. | Miles Pes Miles Pu:

91.5 N 81.0 NW 315.6 NW
163.0 N 114.3 W 59.1 NE
119.0 N 171.5 W 88.0 W

85.0 NW 101.0 NW 1972 W
129.0 NE | 1310 Nw | 164 NW

84.5 + NE | 129.7 N 155.6 W
120.5 N 191.3 N | 260.0 NW

74.2 NE 155.6 N 100.2 N

155.2 NE 91.0 N 192.1 N

90.8 N 68.9 N 235.8 NW |
179.5 N 1261 | NE 144.7 N |
204.5 N 1004 | NE 177.4 W
193.4 ae) Gee CS 125.7 NW
9253 N

N
77.7 N 811.5 NW
N 148.0 sw

427.5 | NW 337.0 N 3156 | NW
36.0 E 535 | NW 591

— a ーー デーーー= ニ ニニ ーー

—_——

| $1101 Jaquisoa(]

0: HE な SI; Jaq MAAN

| s | > | gl

le we | gs IS {uw

| rle |s judy
o lo lalı VNN
G | 91 | SI 494
I | RI | #1 | St WU

Nd OT GNV Ad る 'WV 4 SYNOH HL LV HLNON HOV NI SNOILIHUIA
SHOIHVA NI OHOOH SVM ANIN GAG SANIL FO HOUIN AHL DNIMOHS D ATAVE

TABLE L. THE NUMBER OF TIMES THE WIND REGISTERED A
VELOCITY OF 20 MILES PER HOUR OR OVER FROM
VARIOUS DIRECTIONS.

Month | a | 8R | ¢§ | TOTAL

January
February
March
April
May

June

July
August
September
October
November
December

Total

TABLE M. SHOWING TOTAL MILES OF WIND IN EACH MONTH
FROM VARIOUS DIRECTIONS.

Month

January
February *

March 3 dd. 348.6 |

April | 4574| 544.0]
May

June
July
Avgust |
September
October

November

December

Total 5| 5527. 2672.7 | 3505.4] 9008.7 1829.5

5 Er IE A;
Chart N°: eh

風 - wr
Show. eng the pret atling direction of 2 F- Ps x
7 wes) 2s
the wtn d during the months of the a as
year 2540 (1380) a ez
; N 3b a ¥. oe
N 3b (\
| \
Bi ha kt N36
| 4.24 8
ia | |
FH |
Gi Abel
\ EM | J |
A IR = : e N ER rn R
4 Jan. US R 西 Feb. N 表 西 March. 7 "¢ if
nr R= = |
5 12] S' 17] AN 17]
N 36 N 3b N 3
/ '\ L AN
16 X ) \ | \

6 が EW ae EW Ne“) =
April. = 表 May. X 5 tt 5 June. | メ -
new \ AE | / Ax Vi 4

2 Sm SH
v3 N 36 N 36
/ \
[ウフ N | |
ア / } . )
En J EW NZ E
#7 hy ( 東 7 Aug. i F a5 Sep. ド S ic
At Pa AWN | 月 ん |
\ /
5 A 5 7 S 南
N | 北
N 3b N 3t
/
/ || i
| | / | \
\ | / . Y
we TX A v Nov IV n に Der > h
vo 月 + W & 月 一 it Aor 4

sm S 各 で 間

Chart N6
Sh owtng the ı number of miles of

ER from v carious directions during at

the months of theycar 254.0 (1880)

N 36

Ww = E

= Jen. 束
Sa

Ww E

= April K

|
a
RR | グ

人
Feb \
月 三 |

Si
N 36
W _
a May.
Az
SYR
N 3b

w
” Aug \
n* l

My
No Jt
)
w
„ww Nov
Br
sh

“408

W
® March.
A=

= A Br
Er: - 中
S$ ms +
— / SE -
テテ Be Be 9
N 36

er
EEE
#7 June
A En

N 3b

S 7)

m
a

ie]
ょ ュ

ae aH # H #e
Showing the prevails ng dcrectton cs ee
> BE の ドル 2 人
Sor the wind during lhe year 2540(1580) N 36 A = + 38
rane. oe
/ 中 AA Bh
\
\
/ \
/ \
/ \
/ \
/ \
/
/ \
/ \
/ \
\
/ \

é / 《/ の ()

har! ' 5

~ 77 / 5 s /

2 の the re lalsve nember of miles,
4 = /

の /g ind fron vartous derecttons

= ee Ve

[47 プイ ガイ / オ アル the Year 2540//880)

7 //
/Z/7. IV リ

ipa ug dhe velalive number of times ジア
the we loci ly of Yhe wind was in excess
of se miles per hovr from vartous /
directions during the year 2540 /
(18890) /

A
ノ

Char€N210

fia) - 第
f SC 3 1 て と
Sho ing Pt the general course EEM 第
oflhewind during the year 2520 Se ae
(1880) = ‘
2"< the number of miles of Fl te me
Gendt during the mon ths 4 6
of the year 答 向 =
eae
ee
A a
/ v3
分 >
量 年
\ = 明
Ba
= Ze!
it
中
N
A
| 2\— «l= $ |” me Oh se rar sit ミエ し 。
FASASASA SAP AS AS Aza HA Sy 2, S
Liat TN
上 / \ ン
ミ J \ /
ンーーーー ヤ
514000 ン

Chart N° Fa. FR z& 第
SAowina the Prevailing arvection a A 0 - = fe
an J ToT = ze © S
of wind Mean of 23539and 2540 m. EN
(1§T9and 1880) a oe a >
Sy FS oe
N 36 Nak - 者 年 +5

en N iy je

w : EM E っ \ \ x
wz SS x _ い 月 E
rg Jan. ES 束 し 表 <a March 7 ) di
= ngs
ゞ 需 S 南 SH
N 3b N 4k N ak
IN P2
/ = 1 ae
~ \ \ = ーー 5
に Aprel. \ づ 14 M N E u E
; If srt IN 2 5 June \
Aw | A 4 | 東 Ge | HR
i] 69) |
\] | / \
S w ぐ ca) 6 a
bes re Nab
/N
/| "\
. | \
wf | \ Bw CN Ew ) E
a hy IS] KW Aug AN 東 9 Sept AN 東
月 | / 月 人 | / am
Ss \/
\/
N. jk in” N] Ss #4
/ | N ak Bi
| /\ an!
| | / \ |
w . | ER Ww ( \ Eu B
kh と Po Nor Ki Dee. ti
er 月 hat

AY 27) Ss i で

Chart Nba.

y . . wen
Showing the number of mues [rom
J は = B 1
varvous directions Mean of 2539
and 2540 (1879 and 1880)

N36
N 3b
\
Ww E m
ws Jan. 束 本 Feb
月 一 ーー
S 南 SM
N 3b Nat
Ww E W
April. xk 末 May.
A} A
SH ぐ 南
N jt N 36
Ww | E W
= uly Ei wi Avg
に / 月 ノヽ
/
Si 3m
N tl N 36
N E u
wi Oct eS No
月 十 nt
ゝ 南 S a)

a + 中
y > Bw,
= /
分
Se :

Ew
if we March
Ar
E w
HC W June
月 示
E Ww.
de ay Sept
ral N
EW

7 て 第

Rel “ft Riel Oe
ty = ey Ho

# =. + 2 + oa
SSA PGR Aor | Ae

Ni3t
I?
R
SSH
N 36
AZ に
| K
\
S N)
N 36
( に
] 東
NW]
N 3b

\" だ に
SW

Chart N°7a. 中 ee + 第

Showing the preva cling direction Ps. ote Se
9 ‘ テ デー = ーー = =
of wind Mean of 2539 and 2540/1879 2] IR Te er
and 1880) N \st の 2: Sets ee
| Ape Wiles amas ef ae に
| 内 iw SF +

Chart NSc. bi

wo \ 7
Showing the relative nu mler of meles N 3b Tr. a oe #5
"wind from various directions ST N =’
Mean of 2539 and 2540 (1879 and 1880) ay Xe SM ee =
は Mae ees A

5 x | | 2
5 < fi

Chart ル 99。.

SAowins' the relative number o
limes the velocily ofthe wind ex-

ceeded 20 miles per hour Mean of

2539 and 2540 (1879 and 1880)

Ger \u +t ja + ah

Chart N21 a.

2: Showings the veneral course of the
E 、 ソ る ir

wind Mean of 2539 and 2540 (1879

and 1880)

gr Showing the number of m tles of

wind during each month Mean と

of g33garı d 2540 (1879 and 1880 )

‘ Shas BPRS Ne Sip ae tad AS i+
Sie pr > ご ご S =
z Se& © Sie He Se Ser sie die gic
P
5000 3
P
A

4000

3000

— =”

Chart N2IO c.

Showing he her ht of the baro-
meter from 7A. ae A 3 を みん
2P.M.onthe チ み 2 October, 1580.

StHoBa
zu
ak ty o jun 7

(HF WW FRE gp

S)e43 9

wp ha &

Bort 4
= marke
art
Tomo bie
a He es =
EFAs te
$e 1( 4 YK
Sm NE

“10 2PM.

on theF of October, 1880.

of wind

the veloc ity
Moon the 3"

Chart N 104.

Show;
From 741

さ
3 さき
S ゝ ぐさ
a ; 2S
SS き 32 SS
rhe ee
ROR SS x
KS ドミ N SS «Ss
| S 5 3 Ins さ
iN S 2.8 ミ
SS SS さき き
SIE NE
Is SSS

RAIN AND HUMIDITY.

The total rainfall for the year was 62.899 inches, an increase of almost
exactly 4 inches over that of 1879. The greatest amount in any month was
9.465 inches in June, which is slightly less than the amount for June of last
year, that being also the month of maximum rainfall. The smallest amount in
any month fell in December, being only .016 inches. In 1879 the rainfall was
smallest in November, the amount being, however, more than 2 inches. The
greatest amount recorded in any one day is to be found under date of October
4th although the fall really took place on the night of the 3rd. ‘The fall was
excessively heavy during the short time that it lasted and was accompanied by
a high wind which at once developed into the typhoon of that date to which
reference is made in a preceding part of this report. The maximum number of
days in succession on which rain fell, was 8,—from July 25 to August 1 一 an
also from August 5th to August 12th. ‘The maximum number of days in suc-
cession on which rain did not fall was 17, from January 4th to January 20th;
and from January 4th to February 6th there was only one day on which a
measurable quantity of rain fell, the amount then being but one tenth of an
inch. The tofal number of days on which rain fell during the year was 160,
against a total of 156 for the year 1879.

Of snow,— trace is recorded on February 22nd and also on February 29th,
and on March 2nd the amount measured was 1.386 inches (melted). Snow fell
at no other time in the year.

All of the important facts connected with the rainfall are exhibited in tables
N, O, and P and in charts 11 to 16, constructed from these tables, and others, as were
the corresponding charts in the report for last year. In chart Number 11 the
rainfall for every day in the year is shown, the actual amount of the rain being
equal to the length of the corresponding line in inches. Im charts 12, 13, 14,
and 15 the distribution of the rain according to the direction of the wind is
shown, both as to the number of rains and the actual amount of the fall. The
first diagram in chart Number 16 shows the relative number of rainy, cloudy and
clear days in each month and the second exhibits the total rainfall in inches.
The fall of rain is probably more difficult than any other meteorological pheno-
menon to reduce to any great degree of regularity. A comparison of these
charts with those corresponding for the year 1879 will show, however, that some
degree of periodicity and regularity may be assumed for the rainfall of this
locality. In charts 14 and 15 the results for the whole year are shown and
when placed side by side with those of the year 1879, they strengthen greatly
the conclusion reached in the last report, that by far the greater part of
the rain comes from the North and Northwest. The curves for 1880 have
shifted around towards the North, which is in exact accordance with the shifting
of the wind as shown in the wind charts. When, however, comparison is made

between the curves for the months of the year 1880 with those for the year
1879, considerable irregularity will be observed. This irregularity is most
marked in the summer months; thus, in June 1880 most of the rains came from
the North and Northeast while in June 1879 the greater number came from the
South and Southwest. For the winter months the curves show a difference in

magnitude rather than in direction, althongh they also show a general shifting .

from the Northwest to the North.

The second diagram of chart Number 16 is quite similar in form to the
corresponding curve for 1879. Both show three maximum rainy periods, the first
in february or March, the second, and by far the greatest in June and the third
in October. The inconstancy of the rainfall in the summer months is strongly
shown, however, in that July, which represents a minimum point in the curve
for 1879, almost reaches the greatest maximum in 1880 and September becomes
a minimum instead.

Table R shows the mean percentage of humidity for every day in the year
and table S gives the mean and maximum and minimum for the months. The
mean percentage of humidity for the year was 76.6, the lowest observed being 31.
Although much rain fell in June the maximum percentage of humidity observed
was 97. In the hourly observations, however, for that month a saturated atmos-
phere is recorded on several occasions. In the first diagram of chart Number 17
the variations in the actual force of the vapor for the various months is shown and
in the second the fluctuations in the mean percentages of humidity are exhibited.
Tt will be seen that the real force of vapor is much greater during the summer
months than during the remainder of the year and also that, notwithstanding
the increased temperature, the relative humidity is also, in general, greater

during the summer months.

TABLE N. SHOWING RAINFALL IN INCHES FOR
EVERY DAY OF THE YEAR.

= 8 5 は
0 0 028 0 002) 1.578) .814| trace} .188| .040 0 0
095 0 “1.386 140 0 .233| 2.344 0 | trace| 1.284 0 0
.670 0 006 1.060 0 .018| .184 0 | trace} 1.276 0 0
0 0 0 262) .055 0 0 0 0 | 4.660 0 0
0 0 0 0 0 0 | 1.223] trace} .800 0 0 0
0} o| ol oo| 0| 0 | trace} .100} .052| 0| 505} 0
0 051 0 0 | 3.076} 1.716] .025| .142) trace 0 180 0
0 -650) .286 0 001 0 057 1.095 0 2297 0 O14
0 | .210) trace 0 0 0 720) 441 0 | .006 0 0
0 .620 0 0 583 0 O19} .638| .003 0 0 0
0 | trace} trace] .302 0 031 0 210 0 0 0 0
0 325} .010| .077 0 0 | trace] 446) .006| G9) .070 0
0 0 0 | .075] trace 0 0 0 0 0 | .825 0
0 0 0 .014 0 0 411 0 0 0 145 0
0 0 0 | .400 0 0 | -015 0 0 0 0 | trace
0 560} .510| .880| .240| 1.743 0 0 .595 0 0 0
0 | .090| .097| trace} .040| .102 0 0 0 0 0 0
0 .020| .013 0 0 174 0 070 0 0 0 0
0 | .685| .767 0 0 | .163 0 284 | trace 0 0 0
0 | -920) .730 0 0 | 1.251 0 | trace 0 0 O41 0
-100| .412 0 0 0 874) .805 0 0 006 0 0
trace} 200) .001| .185| .698| trace} .960 0 | trace] .168| 0 0
0 0 942) .635 0 0 0 0 0 400} 070 0
0 0 0 041 0 225 0 | trace 0 100 0 0
0 0 0 0 | trace} 1.228) .180| .610) trace 0 0 | trace
0 | trace 0 0 .220 0 263} .072| .666| 492] ‚031 0
0 580 0 | trace 0 044} .017| .082| .968 0 0 0
0 00 0 | trace 0 | .051| 1.028 012] 0 0 0 0
0 | *0 Mb 805) .050 0 006) .214 660 | 007 0 0 0
ol Aula of mml mml o| of of o| 0
O | essen 1.087-| ...... ZEN vuense 033} 1.568} ...... Fl waver 0

PPSTI] 12 | P86 | 9 |9g8 | & HOPes} ze |g989| T 6T 888TI| 68 |6cO'Tel 8 IRD]
PIO’ | I 5 0 100" | 7 0 0 00 TOD en Jaqwada(]
E20 I 0 0 0 0 087T| 2 PCUUICAN(
TASS | 9S) Wee) TF V0 0 1508: cer | r foo | t loeo | t eosl er 100

0 0 | 6 0 sc | & 99s | 9 OTS | G |98G | F 16897 | OF II Aine.
SIT| ャ 0 0 WTO, he 1202 ee, NSTseil ye wire || gone imma 9unFf
efe | & 0 0. 0 0 12088 | 9 0 0 jee | 8 occ | @ | I Avy
TUTE) ae 大 0 0 |896 | も? | 0 0 alas | Sets hee lieadisl| 6 lady
ee が 【 | 9 0 0 0 0 る OLL 2 0 0 6 が g 6 your yy
oze | 1 o | o | o | o ose} t a Tee Ben | $ Kannaqa.]

| |

MN M MS

VA WAL VOL ANY HLNON
HOVG OM SNOLLIAUIT SIHOIHVA NI QNIA HHL TILIA SIHDNI

NIVG 』O LNOOWY 'IVLOL AHL ONV 1TAd NIVU SINLL JO UHaNAN

MHL DNIMOHS 'O TIAVL

TABLE P. SHOWING THE NUMBER OF DAYS ON WHICH
RAIN FELL, THE TOTAL AMOUNT OF RAIN AND
THE NUMBER OF CLEAR DAYS FOR
EACH MONTH FOR THE YEAR

Month Amount No. of days a Novick 。
which Rain fall| Clear days

January 766 3 20
February 6.105 15 6
March 5.759 16 10
April 5.567 18 8 “
May 5,132 = 42 13
June 9.465 18 4
July 9.359 21 4
u August 6.584 19 1
September 3.286 15 7
October | 9.193 12 8
November 1.867 8 21

December ‚016 3 26

Total 62.899 160 128

TABLE R. GIVING THE MEAN PERCENTAGE OF HUMIDITY
FOR EVERY DAY OF THE YEAR.

. = ba
| > > 2 ss 3 回
a 3 SS = © = =
回 = = = =} a a
= = の & 2 © a 回
> 3 5 = > 80 ぞ ne > [3]

| 5 5 3
2 1818 pa fie Mi ee SB
| A|» en 5 a) x の >) A A

| 59 64 76 54 64 96 97 85 91 86 72 71

| 13 58 48 70 79 83 71 76 78 82 78 88 58
| 14 72 62 61 75 74 ay, 90 79 83 71 58 78
| 15 | 64 73 66 SS 54 gL 78 75 82 79 53 51
| is | 59 | ez | 80 | 94 | 88 | 90 | we | 84 | 87 | 8 | 72 | 50
| 17 70 81 76 66 73 83 77 76 75 84 77 47

18 70 73 87 49 71 93 76 90 79 82 76 58

~~ ee ee

TABLE S. GIVING THE MEAN, MAXIMUM AND
MINIMUM PERCENTAGES OF HUMIDITY
FOR EACH MONTH

Month

Means

Maximum

January
February
March
April
May
June
July

August

September

October

November

December

Minimum

100

a Se

Dw te
eo Gt ue
yp tl tab
| o 8 \ Qo
ya rt).

No
sii
S: < 中
ae
N
ke
yt
き 『

(or gay

に R
ー の 42 の

3
|

ミ
“or

ュー

| |
| |
| |

|
ty
many z I

Luz —— EE _s ee ee ee 0 SE ee Se Oe

BPs +
© og LHS
+ an ful wee
( o & ~ Rob
ya CRP KS |
yg |
IR
Ss
Ss if |
oO: IN 8 |
2 ミミ
ya
SR
S x
き 『

Chart N%9

Si howing relatively the number of times

と rain fellin each month with the wind
in various directions.
X \36
W (Vv EYE
® Jan RK web

July.

Chat N13
Showeng the relatue total amount of Tain

in each month with the wind inverous |

directions.
N 36
Ww 0 a goog
E Ja # wre.
A |
SH
N 3t
先導
a /
wy = W_
| & April. Rm ” May.
f Aw Az
S a
N 36
MM EW
w July. N > ad Aug
AL 月 人
sm
N 36
ie
N
Ww E W
® 7z/. Ww 5 Nov.
AY ir

Ss"

yi =
Eom
A 2 war
OR 0
テ 分 da
IN

E
U A Mareh
月 三
SH 5%
Nae N 3b
ae:
> EW NW
N Ir 5 June 5 AY
\ Aw N
|
|
Wy
s# S
N 3b De
アダ / \
Y Va / Ww
| N 束 vs Sept. し
| デジ 月 な
SH SH
N 36 N 36
pu !
HK IG De a
H=r
S 南 SH

回

0 NE

i "en

ChartNOU4
Showing relatively the number of tines
rain fell の 47 の が lhe year with theuind

tn varus directions

= ee eee ee eee” UL ee eS OS ee, oe

— ee ee

RR +H Un Bs
ae et = |
a (ER +a eh) の た
- FRESH

sg
st" SAR
$ BBA Ph
Sn x

Chart NI5

Showing relatively the number of inches

of rain dunng the year with wind in

various dırectiors

(QR
oe (or BE
BBR AS +
Hi) AE MK ob
KEK A 2
BRK

we \{ Cth

and tbe total rainfall zn inches
each month ジッ

飼い 寺本 tor
Oe tat eae at}
+ AE AS EER eh
4 Ae ke
on Jens 227
Kot

ンク スバ

nthly means of
relative humility

グン ブル

ラン クン

Chart N17

Showing

JE The monthly me
v

Chefa

—_——_—

HOURLY OBSERVATIONS.

Hourly observations were maintained during the months of March, June,
September and December. The entire series of observations is not here repro-
duced as the principal object was to ascertain the diurnal fluctuations. In tables
T aud U will be found the hourly means obtained for this purpose. Table 'T
contains the hourly means for each month and table U the mears for the whole
period of four months. In this the velocity of the wind is omitted and the
mean force of vapor for each hour in the day is inserted. Diurnal fluctuations
are seen much more readily in charts 18, 19, 20, 21 and 22. On the two puges
of chart number 18 are the barometric curves for each of the four months and
also for the mean of all. Concerning the curve for the month of June and also
for the table fur the same month, it ought to be said that it was found necessary
to reject the observations of one day. During a portion of the forenoon of June
7th the barometer readings were undoubtedly serious'y in error anl, as no
means existed for their correction, it was considered best to 1e est them entirely
and in consequence, the barometric observations for the entire day. The hourly
means of the barometer for that month are, therefore, based on observations
during 29 days instead of 30. There were no reasons for doubting the observa-
tions of temperature, humidity ete. made upon the saine day and they have
accordingly been retained. The barometric curves for the four months agree
pretty well as to their maxiinum and minimum points. It is noticeable, how-
ever, that the greatest maximum appears to occur about an hour earlier in June
and December than in March and September. There appears to be considerable
variation in the time of the great minimum which occurs in the afternoon. In
June this seems to be between 4 o'clock and 5 o’clock while in December it is
as early as 2 o'clock. Inasing’e series of hourly observations, however, the
results must be largely influenced by temporary non-perio:lic barometric changes
which sometimes are very considerable an 1 their effect can only be eliminated
by greatly multiplying the number of observations. As it was not possible to
maintain hourly observations throughout the whole year, the months of March,
June, September and December were selected on account of the relation which
they sustain to the position of the stm. An examination of the curve exhibiting
the monthly means from the regular series will show that, in general, these
months were months of low barometric height and the same thing is shown in
the corresponding curve in the report of last year. During the year 1881 a series
of hourly observations will be carried on during those months which seem to be
characterized by a high barometer and these results may, in the future, be com-
bined with those furnished with this report,

The amount of the diurnal movement for each month is as follows;

une Gt ee A
September........... die NIE er
December wen. ee DER
Mean for four months.......... .067 ,,

These numbers have the same relative magnitude as the monthly ranges for
the corresponding months, except that June and September have changed places.
But on June 7th, the barometric readings for which are omitted in the reduction
of the hourly observations, the barometer was unusually low, and had it been
possible to have made use of the records for that day the diurnal range for June
would have been greater than that given above.

The temperature curves show great regularity as to their points of maxima
and minima. The maximum temperature occurs at 3 p.m. in all of the months
ani the minimum at 6 a.m. in March and September, at 5 a.m. in June and at,
7 a.m. in December.

From table I the mean daily range is found to be as follows;

Marche... hace ae 132» M
Junemen; es Pade etme BIS
September... u... „1124
December. ser 1697
Mean for four months.......... 1246

These numbers have the same relative magnitudes as those representing the
inaximum daily range for the corresponding months, from the regular series.

The two pages of chart Number 20 contain curves representing the hourly
means of the percentages of humidity observed during the same four months.
This element is liable to frequent and considerable disturbances from accidental
or non-periodic causes and the curves are, therefore, less regular than those repre-
senting temperature. They exhibit the same general form, however, and follow the
inverse movement of the thermometer very closely. The daily range is as follows;

Märchee. een en 24 per. cent.
>

UN ante SS EN mer,

Beptemberr. een SEN

December .......... RER BE fj = op

From mean curve............. 2D) a ere

It will be seen that the range is greatest when the actual percentage of
humidity is least, and least when it is greatest. It appears from these curves
that there is a time in the early morning, generally from 1 a. m. to 7 a. m.,
during which the percentage of humidity does not vary greatly and that the
fall in the percentage of humidity lags somewhat behind the rise in the thermometer.

Chart number 21 shows the hourly means of the force of vapor for the same
time. hese, like the curves of relative humidity, are somewhat irregular but
are in fair agreement as to their general form. The daily range in no case
exceeds one tenth of an inch and in the curve for the mean of the four months
it is almost exactly one half of that amount.

Table V contains the mean barometer and the mean temperature for every
day of March, June, September and December as obtained from the hourly
observations.

りり

TABLE T. MEAN BAROMETER, TEMPERATURE, RELATIVE
HUMIDITY AND VELOCITY OF THE WIND

| FOR EACH HOUR.
|
kn March | June
] 日 | ー
| pail Baro. Remp. 9 pee Veloci. | Baro. Temp. Hum. Veloci.
| 1| 30.026 43.6 73 | 63 | 29.828 64.3 91 | 30
2| 30.022 | 428 80 | 65 | 29.824 64.1 oy) 28
| 3 | 30.012 yo | ga | vo | 99.893 63.8 92 | 35
4 | soon | 412 | 81 | so | 29.88 63.4 go | 35
| 5 | 30.019 10.7 | x1 5.9 | 29840 | 6320 ale 2030 033
| 6 30.031 404 | 82 | 53 | 29.848 | 63.7 | 9 3.5
| : 7 30.043 40.7 81 | 6.1 | 29.858 64.8 90 | 41
8 | 30.055 42.6 17 | 5.8 | 29.863 | 66.2 SR | 4.7
| 9| _ 30.061 15.2 50 | 7.6 | 29.861 | 68.0 83 | 47
10 30.061 47.4 66 | eo | 29.859 69.5 80 | 49
| 11 30.047 19.4 63 85, 41 29.853 71.0 77 5.3
| 2m 30.026 2312 62 90 | 29.847 72.1 76 5.0
1 30.003 52.5 50 8.3 | 29.83 72.8 75 5.5
| 2 | 29.97% 53.4 As り .8 | 29.829 72.9 76 6.2
3| 29.973 53.7 co | 94 | 29920 | 70 75 | 66
1 29.971 53.0 52 9.5 | 29.815 72.9 75 6.7
| 5 29.972 51.9 65 | 10.0 | 29.815 720. |: kayo aes
| 6| 29.981 50.2 es | 82 | 082 71.1 78 | 66
| 7| 29.988 18.6 73 | 72 | 29.834 | 693 x0 5.6
| 8| 30.002 | 425 75 | 74 | 29.845 08.4 88 | 50
| / 9 30.009 46.5 77 6.6 | 29.856 67.3 85 4.6
| 10| 30.012 45.8 78 6.2 | 29.863 66.5 87 5.0
| | soo06 | 41 79 | 59 | 29860 65.8 89 | 37 |
| 12] 29.998 | 446 78 | 64 | 29854 | 65.1 90 | 38 |

TABLE T. MEAN BAROMETER, TEMPERATURE, RELATIVE HUMI-
DITY AND VELOCITY OF THE WIND FOR EACH HOUR.

September December
Baro. Temp. | Hum. | Veloci. Baro. Temp. | Hum. | Veloci.
29.974 69.0 90 | 39 | 29.018 34.2 70 | 54
29.970 68.5 91 | 38 】 29.952
29.967 67.9 91 3.6 29.947
4] 29.967 67.4 92 33 | 29.044
5). 29.974 67.0 93 | 34 | 29.915
6 ! 29.984 66.7 92 34 | 29.954
7 29.994 67.6 4 | 383 | 29.964
S| 29.009 69.5 87 | 46 | 29.973
7 30.005 717 31 | 52 | 29.979
0 | 30.004 73.7 7 | 53 | 99.976
11! 29.988 154 74 5.8 29.958
2m 29.975 | 750 72 | 55 | 99,996
1} 29.961 77.5 ri | 5.3 29.906
2| 29.947 778 70 | 59 | 29.908
3 29.058 78.3 19) 5. り | 29,909
4| 29.940 77.7 70 6.2 29.922
5} 205 | ze7 | 72 | eo 29.934
6) 29956 75.1 77 4.7 29.950
7 29.969 735 sl 3.9 29.960
8| 29.986 72.5 84 | 8.6 29.907
9 29.991 718 8 | 81 29.972
10 29.992 71.1 87 | 30 29.975
II! 29.991 70.2 |, 29.972
12 29.987 60,5 FU 3.6 29.963

TABLE U. MEAN BAROMETER, TEMPERATURE, RELATIVE
HUMIDITY AND FORCE OF VAPOR FOR EACH HOUR -

FROM THR HOURLY OBSERVATIONS IN MARCH,
JUNE, SEPTEMBER AND DECEMBER.

Baro. | Temp. Hum. Force of vapor.
1shin th 29.948 hy oe ae 0,383
99942 | 524 Ci) 380
29.937 | 7" AM 人 54
29.9581 1 Bl ea SPE IS
と 29.945 | 507 |. | 372
29954 | 07 | 84 | 32
| | 379
381
388
395
403
410
413
29914 | 631° 1.60.77, de
29.910 3; 417
29912 1 |). Bern Se ce 417
99177 | se 6 | 425
299971 BO! 417
29.938 | 581 | 3 | 414
29.950 | 569 | 410
29.957 | 558 | 406
29.961 | 550 | 30 | 405
| 29.957 | 542 | | 398
| 29.951 | 85 | 82 | 392

na u ce ee

—— or =~

TABLE V. MEAN BAROMETER AND MEAN TEMPERATURE FOR
EVERY DAY FROM HOURLY OBSERVATIONS.

March

30.309
30.209
30.181
30.259
30.361
30.389
30.293
30.055

29.894
29.762
29.958
29.973
29.908
30.024
29.869
29.450
30.164
30.291
29.890
29.918
30.018
30.010
20.849
29.774
20.006
20.836

20,724

June

September

December

29.613
29.703
29.900
29.036
29.961

29.963

29.823
29.892
29.886
29.854
29.942
28.944
29.749
29,622
29,657
29.694
29.774
29.986
50.188
30.230
30.008
29.845
29.782
29.862

29.200

60.6

29.956
30,025
30.062
30.010
29.973
29.948
29.921
29.989
30.006
29.832
29.839
30.016
30.092
30.034

29.894

80.054
30.037
29.952
29.839
29.797
29.726
30.101
30.181
29.962
30.078
30.098
29,981

30.147

69.1

71.1

20,985
29.974
30.081
29,049
29.908
29.999
30.053
29.790
29.470
29,501
29.704
30.061
30.100
29,981
29,783
29.876
30.097
30.034
29.810
29.907
29.893
29.880
29.806
29.916
30.110
30.113
29.984
29,822
30.081

50.327

39.1
40.1

56.7

40.6

44.7

fp {oF
+ ue
wh ab) n( 4c 2
riss
or he AE ate
© tema

\ + 第

Chart NIELS
Showing hourly means of zz を >
a portion of Che year.

10 11

が と |
fie

| 5]

6) Z| 8 9 wm Hn re

/ al gi «sel 7| el 9) w vie 7) 2) 3

ni /| 2 3) F S|

| 2) 3| #| 5| 6| 7) 引 9}

|
|
+ + +

|
a
トー

|
—

1
|
†
1
+

u
5

a Ce)

Chart PY TS es 2 第
均 = = =
Showing hourly means of barometer , 時 年 ni +
& = ar
during a partion of the year ME a
3 9
i

December.

時

|
|

s| drasm

|
|

|
W 72)

|
Furt
|

si Pama) KEN |
|

— st am ーー オー ニー ーー ニー
| | |

まま: ほぼ まじ 2

人 re

ee
a r
m ーーーーーーー

ニー ニニ ーー ニーーーーーーーーーーーーーーーーーーーーーー: ーー ーーー

nn
m - - ー =

{ob te (a
x“ feet
+ KR
2
4

Pisa
werke
Start
SE

‚\ 2| 3| | 5|

S
a
=

Charl V2/9.
Showinghourly means of temp -

|
I

nas
it

erature during aporlion of the
sl <| 7

vb tf {or 8
x 4 (eae
4 時 KK 時

FMA BS
2

oe A
2 shi Rik
Ei

rd: Lemp-

Charl V2/9
erature during a portion of the

OP ives .

|
| 2

|
I

6 Z| 8 9 OH

3 # S|

の im 7

| glz| gl 9

December.

“4

= THA TTA
= = | ーー ゴー
ne = = 4 に に に だ に ai
x ー RK en
<r 7 |
4a | __ ie t ee Ce
* | iz
II

7
u
et:

| | |

23»

Mean of Four Months か
|
|

Charl N20

AY howing hourly meansof, “percentage
of Aumidrty during ap orlzon の レス the

I ar.
shia aa.

sk 8S
EEE.)

March.

EEERLEFFEEERREEEFFEEE
Bae

Erna Jn N
Yta)#+ %
Ba)

RR
TENS

0 BEER

I
1

に

。 Mt lo 8
% +HeR+
yy WOR

+S Gs
oe ER
BB ik bhi ae me
Rv 4

tage
of the

レノ カメ

Charl N?°20
Showing hourly meansof percen
ers hehe

N Ss

ダグ humidity

SMR? Ay oo

Mean of Four Months .

Aes Se. FIRE
Re SHALE AE eee we
iS 806 oN SRRE we
Bere va se wel.) ws
MOR SSH SA SAE
NUREZESELSTAENRER
SASCTTRRLZIEN IE ax
KRHENESMDUDSRAME
ee pee eee

の Hr

rite
yk を
7 -
| |
| <
] a
] 5 こ
ド

y ER a a 『 ‘FN Yo Banner.
im a @ Awe Ne oe on, yr vl ons

ーー
D

un it,

;
h-
7 «
\ ;
of
:
を
6
』 5
u i
2
nn

(toe
aa u

|
\ ni wh) AC = s |
ee Ne を =
|wite® [>|
ya BERR 7
aR oe

|
|

2 70\77 \12
peep

|
L

lelslolvireimni2|3|*r

er aR

March.

fal 2ial|elele|7

Chart V?2I/
Showinghourly means of lhe force
ef vapor dnring と 2 の の と の ん of the year.

WERE

Charl NV. 027
Showing hourly means of the force
er vapor dnrıng と 2 portron of the ‚year.

MAGNETIC EXPERIMENTS.

There is no magnetic apparatus at present in the equipment of the observa-
tory, nor does the physical laboratory of the university possess any of sufficient
delicacy for making accnrate magnetic measurements. During June and July
1880, Mr. Fujisawa, a special student in physics in the University, made a series
of observations with the aid of such appliances as could be obtained, for the
purpose of determining the horizontal component of magnetism at Tokio, with
a view to making a comparison between the result and a similar series upon the
summit of Fujinoyama. Although pretending to no great degree of accuracy the
results are, perhaps, of sufficient interest to justify their publication in this
report. From a report furnished by Mr. Fujisawa the following brief account
of the experiments is taken.

Two magnets were used in the determination. The first (A) was 12 cm.
in length, rectangular in section and weighed 68.83 grammes. The secon] (B)
was 10 em. in length, rectangular in section and weighed 57.58 grammes. In
the determination of the horizontal component of magnetism at Tokio the
ordinary method was pursued, the operations consisting in determining the period
of vibration and the deflection from the magnetic meridian of another magnet
placed at a measured distance. The time of a single vibration was determined
by making a chronographie record of the transits of a point upon the end of
the magnet across the vertical wires of a telescope placed at a distance of about
5 metres from the magnet. In this way the period of vibration of the magnet
could be obtained to a degree of accuracy considerably exceeding that attainable
in the other factors of the determination. The moment of inertia of the mar-
net was determined from its figure, its symmetry being such as to render that
method sufficiently accurate. In the determination of the horizontal component
magnet (A) was used; no deflection measureménts were made with the other.

The following is one series of the results of observations for the period of
this magnet ; the time throughout is in sidereal seconds and the location for this
series was in the physical laboratory of the University at Tokio.

July 1 Time of vibration
10.6202
10.6260
10.6262
10.6206
10.6206
MEAN... 10.6227

These, with other results, were combined with the results of deflection observa-
tions, the latter, owing to the imperfect nature of the apparatus, being more
irregular. The mean of five complete determinations gave for the horizontal
component ;

H = 2953
the units employed being the em, gramme and second.
Another series of three complete determinations gave as a mean ;
Houle
And the mean of all gives—
Hi . 2075
Since these observations were completed I have been fortunate in being able,
to secure the results of two other determinations of the horizontal component
of the earth’s magnetism, made in this vicinity and with appliances which are,
in the main, far more perfect than those which Mr. Fujisawa was obliged to
depend upon. For one of these I am indebted to the kindness of Commander
Sampson U. S. N., who made a series of observations during the past summer
in the yard of the U. 8. Consulate at Yokohama. As determined by Commander
Sampson the value is;
29502,

The other is to be found in a memoir recently published by Herr Otto
Schiitt of the Geological Survey who made a determination in Tokio, also during
the past summer. The value there given is;

H 720742

Tt will thus be seen that the value found by Mr, Fujisawa lies between
these two determinations and it may probably be regarded as not far from the
truth.

In August both of the magnets, (A), and (B), were carried to the summit
of Wujinoyama and were vibrated there by Mr. Fujisawa, the period being deter-
mined in the same manner as in Tokio. It was necessary to carry on the
magnetic experiments in the same small hut on the summit of the mountain
in which the instruments for the gravitation determination were mounted and as
the Chronograph was almost constantly in use in connection with the latter, it
was found to be impossible to make more than two complete series of vibrations
with each magnet. As soon as possible, after the return of the party and the
instruments to Tokio, the magnets were again vibrated in order to ascertain what
change, if any, had occurred in their magnetic condition during their absence.

The following are the mean results at different times and places.

Date Place Time of vibration
Julyali CS Ce oe Toon ..10.628
(A) Anish Dr ee Summit of Fujinoyama....10.903
Ayeust 21.81 22... 10km LGA
(dulyibar ser ee Tokio... N a
(1B) SAUTEED) a-nanteaceeeeoee hoes Summit of Fujinoyama.... 9.827
VAMC USD ZA en OT Oss a tae 232 SSC)

It therefore appears probable that the intensity of magnetization had slightly
diminished in both magnets and that the amount of this diminution was some-
what greater in (B) than in (A). If the mean of the times of vibration before
and after the ascent of the mountain be accepted as the true periods, the results
will appear as follows;

ried U) fa La ia ae ca PR PER 10:635

A i the mountain ..............+..10.903
Me MG) ON の ES eels 9.551

; ee the mountain................ 9.827

The relative intensities of the two magnetic fields being inversely as the
squares of the times of vibration of the same magnet, it follows that the value
of H on the summit of the mountain compared with that at Tokio will be:

HVE (AL) eh

While these two magnets agree well in showing that the horizontal intensity
is about five per cent less on the summit of the mountain than it is in Tokio,
yet it must be remembered that the magnets were vibrated in the same place and
under the same circumstances upon the summit. An examination of the loose
fragments of lava found upon the summit reveals the fact that nearly, if not
quite, all of them are slightly magnetic. It is therefore highly probable that
results differing materially from those given above might have been obtained,
had the location of the vibrating magnet been changed. Unifortunately it was
impossible, in the few days during which the party remained upon the summit, to
make suitable arrangements for repeating the vibrations at different points but
it is hoped that a complete magnetic survey of the mountain may be undertaken

in the near future.
-

THE HEIGHT OF FUJINOYAMA.
BY PROFESSOR W.S. CHAPLIN.

I. The following are the principal measurements which have been made of
the height of Fujinoyama,

4322 metres Alcock
2 3987 Fagan
3 3266 Williams
4 3518 Lépissier
5 3729 Knipping (first calenlation)
6 3329 a (second .、 )
ff 3769 Stewart
S 3702 Fenton
9 り 3768 Favre-Brandt
N 10 3823 Nakamura & Wada.
| TGP 9793 Siebold

II. The measurements from which Mr. Knipping has calculated the height
were made, under his direction, by students of Kaisei Gakko (now Tokio Dai
Gaku). The observations were taken at the tenth station on the mountain, and
at Numadzu, a town on the sea-coast distant about 15 miles in a direction
S. 30°. E. from the station on the mountain. Observations were made five
times daily, at 7 and 10 in the forenoon and at 2, 4 and 9 in the afternoon for
fifteen days, and each included a reading of the barometer, the thermometers
and the psychrometer. For the first seven observations on the mountain no
simultaneous observations were taken at Numadzu, so the observations on the
mountain were compared with simultaneous ones taken at Tokio. Mr. Kuipping
says that the observers were skilful and that their work may be relied on.

Two calculations have been made to find the height from these data. The
result of the first calculation is found on page 7 of the 3d Heft of the
Mittheilungen der Deutschen Gesellschaft fuer Natur und Voelkerkunde Ostas-
iens. In making the calculation Mr. Knipping has used an approximate formula.
proposed by Ruhlmann for use in Central Europe. His results are

7 AM. 10 A.M. 2 P.M. 4 P.M. 9 P.M.
3731 metres 3775 3784 3769 3726

He has then taken a mein of the results at 7 A.M. and 9 P.M.. as advised
in Mueller’s Physics, and fonud 3729 metres as the final result.
The second ealenlation was made with Banernfeind’s formula, using the psy-
| chrometric observations, which had not been used before, and which unfortunately

have not been published. He thus found as a mean of the observations at 10
A.M. and 4 P.M. (which are the best hours for measuring heights at that time of
the year according to Bauernfeind) 3829 metres, and as a mean of those at 7 A.
M. and 9 P.M., 3780 metres. An account of this calculation may be found in
a paper by Mr. Knipping read before the Deutsche Gesellschaft &e. July Sth
1576. As Mr. Knipping gives no details of this second caleulation, I have had
the calenlations made again, using as there were no psychrometrie records the
formula given in Guyot’s tables and there credited to Deleros.
as follows :—

The results are

|
|
|

Date TAM. 110 A. | oe amy, | 2 bm 9 Pin: |
1873 |
July 27 3759 3742 3685 |
28 3680 3742 3764 3730 3086
29 30698 3738 3753 3754 3692 |
30 | 3704 | 3794 | 8759 | 3760 | 3706 |
31 | 3687 | 3725 | 3730 | 3774 | 3700 |
Aug. 1 3690 3743 3759 3123 3684
2 3077 3733 3749 3732 3604
3 3608 3733 3741 3712 3066
4 3601 3740 3742 3711 3690
5 3706 3733 3734 3735 3690 |
6 Tal | 3747 3756 3741 3097 |
か 3644 I 3756 3782 3750 3700
8 3706 3749 3785 3728 3079
り 3702 3747 3708 3727 3660
10 36567) 3724 3717 3718 3648
に cae = 3
| Sum 52715 52404 56203 56048 55281 |
Mean 3765 8743 3758 8786 3085 |

To these must be added the height of the summit of the mountain above
the tenth station, which is taken as 61 metres, and the height of the lower barom-

eter above the sea-level, 10 metres; making the heights

3836

Il. Mr. Stewart measured the height of Fujinoyama in 1874, using an
He started from the summit and measured to the sea-level at

omnimeter.

3814

38:

3807

3756

Numadzu, using 97 stations. Unfortunately he has not published any detailed
account of the measurement. I am indebted for the facts which I give to a paper
by Mr. Knipping in the 11th Heft of the Mittheilungen der Deutschen Gesell-
schaft &e. Mr. Stewart found a result of 12364 ft. (3768 metres) and estimated
his probable error at 20 ft.

From my own examination of another omnimeter I am lead to believe that
Mr. Stewart’s estimate of the probable error is too small; in fact, I believe that
if the instrument were used with ordinary care, the probable error would be
nearer 50 ft. than 20 ft.; yet, in the absence of a detailed account of the measure-
ment, it is only possible to guess at Mr. Stewart's error.

IV. Messrs. K. Nakamura and Y. Wada of the Surveying Bureau accom-
panied Prof. Mendenhall’s party to the summit of Fujinoyama in the summer of
1880, and took barometric observations for determining the height of the mountain.
The lower station was at Hara, at a distance of about 13 miles in a direction
S. 12° E. from the summit, and at a height of 3 ft. above the sea-level. Their
observations were taken hourly from 6 A.M. to 6 P.M. during two days, and
included observations on the humidity of the air. After the completion of the
work, as well as before leaving Tokio, the instruments used were all compared
with standard instruments in Tokio, and the proper corrections for instrumental
error were applied to the observations.

Messrs. Nakamura and Wada have given their observations and the result
of their calculations of the height in the Echo du Japon, October 16th 1880.
They have used Williamson’s formula aad tables, and have found as the mean of
all the results 3825 metres.

Mr. Nojiri has recalenlated the height from data furnished by Messrs. Naka-
mura and Wada, and obtained the following results in feet ;—

6AM ZAM. 8AM. OAM. 104M SLAM ee
12311 12377 12532 12593 12608 12577 12608

IVE 2 Pe 3 PM. 4 P.M. 5 P.M. 6 P.M.
12636 12591 12602 12630 12555 12480

The mean of these numbers is 12515 or 3812 metres. The difference
between this result and that of Messrs. Nakamura and Wada arises from the
fact that in the two calculations different tables for atmospheric humidity have
been used.

V. The other measurements given in the table have been made, some of them
with aneroid barometers, some by comparing the height of the barometric column
on the mountain with the mean height of the barometer at the sea-level; and each
of them depends upon a single observation. They are therefore of little weight
compared with an extensive set of simultaneous observations with mercurial
barometers. x

VI. An examination of the individual results in either Mr. Kmipping's, or
Messrs. Nakamura and Wada’s calculations will show that for an accurate measure-
ment but little reliance can be placed on the barometer, variations of 50 metres
from the average in either of them’not being uncommon.

It has been found from long series of hourly and simultaneous observations,
that the heights caleulated by any of the ordinary formule from these observa-
tions are subject to several fluctuations. Generally, a height determined in sum-
mer is greater than when determined in winter; and greater in the daytime
than in the night. The height thus follows roughly the variations in the baro-
metric column, but not in such a way that we are able to find the time when
the formula will give the mean height, when we know the time at which the
barometric column is at its mean height. As both the sets of observations were
made in the daytime and in summer we can say that the heights determined
from these data are probably too great, but we can not say how much too great.

VII. Mr. Futami of the Surveying Bureau has kindly furnished me with data
from which to calculate the height of the mountain trigonometrically. His ob-
servations of altitude were taken at three of the stations of the Geodetical Survey,
one at Kanozan east of Tokio Bay, another at Amagisan in Idzu, aud the third
at Tanzawayama near Oyama. The instrument used was a theodolite reading
by two verniers to 5”; the. barometer and thermometer were read at every
observation. 1 have used only the observations taken at Kanozan and
Tanzawayama.

The height of Kavozan was found by levelling from the sea-shore to be
355.2 metres; anıl the instrament was 1.6 metres above the ground.

In calculating the height of Tanzawayama and Fujinoyama I have used
Bauernfeind’s formula, which may be found in his Elemente der Vermessungs
kunde Vol. 2, page 281, or in the U. 8. Coast Survey Report for 1870 page 160.

The results were as follows

Tl’. above K. K. below 'T’. I, above K. I. above T.
1222.7 m. 1198.2 2218.
1227.5 1208.8 2217
1218.5 1210.8 3439.6 2217
1225.5 1202.5 S441
1214.2
Sums 6108.4 4820.3 6880.6 6652
Average 1221.7 1205.1 3440.3 2217.

T above K = 1213.4

To get the height of Fujinoyama we then add

Height of Kanozan
っ > ogy: Instrument

F above K

or Height of Kanozan
=e „ Instrument
T above K

F above T

The mean of these results is 3792 metres, which I am convinced is more
nearly correct than any other of the results. It falls between the measurement
of Mr. Stewart and those of Messrs. Nakamura and Wada and Mr. Kuipping.

The height will probably be determined soon by the Surveying Bureau, when

355.2
1.6
3440.3

3797.1 metres;
355.2
1.6

3787.2 metres

we may expect to know it with great exactness.

a6 ated nid

Map showing Pelatıve Pos ction Bm Ben
ne ek e RM +
of Tokvo, Fuji-no-yama, Kano-zan, = i BF

2 Re 泌

A mage-san and Ta NZAWA- YAM a.

METEOROLOGICAL OBSERVATIONS ON FUJINOYAMA.

To Mr. R. Fujisawa, special student in Physics in the University, I am
indebted for a carefully prepared report of the meteorological observations made
by the party visiting the mountain in August 1880, the work of which has been
already referred to in this volume. From this report I have selected the foliow-
ing observations and results which will doubtless be of considerable interest to
those who are more or less familiar with the mountain as well as to all who are
interested in the meteorology of elevated points.

The ascent was made by what is known as the Eastern or Subashiri route.
A monntain Barometer (mercurial) was carried by the party and observations
were made at different points during the ascent. Upon this route there are ten
“stations” or stopping places, the last being at the summit. The following
table gives the height of several of these stations above the sea, as deduced from
the barometric observations. It will be remembered, of course, that they only
pretend to be rough approximations as they are based upon a single observation
at each point, the barometric height at the sea level having been assumed to be
30 inches. They may, however, prove to be interesting and useful to persons
making the ascent of the mountain in the future. The numbers given represent
the vertical height in feet between the various stations at which readings were
made.

BE sce ete Ea Ee aD
DUDA MK epee en 2877
WLOINAPBYERNT occ) sexes ees eucssa tees <ecvae 2167
CORT tees een OLD
Becond BatOOg 1240
Hifth Ha UAE foe 1340
ee LA
Wenth (BUONO 55 oi 1140

A number of hypsometric observations were made during the stay upon the
summit. The instrument used included a Fahrenheit thermometer reading to
212°. A careful test of the thermometer in the physical laboratory of the Uni-
versity did not indicate a greater error than one hundredth of a degree at that

point. Barometric readings were taken at the same time, the results of which
are to be found in the following table. In the fourth column will also be found
the corresponding calculated temperature deduced by interpolation from Reg-
nault’s tables. The barometric readings are given in inches, after having been
reduce to the freezing point. The first observation, that made on the afternoon

of August 3rd., was made with a centigrade thermometer. It will be observed
that there is a nearly constant difference of about four tenths of a degree Fahren-
heit between the observed and calculated temperatures.

Date Barometer Hypsometer Calculated.

August 3 3 p.m. 19.350 88°.4 C 88°.2.C
et lps: 19.363 190°.45 190°.86 eo
ke oe DED 19.360 190°.50 190°.86
Bu OBEN: 19.357 190°.43 190°.85
=A EWN He 8) fale 19.377 190°.45 190°.90
‘a span 19.377 190°.50 190°.90
en el pm, 19.366 190.50 190°.87
a PN, 19.358 190°.43 190°.85.
= Pe se 19.352 1909,40 190°.84 が
っ 7.380 pan. 19.352 190°.40 190°.84

“7, Hmm 19.350 1909.40 1909.83

The following table contains the meteorological observations made during
the ascent and upon the summit.

Barometric es Relativ
Date Pressure Pemperature velutive Weather Station
reduced to 32°F (I) Humidity %

August 1880

Ist

6" 307 P.M. 27.310 Subashivi
7" 30’ P.M. 27.305

9» 07 PM. 27.303 ”
2nd

5° 0/ P.M ーー 一 fair | Subashiri
65 307 A.M. _ 一 Mumagayeshi
Sh 307 A.M. 23.880 71 78 cloudy Chiujikiba
10° 8 A.M. 22.849 64.8 80 rainy 2nd Station
11° 40’ A.M. 21.755 61.5 84 cloudy 5th Station
P.M. 20.177 44.9 85 clear Sth Station

3rd

Sun rise 42.0 88 clear 8th Station
7" 30 A.M. 19.303 41.6 95 N er
9 Of A.M. 19,342 83 fair Tent
12> 0/ 19.350 61.5 87 fair “9

1? 45’ P.M. 19.342 57.8 こき: fiir 内

2" 40/ P.M. 19.204 ーー ご fair Kengamine
6" X P.M. 19.341 50.4 66 fine Tent
ge 0’ P.M. 19.373 =+ fine “

4th

5" AM. 19,351 38 = fine Tent

6° A.M 19.351 11.0 98 ” ”

7° A.M. 19.372 44.0 81 3 ”

8° A.M 19.370 16,0 sl ” ”

9° A.M. 19,375 51.2 60 ” „

10" A.M. 19,568 52,5 83 DD ”

11" A.M. 19.867 55,2 61 ry Ri

12" 10.504 55.6 57 „ „

15 P.M. 19.363 57.1 51 ” 5

2 P.M: 19.361 62.6 41 „ „

8 P.M. 10.360 57.6 51 = 5

Barometric ay A a
Date Pressure Temperature Relative Weather Station
reduced to 32°F (F) Humidity &

August 1880

4th
4” PM. 19.355 62.6 40 NE ‘Tent
windy
SY ESM: 19.357 50.2 64 i A
6" P.M. 19.371 45.6 64 5 お
hth
6" A.M. 19.377 37.3 76 nn: Tent
windy
hal 19.363 89.8 71 ” Temple
8" 19.377 89.9 91 a ag
las 19.379 10.4 89 fair "a
10° |, 19.366 41.9 89 fine
11" 19.364 41.1 90 a x
12% 19.360 41.8 90 2 7
TTL Ei fe 19.366 45.1 sl 92 Be
yh 19.358 15.1 83 3
Bu, 19.358 46.3 84 * 3
NN he 19.352 45.0 85 is ”
= tae たこ 1 fine
ee 19.351 44.5 85 windy =H
Chay 19.352 42.8 88 x ee
zu 9.359 2.6 88 fair heavy
UM し / 19.86 12.6 wind 2
Bun. 19.358 40.7 86 fair
) windy
P に fair heavy
gh 9.357 39,9 8 ar heavy
” 19-307 4 wind 22
LOH 19.353 39.1 84 A) a

A minimum thermometer was exposed during the stay upon the summit

with the following results ;—

Nicht ofssrd and NE sns 31°.4
» 。 っ 4th and 5th woe 3202
a5 yy WHAM AOU ra re Bl)

A maximum thermometer was exposed to the sun on the fourth and fifth
and the temperatures reached were as follows ;—

ーー

On the afternoon of the 6th Mr. Fujisawa and Mr. Tanakadate descended
to the bottom of the crater and made three barometric readings at different
points which were as follows ;—

Nonikein Perle... tee 19.773
Southiertin ss, cssetyccisitnescecvsd ren 19.758
er 人 19.777

As before stated, Messrs. Nakamura and Wada accompanied the party to
Fujinoyama and maintained, during two or three days a series of simultaneons
sarometric and Hygrometric observations upon the summit and at the sea-level.
These observations, reduced and corrected, they have kindly furnished for pub-
lication in this report in connection with those made by others of the party.
They are given in the following tables.

BAROMETRIC AND HYGROMETRIC OBSERVATIONS MADE ON
THE SUMMIT OF FUJINOYAMA AND AT THE SEA-

LEVEL BY MESSRS. NAKAMURA AND WADA.

August

1880.

Barometer.

Reading

(English inches).

Attached
thermometer
(Fah. degrees).

Corrected for
index error, capil-
lary action and to

Freezing point.

ここ ここ ここ ここ ここ ここ

ここ ここ ここ ここ ここ ここ ここ

4
4
4
1
4
1
1
1
4
1
1
1
4
5
5
5
5
5
uv
5
5
5
5
5
>
5

9
10
11

9
9
o

Hour.

) AM
7AM

AM
AM

Noon

2 PM
> PM

5 PM
) PM

Ss Al

AM
AM
AM

Noon

PM
PM
PM
PM
PM

) PM

Lower

Station. |

29.957
965
973
989
976
965
965
957
949
952
61
969
905

00
035
035
047
051
O54
057
043
031
043
033
024
028
031
047
075
071
075
071
07 1
067
O45

29.996
30,016
028
036
034
030
031
022
008
010
008
006
006

Upper
Station.

Lower

Station.

18.2
79.
81.
82.1
82.
83.8
86.

ゥ ら ro つら ゅ ふ ー ら

こう くう 5 Ts

>» Do

5 c5
RE.

ゥ ら つ om

ho lo bo mm oc

Upper
Station.

Lower

Station.

29.854
857
860
866
S60
846
R41
833
822
823
821
837
852
916
932
932
939
40
938
933
916
904
914
912
905
O15
952
960
967
955
956
44
939
30
930

Upper
Station.

29 907
916
920
920

BAROMETRIC AND HYGROMETRIC OBSERVATIONS MADE ON
THE SUMMIT OF FUJINOYAMA AND AT TIIE SEA-
LEVEL BY MESSRS. NAKAMURA AND WADA.

Hysrometer.

August Ba
Dry bulb Wet bulb Temperature
ete thermometer thermometer of dew point
1880. (Pah. degrees). (Pah, degrees). (Fah. degrees).
pie - Lower Upper Lower Upper Lower Upper
Day. Hour. Station. Sidton. Station. Are ei Station. | Station.
3 6h AM 78.8 a 76.9 m 75.6 PR
3 7 AM 79.6 hs 77.0 en 75.1 fp
3 8 AM 80.4 2 eu + 74.6 PP
3 9 AM 81.8 4 17.2 35 74.0 ri
3 10 AM 82.7 a 77.5 時 73.9 -
3 11 AM 83.2 A 78.0 ER 74.4 i
3 Noon 84.9 61.5 790 9.5 74.9 58.5
3 1 PM 85.4 er 79.4 5 75.2 4
3 2 PM 85.9 4 79.7 ” 75.4 =
3 3 PM 87.5 en 81.0 = 77.1 お
3 4 PM 87.9 = 81.5 ne 76.7 i
3 5 PM 84.8 時 79.7 „ 76.1
3 6 PM 82.6 50.4 78.8 44.5 76.1
4 6 AM 72.0 42.0 70.8 41.0 79.1
4 7 AM 76.9 44.0 73.0 41.1 70.3
4 S AM 79.0 46.0 74.8 43.1 71.9
1 9 AM 80.2 51.1 76.0 45.6 73.1
4 10 AM 81.7 51.5 75.5 48.5 (OIE?
1 11 AM 83.9 55.1 76.7 47.7 13,
4 Noon 86.0 55.5 77.8 47.1 73.1
4 1 PM 86.5 57.0 78.2 47.1 78.2
4 2 PM 86.0 62.5 79.9 49,2 75.6
4 3 PM 88,0 57.5 81.2 7.5 Tiel
1 1 PM 85.2 62.5 80, 49.0 77.8
4 5 PM 84.5 50.1 80.5 44.0 met
4 | 6 PM 82.3 45.5 79.0 41.2 76.7
5 6 AM 74.2 87.2 72.9 345 72.0
5 7 AM 75.1 39.2 73.3 35.5 72.0
5 S AM » 79.2 39.8 75.5 38.6 72.9
> 9 AM 83.4 40.3 79.5 88.8 76.8
5 10 AM 85.0 40.8 79.5 39.8 75.7
5 11 AM 87.6 41.2 79.7 39.8 75.0
5 Noon 89.2 41.7 81.9 40.2 77.5
5 1 PM 91.5 15.0 838.1 42.0 79.7
5 2 PM 78.1 45,0 77.0 42.4 76.2
5 | 8PM 81.2 46.3 78.1 13.8 75.9
5 4 PM 83.9 45,0 80,0 42.7 717.5
5 5 PM 86.5 41.4 81.0 12.1 79.1
5 6 PM 83.8 42.8 S14 11.0 79.7
6 AM 75.0 39.6 73.4 87.8 7 86.2
7 AM 17.2 11.0 74.4 38.3 7 84.9
Ss AM 79.5 42.9 75.8 10,4 73. 39.1
り AM 81.8 | 45.7 77.6 12.2 74,6 39,8
= 10 AM 83.1 | 46.2 77.5 13.0 78.6 42.2
ー 11 AM 84.0 | 18.2 78.1 13.8 73.9 10,2
の Noon 80.7 18,0 719,6 413.5 75.5 $9.6
= 1 PM 87.8 51,0 80.2 44.6 i. 39.2
2 PM 83.3 3,8 78.8 15.8 75.7 30.4
» PM ROG 10 80.1 45.7 16.7 40.6
ı PM ROG 53.8 80.8 5.0 17.2 20.5
5 PM 85.3 47.3 80.7 48.1 17.6 vo.
6PM 82.9 11.2 79.7 41.1 77.5 18.5

ü ws
er
*
u
ou ・
1
u
im
_
1
『
j
fi
e. i
a”
1
u
ーー
U ー
4 7 f 『
‘
ET
H
y
Hf u
f

rps TES meu aa

~
’
』
A
r +8
1 ei
ue ty
weet |
. も
ee,
2
’ F
hi
x 2
i
J 7
$i 4 TER |
oj
of
A ’ j
7 ree
m a |

ie a wu ー aoe aie

(iw Si yeas
の au if Abn

BAe 【
je N

‘ Um ま
i 1 4

/
本 か

ーー

FIRES IN TOKIO.

BY PROFESSOR K. YAMAGAWA,

In the City of Tokio, fires are so frequent and destructive, that they
doubtless involve greater losses in the shape of property destroyed than all other
causes combined. To one who has not resided in Tokio, it is difficult to
conceive their mignitule and the devastation caused by them. It is not
strange that residents of foreign countries, who have never been in Japan, can
hardly eredit the accounts of these conflagrations which are published; for in
their own countries, houses are, in general, so much more substantial in their
character, that they are not only less combustible, but the progress of a fire is
thus made slow enough to render the task of extinction comparatively easy.
With us, however, houses are generally wooden, and on windy days, when a fire
is once started it is likely to grow, within a short space of time, to one of
extensive magnitude and even with far better means of extinguishing fires, than
we have at present, it would be almost impossible to arrest it until it dies away
of itself on reaching the limits of the city, or by reason of the entire cessation
of the wind. It is said, that in Tokio, the average age of a house is seven years,
that is to say, the whole city is burned every seven years. The assertion is no
doubt exagerated as regards to the greater portion of the city, but if we take the
most populous parts such as Kanda and Nihonbashi wards, the above assertion
is very approximately true. During the reign of the Tokugawa Shoguns, Tokio,
then called Yedo, with double the present population, was the emporium of
wealth of the whole of Japan; and any destruction of property was made good
by the rest of the country, as fast as it occurred, so that not only were fires less
dreaded than would be expected, but in some cases they were rather desired
than otherwise. Paradoxical as it seems, conflagrations used to be called the
“ornaments of the city.” I cannot go into details to explain the complicated
system of Fendalism that bore such strange fruit as this; suffice it to say that
in old times the population was much larger than at present, and the people had
less fear of fires, so that they occurred much more frequently before the late
revolution, one of the results of which was the destruction of the feudal system,
than at present. Below will be found a table of areas of houses burned each
year since the revolution in 1868, The column of areas gives the space actually
oceupied by houses burned during each year and estimated in tsubo, one tsubo being
about 6 feet'square. The total area occupied by houses in Tokio at the present
time, is 3,136,858: tsubo, so that the mean in the table is about 14 % of the total
area occupied by houses, and at this rate it will be about 65 years before the
whole city is burned once. This, however, is a calculation based on the
statistics of the period known to have been more free from fires than any
other, so that this rate may or may not hold in the future. On the one hand
the population is increasing very fast, which fact will make the probability of

fires greater, and on the other hand, the character

Date. Area. R F 2 :
of the building is being improved, and the system of
1868 19,33 fire-brigades is becoming more eflicient. Whatever
1869 40,936 p <
1870 41,577 may be the frequency and extent of fires in future,
th “pared at the present rate, the whole city is burned down
Ole od Ud に = =
1873 73,230 every 66 years, and as this is based upon the experience
1874 12,104 2 : : : ays
1875 20.824 of the last thirteen years, a period during which fires
re gral were less frequent than at any previous time, I think
od vl こん . . . . に
1878 57,855 it is not at all exagerating to say that in former times,
1879 103,213 と a ans
1880 68.684 when fires were much more frequent, the districts in

Fo eee Tokio, where houses are built very closely together, and
otal, 577,067

= = に x in which fires are very frequent, the average age of
Mean. 52,082

a house was about 7 years. The last winter was
remarkable in the frequency and extent of fires. If we take the last four fires

| No. of houses

Date. buried Sh eras exhibited in the subjoined table, the
a total number of houses (a house or number
1880—Dee. 30 2200 of houses or a part of a house occupied by
881— Jan. 26 550) ・ tt .
1881 Tan,.26: | Rune a single family is counted as one house)
„Feb, 11 | 7100 5
a 51 1400 burnt by these fires, of which only one can
ーー 02 & .
be called a great fire according to the
Total. 26200 と >
Rotel meaning of the term used further on, is

26,200. The total number of houses in Tokio is something over 285,000 so
that these four fires alone destroyed about one eleventh part of the city in about two
months of the fire season. This will show that the assertion concerning the average
age of a house in the most populous parts of the city is by no means impossible.

The City of Tokio, formerly called Yedo, was founded by the first Shogun
(the military ruler of Japan during the Feudal period) of the Tokugawa
dynasty in 1590. The site of the city was a mere marshy plain with fishing
villages scattered here and there, but being made the capital of the Shogun
Government, it grew in size very rapidly, and in a few years, became a city
of considerable size. It early became a victim of the flames for in less than 11 years
after its foundation, namely in 1601, the record handed down says that the city
of Yedo was burned entirely. | Eighteen years after another great fire is recorded ;
and twenty-six years after the latter, the part of the city, where at present the
Ginza street and others are found, was burnt, the fire spreading in a southern
direction, burning some eight thousand houses. ‘These three fires are the only
ones recorded; the details, however are not known; hence they are not included
in the following table. Table A contains nearly all fires of considerable
magnitude that have occurred during the last 224 years, that is from 1657 to
1881.* From various records, written and printed, I have selected those fires,
whose length is greater than 15 cho (one cho is about 36U feet) starting from

of considerable extent during the last win- ~

the origin of fire to the farthest point in the burnt district, and the length is
measured by an ordinary scale on a good map of Tokio published by the
Surveying Bureau of the Interior Dept., a process sufficiently accurate for this
kind of work. I have obtained the direction of fires by measuring with a
protractor the inclinations of straight lines to eight points of the compass; when the
lines fall between two points, I take the nearest as the directions of the lines,
and consequently the direction of fires. Table A gives in the first column the
number, in the second the date,f in the third column, the directions measured as
stated above, and the last column contains the magnitudes. In the accompany-
ing map of Tokio, the straight lines used to determine the directions, and
magnitudes are plotted, a number is attached to each so that it may be referred
to Table A. A glance on the map will show to any one, that the greater
number of fires travel from the directions N and NW;; in fact two-thirds of the
fires are from these two directions. Another point noticeable in the map is, that
- the lines seem to converge to the south-east corner of the city which is a neces-
sary consequence of the local arrangement of the city together with the N and
NW winds. Near the point, where the fire No. 17 begins, the moat is crossed
by a bridge continuing the main street running from north to south. This
bridge is called Nihonbashi, the “Japan bridge”, whence all distances in the
empire are taken. The region near the Nihonbashi, especially the part to the
north of the bridge, is the most populous portion of the city and most frequently
visited by fires, as is seen by the number of lines crossing each other in this
district. The longest line in the map is No. 56, and by referring to Table A, we
see that its length is 133 cho or about nine miles: this is perhaps the longest fire
in the record. It occurred on April Ist 1772; the day was extremely windy,
“ and at about noon, a fire originated in a temple in the south western limit of the
city. It spread so rapidly that in less than 24 hours, it reached the northern
limit of the city. While the fire was still raging, the wind shifted to W, and
all that part of the city included between the river and the straight line in the map
was almost entirely destroyed. The following may be regarded as the six most
extensive fires that have raged in the city since its foundation; Nos. 1, 2 and 3
forming one group, Nos. 6, 7, 8 and 9 forming another, No. 15, No. 42, No. 56
and No. 69. In the first two instances, fires of considerable extent occur-
ring very near to each other in time, they have been grouped together and
regarded as one fire, although they are, in fact, entirely independent. In each
of these, more than two-thirds of the city was burned. The first group is justly
regarded as the most extensive. On March 2nd 1657, at about 2 P.M. a fire oe-
curred in the north-west part of the city. The velocity of the wind was enormous ;

* The magnitudes of fires are hest measured by the areas of burnt district; but not having
sufficient data further, I have taken the length as a measure.

+ Months and dayd are reduced from lunar calendar to solar calendar for obvious reasons.

judging from the records and descriptions of the fire, I am inclined to think,
that during most of the time of the fire it must have been more than 60 miles
per hour. It travelled towards the south-east, and in less than 24 hours, it
reached the sea in the eastern side of the river; next day, while this fire was still
doing its work of destruction, another broke out at a point, about half a mile
south-west of the origin of the first. The wind was still blowing with
unabated fury, and after reducing the palace of the Shogun in the castle, and a
countless number of magnificent houses belonging to the Feudal princes, to
ashes, the fire ceased, having reached the sea. But before this terminated,
another had begun about a mile south of the origin of the second, and it also
reached the sea in the south-eastern part of the city. The number of persons
killed by being burnt to death, drowned, crushed, etc. seems to have been some-
thing enormous; old chroniclers delight in putting down this number as 107,046.,
and whatever doubt we may have as to the numerical accuracy of this statement,
it is certain that the number was great; the bodies being so numerous that
identification was impossible, and all were carried to a large excavation made

near the Rio-goku-bridge, and buried there. The temple Yekö-in, where the

great wrestling matches are now held twice a year, was built then near the grave
of so many unfortunates for the benefit of their souls. The second group of fires
took place on March 13, 1668 originating in the western part of the city, and termi-
nating in the sea. While the fire was still burning another broke out, although
the latter was small the length not being greater than 21 cho. Three days
afterward, a fire broke out in the western part of the city, and destroyed a
district of the city 43 cho lone. and two days after the latter, another occurred,
so that in this group, although each one is not very extensive, the four together
covered a great area. The third great fire took place on Oct. 9 1698 at about
10 in the morning, near the present site of the railway station and extended to
the northern limit of the city. The fourth began in the south-western part and
nearly reached the northern limit of the city ; the fifth has been referred to above,
and the sixth originating near the southern limit of the city reached the north
east limit, taking place at about noon Apr. 22, 1806. This destroyed the best
half of the city, and it is said, that more than 12 hundred people were killed in
one way and another. Of these six, it must be noticed that three of them are
from the NW and the other three are from the Sand SW. In fact, these seem
to be the directions to which the high winds are confined (by a high wind is here
meant a wind whose velocity is greater than 20 miles per hour) and consequently
the directions of great fires.

In Table B, all the fires, that I have been able to study, 93 in all, occurring
during the last two hundred twenty four years, and each not shorter than 15 cho,
are tabulated with reference to time and direction, so as to show the number of
fires in a given direction in different months. In table C, will be found the total
sum of the lengtlıs of fires that occurred in different months from various direc-
tions, Looking at these tables, with reference to time we notice a remarkable

absence of fires during the months of July and August, but from September,
they increase in number growing greater and greater until a maximum is reached
in the month of March ; then the number gradually decreases till it becomes zero
in July. This fact is more strikingly shown in Chart A in which the upper
broken line indicates the number of fires in different months, the scale of number
of fires being on the upper half of the left edge of the chart. The chart tells its
own story too well to need any comment. The lower broken line represents the
number of miles of wind during each month, the scale of the number of miles
being written on the right lower edge of the chart. The number of miles is the
average from observations during the years 1879 and 1880, so that the broken
line or the curve may not exactly represent the true average; I am, however, of
opinion, that it is sufficiently near the true average for our purpose. In comparing
these two curves, one is struck with a remarkable general similarity between the
two, showing the obvious truth, that the wind is the main cause of great fires.
Tu both these curves, a maximum occurs in the month of March. The fact that
March is the windiest month in year, no doubt, makes the frequency of fires
during that month the greatest. One important fact noticeable in the chart, is
the remarkable difference between the two curves as regards the month of July,
which can be explained by the following consideration : although winds are more
frequent in July than in the adjacent months, as we have little need of fires
during the hot summer months, the danger of a fire originating accidentally is
greatly lessened and the probability of a great fire at this time is proportionately
diminished. This is undoubtedly the cause of the difference between the two
curves. The curves in Chart B, are very similar to those in Chart A ; except that
the lower curve is drawn so as to show the total number of high winds during
the years 1879 & 1880 (in the chart, the heading ought to read “ total number
during 1879 & 1880” instead of “mean of 1879 & 1880)”. In their general
characteristics the two curves agree remarkably well except in the month of
August, which can be explained by the same consideration as before. As an
additional explanation of the absence of great fires during July, the lower curve
in this chart shows a great depression for July indicating that there is a very
small number of high winds during this month, and consequently a smaller pro-
bability of the occurrence of great fires. In Chart E, the horizontal straight lines
are drawn so as to show the relative magnitudes of all the fires in Table A; in
each, a line representing a fire is drawn so as to fall on a particular point in the
month, representing the day of occurrence and here again lines are crowded
together in the month of March. Charts C and Dare very similar to Charts A
and B: but instead of being referred to time they are referred to directions.
With reference to directions, in Tables B and C, we sce at a glance, that nearly
all the fires have the directions N, NW, S and SW; and that those from the N
and NW are the greatest in number. Referring to Chart C, we begin at E with
no fire, one with SE, and S and SW have respectively 10 and 15. The
curve falls to zero at Wi then rises cnormously at NW and attains a maxi-

mum at N, and then again fallsat NE. The lower curve is a wind curve drawn
by taking mean numbers of miles of wind during 1879 and 1880 from each
direction. The general characters of the two curves are very similar, with a
difference at the SW. In Chart D, the upper curve is the same as in Chart 6
and the wind curve is drawn so as to show the mean number of high winds from
various direction, the mean of 1879 and 1880 being taken. Again the two curves
show a striking similarity with the same difference as in Chart ©. In all the
charts, it must be noticed, that when there is any difference between the fire and
wind curves, it always oceurs on the hot months, and it is believed that a suffi-
cient explanation for this difference has already been given. Now the tables and
charts discussed above establish beyond dispute: Ist the season of great fires
begins in November, and after attaining a maximum frequency on March, ends
on May, comprising a period of seven months, and 2nd the most fires move either
from the N, or NW, or from the S or SW; and 3rd these results are not mere
accidents, but are the necessary consequences of the local meteorology of Tokio.
I will conclude by referring briefly to the causes of fires, and to the loss of
property which they occasion. The origin of fires may be traced to various causes
but among them are two which seem to me of greater importance than others.
The first is the use of kerosine oil, one of new introduction. This being an
article but recently introduced the people are not accustomed to handle it, and
the result is such carelessness, that undoubtedly a great many fires originate in
this way. A number of determinations of the “flashing point” of samples of oil sold
in Tokio have been made in the Physical Laboratory of the University. In nearly
all, the temperature of flashing point was far lower than it ought to be. In some
samples, it was as low as 96° F., and in none of them, was it higher than 120° F.,
the legal minimum in Great Britain, and in some States of the American Union.
For the so called “safety oil”, the temperature of the flashing point was
considerably higher ; in fact, in all samples examined, the temperature was above
120° F. It is evident that the attention of the authorities ought to be called to
this point and proper restrictions placed upon its sale. The second important cause
is incendiarism. There is much difficulty in obtaining an accurate estimate of the
proportion of fires caused by incendiarism. The owner of the house, where a fire
has originated, always shows a reluctance to give a true account of its origin; he
tries to attribute it to a criminal design, whenever he can. Light as is the fine
to an unwilling agent of others’ losses, the moral effect on him is strong enongh
to make him conceal an accident, that happened in his own house. Hence often
the police authority are unable to discover the real cause and the fire is classed in
the catalogue as of unknown origin. But even allowing for this, fires of unknown
origin are so numerous as to lead us to suspect that a great many of them have
originated in a criminal design. The past season has been unsual in the number
of incendiary fires, the rise of prices of necessaries doubtless having driven many
to commit this crime and the police anthority have done their utmost in the way
of arresting criminals; next year will show, whether they have succeded or not.

According to the Fire Insurance Commission of the. Treasury (through whose
kindness, I obtained the table giving the areas of burnt houses for last 13 years)
the average value of one tsubo (about six feet square) of a house is about 30 yen
(a gold yen=about a dollar) and for the last 13 years, an average of 52,000
tsubo of houses has been burnt every year making the total annual loss
1,560,000 yen. This is an average of a period known to have been exception-
ally free from fires ; and besides, this includes nothing but houses. If we were to
include all items of loss direct and indirect, it can not be less than 3.000,000 yen
every year. The loss of houses, and other property by fire is a loss alsolute in
all senses, a loss not only to the city, but to the general community. If we can
absolutely prevent fires by spending 3 million yen every year, it would be a gain to
the country; but I think, a far smaller amount would be quite sufficient to
~ organize a system of fire brigades, that would absolutely prevent any fires of
great extent. -If the people of Tokio were. willing to spend 5 of the total
annual loss at present sustained it would be more than enough for the purpose.

Finally, an estimate of the number of times the city has been burnt
completely during the last 224 years, based upon the foregoing facts, may be of in-
terest. We have the four great fires during the last 13 year, the total length of
which is 8@cho. During this period, the area of houses destroyed by all fires is
677,067 tsubo. The average total area of houses during the last 224 years must have
been about double the present area or about 6,000,000 tsulo, andthe total length
of 93 great fires is 3234 co: so that we may have the proportion :—S6 : 677,067
::3,234:26,623,659. Dividing this. total area of the burnt houses during the
last 224 years, by 6,000,000 tsubo, we get the number required to be 4.4
that is, the city has been burnt every 50 years, which" is not very far from the
estimate, made in the beginning of this paper. は

「

TABLE A.

Month,

Direction,

Magnitude.

1660
1661
1668

79
TABLE A.

Month, Day. Direction. |Magnitude

TABLE A.

Month.

Direction.

Magnitude.

TABLE B.

January
February
March
April
May
June
July

August

September

October
November
December

Total

January
February
March
April
May
June
July

August

oo oOo らら oC 9,918 So

September

October
November

December 105

DD m 9 & ©

+ー
aa らら oF oF Oo "= aS oF: oO

つど oOo So
a © WW Oo De © W mm S ©

—
=
ュー

TOTAL

99 coc ⑤ 59 ら S&S © ©

=
>

0 0 0 0

Total 0 | 1276| 934] 0 | sg| 420] 15 | 3234

Note,—In charts © and D, the totals (number of fires from various directions) are put down

slightly diferent from those in the table, by mistake.
of the curves, the curves are not corrected,

As it does not change the general character

„>
2

ず

ww pst: ve も

“
.

=:
« : I
>>

+"=
Hk | wl xb or &
IHRE Hu

u =|

ar Ba a
レー AS ーー ーー
ih > BP S|

+B Hpi a

a 1.2, Gassen ma Bas

EAN

中
m FRR

へ
月

月

NE
ort uf sha)

ren

of fires

wind in
) .
A Maya Juney J

ng thelast two “hund-
2% Thenumberof miles of
eachmonth (meanof 7879 ANA 1880
April

T° thetolal number

Feb.

—————

Chart A.
ing

_—

Showt
ineach month durt.
Jana

red yEars

Em,

DJ
22009

hundred years

Oe an dan ee ee re ee ee lll

er
#

ーー ーー

all

8 KL
ke es A+ de
geh
SS nit)
ah = BF Ap
+B Mus
Ren
Hoe + de © +
som
> Brest
“ Vf wh a en"
~ork

ご
of mules of wend
onsfmeanof 1379arı

2°* thenumber
if fern tdirecté

wo
indif,
1580)

wo the fast

nb dtreclions durtı

‘Charl LC:

Showin の LX thelotal numbers f fires
ndred years.

ど /

in
/

indi

Chart D.

-
Showing 1 “thelotal number of eT H bee
Sues indifferent directions during Peg eeaote a to
thelast wohundred years. ey 本 っ =
2"*thenumberof highwinds +m 方 Ty Ae eK

_ indifferenldirectonsdnring 1879 2% MASER tk
ane 1880. Fr Ree AK

EA SES Sh SWE We NWS ん ル NER

36} — ーーー —-

4 | a
\
32 | 2 > : : 1

————_ ee

Chart E.

Showing the relalive
Sires which occurred on

of year.

a neludes of
iiferend days

m
dı

ete

本 ku。』

100

.
っ も 電
Zr
A
[4
1
> eo lat
1 a
Vai 1 4

tludesof
rent days

72
e

don dt

therelalivem

rae

phaw ©

tik

f wae

Charl E.
Jires which occurre

of your

Showing

und

Hat A

N
本

222

\
\

)

ーーーーー ーー SA
| | \ N
fx] DA
| /

So spnpusou
UIZIGIYL

ZZ の 72/ 2929 Up SUDA
7

“sapah ,
SIL Mi

ZZ222277 213
WHOL AOALLO AHL AO dEN

PUD Uo pup 247
BAB HOY YG EMI KH AL BY

nr | : N
ッ レー co - 8
er 6 j
4 : / x A
0 5
ぃ
are

a
Er

も
“
‘ibd
3 .
2
1
・
5
|
hyn
s .
,
t
i
'
*
.. m
— ーー
* Pc
で '
as
Pk Muse x
= .
oe a に oa >

MEMOIRS

OF THE

SCIENCE DEPARTMENT,
TOKIO DAIGAKU.

(University of Tokio.)

No. 8.
ee - THE
3 WAVE-LENGTHS
OF

SOME OF THE PRINCIPAL FRAUNHOFER LINES.

OF THE

6 SOLAR SPECTRUM.

T. ©. MENDENHALL, Pu. D.

Proressor or ExPERIMENTAL Puysics IN Tokro DArGAkr。

Du PUBLISHED BY TOKIO DAIGAKU.
TOKIO:
2541 (1881.)

a 5“

Se as

; sj et Der eas
“ sh 4 tr ?

し MA

VP A MEMOIRS is ar
SCIENCE. DEPARTMENT, ©.
TOKIO DAIGAKU, ui yo

(University of Tokio. See el
- No. 8.

— engen

THE

WAVE-LENGTHS

FH a. rr OF : er

の SOME OF THE PRINCIPAL FRAUNHOFER LINES

SOLAR SPECTRUM.
a

T.-0. MENDENHALL, Pu. D

Paorkeson or Farermestir Paysios 1x Tonto DAioAwe。

PUBLISHED BY TOKIO DAIGAKU

TOKIO:

2541 (1881.)

MEMOIRS

OF THE

SCIENCE DEPARTMENT,
TOKIO DAIGAKU.

(University of Tokio.) |
No. 8. ) |

THE

WAVE-LENGTHS

OF
SOME OF THE PRINCIPAL FRAUNHOFER LINES
OF THE

SOLAR SPECTRUM.

T. C. MENDENHALL, Pu. D.

Proressor OF ExPERIMENTAL Puysics IN Tokio Daicaku,

PUBLISHED BY TOKIO DAIGAKU.
TOKIO:
2541 (1881.)

;
x
|

i 8
i
|
|
|
1

ーー

i
{
N =
0
|
N
3 2 x.
\ 」 『
| ’
2 J os Poe ら —

THE
WAVE-LENGTHS
OF
SOME OF THE PRINCIPAL FRAUNHOFER LINES.
OF THI:

SOLAR SPECTRUM.

The following pages contain the results of a series of wave-length measure-
ments male during the months of November and December 1880. It may be
well, in the beginning, to explain why it was considered desirable that these
measurements should be made and the results published.

The length of a light wave of given color has for several years ranked among
the most accurately determined physical constants. In 1868 Angstrém published
his “ Recherches sur le Spectre Solaire” which contains the results of his elaborate
determinations of the wave-lengths of many hundreds of the dark lines. This
work. involving as it did great labor both in observation and reduction, must
always stand among the most perfect of its kind and it is justly referred to by a
well known writer as “characterized by such accuracy and completeness as to
render it worthy of the highest admiration, to be regarded as a pattern to all
investigators.” Indeed if these results could be considered strictly reliable
further work in this direction wonll be almost unnecessary, as there are only a few
spectroscopic researches in which a greater degree of accuracy would be demanded.

In addition to the work of Angstrém numerous contributions of wave-length
measurements have been made from time to time by other physicists. As
regards measurements of the dark lines of the solar spectrum, the principal of
these, as far as Lam aware, are by Ditscheiner, van der Willigen and Mascart.
Many of these measurements have been made since the publication of Angstrém’s
work and, apparently. under equally favorable circumstances. But every student
of spectroscopy is aware of the fact that these measurements differ very con-
siderably among themselves. One only need examine such a table as is given in
the introduction to “Watts’s Index of Spectra,” in which various measurements of
the same line are placed side by side, to be convinced that the lack of agreement
is so great, in many instances, as to throw all of the results seriously in doubt.

Angstrém, in his memoir, compares the results obtained by Ditscheiner, van der
Willigen and Mascart with his own, but Ditscheiner published a revised series at
a later period which differ still more from those of Angstrém.

っ
Z

In order that the extent of these differences may be seen I give below a table
of differences in the wave-lengths of the principal lines, the numbers being
obtained by subtracting Angstrém’s wave-length from that of each of the others,
the unit here and throughout this paper being the generally accepted tenth-
metre. Angstrém’s numbers are his final “definitive values” given in his memoir.
I am not able to refer directly to the original memoirs of either of the others and
in thus quoting at second-hand some errors may be committed. Ditscheiner’s
numbers are taken froma table given in the American Journal of Science for
April 1872 which was copied from his paper by Professor Gibbs and gives the recom-
puted values. The values credited to van der Willigen are from the same source
and those of Mascart are from Angström’s memoir, The numbers under D belong
to Ditscheiner and those under V. W. and M. to van der Willigen and Mascart.

D. VW. M.

B ihe 4.22 —0.5
C 0.2 3.47 — 1.4
D, 2.27 3.31 — 0.83
D, 1.98 3.15 — 1.12
E 1.81 2.44 — 1.69
b, = ih
b, 1.84 3.06

F 1.46 3.26 — 0.94
G Bp 4.12 + 0.30

Of these results, those of Mascart agree most nearly with Angstrém’s tables,
but even here the difference amounts to a maximum of nearly two units in the
case of E. It will be seen that not only do these authorities differ from
Angström but they also differ from each other. In the measurement of wave-
length the most difficult thing to determine is the value of the grating space.
An error in its value will not, however, influence the ratio of the wave-lengths.
From the nature of the differences exhibited in the above table it is clear that they
cannot be accounted for on the supposition of error in the grating space alone,
but that among them all, including, of course, those of Angström. there must
somewhere exist serious errors in angular measurement and doubtless errors in
space determinations beside. When these differences amount, as they do among
these observers, to scveral units, they are sufficient to influence materially much
spectroscopic work in which reference is made to absolute wave-length, according
as one or the other is accepted as correct.

The University having reccived from the makers, early in the year 1880, a
spectrometer of unusual power and exceilence of construction and also a number
of diffraction gratings ruled by Mr. Chapman upon L. M. Rutherfurd’s machine,
of which one or two are unusually good, it appeared to me to be desirable to
make immediate use of it in the measurement of the wave-lengths of several of
the principal lines. It was not thought necessary to extend the work beyond a

few of the principal lines; at least not until a comparison of the results with those
of Angstrém and others should indicate that it was desirable. Angstrém based
nearly all of his work upon micrometric ıneasurements from these principal lines
and if it can be shown that they are at least relatively correct there is strong
presumptive evidence in favor of his micrometric measures. Concerning my
instrumental appliances, a comparison with those used by Augström (I have been
unable to find descriptions of the spectrometers used by the other observers
referred to) will show that in some respects, at least, they are decidedly superior.
In dimensions and fineness the grating far exceeds his, giving much greater
dispersion and, doubtless, superior definition and the angular measurements
were capable of being male with a greater degree of precision. In short, even
if the value of the grating space was unknown, I ought, with reasonably good
observation work, to be able to decide in favor of one or the other of these
discordant sets of results, or, possibly, against them all.

The work was done during the months of November and December on
account of the unusually clear weather which nearly always prevails here at that
season ; the prevailing low temperature at that time of the year is, however, an
objection.

THE SPECTROMETER.

The general construction and relative dimensions of the various parts of the
instrument are shown in the accompanying photograph. The instrument was
made in the winter of 1879-SO by Messrs. Fauth and Co. of Washington D. C.,
U.S. A. The mechanical execution of the work was excellent in every respect
and highly creditable to the makers, whose reputation for the production of
instruments of precision is well established.

For many valuable suggestions concerning the details of its construction
and for a careful examination of the instrument after its completion and before
it was sent to Japan, I am indebted to Professer C. A. Young of Princeton
College, and I wish to acknowledge with gratitude this, among the many obliga-
tions under which he has placed me and, indeed, every student of spectroscopy.

The dinfensions of the instrument are nearly as follows ;-—

IE TE EN 30 cm.
Length of View & Collimating Telescopes.............. 35 cm.
Aperture ,, » a NE N Men:
DE RER ni ee sco 16 DECOM:

The circle is divided to spaces of 5’ each. It is read by two microscopes
diametrically opposed to each other, the eye pieces of which are provided with
micrometers, one turn of the screw being equal to 1. The micrometer screw
heads are divided into sixty parts, thus carrying the reading to single seconds.
‘The various parts are capable of almost every desirable movement relative to each
other. The lower arm., supporting the collimating telescope, swings completely

around the vertical axis carrying all that is above along with it. The second
arm, supporting the view telescope, swings around independent of the collimator
and circle but carrying with it the reading microscopes which are attached to it
by a strong cross beam at right angles. The circle has a movement independent
of everything below it and the table upon which the grating is placed also rotates
independent of the circle. ‘The position of this table may be read, when necessary,
by means of two verniers placed opposite to each other. All of the movements
are, of course, provided with the necessary clamp and tangent screws. It will be
observed that the arms supporting the view and collimating telescopes are
provided with balancing weights so that both case of motion and stability are
secured. The eye piece of the view telescope is provided with a micrometer, as
is also the collimator for measuring and adjusting the width of the slit. A
striding level is provided for the telescope and collimator and levels are also
attached to the arm of the collimator and to the grating table. The latter has
screws for leveling and for centering.

The instrument was mounted upon a stone pier in a small building erected
for the purpose, the inside walls of which were blackened. The direction of the
collimator was nearly North and South. Through a small opening in a shutter
opposite the collimator a beam of light was thrown from a heliostat mounted upon
a pier just outside.

ADJUSTMENTS,

It will be seen that all necessary adjustments can be made with little
difficulty. The axis of the instrument is made vertical by means of the levels.
The telescope and collimator are adjusted for height and direction by means of
the vertical pillars upon which they rest, which have movements in a vertical
direction and about vertical axes. The telescope and collimator and the grating
table are then leveled. The adjustment of telescope and collimator for parallel
rays is easily accomplished by means of the collimating eye-piece, using the sur-
face of the grating as a reflector. The advantage in using this form of eye-piece
in this work is very great. Each telescope is adjusted to its focus independent
of the other and when it is desirable to render either telescope or collimator
normal to the grating it is accomplished accurately, quickly, and without the aid
of the circle. The slit and cross wires are adjusted to the line of collimation by
turning the telescopes in their supporting Y’s. The face of the grating is
rendered vertical by means of adjusting screws arranged for the purpose, using
the reflected image of the cross wires. To make the lines of the grating vertical,
a diaphragm is placed over the slit so as to reduce greatly its length. A very
narrow spectrum results and it is easy to adjust the grating so that this appears -
in the same position on both sides of the collimator.

The method adopted for making the observations, being that of normal
incidence, is discussed in detail in another place. A setting was made upon the

cc Oe u

line on one side, both microscopes were read and the telescope was moved to the
other side and set upon the same line and the reading taken again. The
temperature of the grating was noted after each setting. This was always
repeated several times and then the circle was unclamped and turned through an
angle of from 20° to-60°, the grating was again made normal to the collimator
and the angular measurements resumed. The ease with which the grating could
be adjusted by means of the collimating eye-piece reduced the labor of working
over different portions of the circle very greatly and, besides, it was easy at any
time to examine this adjustment so as to be sure that the grating had not been
disturbed from its position. At frequent intervals during the measurements the
instrument was readjusted throughout, in order that no constant errors due to
false adjustments might influence the results.

THE GRATING.

The University possesses four diffraction gratings ruled by Mr. Chapman.
Three of these are upon metal and one is on glass. I am informed by Mr.
Chapman that the composition of the metal is of copper and tin in the ratio of
17 parts of the former to 8 of the latter.

The metal gratings are square, each side being 7.5 cm. in length. The
thickness is 7.7 mm. The ruled surface covers about 20 square centimeters.
The finest of the three consists of about 30000 lines, nominally ruled at the rate
of 17296 to one English inch. This grating has great dispersive power, therefore,
separating the D lines in the spectrum of the second order a trifle less than 5’,
and in addition to this the definition is most excellent. In consequence of these
facts this grating was selected as being the most desirable for the work of precise
measurement of wave-lengths and, with the single exception noted in the results
given, the measurements were made upon the spectra of the second order. There
are one or two peculiarities about the spectra produced by this grating, found
also to exist in a less degree with the others, which are worthy of a moment’s
notice. ‘The most notable of these is the inequality in brightness of spectra of
the same order on opposite sides of the normally incident rays. This is notice-
able in spectra of both the first and second orders but is much more marked in
the latter.

The intensity of illumination at any point in a plane at right angles to the
lines of a diffraction grating is generally expressed in the shape of a function
consisting of the product of three terms. When a certain relation exists between
the second and third terms, spectra of a certain order will disappear entirely and
when this relation exists approximately, these spectra will be correspondingly
faint.

The order of the spectra thus affected is determined by the relation between
the whole grating space and the reflecting surface which remains after the groove
is cut ont, or between the opaque and transparent spaces in a ruling upon glass.

It is, therefore, not uncommon to find spectra of certain orders much fainter than
others. But as far as this cause is concerned, it affects alike spectra on both
sides of the incident beam and the explanation of the inequality referred to must
be looked for elsewhere.

It seems natural, of course, to look for the cause in some lack of symmetry
of the thin reflecting surfaces, in reference to the normal line, produced in the
operation of cutting the lines in the metal.

To establish the existance of this an examination of the grating was made
by means of polarized light. It is weil known that if light, polarized in the
plane of incidence, be reflected from a metalic surface the reflected beam will
remain polarized in the same plane. If the plane of first polarization is not
identical with that of incidence the reflected ray is elliptically polarized and if
examined in the ordinary way it will. in general, present the appearance of being
polarized in a plane inclined to that of first polarization. The grating was
arranged so as to reflect a beam of light polarized in the plane of incidence and
in such a way that it could be shifted about in its own plane and also turned
about a normal axis. It was at once observed that when the light was reflected
from the ruled surface of the plate, the ruled lines being nearly parallel to the
plane of incidence, a decided change in the plane of polarization took place.
When the grating was turned through an angle of 180° about a normal axis the
inclination of the plane was on the opposite side. When the lines were at
right angles to the plane of incidence no change of polarization could be observed.

From this it seems probable that in the process of cutting the grooves in the
metal each narrow reflecting surface had been slightly tipped out of the general
plane, towards that part of the surface already ruled and this would evidently
have the affect of increasing the brightness of the spectra on that side at the
expense of those on the other. Another grating was then examined in the same
manner, the relative brightness of the spectra on different sides not being known.
The change in the plane of polarization was not so marked in this as in the other,
but it was very readily distinguished. Im this the grating space is twice as
wide as in the first and it is evident that the disturbance of the plane of reflection
is not likely to be so great. It was easy to predict, however, on which side of
the normal the spectra would be faint and, in the hands of an observer who was
entirely ignorant of the examination that had been made, this prediction was at
once verified. The microscopic examination of these metallic gratings is difficult
owing to the difficulty of illumination when an objective of sufficiently high
power is used. A good illumination was finally secured, however. by throwing
the light in a plane nearly parallel to the surface of the grating. When this
was done the appearance presented was exactly in accordance with what had been
anticipated. When the light was received from the side towards which the narrow
reßecting surfaces were tipped, as indicated by the polarization experiments,
these surfaces were easily and distinctly visible as bright lines; when the grating
was turned in its own plane through 180° from this position it was only with the

greatest difficulty that the lines could be distinguished at all. Finally another
observer was requested to determine the effect, if any, upon the relative bright-
ness of the spectra on different sides of the collimator, produced by varying the
inclination of the grating to the collimator and the results agreed precisely with
the foregoing hypothesis.

There is another defect in this grating which seems to me to be likely to
exist in all gratings of a similar space and magnitude. It is the faintness of the
illumination at the two extreme ends of the spectrum, especially at the violet end.
In the American Journal of Science for November 1880., Professor Young, in
describing the excellent qualities of a similar grating, makes the same complaint
in regard to the violet end of the spectrum. Iam inclined to the opinion that
the cause of this is to be looked for in the minor periodic variations in the third
factor of the expression for intensity of illumination, referred to above. Besides
the great maxima of this function which determine the location of the spectra of
the various orders, there are numerous others of extremely small magnitude
compared with the first. They are distributed throughout the whole range
within which the great maxima are visible and, as they contain the spectral
components of white light, they will tend, notwithstanding their extreme faint-
ness, to produce a general illumination of the whole field, so as to destroy, to
some extent, the blackness of the background against which the ordinary
spectra are seen. Fraunhofer called these “spectra of the second class” and
Angström refers to them in his memoir in a discussion of the relative merits of
gratings in which a given number of lines is made to cover a great or a small
space.—The grating space being constant the number of these minor spectra in
a given space will be proportional to the number of lines. But as the brightness
of the great spectrum at any point varies as the square of the number of lines it
would seem that the one could never overtake the other. When we consider,
however, that the minor spectra contain all of the components of white light it
seems probable that where, as in the extreme ends of the ordinary or great
spectra, the light is, at best, exceedingly faint, that of the minor spectra may be
sufficient to nearly overpower it and to render observations difficult in these
regions. If this be correct it must be admitted that, while in general there is
great advantage in the use of gratings of great fineness, for the examination of
the extreme regions of the spectrum a smaller number of lines covering the same
space would yield better results. In the present instance it will be observed that

_ neither the A nor the H lines were included in the measurements as they could
not be observed with sufficient satisfaction to make a measurement desirable.

METHODS OF OBSERVATION,

In making the measurements necessary to the calculation of the wave-length
of any line, the essential parts of the apparatus, aside from the graduated circle,
are the slit with collimating lens, the grating and the telescope. In general
there will be three different arrangements possible.

The grating may be fixed so that its plane is normal to the collimator ; or it
may be normal to the telescope; or it may be normal to neither. ‘These arrange-
ments give rise to several different methods of making and reducing observations,
each of which seems to possess some advantages peculiar to itself. But it
becomes necessary to choose from among these that which, under the circum-
stances, proves to be the most desirable and, in order to present more clearly the
reasons which led to the selection made in the present instance, it may be
desirable to present a brief discussion of the principal methods which may be
adopted. For this pnrpose,—let

# = Angle between the telescope and the normal to the grating.
¢ = Angle between the collimator and the normal to the grating.
s = The grating space.

Then, in general, we have.
)=s (sin # + sin の )
the spectrum being of the first order.

Ist. Suppose $=0 that is, the collimator is normal to the grating. The
formula reduces to
A=s sin 0
and the operation consists in measuring the angle # which is best done by setting
on the line, first one side of the normal and then on the other, and taking half
of the angle passed over.

2nd. Suppose # = 0—or the grating is kept constantly at right angles to
the telescope. The resulting formula and process of measurement will be similar
to the preceding.

3rd. The telescope and collimator may be kept at a constant angle with
each other and the grating moved. The observations and reductions are easily
made in the following manner.

Let a = angle between telescope and collimator.
b = twice the angle through which the grating must be turned
in order to bring the line and the image of the slit upon
the cross wires of the telescope.

0 and ¢ = as before. 一

and the general formula reduces to

b a
N In ー
2 s sin COs 5

4th. A method of observation and reduction was used by Angstrém in which
the grating was made nearly normal to the collimator, the deviation from strict
perpendicularity being so great, however, that it conld not be neglectel in the
calculations. This method was not used by the writer, because it is especially
snitable for transparent rather than reflecting gratings, “and also because it was
perfectly easy to make the grating so nearly normal to the collimator that it
might always be assumed to be accurately so. On account of its being the
method used by Augström and in order that it may be compared with that
adopted in these observations, it is here included with the others.

Snppose the grating to be of glass and to be adjusted so as to be as nearly
as possible normal to the collimator. Readings are then taken with the telescope
in three positions; when set on the line on one side; on the image of the slit; and
on the line on the other side.

Let a, S and a,, be these readings
let --._—' = y and pe Sr)

then the formule for reduction used by Augstrém are ;—
2=s sin; cos (0+ )

cos 7
1 一 cos7

and tan 6 = . 0 approximately.

These formule are easily developed. As before

let 6 = angle between collimator and normal
er % telescope ,, » in one position.
Be 5 4 ji „ in the other position.

then we have

A= s(sin#@—sing) (1)

d
=” ı=s(sin d+sind) (2)
et. 9-15
Wat = a > ay = ee
の ーー
and é+¢= ーー

Adding (1) and (2) and substituting these values, we have: 一
A=s8 sin 7 cos (0 +4- p)
Also subtracting (2) from (1) and reducing in the same way,—

sin d= cos 7 sin(d +)

The angle 2 will always be small; in Angströms 35 observations upon the E
line, the only ones given in full in his memoir, its value was generally less than
3, reaching nearly 11’ in one case however. Assuming d=sind and cos 6 = 1
and expanding the above, we have

from which & is computed.

The relative merits of these methods in any particular case will depend
largely, of course, on the construction of the instrument anl the nature of the
grating with which the work is done. Thus, in the present instance, the instru-
ment does not adapt itself readily to the use of the second method and I do not
know that it possesses any particular advantages over the first. The grating used
prevents the adoption of the fourth method, besides it is unnecessary to resort to
the approximation which forms a part of the process of reduction in the use of
that method. The choice, therefore, seems to lie between the first and third.
From a theoretical stand point the third method is by far the most tempting of
all and it certainly offers many advantages. Among these may be mentioned
the fact that the mass of matter to be moved in the process of making a measure-
ment is reduced to a minimum, being only the circle and the grating upon it.
This motion is also generally much smoother than that of the telescope or
collimator and hence, more completely under the control of the observer. Thus
the settings will be easier and possibly more precise. Another great advantage is
that as the grating has a movement independent of the circle the method of
repetition may be introduced. The angle a is measured very easily by first mak-
ing the grating normal to the telescope, by the aid of the collimating eye-piece, and
then turning the grating until the image of the slit is bisected by the cross-wires.

In spite of the numerous attractions offered by this method it was very
reluctantly rejected, after several hundreds of observations had been made for
purposes of comparison between it and the first, and the latter was accepted as,
on the whole, more accurate. There are several objections to the third method,
not at first apparent, but which became evident during its use. One of these
was that it was found to be impossible to rely upon the constancy of the angle
between the telescope and collimator. Sometimes a change would appear to
occur suddenly and the magnitude of the change was often so great as to produce
a decidedly sensible effect on the resulting wave-length. I attribute these changes
to the effects of change in temperature upon various parts of the instrument; a
sudden shifling taking place when the stress due to expansion reaches a certain
amount. Of course the real movement is exceedingly minute but it is sufficient
to produce a perceptible effect on the angular measurements. The existence of
these changes renders frequent measurements of the angle between the telescope
and collimator necessary. As the collimating eye-piece is a necessity in these

ENT EL

measurements a change must be made if higher powers are in use in the line
observations.

Again, it was found that it was unadvisable to make numerous repetitions
of an angle. except when the temperature was unusually constant, for to make
the temperature correction it was desirable to know the temperature correspond-
ing to each setting rather than to depend upon temperature observations made at
the beginning and end of a series. Finally the measured angle, that throngh
which the grating was turned, was much smaller than the corresponding angle
in the first method and in consequence the results, which depend upon the sines
of these angles, are affected to a greater extent by small errors. A glance at
the formula shows that it is desirable to make a as small and b as large as
possible. With a reflecting grating a cannot be diminished beyond a certain
limit; that is, with a spectrometer of ordinary construction. An instrument of
what may justly be called “extraordinary” construction has recently been made
by Alvan Clark & Sons for Princeton College U.S.A. In this instrument the
collimator and telescope are one, as far as the use of the same tube and object
glass goes. It is obvious that this ingenions plan gives the maximum efficiency
in certain operations, and if some optical difficulties which present themselves
can be entirely overcome this instrument will doubtless take rank as the most
accurate existing means of investigation in this direction.

Concerning the first method, which was finally adopted, there are one or two
difficulties, which appear to he of considerable moment at first, but which are
found to be manageable in practice. With a reflecting grating it is, in general,
necessary to swing the telescope through at least two thirds of the whole
eirenmference in making a single pair of readings. While in this case the
mass to be moved is mnch greater than in the previous method, yet the excellence
of the construction of the instrument is such that very little is lost in the way of
smoothness of motion and delicacy of setting and much is gained in the firmness
and stability of the instrument by the considerable weight of its various parts.
Repeated examinations have shown that the various movements are entirely
independent of each other and the motion of the telescope never produces any
sensible disturbance of the other parts. Dy the use of this method, advantage
cannot be taken of the principle of repetition as it can in that of oblique incidence.
But it is believed that most of the advantages of repetion have been secured and at
the same time the objection to it has been avoided. The ease and rapidity with
whieh the grating conld be made normal to the collimator was such that
measurements could be readily made over different parts of the circle in succession
and thes the errors of graduation eliminated as far as possible. "The temperature
correction, which is very important, was secured by reading the thermometor
immediately after cach setting upon a line. Finally, it may be urged that we
can never le ceituin that the grating is absolutely normal to the collimator.
This is quite true, but it is also true that the error in this adjustment may be
considerable without sensibly affecting the result.

To investigate this question suppose that a reflecting grating is used and
that we look in the direction of the collimator, towards the grating;

let #= angle between telescope and normal on the right

= 3 ” ” ” ” 73 an SKE,
— Zr > collimator ,, 時
の っ “ - ,» telescope on the right
b= 2 29 ” ” 39 ” は / left.
Men en 5 中 = for strictly normal incidence.

then b = ?—aandbh=/’+a

and we have for the spectrum of the first order—

= sin (b + a)+ sina (R)

nin の 1 テン

= sin (b, — a) — sina (L)

the normal to the grating being supposed to be thrown slightly to the left.
Differentiating and supposing a = O we have
db = — (1 + sec b) da
db, = (1 + sec b,) da に

b+ 5b ;
and the error in taking ort ! for b, will be
a

(sec b, — sec b) da

An examination will show that if b, be 45° and the grating be as mich as 5’
away from its true position this error will be less than half a second. To
facilitate and increase the accuracy of this adjustment the brass jaws of the
slit were blackened and then when the collimating eye piece was placed in front
of the slit and properly illuminated a very clear and well defined image was
produced by reflection from the grating. It was found that the adjustment
could be made very quickly and within one or two seconds. It is safe, therefore,
to assert that the results are entirely free from any sensible error originating
in this adjustment.

CORRECTION FOR TEMPERATURE,

To make the reductions for temperature it is necessary to know the tem-
perature at which settings were made and the co-efficient of expansion of the
material of which the grating is composed. It is believed that the first has been
accomplished with a degree of accuracy exceeding that in the case of any series
of measurements hitherto made. Angström admits and deplores the difficulty

ees, eee eae err

or impossibility of knowing exactly the temperature-of his grating. Tn the use
of a transparent grating it is not easy to see how its temperature could be
measured, but there exists the compensation that the co-efficient of expansion
of glass is so much less than that of a metallie reflector that it becomes less
important. In these measuremeuts a small rectangular vessel of thin copper
was sealed to the grating which itself formed>one side of the cup thus formed.
This was filled with water in which a small thermometer was constantly immersed.
As the changes in temperature were never very rapid the reading of this thermo-
meter can differ but little from the true temperature of the grating. The
thermometer was compared with a standard and table of corrections prepared.

As far as I conlıl learn in the beginning, nb determination had been made
of the co-efficient of expansion of the alloy from which these gratings are made.
A series of observations was therefore undertaken, for the purpose of ascertaining
this from the grating itself by means of angular measurements made upon“ the
sume line at different temperatures. If we take the equation—

A= FID

in which 4 represents wave length, s the grating space and 4 the angle of devia-.
tion, we obtain-by differentiating —
ds ~

— =— wtbdb
4 s
When the increment of temperature is 1° the left hand member represents the
co-efficient of expansion. It will thus be seen that this can be determined with-
out any knowledge of the wave length of the line set upon or of the value of
the grating space. In the first series of experiments an attempt was made to
carry the grating alone through a range of 5° or 10° of temperature, that of
the surrounding air remaining sensibly constant. The grating was made to

"take the place of one side of a wooden box which was filled with water whose

temperature was changed as desired. It was found to be difficult to prevent a
Slight shifting in the position of the grating, due to the changes in temperature

of the box and its support on the table of the spectrometer, and the temperature

could not be maintained constant for a sufficient length of time to enable obser-
vations to be made on both sides of the collimator so that the effect of this
slight change in position might be eliminated, And besides, the difference of
temperature between the two faces of the grating caused it to warp slightly
which injured the definition if it did not otherwise interfere with the acenracy
of the results. In a brief article in the American Journal of Science for
March 1881, I described this mode, of making the determination and gave the
results of a series of measurements. The objections to the methods were at
that time fully appreciated and a more reliable process was indicated. The
opinion was also stated that the result there given was move likely to be too
high than too low. In fact when I came to apply this correction to the regular

observations, especially in the case of one or two lines in the measurement of
which great care had been taken, it proved to be a correction which would not
correct. .It therefore became necessary to make a redetermination of this con-
stant. It was found upon trial to be possible to raise the temperature of the
room in which the spectrometer was mounted, as high as 18° or 20° C., when
the temperature outside was much lower and to keep it nearly constant for a
considerable length of time. A series of measurements was therefore made
upon a fine line, near to b,, while the temperature of the room was low and
then afterwards it was raised to about 18° and another series made upon the
same line. This plan avoided the principal difficulties met with in the preceding
and the result is, without doubt, much more reliable. ‘The lower temperature
was nearly 4°, the results being reduced to that, and the upper differed little
fiom 18°, to which the measurements at the high temperature were reduced.
Thus there was a range of 14° which is greater than that of the previous deter-
mination. ;

From these observations the following was obtained ;—
db = 34.890203
b = 44° 56 50!
and for the coefficient of expansion -
e = .0000189

the probable error of which is less than one percent of the whole. The result
of the first determination was,— e = .0000202 and without doubt there existed
some source of constant error which was not detected.

The direct method of applying this correction would be, of course, to
determine the value of the grating space for such degree of temperature at which
an angular measurement was made. It has been found to be much more conve-
nient, however, to apply it directly to the angular measurement itself. If b and
b, be the angles for the two lines we shall have, since e is constant,—

cot bdb = cot b, db,

or db, = ——.d

Thus having determined the value for one line, by observations at different tem-
peratures, that for any other line can be easily computed.

In this way all of the angular measurements have been reduced to a common
temperature of 18°, which is also the temperature assumed for the grating space.

CORRECTION FOR BAROMETRIC VARIATION,

The observations may be assumed to have been made, in general, under a
Standard pressure of 760 mm. With a single exception, it has happened that

the measurements were made upon days during which the fluctuation of the
barometer from the normal was very small, generally not more than one or two
millimetres at the time of making the reading. The exception was on Novem-
ber 26, when the height was abont 10 mm. below the normal at the time of
making some measurements upon the E line. These measurements have been
corrected for this difference but this correction has not been introduced in any
other other instance as it has not amounted to more than a small fraction of a
second, and its effect will doubtless be eliminated from the general mean, owing
to the fact that these small fluctuations were about as often on one side of the
normal height as on the other.

THE VALUE OF THE GRATING SPACE

By far the most difficult quantity to determine with accuracy is the value
of the grating space. When angular measurements can be made with the
precision which it is believed has been reached in these determinations, the
comparison Of the space of one grating with that of another can be made with
a degree of accuracy probably exceeding that attainable by ordinary methods.
For this purpose the gratings need not be brought together, the length of a wave
of light being a common measure easily applied to both. The value of the
space of one grating may be made to depend, therefore, upon that of another, the
absolute length of which has been ascertained with great care, provided both
have been employed in precise angular measurements upon one and the same
line.

Instead of this a direct comparison of the grating space with a standard
unit of measure may be undertaken. Where the first method is possible it is
vastly more convenient and, in general, more accurate than the last. Although
at least two Metres of well determined length are available in Japan, no com-
paring apparatus suitable for such work could be obtained and it was, therefore,
not only desirable but necessary to rely, if possible, upon the first method for
the determination of the grating space. During the progress of the observations
I had hoped and expected to be able to make such a comparison between this
grating and another whose space had been determined to a great degree of
precision. Mr. ©. S, Peirce of the U. 8. Coast Survey, has been engaged in a
series of observations of great interest and value for the comparison of a wave-
length of light with a standard metre. The operation includes, of course, the
accurate measurement of the angle of deviation of a certain line, produced by
a grating whose space is definitely known.

Unfortunately I have not been able to procure the results of this investiga-
tion, indeed I believe the final comparison of the grating with the metre has
not been entirely completed. I am therefore unable to give the value of the
grating space with that degree of certainty which was at first expected and

which would be so desirable, Jt must be remembered, then, that the results do
not possess that value as absolute wave-lengths which I believe would otlierwise
belong to them. Their relative valne is not altered, however, by this fact and
I hope to have furnished the data by means of which, at any futnre time when
it becomes possible, such slight corrections as may be necessary may be easily
applied. Tt is believed that the value of the grating space assumed is very
close to the true value and that the corrections to be applied hereafter will be
very small.

The following are the facts in regard to the grating space as far as at
present known.

The erating is marked by the maker “17296 lines to one inch.” Tam
also informed by Mr. Chapman that it is attempted to keep the temperature as
nearly constant as possible at 65° F. during the process of ruling, but that as
the ruling of a large grating ocenpies several days the temperature often fluctuates
{wo or three degrees above and below that point. The temperature to which
these observations have all been reduced is 18°, which is very nearly that at
which the ruling was made. This number, 17296 to one inch, is equivalent to
680.957 to one millimetre and were the inch of the machine known to be correet
this number might be accepted. But Mr. Chapman says that one English inch
is undoubtedly less than 17296 of these divisions by one or two divisions.

In his ‘Measurements of Gravity at Initial Stations in America and
Enrope”* Mr. Peirce has given the result of an elaborate examination of a centi-
metre ruled upon Mr. Rutherfurd’s machine. The concluding sentence of this
discussion is as follows ;—

“This centimetre is equal to 68093 teeth of Rutherford’s machine, and as
it is ghee too long, we conclude that 6809 teeth make a centimetre at ordinary
temperatures, say about 18°.” As this temperature agrees with that at which
the grating was ruled and also that to which the observations have been reduced,
it is the most acenrate determination at present available and in all of the
following reductions the spaces have heen assumed to be 680.9 to one millimetre.
This number is not free from donbt, however, as is evident from the fact that
the temperature at which this particular grating was ruled may not be exactly
identical with that of the above comparison and besides there is room for doubt as to.
the accuracy of the decimetre of comparison and Mr. Peirce justly remarks that “all
exact measures of length made now must wait for their final correction until the
establishment of the new metric prototype”. It is likely therefore that the
assumption made is the best possible under the circumstances.

In order that it might be possible to correct these results, if necessary, by
means of the method of comparison of angular measures, I have endeavored to
connect the series with Mr. Peirce’s angular measures for the comparison of a
wave-length with a metre. For this purpose I have referred to the only published

* Appendix No 15—-Report of 1876—U. 8. Coast Survey.—Printed at Washington—1879.

account of his results that I know of, which is to be found in the American
Journal of Science for July 1879. Unfortunately, the data given there are

not sutticient to enable me to determine with absolute certainty the line set upon.

Mr. Peirce made two series of observations; in the first he measured the deviation of

aspectral line, “van der Willigen’s No 16,” “using a certain gitter of 3404 lines to
the millimeter”. I have not been able to refer directly to van der Willigen’s neas-
urements to ascertain the position of his “No 16” but as Mr. Peirce’s grating space
was double mine I have endeavored to locate it from his measures. [ have assumed
that the work was done in the spectrum of the fourth order. Although nothing is
said of the order of the spectrum in the article referred to, it is clear that no other
supposition would be admissible. ‘This line should, therefore, have the same devia-
tion in the spectrum of the second order of the grating used in these measurements.
Mr..Peirce says nothing about the temperature at which his measurements were
male so that there is room for some variation on this account.

Upon examining the spectrum at this particular deviation there appeared
to be no line at that point which seemed to me to possess sufficient distinctness
and prominence to have been selected for this important duty. A well defined
and sufliciently prominent line very close to the indicated point, a little nearer 6,,
seemed to be most probably the line upon which the settings had been made.
This line was accordingly selected and a considerable series of careful measure-
ments were made upon it. Assuming this to be the same line used by Mr.
Peirce, the difference is not greater than can easily be accounted for. His
result is

b= 44? 67 ar)

and my own, the mean of 34 measures which will be found in detail in the fol-
lowing pages, was ;—
b = 44° 56’ 507.35

the difference being 19.4

Mr. Peirce does not state the temperature to which bis results were reduced.
With my own grating the increment of are for this line was 37.89 for 1°C, so
that it is clear that there is a possible variation in temperature easily sufficient
to account for the difference in the angular measurements, especially when it is
remembered that the gratings may have been ruled at different temperatures and
that they are composed of substances widely differing in their co-eflicients of
expansion.

The second series of measurements made by Mr. Peirce was upon another
line near the first and with “another much finer gitter.” As this is a region
crowded with lines and as the fineness of the grating, was not definitely stated,
it was impossible for me to decide certainly as to the line made use of, so it was
determined to rely upon the first results, which are undoubtedly of sufficient
accuracy, assuming that the right line has been found.

If the exact value of Mr. Peirce’s grating space was known, that of the grating
used in these measurements could be at once ascertained with a degree of accuracy
equal to that of the angular measurement of this line and whenever the final com-
parisons are completed it will be easy to apply the necessary corrections.

MICROMETRIC MEASUREMENTS.

Micrometric measures were made in a few cases in which the lines were
so near to the reference line that the effect of a possible error in the value of the
micrometer screw or in the adjustment of the grating would be small. In order
to determine the value of the micrometer screw twenty measurements were made
of the distance between E and },, this distance having been carefully ascertained
by means of the circle. The space between these two lines was divided into
two nearly equal parts which were separately measured, in order to bring the
measurements into the centre of the field as nearly as possible. The resulting
value of one division of the head of the micrometer screw was applied to the
measurement of b,, b, and b,, referred to b, and also to D,, referred to D,. The
extreme distance measured in this way was less than 11’ and it is evident from
the investigation already given that the results will be free from error due to
grating adjustment and also that they would be but little influenced by tempera-
ture although the latter correction has been applied.

RESULTS.

The following are the results of the angular measurements made upon
what was presumed to be van der Willigen's. No. 16, for the purpose of ascertain-
ing the relation of the grating to that of Mr. Peirce already referred to.

As the result of this series may possibly be of great value in the final
adjustment of the absolute wave-lengths, it is thought proper to give all of the
original measures, together with the temperature at which each was made, and
the value of each when reduced to the standard temperature. The numbers
will also serve to give an idea of the general accuracy of the angular measure-
ments as well as of the sufficiency of the temperature corrections. It ought to
be said, perhaps, that for the most accurate angular work this line is not well
suited; a thinner line would undoubtedly give a more uniform series, But the
fact that the line is easily seen, even under inferior optical power, is decidedly
in its favor, as far as its general utility as a “reference line” goes. The measures
were made in the spectra of the second order.

VAN DER WILLIGEN’S No. 16 (?)

Date b observed Ah b reduced to 18°
December. 19 EVANS (Maes a | 229 44° 46’ 49”.0
” es) 6°.8 0
” „ 9 26”. 8°.6 a 50".2
3 ON. 10°.15 oe
A DD 11292 eg
$3 mee Nee 7 12°.9 050
Pr tg EO NO 13°.8 ae MOOS
” » 22 17.3 15°. » 5 497.9
5 44° 56’ 497.9 18H ate Pos
= ANZ 183 ee DOGS
a yh i A942. 18°.1 AIG
= 4925045 19 ol
5 pata 20S! 17°.4 pl yy OT
FA ag we eS Nicks u a SO
Pr yt B48 16°.9 oe ee eG
Dec. 20 44° 56! 507.6 18°.3 44° 56! 517.7
” AN as 50.5 18° Le} 507.5
Wee ee) ee.) aka ea
FS eee IS 17°.4 sy bgp ten 4

” 230,27 54”.6 re Si 50.8
‘i 。 5476 16°.7 。 っ 497.6
a hinge BONS 16°.5 yy 50.6
Dec. 21 44° 57! 33.6 6°.5 44° 56! 497.8
” ” jr 24".5 8°.9 ay Jay 497.5
内 fs, ORS 10°.6 yy 507.4
" 129.3 9
‘i 407 名 13° a a
ek „ 56’ 567.5 16°.3 en
4 OD 16°.4 AM
9 vn 56.6 16°.4 ee Bere
の » yy 5276 1702 nn 497.6

デ

Between the date of the last observation and the next to be recorded the
spectrometer was used for prism work and in consequence the grating had been
removed, and an entire new adjustment in every particular had been made. The
close agreement of the following three observations with those preceding furnishes
strong evidence of their freedom from any constant error due to adjustments.

Date b observed T b reduced to 18°.
Dec. 27 AVS are (OMCs 1525 44° 56’ 50.8

5 Do! 594 1925 m A

73 Pe oy ent Us 159.3 N,

Mean of 34 measurements.

b = 44° 56’ 50".35 + 07.08

Among the various values for the wave-length of this line, published by
different authorities, are several that differ so widely from the others that it
seems difficult to account for the discordance in any other way than by assuming
that different lines have been measured in different cases. This region of the
spectrum abounds in absorption lines and mistakes might easily be made in the
selection of the proper line. In the present instance the line selected is believed
to be that upon which Angström made his measurements. It is a line, some-
what darker and more prominent than the others, on the edge of the group in
the direction of least wave-length, as is indicated in his map. A very fine line
is often visible, still further in the same direction and distant from B something
less than one unit of wave-length.

This is the only line in the series in which all of the measurements were
made in any other than the spectrum of the second order. The first order was
preferred in this case on account of the faintness of this part of the spectrum

on one side of the normal to the erating.

Date No. of obs. Value of 7 A
Dee. 1. 4 De 3a 6866.84
3 ote ian moe eros
be 3 Joe ee ee elt
Dec. 2. 4 SS 80
a Fae ONT, Eee)
3 ie! » .84

Mean of 21 observations.
7 = 6866.82

This line is broad and in preference to attempting to set upon the middle
of the line in every instance, the edge of the line was selected and the results

mee 。

LE OE IN: cm ly et ーー int ie

“a

_
21
afterwards reduced to the middle by means of a mierometrie measurement of the
width of the line.
Date No. of obs. Value of b A
)
Nov. 30 4 GER. 1 28EH 6561.64
on 2 eh eee Ui} ep)
| Be 3 EN, Sm
+ 3 ragt. BER
np = ECHTEN A.) 200
4 OO Nay OW
x Mean of 20 observations—
| 4 = 6561.62

Another series of 13 measurements was made in the spectrum of the first
order, in which the line is more neatly defined and the set was made upon the
middle of the line. The mean of the results agrees almost exactly with the
preceding. being

A= G56 EG

SS << ae

D, and. D;

These lines are of such frequent use in wave-length measnrements as furnish-
ing, in the distance which separates them, a sort of unit of comparison, that it
was thonght advisable to make a more extended series of measurements upon
them than upon any of the others. D, which is that of greatest wave-length,
was measured directly but the wave-length D, was determined by means of

micrometer measures from D,.

ーー

Date No. of obs. valne of b /
Nov. 18 8 323 37.6 5894.79
reget 5 re N,
EN! 3 ete Shed ae
1 3 ae ee ae
Fr うり お u:
3 ERST nr
1 2 3 ~ 36".8 Ay |
> oe 3 ens Bye 18
FR 3 Pig BPM sy soe
La : rg Be 6

=
=

7 en
Fr ” o er „ 36 2 er}

Mean of 40 observations—
4 = 5894.780

Forty measurements were made with the micrometer for the distance of D,
from D,. The lines are too wide for very exact measurements but the distance
between their centres thus obtained is donbtless sufficiently exact.

It is
6 = 4 46"8 + 0"2
from which we have
be
and A = 5888.68
From these results it appears that
D,—D, = 6.10

According to some previous measurements already referred to this quantity
is as follows:

Mascart 6.3
Ditscheiner 6.4
van der Willigen 6.14

Angström furnishes two results differing slightly from each other. From
his tables of wave-lengths it appears that the distance between these lines is
6.01, but on another page of his memoir he determines the wave-lengths
of the two lines according to what he calls the “method of coincidences” and
from those values we have,

D,—D, = 6.08

These last numbers he evidently considers the most accurate for he adopts
their mean for his wave-length of D in his final table of “definitive values”.
The result given above differs very little, therefore, from what may be accepted
as Angström’s best measurements. It is impossible, of course, that any error
will be found to exist in the grating space which will affect this result in the
second decimal place: but a small error in that figure may exist consequent
upon errors in the micrometric measures.

This is a double, one line being due to iron and the other to caleium and
iron. The measurements were made upon the latter. which is that of the great-
est wave-length. It was upon the E line that Angstrom made his most
careful measurements and he presents in full, in his memoir, the series of 35
observations. But these appear to have been made upon the mean of the two
lines, at least the result is exactly the mean of the wave-lengths of the two lines
given in his table and in his series of micrometric measures values differing
slightly from these are assumed. It is a little difficult to understand how
accurate angular measures could be made upon the mean of two lines, or, at
least, why one of them should not have been selected, provided there was suffi-
cient dispersive power to separate them, which must certainly have been the
ease. It is worth noticing that the wave-length obtained by Angstrém for the
mean of these lines agrees precisely with that of the line of greatest wave-length
as determined by the following measures.

Date No. of Obs. Value of b. 2
Nov. 26 > Aa IT 5269.13
ee 3 の an erg 2:10
Nov. 27 4 =f 5 et | er
ail ape 3 N ae oe u” 14
” ” 3 ” ” 975 29 .14
29 2» 3 ” ? 9”.4 „ 13
7 iB 22 3 22 39 Me 1 „ .13

Mean of 22 observation.—

= 5269.13

b, b, by and by

As regards these lines the nomenclature of Angstrém’s table and map has
been followed. According to Angström and 'Thalen, 5, and b, are to be attributed
to magnesium, 4, to nickel and iron and 4, to magnesium and iron.

The first, b,, is a broad line or band but near the edge towards the blue is
a narrow, well defined line upon which the setting was made. ‘The distance from
b, to E being less than 1° the position of the line was determined in reference to
E the measnres being made, however, with the circle and not with the micro-
meter. It is believed that no error existed in the adjustment of the grating of
sufficient magnitude to influence the results, especially as the grating was read-
justed several times during the measurements.

Date No. of Obs.
Dec. 7 7
Sd Bas 4
Dee. 5 4
5

”

ここ

” 55

Value of b の (
44° 53’ 217.6 5182.40
54 nae ae ae
nn ie} 40
‘ Pe NG ree
2 40

Mean of 21 observations.—

i=5

The remaining three lines of this group b,. 4, and b, were

metrically, by referring them to b..
with the following results ;—

182.40

measured miero-

Twenty measures were made on each line

line value of b yk
b, 44° 46’ i879 5171.47
b, 44° 43 40”8 5167.73
b, 44° 497 387.6 5166.16

One of the objects in measuring the lines in this group was to verify their

dlistances from each other that they might be used as standards of comparison in

that part of the spectrum, as the D lines are so frequently.

In one or two

instances, in well known works on spectrum analysis, the mistake has been made

‘ of applying the D lines as a unit of measure in the violet end of the spectrum

and failing to consider the abnormal
spectrum,

extension of that end of the prismatic

The use of the b lines may prevent so great an error as well as be

exceedingly convenient on account of their prominence and the convenient por-

tion of the spectrum in which they are found.

length we have the following:

b, — by
b, — b,
b, — by

From Angstroms results the same

た 03
be ne b,
b, x が

For their differences of wave-

= TO93
3.74
1.59

measures appear as follows ;

= 10:94
3.68
1.60

and the two series agree exceedingly well.

25
F

The following are the results of the measures upon the F line.

Date No. of Obs. Value of 6 2
Dec. 4 4 21 26" 230" 4860.17
a 3 “ oe 20" ae!
a ee 3 a 02903 ts)
が OD 3% 時 | Aa oul Wi
ait 3 i ae sO ies SET
Be ay 4 < sete neu

Mean of 20 observations

2 = 4860.16

Owing to the inferiority of the spectrum on the west side of the normal,
which has already been discussed, this line could not be seen there with sufficient
distinctness to make direct measurement desirable. The construction of the
instrument rendered it impossible to work upon it in the spectrum of the first
order, and, as it was but little more than 5° from F in the second order it was
thought that the most accurate determination would be made by referring it to
that line which was accordingly done. It is recognized, of course, that errors
in grating adjustment, which will influence the resuits of this measurement by
an amount proportional to the difference of the secants of the two angles, are of
greater importance here than in any other instance; but this difference is still
small and the error must be greater than is supposed to have existed in any case
in order to produce a sensible effect. Besides, this adjustment was made five or six
times during the measurement and the influence of accidental errors must be
eliminated to a great extent. The uniformity of the results proves that such
errors must have been extremely small.—The following are the results.

Date No. of Obs. Value of b メ

Dec. 6 6 35° 54’ 467.2 4307.19
oe 4 ae A „238
» 39 4 ” ” 457.6 » A8
ee, 3 ee te awe ie ee
ae 4 a FRE ul: DB

Mean of 21 observations

A= 4507.19

RESUME.

Below are given the collected results of these observations. In the second
column will be found the corresponding wave-lengths according to Angstrém and
in the third their differences.

M A A—M
B 6866.82 6867.9 18
Cc 6561.62 6562.1 48
D, 5894.78 5895.17 39
D, 5888.68 5889.09 Al
E 5269.13 5269.59 AG
bh, 5182.40 5183.10 70
b, 5171.47 5172.16 69
ds, 5167.73 5168.48 75
db, 5166.16 5166.88 72
F 4860.16 4860.72 56
G 4307.19 4307.25 06

It will be seen that the difference in no case amounts to a whole unit.
Much of it, indeed, might be destroyed, by a more accurate knowledge of the
grating space in each case. It will be observed that a sudden increase in the
differences occurs when the group of 5 lines is reached. I think this finds an ex-
planation in the fact, referred to in giving the observations on b,, that the setting
was made upoa what appeared to Le a well defined part of the line, towards the
side of least wave length. As the other three lines are referred to this, it is to be
expected that all will be affected alike and this is the case. It seems likely that
Augström set upon the middle of the line. The part of the line referred to in
these observations seems so well defined that I prefer to retain that as the point
of departure rather than to reduce to the middle of the line, although the
uniformity of differences would be greatly increased by such a reduction. I,
know of no reason for the small difference in the results for G, other than the
difficulty of making measurements in that part of the spectrum which would
doubtless affect both sets of observations more or less. If Angstrém had used
the same part of b, for a point of reference that has been adopted here, I am
convinced that for all of that portion of the spectrum in which the best measure-
ments are possible, the series of differences would have been very nearly uniform
and easily accounted for by an inconsistancy in the determinations of the grating
space. As before stated, the ratios of these numbers ought to be independent of
the value of the grating space and in order that a comparison may be made
upon that basis, ratios have been computed for both series, using the value of D,
asa unit. They are as follows.

27
RATIOS
al
M A A—M

B 1.16490 1.16485 一 .00005
C ‚111312 1.11313 ‚00001
D, 1. i 0
D, ‚998965 998965 0
E 893864 893882 .000018
b, ‚879151 ‚879212 ‚000061
b, 877297 877355 .000058
b, ‚376662 ‚876731 ‚000069
b, 876396 876460 ‚000064
F 824485 824526 .000041
G .730679 730640 ‚000061

The agreement here is sufficiently close, it seems to me, to justify the
assertion that any other wave-length measurements that differ widely or irreg-
ularly from these must be incorrect.

MEMOIRS

OF THE

SCIENCE DEPARTMENT,
| TOKIO DAIGAKU

(University of Tökiö)

No. 9.

J. A. EWING, B.Sc., F.R.S.E.,

ProrEssor or MECHANICAL ENGINEERING AND Paysics IN THE University OF TOki6, {
< x rT 1
VicE-PRESIDENT OF THE SEISMOLOGIGAL SOCIETY OF JAPAN,

- 4

PN. 。 a

MEMOTRS

SCIENCE DEPARTMENT,

TOKIO DAIGAKU

(University of Tolxio)

INT On 9:

EANTHOUAKE MEASUREMENT

J. A. EWING, B.Sc., F.R.S.E.,

Proressor OF MECHANICAL ENGINEERING AND Puysics IN THE University or TOki0。

VicE-PRESIDENT OF THE SEISMOLOGICAL SOCIETY OF JAPAN,

み クジ >
JUN 23 1884

eo
Le

ーー

PUBLISHED BY TOKIO DAIGAKU
TOKIO:
2543 (Japanese Era.)
1883 A. D.

PREFACE.

The following Memoir has grown out of work which, with the generous
assistance of the President of the University, I began in the winter of 1879-80,
and which has been continued until now. The erection of instruments adapted
to the absolute measurement of earthquake motions was quickly followed by the
registration of numerous earthquakes. These occur in Tokio with a frequency
which gives resident students of seismic phenomena an advantage difficult to
overrate; and it is therefore no wonder that the instrumental side of Seismology
has had its most considerable development in Japan.

The essay is chiefly an account of my own instruments, methods, and results;
but, with the view of making it as complete a treatise on Earthquake Measure-
ment as the present state of the subject allows, I have included many notices of
apparatus designed by other contemporary workers, as well as of older forms.
In referring to the work of other observers, I have endeavoured in every case to
give the fullest credit for novelty, where there is novelty. It may be added,
with respect to recent seismological work accomplished in Japan, that the five
volumes of the Seismological Society’s Transactions now published contain
original accounts of most of the instruments described in this paper, and should
be referred to by any one interested in the history of the subject.

My special acknowledgments are due to Mr. Kato, the President of the
University, to Mr. Hamao, the former Vice-President, and to Mr. Hattori, the
present holder of that office, for providing me with the means of establishing a
Seismological Observatory. The frequency and success with which earthquakes
have been recorded are in great measure owing to the attention bestowed on the
instruments by my assistant, Mr. K. Sekiya, in whose care I am happy to be
able to leave the apparatus on my approaching departure from Japan. To
Mr. Fukuda J am indebted for the preparation of most of the Plates which give
illustrations of the seismographs I have used. Plates XII-XX are exact copies
of actual earthquake records, executed with a fidelity which does great credit
to the draughtsman, Mr. Suzuki.

J. A. EWING.

Tue University, Toxo,
May 2nd, 1883.

CONTENTS. |

BEER AT an Brenn adsense ang eer FUTTER EN LEEREN. ix
- NETTE. Dre POL CD RT で いわ OPC PETER xi |
| CHAPTER I, |
THEORETICAL CONSIDERATIONS REGARDING EARTHQUAKE MOTION.
| ART. PAGE
| 1. Vibrations proceeding from a centre within an elastic substance ...... 1 |
| Bela OP DPONSE 2... essen taste N AER 2
| 3. Effect of boundaries. Reflection and Refraction.……………………….……….. 3
4. Example: Case of an Earthquake reflected at the surface of a sea |
OU は たつ の 人: イチ PT PCLT PE ERE FARR 4 |
5. Reflection and Refraction at the boundary of two solid mediums...... 5 i
1 6. Waves diginating in or transmitted into an eolotropie solid ...,..... 6 |
. 7. Effect of a stratum in reduplieating waves ............,.ccc0ccc ces seen 6 N
{ 8. Effect of obstacles. Earthquake Shadows and Diffraction ............ 8 N
| ER EL eg Er N DI LTE TEE FERREREENSTRHRUTE RO 9
6 NEE 97 INI DOLIEDY CIBBLIGIGN <1 -\cc< yee Godeses as ivv.cas sos adesegavsarneeaaqaes 10
; 11. Movements of a particle during an Earthquake ........................... 11
12. Motion at the surface of the soil due to a normal wave.................. 12
13. Preliminary Statement of the Results of Observations ........ ......... 18

CHAPTER II.

INSTRUMENTS FOR OBSERVING HORIZONTAL MOVEMENT: —

THE HORIZONTAL PENDULUM SEISMOGRAPH,

14. Measurement of earthquake motions by reference to a Steady Point... 14

15. Kinetie condition yielding a Steady Point ............ .00..0...0....04... 14 |
16. Static condition requisite for a Steady Point... 15 |
17. Horizontal Pendulum Seismometer ............422 スス ee 16
18. Friction Error, Inerense of Effective Inertia by the use of a second |
mass pivotted on the first at the Steady Line........................... 17 |
19. Determination of the Effective Inertia and Steady Line ............... 18 |
4 i

PAGE

Multiplication and Registry of the motion ............... 0.0... .eccce eee 18
Multiplication by an independent lever lg ココ ーーー 19
Record of successive displacements in conjunction with the time ...... 20
Actual Horizontal Pendulum Seismograph .…………………………….…………………… 20
Improved form of Horizontal Pendulum Seismograph .................. 22
Horizontal Pendulum Seismograph with Ring Bob _..................... 24
Relation of the Friction Error to the dimensions of the frame and to

the suspended mas... に に ミニ ーー の oo は ここ 24
Horizontal Pendulum with flexible tie and pivotted strut ............... 26
Horizontal Pendolum without Jonnfg............ 00 000 000 ee 27
Horizontal Pendulum with two degrees of freedom........................ 28

CHAPTER III.
OTHER INSTRUMENTS FOR OBSERVING HORIZONTAL MOVEMENT.
Mass free to roll: 22) cies coe doce 2 ee eet ee na teee ee ne ee

Rolling Sphere Seismograph ーー ーー こら

Friction in the Rolling Sphere Seismometer ............0.0ummunmensenenn 92

Other Rolline Sphere Scismographs....... 0. .cap ne. ミン ンー 33
=
Rolling ‘Cylinder Seismograpl ... 2... 。、 ニー ニー ニドム no 34

The Common Pendulum ......... U... scwoein coesek en te eR

Actual Pendulum Seismometers .................. きい とも CE 37
Long Pendulum Seismograph ーー RER 38
Influence of friction on the Pendulum and other Seismometers ...... 40
Short Pendulum with considerable friction ........0........................ 41
Methods of making a short pendulum astatie ーー …… 42
Duplex Pendulum Seismometer.. ーー ト トト いっ ニッ ーー … ツ ーー

Duplex Pendulum with a single bob ........4....00.0.0 200 sane: on ee

Forbes’ Inverted Pendulum Seismometer .................... He DEE)

CHAPTER IV.

INSTRUMENTS FOR OBSERVING VERTICAL MOVEMENT.

Loaded Spiral Spring EE 47
Horizontal Bar with flexible support and loaded end..................... 4
Horizontal Bar with stretehed supporting spring and loaded end...... 48
Astatie Horizontal Bar, without liquid... ............zuzu02 222220202000 eet ees 49
Actual Vertieal-Motion Seismograph ..... wok lis eee ees: Paras 50

Hydrometer Vertieal-Motion Seismograph ii 51

|
i )
Vil |
| CHAPTER V. |
| RESULTS OF OBSERVATIONS. |
Kur Pp = |
bree LL SEO N en re KO TR Ru re DD cen tide rt tete revs syeas. OD
; 51. Records on continuously moving plates ............ EL の PT 53
| UM DIES iLO) Ca) Seo ate se ee ee ee Ae 54
Da. Barthguake of Wehruaryı 7th; 1881 aa. a ne 55
| 54 | Karihqunke of March’ 8th) 1881 2... lin. 56
| DEIBRATCHEBARBN O2 Marchiiet; TOBDE 5.) お た 27 の 58
He Basthaunke/or March: CUES 1882. nenne (on! age deraneaeaangee 59
: Den beaktnquake Ob March L9G, 18027... Zen crept esac パーン ーー… ーッ ーー 60
| ape Barthguake.of Aupusb 18th; 1882... ue ニー アン ーッ ドー 60
1 59. Earthquake of September 29th, 1882. .…………/…………………………………………………… 61
Oui Parthauake OLA PTL 2drds 1883 9. .c certains sacaieccs nasse ann esse nennen ree 61
SEMESTER OM TESS an Mea cane cnet ent te 2572 ペー ベッ
PON CEG IP ENG. UNE age RR scare Er ER
CHAPTER VI.
1 MISCELLANEOUS INSTRUMENTS.
4 Be KU Tre ON RI DE N nenn len nass sehen RE.
PER TRUM IY ohare dae Ne areas eck ea hecho sip eabe tee Ou eaecg a ON
65. Mallet’s Ball Seismometer .......... ee: 65
66. Pendulums intended to swing ......... une ernennen 69
67. Perry and Ayrton’s proposed Seismograph te
68. Electric Seismoscopes ......... 70
See PA TIEFLB DAWMIG-ADDATALUR) Vin cis あの メル ツー いい ウン ッ se 1
ST OTT CRATE yi Ain Aa ne a eee Rates eal の で CC わい 72
71. Rossi’s Seismoseope and Microseismie Apparatus し ;
EEE CE WEIBTOBSIEIMOBLODRRT «Hin een Vase ade si sds o2ss\csvcavaevetel tous! 200) CO
73. Optieal method of multiplying the displacement of a pendulum ...... 74
9 74. Measurement of earthquakes by reference to their Effects on Buildings 75
CHAPTER VI.
SUMMARY AND COMPARISON OF INSTRUMENTS AND METHODS.
75. Uselessness of all except Steady-point Seismometers 2.2... 76
76. Classification of Steady-point Seismomdters............. nee 76
77. General principle of Steady-point Seismometers に に ーーー ut

ART.

viii

PAGE
Methods of registering the movements of the ground relative to the

Steady Point ......... sda aden oobi, Cena set ee Oe eae eae ae 77
Objections te static Records... 2er ee 77
Records of displacement in conjunction with time ....... ca の 78
Constructive details: !....-...00. 2 en «oer econo don ee 79
Requirements of a Seismological ObservatorY.…… せ … せ …… せ ………… ee 81
Determination of the Direction and Velocity of Transit.................. 83
Velocity of Transit of Artificial Earthquakes Ense are nase TR 86
Experimental Tests of Beismographs.……… せ せ せ …… ーー ドー 86
APPENDIX (ro CmAprrR III): —

Astitio suspension by Winkwork ーー 89

u
1
q
1
_-
d

IS! OF FLATES.

Prare I, Figs. 1-3. — Horizontal Pendulum Seismograph, original form.
Prare 1, Figs. 4-6. 一 Horizontal Pendulum Seismograph, improved form.

Prare IL, Figs. 7 and 8. — Details of Horizontal Pendulum Seismograph.

Prare IV, Figs. 9 and 10. 一 Horizontal Pendulum Seismograph with ring bob.

Prare V, Figs. 13-18.— Horizontal Pendulum without joints.

Prate VI, Figs. 19 and 20.— Horizontal Pendulum with two degrees of freedom
(Gray).

Prare VII, Figs. 11,12, 21, 22, 23 and 33.— Conical Pendulum; Rolling Sphere;
Rolling Segment; Rolling Cylinder; Vertical Seismometer (Gray ).

Prare VIII, Figs. 24-28. — Long Pendulum Seismograph.

Pirate IX, Figs. 29 and 30. 一 Duplex Pendulum Seismograph.

Pirate X, Figs. 31 and 32. — Astatie Pendulum Seismograph.

Prare XI, Figs. 34-37. — Vertical-Motion Astatie Seismograph.

Pirate XII. —Record of Earthquake of Feb. 7th, 1881, given by the Horizontal
Pendulum Seismograph.

Pirate XIII.—Reeord of Earthquake of March 8th, 1881, do. do.
Prate XIV.—Record of Earthquake of March Ist, 1882, do. do.
Prarte XV.—Record of Earthquakes of March 11th 1882, do. do.
Pirate XVI.—Record of Earthquake of March 19th, 1882, do. do.
Prare XVII. —Record of Earthquake of Aug. 18th, 1882, do. do.
Prare XVIII.—Reeord of Earthquake of Sept. 29th, 1882, do. do,
Prare XIX. —Record of Earthquake of April 25rd, 1883, do. do.

Also Statie Reeord of the same Earthquake given by the Duplex Pendulum
Seismograph.
Prare XX.—Record of the same Earthquake given by the Long Pendulum
Seismograph.

Prare XXT, Figs. 38-45.—Circuit-closer (after Palmieri); Diagram of Connec-
tions; Sensitive Cireuit-closer (Milne); Sensitive Circuit-closer (Rossi);
Time-taker (Milne); Driving gear for Record Receivers.

Pirate XXII.—Experimental Tests of the Horizontal Pendulum Seisinograph.

Prare XXIII, Figs. 46-53.—Methods of Astatie Suspension by Linkwork.

INTRODUCTION.

Any sudden mechanical disturbance occurring within or on the surface of
the earth sets up a state of elastic vibration in the substance of the crust, which
is propagated with diminishing intensity throughout the neighbourhood of the
source, in the form of waves of compression, or distortion, or both, This motion
constitutes an earthquake.

The disturbance may also produce, as a secondary result, gravitation waves
in the water of seas or lakes. These admit of easy measurement by means of an
ordinary tide-gauge ; and in the present paper we have to do only with those
vibrations which owe their diffusion to the rigidity and compressibility of the
solid, or the compressibility of the fluid portions of the earth’s erust.

The name earthquake might fairly be applied to all elastic vibrations in sea
or land, without reference to their origin, whether that be the erumpling, tearing,
or slipping of strata, the eruption of a voleano, the collapse of a subterranean
cavity, the explosion of a mine, the rumbling of a carriag
any one of the thousand other events which might be named as causes of a sudden

e, the tread of a foot, or

disturbance of equilibrium, A reasonable but not strictly definable limitation
restricts the use of the term to those comparatively large motions which extend
over a considerable area, and whose immediate causes are natural and, in general,
somewhat obscure.

The comparative suddenness of the movements to which the name earth-
quake is applied serves to distinguish them from other much more gradual elastic
disturbances which are continually taking place in the earth’s crust. It may be
asserted with confidence that tidal deformations of the solid substance of the
earth are caused by the varying attraction of the moon and sun, although their
presence cannot be actually observed because they are masked by other and
irregular movements of the same or a higher order of magnitude. ‘The existence
of these has been demonstrated by the remarkable experiments which Messrs.
George and Horace Darwin have carried out at the instance of a committee
appointed by the British Association for the purpose of investigating the lunar
disturbance of gravity. ‘Their observations show that even when tremors due to
local traffic are eliminated, the solid ground is never really at rest. An increase
of air pressure over any district causes there a dimple or elastie depression of the
soil, while a relaxation of pressure allows the earth’s surface to bulge upwards.
The changing load due to the heaping up and withdrawal of water by tidal
action makes the ground beneath sink and rise. Other more superficial strains

xii

in the soil are caused by alternations of wet weather with drought, and by the
absorption and radiation of solar heat. These and perhaps other causes combine
to produce an incessant yielding of the earth’s crust, so considerable and at the
same time so irregular that the attempt to distinguish in it a periodie part
directly due to lunar attraction has been abandoned as hopeless.

Although the movements just described are due to the elasticity of the earth,
it would be a misnomer to call them earthquakes. The far more rapid tremors
caused by a sudden mechanical disturbance are easily distinguishable from them,
and require wholly different appliances for their detection and measurement.

After freely applying the restrictions which have been indicated, we are left
with an immense number of disturbances to which the name earthquake is
practically applicable. Their magnitudes vary within the widest possible
limits, from the scarcely perceptible movements familiar to all residents in an
earthquake country, to the convulsions which have destroyed cities and changed
the face of a continent. From these, in some form, no part of the earth’s surface
is entirely exempt, and in certain favourable districts they oceur with almost
daily frequency.

Earthquake Measurement, the subject of the present paper, consists in
determining as fully and exactly as possible the character of the motions which
make up an earthquake. If this determination could be pushed back so far as
to include the initial motions of the portions first disturbed, it would involve
diseovery of the originating impulse. It cannot be said that this result has
hitherto been achieved, except in a very partial degree; and our knowledge of
the origin of earthquakes consists chiefly of deductions as to what may be expected
to result from the earth’s gradual approach to a state of thermal and mechanical
equilibrium, along with inferences from what geology tells about ancient and
contemporary disturbances of the crust. This department of Seismology lies
outside the seope of the present paper. It has been deyeloped in a masterly
manner by Hopkins in his Report to the British Association on the Geological
Theories of Elevation and Earthquakes.* In the same paper he has applied the
theory of waves in an elastie solid to the case of terrestrial disturbances, A brief
restatement of the latter part of the theory of earthquakes will not be out of
place here, since it both teaches the earthquake obseryer what to look for and
guides him in the interpretation of his results. It will accordingly be found in
Chapter I; after which instruments for earthquake measurement will be described,
and the results of actual observations be stated and examined,

* British Association Report for 1847, pp. 33-92. 9

6,2

CHAPTER I.

THEORETICAL CONSIDERATIONS REGARDING
EARTHQUAKE MOTION.

$ 1. Vibrations proceeding from a centre within an elastic substance.

When a single sudden disturbance of the most general character occurs at
any place in an indefinitely extended homogeneous isotropic elastic solid, two
waves or states of vibratory motion will proceed outwards along straight lines
from the origin, with two different velocities. If the space in which the dis-
turbance originates is very small compared with that into which the waves
subsequently diverge, we mi Ly without sensible error consider the wave-fronts as
spherical surfaces hi aving the origin for centre. By wave-front is meant any
surface forming the /ocus of particles which are simultaneously moving in the
same phase. The line of transit, or direction along which each wave travels, is
normal to the wave-front. One of the two waves consists of compression and
dilatation of the material in the direetion of the line of transit. This, which is
propagated with the higher velocity of the two, is ealled the Normal wave: its
passage causes each particle to move in the direction of the line of transit of the
wave. The other wave, which travels more slowly, consists of distortion of the
material unaccompanied by any change of volume. It is called the Transverse
wave, and its passage eanses each particle to move at right angles to the line of
transit of the wave. The path in which a particle moves must be rectilinear so
far as the normal wave is concerned; but the transverse wave may give rise to
motion in any curve lying in the plane tangent to the wave-front. In neither
case is it necessary that the path should be elosed. This will happen only when
the originating disturbance is eyelic : more generally a particle will be found
permanently displaced from its initial position after the wave has passed,

We can easily imagine an originating impulse of such a character as would
give rise to either wave alone, but in general both waves will be produced.
Again, the properties of the disturbed medium may be such that one or other
wave is impossible: thus a transverse wave cannot he generated in or transmitted
by fluids owing to their want of elasticity of form; on the other hand an ineom-
pressible medium cannot be the seat of normal waves. ‘The transmission of sound
by the air illustrates the existence of normal waves in a substance capable of
them only ; while the transmission of radiant energy by transverse waves in the

luminiferous ether exemplifies the other case.
1

SN 2. Velocities of Transit.

The velocity of transit of a wave through an elastic substance is
IM
Neat
i

where 。 is the density of the material and M is the modulus of elasticity for the
particular kind of strain of which the wave consists. For transverse wayes in an
isotropic solid M is the modulus of rigidity, which, following the notation of
Thomson and Tait (Nat. Phil. Chap. VII.) we shall denote by ». The strain
involved in the transmission of the normal wave consists in the compression and
expansion of spherical shells. At any point not near the origin, this approximates
very closely to simple longitudinal strain (extension or compression in one diree-
tion without change of lateral dimensions). For this the modulus is た 上 まめ
where た is the modulus of volume elasticity, or reciprocal of the eubie compress-
ibility, and n has the same meaning as before. The quantities n and た have no
ascertained relation to each other, and must be separately determined by experi-
ment for any given substance,

As both た and » are essentially positive the modulus for the normal wave
is the greater of the two, and hence the normal wave always travels faster than

the transverse wave, the ratio of their velocities being ————*—. The two

waves will become more and more separated the farther they spread,

When, as will often be the case, the originating impulse is not single but
consists of a number of successive displacements, there will be two groups of
waves, normal and transverse. Each group will spread in the form of a spheri-
cal shell whose thickness is equal to the velocity of propagation of the corres-
ponding type of waves multiplied by the time during which the originating
disturbance has lasted. At a sufficient distance from the origin these two groups
will, on account of their different velocities of propagation, he completely
separated ; and a particle there will make first a series of oscillations in the
direction of the line of transit of the wave, and immediately afterwards a second
series in directions perpendicular to the first. A particle more distant from the
origin will experience two corresponding sets of moyements with an interval of rest
between them. On the other hand, at points nearer the origin the two groups
will be more or less superposed, but the initial displacement of a particle any-
where will be along the line of transit, provided that any normal wave is
transmitted whose origination did not ocenr later than that of the earliest trans-
verse waye. The initial movement may be either away from or towards the
origin, according to whether the initial strain there is one of compression or
dilatation. Provided that none of the displacements are large and the medium
is perfectly elastic, the speed of transit of all the separate waves in each group
will be the same.

ーー

$ 3. Effect of boundaries. Reflection and Refraction.

When a wave, whether normal or transverse, meets a surface separating the
medium in which it is travelling from another medium, it will in general suffer
reflection and refraction. The lines of transit of the reflected and refracted
waves will lie in the plane which contains the line of transit of the incident wave
and the normal to the separating surface.

The simplest case is that in which both media are fluid, and consequently
the incident, the reflected and the refracted waves are all of the normal type. In
the figure, let AB be the bounding surface between two fluid media, and OP
the direction of propagation of the incident wave. MN is drawn normal to
the surface AD.

N ノ @
4 aK B

A wave is reflected back int» the first medium along the line PQ, and
another wave is transmitted into the second medium along a certain line ?’Q..
The angle OPM or # is called the angle of incidence, Q,?N or 9, is called the
angle of refraction and QPM the angle of reflection. QPM is equal to OPI,
while 9, and 9 are connected by the equation

V Ne
sind sin UM
where V is the velocity of transit of the incident wave, and V, is the velocity of
transit of the wave transmitted into the second medium.

In the case represented in the figure V, is less than V, and there will be a
refracted wave whatever be the angle of incidence, If the velocity in the second
medium were the greater of the two, the refracted wave would be bent away from
instead of towards the normal, and by increasing the angle of incidence sulli-
ciently we should have PQ, coinciding with PB. This would occur when
i 1

take place; the incident wave would be restored to the lower medium without

7 = sin At this or any greater angle of incidence total reflection would

loss of energy, and with no other change than a change of direction and a change

of phase,

In the more general case, when a refracted as well as a reflected wave exists,
if we call the amplitude of displacement in the incident wave unity, the ampli-

tude of the reflected wave is

Py cot 7
p on
p _cot 0

where » and p, are the densities of the first and second medium respectively.”
Also the amplitude of the refracted wave is
2 sin 7

sind,
Py 4 _cot 0,
p cot 7
sin 0 .
For — we may write ‚ and
sin 0, . 1
cot 7 V Vi
= ーー = tan? 0.
cot 7 v, WM ; (7 1) si

$4. Example: Case of an Earthquake reflected at the surface of a sea or lake.
Observations of the velocity of sound have shown that the speed of transit
of elastic waves is about 1435 metres per second in fresh water, and about 332

7
metres per second in air. Hence Y for an earthquake reaching the upper sur-
1
face of a sheet of water is 4.3, and
cot 7

= 4.3 v1 + 0.947 tan? 7.
cot の

‚pi . » a,» Ka, = デュ . .
he ratio of densities ‘' is about 0.0013. These data give for the ratio of the
アル

amplitude of the refleeted wave to the amplitude of the incident wave the value
1 — 3300 Y 1 + 0.947 tan? 7

1 + 3300 v 1 + 0.947 tan? の
By writing this
3,

I -+ 3300 1 1 + 0.947 tan? の i
we see at once that the amplitude of the reflected wave is a minimum when the

path of the incident wave is perpendicular to the boundary, and inereases to
unity as the angle of incidence approaches 90°. Even in perpendicular incidence,
however, the reflected wave has an amplitude of 0.9994, that of the incident
wave being unity, From this, and from the fact that the energy of the two
waves is as the square of their amplitudes, it appears that an earthquake is in all
cases reflected from the free surface of water with exceedingly little loss.

* Nee Green's Mathematical Papers, p. 238, or Lord Rayleigh’s ‘* Sound ’’, § 270.

The amplitude of the wave transmitted into the air is
8.6
0.0013 + 4.31 1 + 0.947 tan? 0”
which is a maximum (equal to 2 very nearly) when the wave strikes the boundary

normally, and diminishes to zero as @ is increased to 90°.

The example will serve to show how, on account of the great difference in
density and elasticity of the two mediums, while an earthquake is almost per-
fecily reflected when it reaches the atmosphere, it nevertheless may cause acrial
vibrations of considerable magnitude, which, if the waves follow each other with
sullicient frequeney, will be perceived as sound,

$ 5. Reflection and Refraction at the boundary of two solid mediums.

The changes which elastic vibrations undergo at the surface of separation of
two solid substances are of a more complex kind than those described above.
The incident wave may be of either the normal or the transverse type. In either
case there will, generally, be two reflected and two refracted waves, one of each
pair being normal and the other transverse. The relation of yelocity and direction
is most simply expressed by saying that the velocity with which the trace of each
of the five waves (one incident, two reflected, and two refracted) moves over the
plane of separation of the mediums must be the same for all. The velocity of the
trace is the velocity of the wave divided by the sine of the angle which the line of

transit makes with the perpendicular to the boundary, and henee 。 has the
sin

same value for every one of the five waves eoneerned. In the figure, let OP be
as before the direction of incidence, AB the boundary and ALN perpendicular to it.
From P set off a distance PR equal, on any convenient scale, to the velocity of
transit of the incident wave divided by the sine of the angle of incidence. On
PR as diameter describe a eirele. From P as centre describe ares with radii equal

successively to the velocity of (1) normal waves in the first medium, (2) transverse
waves in the first medium, (3) normal waves in the second medium, (4) transverse
waves in the second medium. Let these ares cut the eirele in Q, Q’, Q,, and Q,/
respectively. Then PQ, PQ’, PQ,, and PQ/ represent in direction as well as in
velocity of transit the four waves which result from the incident wave OP. In
the figure the ineident wave has been assumed to be of the normal ty pe: the same
construction with a different length of PR will apply to determine the disturb-
ances produced by the incidence of a transverse waye.

In special cases the system of waves which proceeds from the boundary
will be less complicated: thus, in the ease of perpendicular incidence, each incident
wave will give rise to only one refracted and one reflected wave, both of the same
type as the first. And, generally, any transverse wave in which the displace-
ments are at right angles to the plane of incidence will be refracted and reflected

as a transverse waye only.

$ 6. Waves originating in or transmitted into an wolotropic solid.

When the medium in which the disturbance originates is »olotropie with
respect to its elasticity three different modes of vibration will in general be set up,
and propagated with three different velocities.* And consequently when waves,
of whatever type, impinge on an wolotropie medium, each incident wave will in
general cause three refracted waves to be transmitted in different directions and
with different velocities.

ST. Efieet of a stratum in veduplieating waves.
Let ABCD be a stratum with plane parallel sides lying between two other

medinms, in one of which (that on the left of the figure) a plane wave is advanc-
ing in the direction OP or O'R.

A D

* Sir W, Thomson, Ene. Brit. Ninth Edition, Art. ‘‘ Elasticity ’’, Chap. XVII.

At any station S within the stratum or S’ in the medium beyond, the first
wave to arrive will be the direct result of the refraction of the incident wave at
R. A little later a second wave will arrive, due to a different portion of the
same original wave-front, which has struck the stratum at P, been partially re-
flected at Q and again partially refleeted at R. The interval of time between the
first wave and this one will be the tine taken for the wave to travel in the
stratum from p to Q and from Q to R, p (determined by drawing a perpen-
dicular from 2 on PQ) being the point already reached in the line PQ by the
wave-front アガ when it strikes the stratum at R. And by extending the
same construction it is obvious that a third wave will reach the observing station,
after suffering four reflections, a fourth after suffering six reflections, and so on,
the interval between the successive arrivals being the same as that stated above,
namely

ーー V,
where V, is the velocity of transit through the stratum.

We thus find at any point to the right of AD, as the result of a single in-
cident wave, an indefinite series of waves of the same type, having the same
direction, with a constant period て and with amplitudes diminishing in a geome-
trical progression. Writing 6 for the perpendicular thickness of the stratum, and
7 for the angle which the waves in it make with the normal to its sides, we have
2b sin 70,

. H ence
cos 7

b
PQ = - and Pp ーー
cos 4, ?

2b cos 6,
So ae |

A
lI

2h
¥,

perpendicular to the stratum. The amplitude of each wave is less than that of

which is a maximum, equal to when the path of the incident wave is

its predecessor in the ratio nn’: 1, where n: 1 is the ratio in whieh a wave in the
stratum is reflected at CD, and n’: 1 is the ratio in whieh a wave in the stratum
is reflected at AB.

Asa simple example, take the case of a single wave of compression and
dilatation ascending vertically through a sheet of still water of a uniform depth
of 10 metres, A series of doubly reflected waves will follow the first at a

ee , 2X10 Lys に - F >
uniform interval of 。 or. of a second. The rate of decay of amplitude in
1435 12 .

the series will depend almost wholly on the degree of imperfection of the internal
reflection at the bottom of the water, since reflection is almost perfect at the upper
surface (§ 4). The ratio of the internally reflected to the incident wave at the

. . . 7 . ;
bottom will be considerable provided that f is not nearly equal to Y (§ 3), and
pP i

in actual cases it will probably be about 4 5. In these conditions a continuous
sound may be produced with rapidly diminishing loudness, and be audible for an

appreciable time. More commonly, however, the secondary vibrations due to
double reflection in a stratum will have too long a period to give rise to a con-

tinuous sound,

If the sides of the stratum are not parallel, the period and also the direction
of the secondary waves observed at any station in or beyond the stratum will
change progressively, and a like effect will be produced even by a uniformly
thick stratum when the front of the incident wave is not plane but curved, The

figures given above illustrate these cases.

S 8. Effect of obstacles. Earthquake Shadows and Diffraction.

When waves travelling through an otherwise homogeneous medium impinge
upon a space occupied by matter whose mechanical properties (as to density,
compressibility or rigidity) differ from the properties of the medium round it, the
foreign body moves so as to act as a new source of disturbance, and secondary
waves are thrown off from it in all directions. These are of course generated at
the expense of the energy of the original system of waves.

When the dimensions of the foreign body are sufficiently increased, reflec-
tion and refraction take place at its surface according to the principles already
stated. If we suppose it to be impervious to waves of the particular class which
strike it, it will act as an obstacle to shield from disturbance a portion of the
medium lying beyond it, and may therefore be said to cast an earthquake shadow.

The geometrical outlines of the earthquake shadow will of course be deter-
mined by drawing lines parallel to the direction of the incident waves and
tangent to the sides of the obstacle: but neither will this region be completely
shielded from disturbance nor will the surrounding portions of the medium be so
intensely disturbed as they would be were the obstacle absent. Im other words

& Pert

there will in general be a great deal of diffraction: the limits of the shadow
will be ill defined, and unless the obstacle is very large or the observing: station
very near it, a complete absence of disturbance will rarely be found. ‘This is because
of the comparatively great wave-length of earthquake vibrations: to cast a clear-
ly defined shadow the dimensions of the obstacle measured transverse to the path
of the incident ray must be great compared with the wave-length of the constitu-

ent vibrations—a condition difficult to find fulfilled when we are dealing with

the long-period waves which make up an earthquake.

$ 9. Energy of Vibrations.

The energy which must be expended in producing an elastic wave, and
which is carried by the wave as it spreads, depends amongst other things on the
form of the wave—that is to say, on the relation of displacement to time during
the passage of the disturbance. In a wave of the simple harmonic type, the
energy per unit of area of wave front is

IE p Va

sense

where # is the mean density of the medium, V the velocity of transit, @ the
greatest displacement of a particle from its undisturbed position, and z the period
of the wave. When a spherical wave spreads in a uniform and perfectly elastic
medium, its energy remains constant but is distributed over an area which
increases with the square of the distance + from the origin, and since the other
quantities are constant (except very near the source), the amplitude of vibration
will diminish as the distance from the origin increases, or

の oc o>.
When, however, the disturbance oceurs in a stratum whose properties differ much
from those of adjoining strata, so that little of the energy of the earthquake is
transferred to them, and the vibrations spread nearly in one plane instead of
spherically, the energy per unit of area of wave front will diminish nearly as the

distance from the origin increases, and
1

@ coe approximately.
There is some evidence that a disturbance occurring on the surface of the soil is
in certain cases propagated to surrounding parts of the surface more nearly in
this manner than by spherical distribution.

When reflection and refraction occur at the boundary of two mediutos, the
energy of the incident wave will be shared by the refleeted and refracted waves
without loss, except in the special ease where the adhesion of the two mediums is
insuflicient to prevent relative sliding during the passage of the vibrations, when
a portion of the energy will he dissipated by friction at the boundary, A similar
dissipation will occur at a fissure in a single medium, if the parts separated by
the fissure do not eohere firmly enough to prevent relative motion. Another
cause of loss of energy by the principal waves of an earthquake is the heteroge-

neity of the vibrating medium, in which every local variation of mechanical

quality causes the places where such variations occur to be originators of secon-
dary waves (§ 8).

S 10. Eject of Imperfect Elasticity.

The fact that the elasticity of most of the substances which make up the
earth’s erust is imperfeet operates powerfully to diminish the energy and therefore
the amplitude of earthquake motions as the vibrations spread from the souree,
and consequently to limit the area throughout which any one disturbance is felt.
The earthquake energy is expended in overeoming internal friction within the
vibrating medium, and takes the form of heat. This internal friction is in
general of two kinds: in one—the viscosity of fluids—the resistance to relative
motion of adjoining parts depends essentially on the velocity of that motion,
being, for low velocities, proportional to the velocity. The other is of a static
character ; it resembles the friction between solid bodies in being nearly indepen-
dent of the velocity with which the changes of form take place, and, in any one
substance, depends chiefly on the extent of the deformation. If we cause a
viscous fluid to undergo a eyclic change of form or volume, an amount of energy
will be dissipated which will be (approximately) proportional to the rapidity of
the change, and will be zero if the change takes place infinitely slowly. If we
perform the same operation with an imperfectly elastic solid, an amount of
energy will be dissipated which will be nearly independent of the rate of de-
formation, and will not vanish when the change takes place infinitely slowly, but
will be greater in proportion to the whole amount of work involved in the
operation the greater the amplitude of deformation is, since the body’s elasticity
is more nearly perfect for small strains than for large ones.

The effect of viscosity on vibrations propagated by a fluid has been examined
mathematically by Stokes*, who has shown that besides reducing the amplitude
more rapidly than the spreading of the wave involves, it makes the velocity of
propagation less than it would otherwise be. Both effeets are greater for waves
of short period than for waves of long period.

The writer is not acquainted with any investigation of the effect of imperfect
elasticity on vibrations transmitted by solids. If, as appears probable, its effeet is
to reduce the velocity of propagation, as well as the amplitude, we should expect
to find this reduction less for waves of small amplitude than for waves of large
amplitude, since the strains involved in waves of small amplitude would lie
within or nearer to the so-called limits of elasticity, inside of which the substance
obeys Hooke’s Law, And further, in a group of waves whose amplitudes were
unequal, the tendency would be for those whose amplitudes were large to lose
more energy than the others in proportion to their whole stock, and hence
differences of amplitude would be reduced. Thus a series of waves origi-
nated by double reflections in passing through a stratum (§ 7) would tend to have

* On the Internal Friction of Fluids in Motion, ete. Cambridge Transactions, Vol. VIII:
or Reprint of Papers, Vol. I, p. 101.

their amplitudes to some extent equalized by passing subsequently through a
medium whose limits of elasticity were so narrow as to be exceeded by the strains
the waves caused. The elasticity of the medium might however be nearly perfeet
with respect to the very minute vibrations which accompany or originally follow
the principal shock, and these would therefore be transmitted with comparatively
little loss, and (if the above conclusion is correct) with a velocity greater than that
of the principal movements. Moreover, it appears probable that a medium such as
clay or semi-solid mud will transmit vibrations of short period more nearly like
a true solid, and vibrations of long period more nearly like a viseous liquid, and
that the former will be propagated with greater velocity than the latter. It is a
fact of common observation that the sound which in many cases accompanies an
earthquake is usually heard before the arrival of the separately perceptible un-
dulations of the soil; and the writer’s observations have shown that the records
given by seismographs frequently exhibit minnte vibrations of short period at the
beginning of earthquakes, either entirely preceding or superposed on the earliest
principal movements, whose periods are longer and which continue to be traced
after the short-period movements have died out.

$ 11. Movements of a particle during an Earthquake.

The actual constitution of the earth’s erust is so far from being homogeneous,
even on a large scale, that any disturbance propagated through it must be greatly
modified by the numerous reflections, refractions, and diffractions which occur
along its route. Even when a single impulse proceeds from a single originating
point, it is obvious from what has been said above that in general many waves
both of compression and of distortion will reach an observing station distant from
the souree, travelling in different lines. The direction of the principal normal
impulse will not in general be that of a line joining the observing station with
the souree, Its line of transit will usually be bent in a vertical plane, both
sharply by passage through surfaces which separate different strata, and gradually
on account of continuous changes in the elasticity and density of a single stratum,
produced, in the upper layers especially, by gravity and by the presence of mois-
tnre in the soil. And in many cases this bending will take place in a horizontal
plane also, with the result that the azimuth of the principal normal displacement
which oceurs at a distant station will not coincide with the bearing of the origin.

When, in addition to the considerations which have been indicated above, it
is borne in mind that the place of origin of an earthquake may itself be widely
extended (as, for example, in shocks which are caused by the formation of a fissure
or “fault” of some length), and that each part of the initially disturbed region
will generally originate not one but many successive impulses, it is clear that the
movements which may be expected to occur at an observing station are complex
to an almost indefinite degree.

To completely determine the motion of the ground at any one point we

must observe, from the beginning to the end of the disturbance, the displacement

of a particle (measured from the position oceupied before the disturbance began)
both in magnitude and direction, in conjunction with the time, To this end it
is usual, and for most purposes desirable, to treat each displacement as the re-
sultant of three rectangular components, one vertical and two horizontal. Tf the
relation of the displacements along three axes to the time be recorded, we possess
complete information as to the movement of the ground at the observing station,
By combining the records for three axes we may of course deduce the absolute
direction of motion of a particle, its velocity, and its rate of acceleration at any
instant during the earthquake ; while by comparison of the records taken at one
station with corresponding records taken at other stations it is (theoretically, and
in a few cases practically) possible to determine the velocity and direetion of

propagation of the earthquake waves.

N 12. Motion at the surface of the soil due to a normal wave.

At any point well within a uniform vibrating mass the direetion of motion
of a particle during the passage of a normal wave coincides with the line of
transit of the wave (§ 1). But when a wave of compression and dilatation ap-
proaches a boundary of the medium in which it travels, in any other direction
than perpendicular to the boundary, the movement of particles there will no
longer coincide with the line of transit of the wave. If the neighbouring sub-
stance, from which the medium in question is separated by the boundary, differs
from it in density or elasticity, the portions of the original medium which lie near
the boundary will be either more or less free to expand and contract laterally
during the passage of the wave than other portions. To take the important prae-
tical ease, a portion of the soil at or close to the surface of the earth is more free to
expand laterally when it is compressed by an obliquely incident normal wave, than
other portions lying in the same wave front, and henee the strain at the surface
caused by a plane wave (or spherically divergent wave from a distant source)
will not be a simple longitudinal strain, but, since a state of compression coexists
with a state of displacement away from the origin, the displacement of a surface
particle during the passage of the phase of compression will be upwards as well
as outwards along the normal to the wave. The direction of the particle’s motion
will therefore be more nearly vertical than the line of transit of the wave—a fact
which prevents us from inferring the direction of transit from a knowledge of
the direction in which a particle at the surface is initially displaced.

my
Py,”
7 \
シー ンズ バ 8

Thus in the figure, when the shell S is compressed by the passage of a
normal wave whose source is at O, a particle P at the surface is displaced along
Pm because of the lateral expansion of the soil, as well as along Pn; and again
when the shell AB is dilated, the particle is displaced downwards as well
as along PO, and hence its direction of oscillation lies between Pn and Pm.

$ 13. Preliminary Statement of the Results of Observations.

Automatic records given by seismographs confirm what has been said, on
theoretical grounds, as to the complexity which earthquake motions may be
expected to present. They show that, as observed at a station on the surface of
the earth, an earthquake consists of a very large number of successive vibrations
—in some cases as many as three hundred have been distinctly registered. These
are irregular both in period and amplitude, and the amplitude does not exceed a
few millimetres even when the earthquake is of sufficient severity to throw down
chimneys and erack walls, while in many instances the greatest motion is no more
than a fraction of a millimetre. The periods of the principal motions are usually
from half a second to a second, but, as has been already said, the early part of the
disturbance often contains vibrations of much greater frequency, The earthquake
generally begins and always ends very gradually, and it isa noteworthy fact that
there is in general no one motion standing out from the rest as greatly larger
than those which precede and follow it. The direction of motion varies irregu-
larly during the disturbanee—so much so that in a protracted shock the horizontal
movements at a single station occur in all possible azimuths. The duration, that
is to say the time during which the shaking lasts at any one point, is rarely less
than one minute, often two or three, and in one case in the writer’s experience
was as much as twelve minutes,

A more particular account of the actual results of earthquake measurement
will be given in a later chapter. The facts just stated are deduced from the
writer’s observations, made at the University of Tokio, and they describe, in
brief, the characteristics of the moderate earthquakes which occur with great
frequency in the Plain of Yedo, occasionally with destructive effect, but much
more commonly without doing any damage or attracting more than a momentary
attention from the inhabitants. ‘This outline statement of actual results has been
presented here as affording data by which we may more readily discuss the practi-
cal value of various seismometers. Much time and labour has been wasted in
seismometry through false preconceptions as to the character of earthquake
motion. Many instruments have been constructed under the idea that an earth-
quake consists essentially of one relatively large impulse, easily distinguishable
from any minor shakings which may accompany it. How far this may be true
of great earthquakes, or of earthquakes in other localities, it is as yet impossible
to say; nothing could be less true of the earthquakes we experience in Japan,
whose frequeney and manageable size make them good subjects for measurement.

CHAPTER IL.

INSTRUMENTS FOR OBSERVING HORIZONTAL MOVEMENT :—

THE HORIZONTAL PENDULUM SEISMOGRAPH.

$ 14. Measurement of carihquake motions by reference to a Steady Point.

Instruments designed for the measurement of earthquake motions aim, with
few and unimportant exceptions, at giving a point which does not move during
the disturbance, and which will therefore serve as a datum with respect to which
the movements of the earth may be determined. No actual instrument gives a
vigorously steady point throughout a prolonged disturbance, but certain devices
which will now be deseribed achieve the desired object nearly enough for prae-
tical purposes, and it is by them that our exact knowledge of earthquake motion,

so far as we have any exact knowledge, has been gained.

$ 15. Kinetie condition yielding a Steady Point.

If any one point of a rigid body suffers a displacement in the line joining it
with the centre of inertia of the body, the whole mass will share the same motion,
which is then one of pure translation; but if the point suffers a small displace-
ment in any dircetion at right angles to this line, the motion of the body will
consist of rotation about an axis (called the Instantaneous Axis) which intersects
the prolongation of the line joining the displaced
point with the centre of inertia, and is perpen-
dicular to the plane containing that line and the
direction of displacement. Points situated in the
instantaneous axis remain at rest, and all other
particles of the body move through distances pro-
portional to their distance from the instantaneous
axis. Thus, in the figure, let the mass sketched,
whose centre of inertia is at G, be pivotted at
P. When P is moved in the direetion GP or
PG, all points of the mass are equally displaced:
but when P moves through a small distance in
any direction PX perjendieular to PG, the mass
will revolve about a certain axis ZZ, which is the
instantancous axis for the assumed (infinitesimally
small) displacement. 77 is in the plane GPY,
PY being drawn perpendicular to PX and PG,

and it euts the line PO produced at a point Q such that
I?
PG

where % is the radius of gyration of the mass about the axis ?Y. Q is called the

centre of percussion relative to the axis PY.

If the mass has kinetic symmetry about the line PG, so that た has the same
value for all axes passing through P and perpendicular to P’G, then the instant-
aneous axis corresponding to any other displaccment of 2? perpendicular to 7
will also pass through Q. Thus, for example, a small displacement of P along
PY will cause rotation of the mass about the axis JJ. In this case the point Q
will (so far as its motion depends merely on the inertia of the mass) remain at
rest during any small displacement of the point P in any direction in the plane
XPY; and Q may therefore serve as a Steady Point by reference to which such
displacements may be measured.

The fact that in earthquake motions the displacements of the pivotted point
P are not infinitesimally small does not affeet the result just stated to any prac-
tical extent; it is suflicient that the displacement should he small compared with

the length OP—a condition easily satisfied in practice.

$ 16. Static condition requisite for a Steady Point.

If no forces acted on the mass other than those producing displacements of
the pivetted point P, the statement of the last paragraph would be true without
reservation—that any single small movement perpendicular to ’G would leave
a certain line in the mass at rest, and that in the case of a mass with kinetic
symmetry about /’G, a certain point Q would be left at rest whatever motions of
P were caused to take place in the plane perpendicular to PG, provided these
motions were small.

But in attempting to apply this principle to earthquake measurements we
must take account of the existence of other forces which inevitably act on the
pivotted mass, notably of gravity. Assuming the mass to be in equilibrium
before a displacement of 2 occurs, it is obvious that, in order to the maintenance
of (/ as a steady point, the equilibrium must be neutral with respect to the dis-
placements now under consideration. If it be unstable the displacement of P
will cause an increasing disturbance of the mass as a whole. If it be stable the
assumed displacement will cause oscillations which, if the displacement of P
occurs periodically, may attain a magnitude so great as to deprive Q of all claim
to be called a steady point.

If, for example, the point P at which the (otherwise free) mass is pivotted be
fixed to the surface of the earth, the mass may be placed in equilibrium in two
positions—either as an inverted pendulum with G vertically above 7’, or as a com-
mon pendulum with G vertically below P. The former arrangement gives

unstable equilibrium and is of course impracticable for seismometry without

important modification. In the latter case the equilibrium is stable, and any hori-

zontal displacement of P will introduce a couple due to gravity, tending to make
2 follow the movement of P. The pendulum swings, with the result that Q, far
from remaining at rest, may sometimes acquire a movement much greater than
that of P itself.

We shall see later that by combining a common with an inverted pendulum
we may obtain a steady point in neutral equilibrium, capable of being used for
the measurement of motions in any horizontal direction, and that the same result
may be arrived at in other ways. ‘To obtain a steady point with respect to
movements in one direction only is, however, a simpler problem, the solution of
which will be deseribed first.

$ 17. Horizontal Pendulum Seismometer.

Instead of having only one point fixed, let the body be pivotted in such a
manner that a certain line (say the line POP in the figure below) is constrained
to remain at rest, freedom being left to rotate about this line. Further, let this
fixed axis be vertical and attached to the earth’s surface: the body then forms a
horizontal pendulum,

Zr Zi
Ip 1
re,
1 le
IS ig
% HS

Then if an earthquake movement occurs either vertically, or horizontally
along OG, the pivotted mass simply suffers displacement as a whole. But if the
ground moves through any small distance horizontally and perpendicular to OG,
the mass rotates through a correspondingly small angle about the vertical axis

i?
OG
equilibrium of the mass being neutral, no unbalanced forces are brought into

II cutting OG produced in Q, so that OQ = as before. More than this, the

action by the change of position, and there is consequently no reason, statie or
kinetic, why the line 77 should move either during or after the displacement of
the axis of support. It remains unaffected by the disturbance, and any point of
it may be taken as a datum by reference to which the motion of neighbouring
hodies fixed to the earth’s surface may be determined.

To obtain a complete measurement of any horizontal motion of the ground,

it is only necessary to use a pair of similarly pivotted and independent masses,
which may most conveniently be set at right angles to each other, so that each

will give a steady point (or rather line) with respect to that component of the
actual horizontal motion which causes bodily translation of the other.

We cannot, however, apply the same device, without an important addition
which will be deseribed in a later chapter, to the measurement of vertical motions,
since for their registration the mass must be left free to move in the direction in
which gravity acts.

$ 18. Friction Error. Increase of Effective Inertia by the use of a
second mass pivotted on the first at the Steady Line.

Tn an actual instrument the friction at the joints by which the mass is
pivetted on the line PP, and also at the recording index, when one is used to
give a permanent trace of the disturbance, prevents the line 77 from remaining
strictly at rest. The friction at the joints will tend to make the mass acquire
during displacement less angular motion than is necessary to keep 77 at rest ; and
hence the practical instantaneous axis, which does remain unmoved, will lie far-
ther away than 77 from the axis of support. Consequently a measurement based
on the assumption that the line 77 is steady will err by being too small. Friction
at the point of the recording index will introduce an error of the same kind.
If these errors were uniform it would be easy to ascertain and allow for them,
but they will in general vary in different movements and even in different parts
of a single movement, since they depend on the ratio of the forces due to friction
to the resistance which the pivotted mass offers to angular acceleration. Hence
the friction error (with a given arrangement of joints and marking pointer) will
be great when a displacement occurs slowly, but comparatively small in sudden
disturbances, and practical experience has shown that in many cases earthquake
movements are so slow as to make the elimination of considerable friction errors
a matter of great difficulty. The effective inertia of the suspended mass must
be inereased, and, relatively to it, the friction reduced as far as possible.

The most effective position for the material of the pivotted mass is at the
instantaneous axis, and accordingly we may improve the steadiness of the line 77
by pivotting there in neutral equilibrium a second mass or bob, free to turn about
the line //, and therefore equivalent to an equal quantity of matter concentrated
at that line itself. Moreover, the mass of the original body pivotted at PP
(which is less advantageously distributed) may then be diminished as much as we
please, or rather as much as is consistent with rigidity. In doing this, care must
of course be taken to preserve the kinetic relation between the axes アア and II
unchanged. And if the mass of the original pivotted body or frame, as we may
now call it, since its function is merely to furnish an axis of support for the bob,
he very small compared with the mass of the bob, a small error in the position
of II will not affect the steadiness of that line sensibly. With this disposition of
parts we secure the maximum of effective inertia for a given total mass and a
given distance of the instantaneous axis from the axis of support.

$ 19. Determination of the Effective Inertia and Steady Line.

In estimating the effective inertia and steadiness of a system composed of a
frame, or piece pivotted to the earth, and a bob, or piece pivotted to the frame at
its corresponding steady axis (passing through the centre of percussion), it is
convenient to remember that the frame is equivalent, kinetically as well as
statically, to a pair of masses, one concentrated at the axis of support and the
other at the centre of percussion.

Calling

M, the mass of the frame ;
M, the mass of the bob ;
r the distance of the centre of gravity of M,from the axis of support ;
v the distance of the centre of percussion of the frame from the axis
of support;
then M, is equivalent to a mass = M at the centre of pereussion, together with
=
a mass (1 一 ) M, at the axis of support. Of these the former alone contri-
butes to the steadiness of the system. The bob is equivalent to a particle of
mass M, concentrated at the axis about which it is free to turn on the frame; and
the whole effective inertia (or in other words the mass referred to the steady line) is

1 tM,
r

If instead of being pivotted the bob were fixed to the frame, we should require
to treat the whole mass as “frame”. The effect would be to shift the steady
line outwards, since the moment of inertia of the system about the axis of support
would be increased. The difference in general effect between a frame with
pivotted bob and a frame consisting of supporting piece and bob rigidly connected
is considerable if the transverse dimensions of the bob are comparable to its
distance from the axis of support, and in that case we do well to pivot the bob;
but when the bob is small and the horizontal length of the frame is great,
pivotting the bob brings little advantage to counterbalance the complication of
parts which it involves.

A convenient experimental way of finding the steady line of the frame is to
hang it up vertically, with the ordinary axis of support placed horizontally. Then
let it be swung through small ares as a common pendulum, and its period (て)
measured, The equivalent simple pendulum is one whose length is 7’, the
distance from the axis of support to the centre of percussion, and therefore

$ 20. Multiplicaticn and Registry of the motion.

In those earthquakes which both from their frequency and moderate charac-
ter are best suited for registration, the actual motions are usually so small that to

obtain distinct indications we require some mechanical multiplication of the
displacement of the ground relatively to the steady line or steady point. In the
horizontal pendulum seismometer the simplest way to effect this multiplication is
to make a portion of the frame (§ 18) consist of a light rod projecting to a con-
siderable distance beyond JJ, and to use its end as a pointer which inseribes the
record on a fixed or moving plate of smoked glass. Then when a small displace-
ment of the axis of support アア takes place perpendicular to the plane of PGP,
while 77 remains at rest, the marking point moves in the opposite direction
through a distance equal to the displacement of the ground multiplied by

Distance of Marking Point from Instantaneous Axis

Distance of Instantaneous Axis from Axis of support.
But at the same time the plate on which the record is traced has been carried,
along with other neighbouring objects on the earth’s surface, in the direction of
and by an amount equal to the displacement, and hence the length of the line
traced on the smoked-glass plate is the sum of these two movements, and is equal
to the actual displacement multiplied by
Distance of Marking point from Axis of Support

Distance of Instantaneous Axis from Axis of Support.
Or, more simply, we may conceive a displacement opposite to that of the ground
to be superposed on the actual motion of every point. Then the axis 77 is to be
taken as moving while objects fixed to the earth remain at rest, and it is at once
evident that the above is the ratio of multiplication.

In the arrangement just described the pointer is rigidly attached to the

frame, and must therefore be included as forming part of the frame in caleulating
or finding experimentally the position of the steady line.

$ 21. Multiplication by an independent lever.

Aithough in the horizontal pendulum seismograph a multiplied record is
very conveniently got by using as multiplying lever the pivotted frame of the
pendulum itself, it is frequently desirable in other instruments (and even in the
horizontal pendulum, if the length of the pendulum is considerable) to magnify
the motion by the use of an independent lever, pivotted to the ground, and having
its short arm connected to some point in the suspended mass, while its long arm
carries the tracing point.

In order that such a lever may not affect the steadiness of any seismometer
to which it is applied, it should in strictness conform to the following condi-
tions, the first static and the second kinetic :—(1) together with the suspended
mass to which it is applied, it must form a system whose equilibrium is neutral ;
(2) its point of connection with the principal mass must be related to its own
dimensions and point or axis of support in such a way that, when a disturbance of

the ground takes place, no stress due to the inertia of the parts will be introduced

The simplest way of securing the second condition is to make the point of

connection the centre of percussion of both pieces; in other words, to make the
multiplying lever touch, at its own centre of percussion, the steady point or steady
line of the scismometer to which it is applied. If, however, we prefer to apply
the multiplying lever elsewhere than at the steady point of the seismometer, we
may still fulfil the kinetic condition rigorously, by arranging the points of sup-
port and of contact so as to satisfy to the following proportion :—Let P, as before,
be the point of support and Q the steady point of the principal piece or seismo-
meter proper; and let p be the point of attachment to the ground, and 7 the
corresponding centre of percussion of the multiplying lever,—then the point of
contact between them must occupy the same relative position with respect to ヵ
and q in the lever as it occupies with respect to P and Q in the principal piece.

In practice, however, we may avoid the necessity of considering precisely the
static and kinetic qualities of the multiplying lever by simply making it so light,
relatively to the principal steady mass, that its own weight and inertia do not
sensibly affect the result. A practicaily far more important consideration in its
construction is the friction of the marking point, whose bad effects increase with
the ratio of multiplication, and which, unless proper steps are taken to reduce it,
interferes very seriously with the correctness of the indications.

SN 22. Record of successive displacements in conjunction with the time.

To observe the numerous successive displacements which make up an earth-
quake, it is only necessary to give the surface on which the writing pointers trace
their movements a continuous motion, the best direction for which is directly
away from the pointer and at right angles to its expected displacement. In some
instruments the surface to be written on is set in motion by means of the earth-
quake itself—the disturbance being caused to start a clock or other motor. This
plan is subject to the disadvantage that at the beginning of the earthquake a
portion of the disturbance, of uncertain length, occurs without leaving more than
what may be called a static record. When the horizontal pendulum is used as a
seismograph it is quite practicable, and for many reasons preferable, to keep the
surface moving under the pointers continuously and uniformly by elock-work, in
expectation that a shock may begin at any moment. A sheet of smooth glass,
covered with a thin layer of soot by being held over the flame of a smoky lamp,
forms a convenient surface for the reception of earthquake records, By making
the pressure of the marking point on the glass no greater than is needed to break
the film of lamp-black, the friction between it and the plate may be kept so
small as to introduce almost no error; and, after an earthquake, permanent
copies of the record may easily be got by varnishing the plate and using it as a
“negative ” from which to print photographs.

$ 23. Actual Horizontal Pendulum Seismograph.

A seismograph consisting of a pair of horizontal pendulums, which indicate
two components of the horizontal motion of the earth, was designed by the writer

™

in 1880* according to the principles set forth in the above paragraphs, and has
been since then in almost continuous use at the University of Tokio. Its con-
struction will be readily understood by reference to Plate I, where fig. 1 gives
an elevation and fig. 2 a plan of one of the horizontal pendulums, and fig. 3
shows in outline a plan of both and their position with respect to the moving
glass plate. Fach of the two precisely similar horizontal pendulums consists of
a light but rigid brass frame a, a long pointer of straw 5, and a solid eylindrical
brass bob c. The frame a and pointer b are free to rotate together about a
vertical axis dd (which is fixed to the earth and forms the axis of support),
being pivotted by a conical steel point which works in a conical hole at the
bottom, and by a conieal hole at the top into which presses a set-screw carried
by an inverted stirrup which is in one piece with the base-plate. In fig. 1,
one side of this stirrup is, for the sake of clearness, supposed to be removed.
The bob e is pivotted to the frame a in a similar manner, so as to be free to
rotate about the axis ee, The pointer b is attached to the frame a at f, by a

jeint which allows it to rise and fall vertically, but not to move in a horizontal

plane except along with a. The frame a and pointer 6 together are proportioned
so that ee is the steady line ($ 17) relative to dd as axis of support. Hence
when a small horizontal displacement of the ground occurs transverse to the lever,
ce would (were it not for friction at the joints and at the marking pointer) remain
stationary, even without the help of the bob c, and the inertia of the bob, pivotted
as it is on the line ee, enormously increases the tendency of that line to keep
steady. For these movements therefore the line ee, and indeed the bob as a
whole, serves as a standard of rest, and they are recorded on a magnified scale on
the smoked-glass plate g, which is lightly touched by a steel point at the end of
the straw b. Horizontal movements parallel to the length of this lever leave it
unaffected, but are recorded by the second lever, similar to the first, and set at
right angles to it. The two are placed so that their marking points touch a
circular glass plate 7 at different distances from its centre. By means of a clock

* See Proc. Royal Society of London, No. 210, 1881, p. 440; also Trans. of the Seismological
Society of Japan, Vol. II, p. 45. It appears that the earliest attempt to apply the horizontal
pendulum to the measurement of earthquake motions was made by Prof. W. 8. Chaplin, of the
University of Tokio, about 1878, His apparatus consisted of a wooden rod, free to turn about a
vertical axis, and carrying at its end a rigidly attached block. It was intended that the motion
of the earth should be recorded by a tracing point fixed to the block, writing on a smooth surface
fixed to the earth below it. There was no multiplication of the motion, and either for this
reason, or because friction was not sulticiently avoided at the joints and pointer, no results were
ever obtained, and the apparatus was ebandoned. The instrument described in the text was
constructed in the summer of 1880 and recorded its first earthquake in November of the same
year (see Transactions of the Asiatic Society of Japan, Vol. IN, p. 40). So für as the writer is
aware, no continuous record of earthquake motion had been previously obtained hy any observer,
Some time before its exhibition to the Seismological Society the apparatus was shown to Messrs,
J. Milne and T. Gray, who immediately adopted it (with the writer's consent), and have since
employed it, with slight changes in details of construction, to measure artificial as well as natural
earthquakes. Later, Mr. Gray communicated a paper to the Seismological Society * On Stendy

Points for Earthquake Measurements "' (Trans., Vol. III, p. 1.), in which he described a num-

ber of other contrivances embodying the same dynamical principles (see below, きき 27, 20 and 31).

(see § 24) the plate is kept revolving continuously in the direction of the arrow
about a vertical axis fixed to the earth, and so long as no earthquake occurs each
pointer traces over and over again the same circle on the revolving surface.
During a disturbance each records the component transverse to itself of all the
successive horizontal motions which may occur. The base-plate of each lever,
which must of course be rigidly fixed to the earth, is clamped by means of three
levelling-serews and a holding-down bolt in the centre to the top of a wooden
post, firmly driven into the ground and cut off a few inches above the surface.
Three convergent Vs, or a pyramidal hole, a V-slot and a plane surface are cut
on the top of the post for the levelling-serews to press into, Either arrangement
gives what Thomson and Tait * call a geometrical clamp, and, while affording
perfect definiteness of support without any nice fitting, allows the apparatus to
he removed and replaced in precisely the same position whenever that may be
desired.

To reduce as far as possible the friction of the marking point on the glass
plate, a light spring ん is added (adjustable by the serew 7), which carries a portion
of the straw pointer’s weight by means of the silk fibre 7.

§ 24. Improved form of Horizontal Pendulum Seismograph.

Experience gained with the instrument deseribed above led to a considerable
modification of details, chiefly with the view of diminishing friction, and of
rendering the apparatus less liable to get out of adjustment. The resulting form
is shown in Plates II and IIT. In Plate II, fig. 4 is a general plan of the
seismograph and driving clock complete ; fig. 5 an elevation of the glass plate
and sectional elevation of one of the pair of pendulums; fig. 6 an elevation of
the driving clock, showing the speed-governor in section. In Plate III, fig. 7 is
an isometrical drawing of one of the pendulums, and fig. 8 shows the method by
which the glass plate receives motion from the clock. As before, each pendulum
consists of a frame pivotted, about a vertical axis, to a fixed stand, and furnished
with a massive bob which is pivotted to the frame about a line which is the
instantaneous axis corresponding to the axis of support. A light prolongation of
the frame forms the multiplying lever, whose motions are recorded on a con-
tinuously moving smoked-glass plate.

Each frame a (figs. 5 and 7) is a light triangle of steel, and the bobs are
truncated cones of cast iron, pivotted to the frames in a manner which the figures
sufficiently explain. The axis of support of each frame is defined by the points
of two hard steel screws 6 and c, of which the upper one b works in a V-slot in
a piece of agate, strongly supported by two wooden uprights fixed to the base-
plate. One of the wooden uprights is removed in figs. 5 and 7 to allow the
frame and other parts to be seen. The lower pivot e works in a conical socket
of hard steel and points in the direction of the thrust (an improvement suggested

to the writer by Mr. T. Gray)—that is to say its axis, if produced upwards, would

* Nutural Philosophy, Vol. I, § 198,

pass through the point of intersection of a horizontal line through D and a verti-
cal line through the centre of gravity of the whole suspended mass. The serew
e is fixed to the steel frame once for all, but the upper serew 6 is adjustable and
is provided with a jam-nut, in order that the axis of support may be brought into
parallelism with the axis of the bob. The V-slot and conical holes into which
the various points press are cut with a more obtuse angle than the points which
enter them. The pointer is a light reed furnished at its writing end with a steel
point, consisting of a piece of bent watch-spring filed.sharp and glass-hardened.
Its length is such as to give a record which is four times the actual displacement
of the ground. The pointer is jointed to the steel frame at d (fig. 7) and the
greater part of its weight is borne by a spiral spring of fine wire, hanging from a
bent arm fastened to the frame by a pinching-serew at e, which allows the
tension of the spring to be adjusted by raising or lowering the end of the arm.
The spring supports the pointer near the joint and somewhat below the central
line, an arrangement which permits the pressure of the marking end on the glass
plate to be kept sensibly constant throughout a considerable range of up-and-
down movement of the plate.

The glass plate is carried by a vertical steel spindle pivotted in a wooden
stand between an agate cup at the bottom and a steel serew at the top, as is shown
in figure 5. ‘To enable the plate to be readily removed and a fresh one put in,
it is attached to the spindle thus :—a cireular brass plate f is permanently fixed
to the spindle, and carries three inverted levelling-screws on whose points (which
are covered with chamois-skin) the glass plate g rests. Next comes a soft leather
washer and above it a stout brass washer A, which is pressed down on the plate by
a wedged-shaped key 7 fitting in a slot in the spindle. The levelling-serews in f
allow the plane of the plate to be easily adjusted: a rough adjustment is enough,
since the jointed pointers are able to follow any up-and-down movements which
the surface of the plate may make as it revolves.

The driving clock is an ordinary train of wheels driven by a weight of ten
kilos., which is wound up once a day and has fall of about five metres. The speed
is regulated by a governor whose construction will be clear from fig. 6. It con-
sists of a vertical spindle carrying a pair of jointed rods, each provided with a
pair of brass balls, and each carrying a fan which dips into a trough of oil when
the governor revolves. The centrifugal tendency of the balls is partially resisted
by a pair of springs which tie them to the spindle. The governor receives its
motion by rolling contact between a dise fixed on the vertical spindle and another
dise on the last axle of the clock-train. The clock drives the glass plate by a
roller k, whose axle is made flexible at one point (1) in order that it may adapt
itself to the varying height of the surface. This is done by cutting the axle at /
and connecting the ends by a single turn of steel wire, which forms a convenient
substitute for a Hooke’s joint. The axle of the roller passes through a guide m
(fig. 8), and close to the slot in which it works is a second shorter slot n. By

lifting the axle into this, た is raised out of contact with the plate, whose motion

is accordingly arrested although the clock continues to run. One turn of the
plate takes from a minute to a minute and a half.

In the figures the whole apparatus (except the clock, which is separately
supported) is represented as mounted on a wooden patform on the top of a post
stuck in the earth. Recently it has been found better to use a low solid stone table,
as the warping of a wooden support is apt to produce a radial creeping of the
pointers on the plate, and consequently a gradual widening of the lines they
trace as the plate revolves.

In practice the horizontal pendulums have their axes of support slightly
inclined forwards, to give a small degree of stability. This, indeed, must be
given in all seismometers, partly to prevent the equilibrium from becoming
unstable through any slight accidental change of adjustment, and partly to pre-
vent any excessive displacement of the so-called steady point by the accumulated
effects of frictional and other disturbing forees during a prolonged shaking.

$ 25. Horizontal Pendulum Seismograph with Ring Bob.

Another form of horizontal pendulum seismometer, representing a stage in
the development of the instrument intermediate between the two forms already
described, is shown in Plate IV, where figs. 9 and 10 give a sectional elevation
and plan respectively. The construction will be evident from the drawing
without explanation, There are, as before, two similar levers set at right
angles to each other, which write their records on the same reyolying smoked-
glass plate, a portion only of which appears in the drawing. The distinguishing
peculiarity of this form is the shape of the bob, which is a hollow cylinder wide
enough to enclose the upright piece which affords a vertical axis of support.
By this arrangement the two axes may be brought near together without
necessitating the use of a small bob. The ring form has also the advantage of
having a relatively great moment of inertia about its own axis, which gives it
much power to resist the tendency of friction at its joints to set it in rotation
during a displacement of the ground. This merit, however, does not counter-
balance other practical defects and inconveniences in this modification of the
instrument, which is on the whole decidedly inferior to the later form described
in the preceding paragraph. The muitiplying ratio of the instrument figured in
Plate IV is seven to one,

S 26. Relation of the Friction Error to the dimensions of the frame

and to the suspended mass.

It is interesting to enquire briefly how the frictional error is likely to be
affected by changes in the arrangement of the parts. For this purpose we may

conveniently separate the whole frictional resistance into three parts—( ヶ ) that due
to the marking pointer, (3) that due t» the vertical component of the statical
pressure on the pivots, (7) that due to the horizontal pressures on the pivots, which

25
are exerted to balance the statical moment of the overhanging masses. The
acceleration of the axis of support during an earthquake is usually so small that
in considering the influence of friction at the pivots we may without sensible
error restrict ourselves to the statical forces,

It is clear that the pressure on the pivots due to the moment of the over-
hanging weight will be reduced (other things being unchanged) by increasing
the height of the upper pivot; this is one of the particulars in which the instru-
ment of § 24 is better than those of § 23 and § 25. On the other hand, this
pressure will be increased by lengthening the frame, in simple proportion to the
length. But the vertical pressure on the pivots will not thereby be changed,
aud the moment of inertia of the system about the axis of support will be
increased in duplicate proportion to the inerease of length; hence if the relation
of frictional couple to pressure on the pivots be constant, a considerable increase
of steadiness would be given by lengthening the frame horizontally.

Again, if without changing the length or height of the frame we vary the dis-
tribution of mass between it and the bob (M, and M,, to return to the notation of
§ 18), it is clear that the vertical pressure at the axis of support, giving rise to (3),
will be simply proportional to M, + M,, while the horizontal pressure giving rise
to(7) will vary as + M, + 7’ M,. The effective inertia is, as we saw in § 19, to be
measured by = M,-+- M,. It appears therefore that so far as it depends on
horizontal forees at the pivots, (7), the frictional error is not altered by any
change in the distribution of the mass between the bob and the frame, but that
as regards (3) the most favourable distribution will be to put as nearly as possible
all the mass into the bob. In other words, the distribution actually made in the
three instruments already described is the best not only as regards economy of
material, but (what is of much more importance) as regards reduction of the
frictional error.

Next, if we suppose the ratio of the masses and the dimensions 7, r’ and the
height of the frame to be unchanged, it is a question of much practical impor-
tance how great an amount of matter may advantageously be suspended. If the
sliding surfaces of the pivots were quite rigid, and the coefficients of friction
unaffected by changes of pressure, any increase of mass would produce a propor-
tional increase of friction at the pivots, leaving however the part of the whole
friction designated above by (a) unafleeted. The whole friction would therefore
increase less rapidly than the effective inertia, and consequently any addition of
mass would be an advantage as regards steadiness.

But when additional weight is applied the pivots are compressed and there-
fore the amount of their sliding motion, relative to a given angular movement
of the frame, is increased, Hence the resisting couple due to friction at the
pivots increases in a much more rapid ratio than the effective inertia, when the
weight is increased, and thus, although any addition to the weight gives

an advantage in opposing («@), it introduces a disadvantage by giving more than

a
>

a proportional increase of (3)and (7). Hence in any given case (other things
remaining unchanged) there will be a certain value of the suspended weight
which will give a maximum of steadiness. To attempt to calculate this maxi-
mum theoretically is out of the question, and no more than a very rough
experimental determination is practicable. The desirable amount of weight will
be small if (a) is a small part of the whole resistance. The writer’s experience
points to the conclusion that in practice the greatest steadiness is to be got by using
a mass of no more than 1 to 2 kilos. as the bob of a horizontal pendulum seismo-
graph. A mass greatly heavier than this, while it has more power to move its
tracing index, tends so considerably to crush the hard steel points by which the
supporting frame is pivotted, that the frictional error introduced by the pivots

more than counterbalances the advantage gained,

)

S 27. Horizontal Pendulum with flexible tie and pivotted strut.

Instead of using a rigid frame to carry the massive bob, we may suspend it
by a single tie and strut, and by making the tie flexible we may avoid haying a
joint at its place of connection to the fixed support. Thus, in the sketch, M is
the heavy mass hung from a fixed post P by a
wire 7’ which ties it to the post at a, and held out
so as to form a horizontal pendulum by the rigid
strut S pivotted in a conical socket ). The mass
M may of course either be fixed to S, in which
case the steady line is at the axis of instantaneous

rotation of the system, or it may be pivotted at
the axis of instantaneous rotation of S alone (the
mass of 7 being negligible). A prolongation of
S beyond M forms a convenient writing pointer.
A pair of similar instruments serve to register
the two components of the horizontal motion.
The necessary freedom of rotation about the axis
of support ab is afforded by combined twisting

hwo i and bending of the tie at a, whose elasticity gives
a small amount of stability to the system. If
that is too inconsiderable for practice it may easily be increased by advancing the
point a slightly in front of the vertical line through 7. Except for imperfect
elasticity, there is no frictional resistance at a. The strut need not of course be
horizontal, but by placing it so we reduce the stress on it, and consequently the
friction at its pivot, to a minimum.

This modification of the horizontal pendulum is due to Mr. T. Gray, who
has suggested it in his paper “On Steady Points for Earthquake Measurements ”,
communicated to the Seismological Society of Japan shortly after the instrument
described in § 23 had been exhibited. Figs, 11 and 12, Plate VII, show an
elevation and plan of the instrument as deseribed by Mr. Gray in the Philosophi-

cal Magazine for September 1881. There each bob consists of two masses ALL
fixed to the ends of a cross bar b, which is suspended by a wire ¢ from an adjus-
table projecting arm c clamped to the post ?. One of the bars b is bent to
make it to pass clear of the other. The struts a a are short compared with the
length of the tie, and consequently bear only a small fraction of the weight.
The multiplying arms 77 ave inclined to the directions of displacement and
parallel to each other, so that their records may be traced side by side—an arrange-
ment which allows the two components of the motion to be more readily com-

pared and compounded.

$ 28. Horizontal Pendulum without joints.

The method of suspension just described leaves one joint at which there is
rolling contact and consequently friction, namely the socket of the strut; but
there is no reason why we should not use a flexible piece instead of a joint there
as well as at the other end of the axis of support. To do this it issonly necessary
to fork the strut, let its end project beyond the axis of support, and tie it back to
that axis by an elastic wire, or preferably by a thin ribbon of tempered steel,
placed with its flat side vertical. A fixed vertical pin standing in the axis of
support and capable of sliding freely in a horizontal slot in the strut, is added to
prevent any bodily translation of the stiut to one side or the other during an
earthquake, so that the hanging mass may have no freedom to move otherwise
than by rotating about a nearly vertical line joining the upper and lower points
of attachment.

This idea has been practically carried out in the instrument shown in Plate
V. Fig. 13 is an elevation showing one of the two pendulums, and Fig. 14 is a
sectional plan. ‘The post P, firmly stuck in the earth, carries two horizontal
levers J, and ZL, set at vight angles to each other, to record two rectangular
components of horizontal motion. The bobs 7, AT, are fixed to the rods Z, Z, and
tied to a pair of small vices at the top of the post by fine steel wires T’T. Fig.
15 is a plan of the top of the post, and shows the arrangement of the vices, which
are fixed in a manner which gives them two (horizontal) degrees of freedom of
adjustment. At the back end of each rod Z is a fork (shown on a larger seale in
figs. 16, 17, and 18) which consists of two parallel cheeks of brass aa termina-
ting in a vice b which is clamped by the serew-bolt and nut c. A split upright
pin p is fixed to the post, and a short thin flat band of very flexible steel is
clamped between it and b, the end of Z. This is kept in tension by the thrust
exerted by Z, and when the horizontal pendulum swings the spring bends at or
close to a vertical line in the centre of the pin. A sector of the pin, facing
towards a, is cut out to give the spring room to bend about the axis of the pin
(see figs. 16 and 18). The split sides of the pin p are pressed together by the
nut n and so caused to hold the spring clamped between them, In fig. 18 one
of the cheeks a is removed and the pin p shown in vertical section: the flat

spring appears there and is lettered s. Only the upper portion of the pin p is

screw-tapped: the lower part is a smooth cylinder, and the cheeks a a pass just
clear of it on either side, their distance apart being adjustable within small limits
by the serew-bolt and nut d. Hence no horizontal translation of Z can oceur, and
the only freedom of motion possessed by the suspended rod and mass is freedom
to revolve about a nearly vertical line joining the upper end of 7 with the axis
of the pin p. A pair of long bamboo reds R, and RR, serve to multiply the
motion, and record their displacements side by side on a revolving smoked-glass
plate G by means of the hinged pointers ¢, and t,. These are short pieces of
straw tipped with steel, and each is provided with a little balance-weight w,
behind the hinge, which lightens the pressure of the pointer on the plate and so
reduces the friction. To bring the records parallel, the rod R, is set at right
angles to Z,, and a counterpoise W serves to bring the centre of gravity of the
system back into the line of Z,. The carth’s motion is multiplied seven times
in the record. The smoked-glass plate is similar in its arrangement of support
and driving gear to the one shown in Plate II, but is of considerably larger

size in the present case.

$ 29. Horizontal Pendulum with two degrees of freedom.

In the horizontal pendulum seismometers.hitherto deseribed, each piece has
been capable of showing motion in one azimuth only— motion, namely, perpen-
dicular to the plane containing its axis of support and its steady line; and a pair
of pieces has therefore been necessary to determine horizontal movements in
general. Mr. Gray* has adapted the horizontal pendulum to the determination
of motion in any azimuth by making the pivotted supporting frame in two pieces
placed in planes at right angles to each other, and jointed together in a vertical
axis. The arrangement will be clear by reference to Plate VI, figs, 19 and 20.
There, as in Plate IV, the bob isa ring. The primary axis of support is aa
fixed to the earth. The pivotted frame which supports the bob consists of two
parts A and B, jointed to each other with freedom of relative rotation about an
axis bb, which stands to aa in the relation of axis of instantaneous rotation for
the frame A. The bob is pivotted to ee, which bears a like relation to bb for the
frame B. Multiplication of the earth’s motion is effected by an independent
lever, arranged vertically. It passes through a universal joint fixed to the base
plate and has a small counterpoise near its upper end, where it touches c below
the axis of the bob. Obviously the bob is not disturbed by a small horizontal
movement in any direction. ;

In the next chapter other devices will be described, whieh allow the hori-
zontal part of earthquake movements to be registered by affording in some cases
a steady line with respect to motion in one azimuth, in others a steady point
with respect to motions in all azimuths, Of instruments giving a steady line
none is more successful and generally practicable than the simple horizontal
pendulum ; but of those which give a steady point it is probable that many yet to
be described are preferable to the horizontal pendulum with two degrees of freedom.

. © On Steady Points &c.’’—Trans. Seismological Soe. of Japan, Vol. III, p. 5. An essen-
tially similar instrument is figured in the Philosophical Magazine for Sept. 1881.

CHAPTER IIiI.

OTHER INSTRUMENTS FOR OBSERVING HORIZONTAL MOVEMENT.

$ 30. Mass free to roll.

From what has been said in §§ 15 and 16 it will be obvious that any
method by which a mass can be supported with one (or two) degrees of freedom
to move horizontally, and in neutral or feebly stable equilibrium with respect to
such movements, can be adopted in the design of a horizontal seismometer, the
centre of percussion of the mass relative to its axis (or point) of support being
taken as the standard of rest.

One method of support, in which the equilibrium can be made as feehly
stable as may be desired, is given when a mass rests by means of a curved base
on another plane or curved surface fixed below it, the mass having freedom to
roll. The steadiness of the instantaneous axis may be increased by supporting a
second mass there, with appropriate freedom of motion relatively to the first mass.

In the figure, let PQ be a portion of the base of

9\ the rolling mass, whose centre of gravity is at @, and

Ai \ P(J' a portion of the fixed surface on which the mass is

Gis. free to roll. Let O and の he the respective centres of

e eurvature. In the initial position of equilibrium the

2 oe - > { line OO" is suyposed to be vertical and therefore con-

/ tains G. Now let the mass he displaced through a small

are PQ, and let PQ! be made equal to PQ. The dis-

| placement brings Q into contact with Q’, and causes the

| radius QO to form a prolongation of the radius O'Q’,

| Hence a line QR, diawn so that the angle OQR is equal

to PO'Q’, will be vertical in the new position. The
or

equilibrium, with respeet to the assumed displacement,
will therefore be stable if G is below FR, neutral @ if coincides with 7%, and
unstable if @ is above R.

Writing x and 7 for the radii of curvature OP and OP’ respectively, and
s for the are through which the point of contact moves, we have in the triangle
ORQ

ad, A Beh
OR : r = sin 一 :sin - 。
=

=r:r-+»,
if we assume s to be exceedingly small compared with the radii.

Hence PRi rae sr +r,

LESS

30
and the condition which makes the equilibrium neutral in very small displace-
ments is that
R PG = Dias AIAN
If the surface of the support is concave, as in the

adjoining figure, the condition giving neutral equili-

brinm is

Pane
Nr while if the surface of the support is convex, and the
LEN base of the rolling mass concave, as in the lower figure,

NN the condition beeomes

The seismometrical application of this method of

getting neutral equilibrium appears to have been limited
2 to the case of a spherical or cylindrical mass rolling on a
plane surface. In that case the formula gives PG = r,

a condition which is actually fulfilled by a homogene-

| \ ous spherical ball or solid or hollow cireular cylinder of
| ! uniform thickness; and the equilibrium is then neutral

| for large as well as for infinitesimal displacements, a

| fact which makes rolling spheres and cylinders suitable

| for the registry of large earthquakes. In some applica-

| tions which have been made, the spherical or cylindrical

ee aA mass has been used alone, its centre of percussion or
instantaneous axis forming the steady point or steady
line with respect to which measurements of the motion

of the supporting table have been taken. In other cases

of ・
the ball or cylinder has been used to support a second
mass which furnishes the greater part of the effective
3 inertia of the system.

$ 31. Rolling Sphere Seismograph.

The latter plan was adopted in what appears to have been the earliest seis-
mometer of this class—an instrument designed and used to measure earthquakes
in Japan by Dr. G. F. Verbeck in 1876-7. In Dr. Verbeck’s apparatus the
supporting table was a marble slab, with its upper surface ground plane and
carefully levelled. On this four balls of rock-erystal were placed (three would

have been sufficient) each about one inch in diameter, and on them was laid a

massive block of hard wood whose weight greatly exceeded that of the balls.
In the centre of this a vertical tubular hole was hored, in which a weighted
pencil, sliding freely, was carried, its point resting on a piece of paper fixed to
the slab below. This gave a static record of the relative motion of the block and
the slab, without multiplication.”

A similar instrument in which the base-plate is of glass, and the record is
given by a steel needle writing on a smoked-glass plate, has recently been
described by Mr. C. A. Stevenson to the Royal Scottish Society of Arts, and has
been applied by him to the measurement of earthquakes in Scotland.

It appears to have been assumed by these observers that (except for friction)
the block resting on the balls would remain at rest dnring a horizontal displace-
ment of the base, A little consideration, however, will show that when the base-
plate moves in any direction the block will in all cases move through a small
distance in the opposite direetion. The centre of pereussion of a spherical ball
taken separately is at a height of seven-fifths of the radius from the bottom, and
hence if we wish to keep the block at rest it must be supported at that height
instead of at the top of the balls. Placed where it is, it has the effect of raising
the centre of percussion or steady point of the system considerably, but that
necessarily remains below the top of the balls. We may conveniently examine
the kinetie qualities of the system by considering the upper block as equivalent
to a number of particles, one situated always at the top of each ball. When
three balls are used the mass of each of these particles is to be taken as one-third
the mass of the block. Calling the mass of the block 3M, that of each ball
m, and the radius of each ball +, the system is kinetically equivalent to three
pieces standing upright, the moment of inertia of each about the base being
Im "+ 4M},
and the height of the centre of gravity of each from the base is

mr + 2Mr

m+M
Hence the centre of percussion of each is at a height above the hase equal to
Imı* ト 4M>? r(im 4 4M)
“mr + 2Mr ae att m -- 2M

This, which is > Zr and <2r, is the height of what we may here call the
steady plane of the system above the base, and this position should be taken as
the standard of rest in reckoning the displacement of the earth’s surface. When
the base moves through a distance s, the top of each ball (and therefore the block)

moves oppositely through a distance

3 1
( zm |,

| im + 4M )

* Dr. Verbeck tells the writer that he designed, but never used, a lever arrangement by

which a multiplied record of the earth's motion would have been given.

ESS

and, if there were no friction, the records given by a tracer fixed to the block
should be reduced by this amount. In Dr. Verbeck’s instrument the mass of
the block was so great relatively to the mass of the balls that the impulse which
their inertia gave to the block-must have been negligible: in other words the
steady plane must have been exceedingly close to the top of the balls. The
effective inertia of the system (with three balls) is

3 (m + 2M)"

im +4M —

SN 32. Friction in the Rolling Sphere Seismometer.

Frietion, or rather rolling resistance, constitutes a serious source of error in
the indications, and has of course the eflect of making the recorded displacements
smaller than they should be by an amount depending, in a given instrument, on
the rate of acceleration of the earth’s surface durin

g the disturbance. According

to Rankine (Applied Mechanics, $ 682) the rolling resistance may be taken as a
couple whose moment is found by multiplying the normal pressure between the
rolling surfaces by an arm whose length (which we will write 1) depends on the
nature of the surfaces. To simplify matters we may consider separately these
two cases: (1) mass and weight of balls negligible compared with block, as in
Verbeck’s apparatus; (2) no block, the inertia and pressure being due to a spheri-
cal ball alone.

In (1), calling the mass of the whole block M, the pressure at the top and
also at the bottom of the balls is Mg and the couple due to rolling resistance at

these two places is (2 + 1’) Mg, where i and 7 apply to the two pairs of surfaces.

erie e : 5 (7 +U) Mg F
This is equivalent to a horizontal foree —--—,— tale) applied to the block, and
ay
: (7 十 の
produces an acceleration = 9,
=r

In (2), calling m the mass of the ball, the effective inertia (or mass referred
to the steady point) is 2m, and the couple due to rolling resistance is Img.
ling

This is equivalent to a force — J applied at the steady point, and produces a
5

3 2 lq :
horizontal acceleration there equal to —~. Hence it appears that both arrange-
"A

ments are affected by rolling friction to the same extent, provided we use balls
of the same radius in both, and the same two materials at each place where
rolling contact oeeurs.

Rankine gives the following values for the arm / expressed in feet :—
0.006 (Coulomb)
Lignum- Vite on oak ...... Menten «0.004 お
Cast-iron. om Cast-Iron 4.2 cc sseeeeee! 0.002 (Tredgold).

If we suppose that by choosing suitable substances for fixed slab, balls, and

Oak on oak

block, the value of 7 and I’ can be reduced to, say, half of the least of these, or

0.001, we find (taking + = $ inch, as in Dr. Verbeck’s apparatus) the accelera-

. . . Ye . ‘ . .
tion of the block by rolling friction to be oe In a later chapter it will be

shown, as the result of direct observations, that the horizontal acceleration of a
particle on the earth’s surface does not reach this value except in earthquakes of
considerable intensity, and even then falls short of it during a great part of each
displacement. Hence, apart from the friction of the writing pointer, an instru-
ment of this kind is likely to possess so much resistance that it will fail altogether
to record small earthquakes, and give a very imperfect record of others which
are great enough to leave some traces of their passage. A frictional acceleration
amounting to one-fortieth of the value of gravity is altogether out of the question
in a seismometer, and it appears that even by increasing greatly the radius of
curvature, a rolling mass seismometer cannot easily, if at all, be brought to the
state of efficiency which is reached without difficulty by the horizontal pendulum
and other devices, and which is indispensable to accuracy in the measurement of

small earthquakes,
§ 33. Other Rolling Sphere Seismographs.

A single solid spherical ball has been used by Mr. Gray and is figured in
fig, 21, Plate VII. The following description of it is, with the figure, taken
from the Philosophical Magazine for September, 1881 :—

“ A sphere of lead,* iron, or any other heayy substance rests on a flat plate
B made truly plane and furnished with three levelling screws Z. Anarm 4, fixed
to the base B, is so formed that a circular ring fixed to its end is held in a
horizontal position with its centre vertically above the highest point of the
sphere. This ring carries a species of spring universal joint, consisting of four
very light bent springs, j, arranged at right angles to one another and meeting
in a small round dise, 6, at the centre. The lower end of the lever, /, passes
through this ring b, and is fixed to it at such a point that its lower end, which
is rounded, just fits a small hole in the top of the sphere S. Between SS and
} a small sphere, s, is fixed to the lever J, and is so proportioned that the lever J,
when pushed at /, tends to rotate around a point a little above its lower end,
thus diminishing the push on the sphere S. The springs 7 serve to allow the
lever ! to turn in any direction, and are made so light that they can only make
the ball roll with a very long period. When thus proportioned they serve
the purpose of a universal joint, and at the same time give a little stability
to the parts, thus preventing the plate 7, if it be put in motion, from causing
the ball to roll over. The lever 7 is a rod of bamboo which is at the same
time very light and rigid; at the upper end the red is flattened and hinged just

above the bend by a piece of tough Japanese paper glued to its upper side...

* Lead would be an exceedingly unsuitable material on account of its viscosity, which by
allowing it to Hatten out at the hottom, especially after a long time of rest, would give great

resistance to rolling and an objectionable stability for small displacements. (J.A.E.)

The multiplication given by an instrument of the sort may be determined experi-

mentally; or it may be approximately caleulated by taking a point O at a
height equal to seven-fifths of the radius of the large sphere S as nearly steady.”
In other instruments used by the same observer a segment only of the
rolling sphere has been employed, with an independent mass pivotted at the
centre of percussion, which is arranged to be at such a height above the centre
of the spherical segment that the equilibrium of the system is nearly neutral.
“This method of construction allows the radius ef the sphere to be much in-
creased ; but it introduces a difficulty of adjustment, a complication of parts, and
generally a slight want of symmetry, which causes a little uncertainty in the
interpretation of the records.’* An instrument of this class is shown in fie. 22,
Plate VII. There the point P should be the centre of percussion of the rolling
segment S, to secure which the weight JV can be moved up or down. By varying
the mass of the pivotted ring だ the centre of gravity of the system can he brought
below the centre of curvature just enough to give a small amount of stability.

A simpler arrangement, and one yielding nearly as much effective inertia in
proportion to the pressure on the base, is a solid mass with as condensed a form
as possible, rigidly attached to a light spherical segment which rolls on a fixed
plane base. The centre of gravity of the whole piece must be a very little
below its centre of curvature (unless a small amount of stability be given by
other means), and its contre of percussion, which lies only a little way above the
centre of gravity, is of course to be taken as the steady point. In the instrument
deseribed above the pivotted ring is equivalent to a particle of the same mass
situated at 7°: the arrangement now mentioned, whieh is much easier of construc-
tion and adjustment, approximates closely to the former in proportion as the
dimensions of the bob (here fixed, instead of being pivotted) are reduced, without

reduction of its mass.

§ 34. Rolling Cylinder Seismograph.

The seismometers of § § 29-31 are intended to record horizontal motion in
any direction, ‘To nestrict the record to one azimuth it is only necessary to use a
rolling surface without curvature, and therefore without freedom to roll, in a
line perpendicular to the direction of the motions which are to be registered.
Thus a pair of circular cylinders placed at right angles to each other on a plane
hase may be used to give separately two rectangular components of earthquake
motion perpendicular to their respective axes.

This plan has been adopted by Mr. Gray in an instrument described before
the Scismological Society of Japan (Trans., Vol. III, p. 143) and again in the
Phil. Mag. for September 1881. Fig. 23, Plate VII, shows the construction.
The eylinders are marked e and ec’. A bent arm A, which is fixed to the base-
plate and passes over one eylinder, supplies fulerums for a pair of multiplying
pointers 2//. ‘The eylinders are made hollow, and hence the steady line of each

* Gray, loc. cit.

Dn et tne

SS oe Au

ト

35
comes near the top. Calling 7, the external, and r, the internal radius, the
height of the steady line above the base is

3 7 +,

To give a small degree of stability, Mr. Gray suggests the use of a small
cylinder rolling freely inside each of the larger ones. A spring, or a light pendu-
lum hung from above and gearing with the axis of the cylinder, might be used,
“An interesting modification of this machine might be made by placing two
equal cylinders on a horizontal plate with their axes parallel, and placing on
them a second horizontal plate so that its upper surface should always be in a
plane through their instantaneous axes. This eyuld be alone by causing the
plate to rest by means of arms on two pairs of smaller cylinders of proper
dimensions, so disposed that each of the larger cylinders should bear at each end
one of the smaller eylinders coaxial with the larger cylinder and projecting from
it. A third eylinder placed on the upper plate with its axis at right angles to
those of the first two would, for small motions of the earth, have a line in itself
which would remain approximately at rest.” It must, however, be admitted that
this elaborate plan for a seismometer with two degrees of horizontal freedom has
little to recommend it.

For large motions Mr. Gray proposes to arrange the multiplying levers so
that when the displacement exceeds a given value they will go out of action, and
a direct writing appliance take their place. “A very simple method of writing
large motions would be to attach a fine point to the end of the cylinder at its
centre, and allow this point to write on a plate placed in front of it and fixed to
the hase-plate.” This would give a record equal to rather more than half the
actual motion,

Compared with the rolling sphere, the rolling cylinder seismograph has the
advantage of easier construction ; to turn a eylinder accurately being a matter of
far less difficulty than to turn a sphere. It is however, equally with the rolling
sphere, liable to the great practical objection that its frictional resistance is so
great as to prevent it from properly recording a slowly changing motion of the

ground,

$ 35. The Common Pendulum.

It has been already pointed out that a mass hanging pivotted by one point
to a fixed support, so that its centre of gravity lies vertically below its point of
support, fails to act as an absolute seismometer on account of the stability of its
equilibrium. The arrangement forms, of course, a common pendulum, and
is liable to the objeetion that during any prolonged shaking of the support
the pendulum acquires a swing whose amplitude may be, and often is, greater
than the motions of the ground.

If, however, we make the length of the pendulum very great, its equilibrium

will be as nearly neutral as, in fact, it is ever desirable that the equilibrium of a

seismometer should be. If its period of swinging is much greater than the periods
the earthquake waves, these will not generally communicate any immoderate
motion to it, and even if it swings considerably there will no difficulty, provided
the motions are recorded on a continuously moving plate or drum, in distin-
guishing the long-period waves due to the swing of the pendulum from the
relatively much shorter waves due to the motion of the ground. The latter are
superposed on the former, and in measuring them the undulating path traced out
on account of the swing is to be taken as the datum line, or line of no displace-
ment, instead of the line which would be traced if the pendulum had remained
rigorously at rest.

The tendency of a pendulum to acquire a swing when its point of support
is shaken may be reduced and even completely obviated by introducing frictional
resistance, which dissipates, more or less completely, the energy communicated to
the hanging mass by the successive horizontal impulses to which it is subjected.
This action necessarily occurs to a certain extent in all seismographs, a certain
amount of friction being inevitable, especially where the earth’s motion is con-
siderably multiplied by a recording lever. The friction may of course be
purposely increased, and by doing this even a short pendulum may be prevented
from acyuiring a swing during a prolonged disturbance. But any introduction
of friction involves a sacrifice of accuracy in a seismometer, and the effect of
adding enough resistance to prevent a short pendulum from swinging is that it
altogether fails to record minute earthqnakes, and gives unduly small indications
during more considerable shocks.

Many observers have used for the measurement of earthquake: motion a
pendulum provided with a tracing point or other kind of index by which its
greatest displacement from the vertical is registered. Now it is clear that if the
pendulumm is liable to be set swinging during an earthquake such indications are
wholly without value. They depend not only on the amplitude but also on the
duration of the disturbance, and, very directly, on the period of the waves.
They may be, and often will be, many times greater than the motion of the
ground,

On the other hand, if the pendulum is so much retarded by frietion that
the energy which it receives by any one displacement of the point of support is
dissipated before the next displacement occurs, no accumulation of small motions
can oceur. The pendulum will not swing, and its greatest displacement cannot
exceed, but will actually be always less than the greatest motion of the ground
during the disturbance,

If we attempt to obtain, by means of a pendulum, a static record of earth-
quake motion —that is a record inscribed on a stationary plate (or registered by
some other equivalent contrivance) instead of on a continuously moving plate—
then, since we have no means of distinguishing (in the record) earthquake
motions proper from the swing of the pendulum, we must resort to the arrange-

ment just mentioned. That is to say, the pendulum must have enough frictional

aie os a

resistance to prevent it from aequiring a swing. If, however, we allow the motions
to be traced on a continuously moving record-receiver, any moderate swinging
of the pendulum, provided that is of much longer period than the earthquake
waves, is not seriously objectionable, and we may then reduce the friction as far
as possible. The same remarks apply with ejual force to all forms of seismo-
meter in which a so-called steady point is sought after. In all such instruments,
static records have no meaning or value unless swinging is entirely prevented.
This ean be done by introducing enough friction, but to do it, especially when
the stability of the free mass is considerable, involves a great sacrifice of accuracy
—a sacrifice whieh there is no need to make when a continously moving record-
receiver is used.

SN 36. Actual Pendulum Seismometers.

Mr. R. Mallet, writing in 1858, refers to the common pendulum, as “the
oldest, probably, of seismometers, long set up in Italy and southern Europe. A
pendulum, free to move in any direetion, carries below the bob a style partly
immersed in a stratum of fine dry sand, spread to a uniform thickness over the
concave surface of a cireular dish placed beneath, marked to the cardinal points,
whose centre is beneath the point of suspension of the pendulum when at rest,
and whose concavity is that of a spherical segment of a radius equal to the
length of the pendulum and style, plus rather more than the depth of the
stratum of sand. It was supposed that the style would mark a right line when
seen in a plane vertical to the sand-bed, and in the direction of the shock.’’*

Pendulums provided with sliding pencils writing on paper fixed below, or
with some other contrivance for giving static records, have also been used at
Comrie, in Scotland, by a committee of the British Association ; in Japan, by
Dr. Verbeck (in 1871) and subsequently by Wagener, Knipping, Milne, Gray,
and others; at Rome by Secchi; at Manila by Fanrat; and no doubt by many
observers in other places. In one of Mr. Milne’s arrangements a long pendulum,
hung from the roof of a house, carried two vertical sliding styles which rested on
the surface of two strips of smoked glass on a table below. The strips of glass
were started into motion by an earthquake disturbance (which released a catch)
and passed under the styles, moving in lines at right angles to each other. This
gave a continuous record of the motion of the pendulum, without multiplication,
beginning at an uncertain interval after the beginning of the earthquake, and
lasting for only a few seconds.** Dr. Wagener’s apparatus (successfully used
by Mr. Knipping in Tokio from 1878) was a pendulum 3 feet long supported by
a rigid frame, and provided with a multiplying lever by which the relative
displacement of the earth and the bob of the pendulum was magnitied 24 times.

* British Assoc. Report, 1858, p. 73.

† British Assoc, Report, 1841, p. 46.

t Proc. Royal Society, Vol. XXXI, p. 460.

** Trans, of the Seismological Society of Japan, Vol. ITI, p. 12.

The lever was pivotted to a fixed support by a ball and socket joint, and con-

nected to the pendulum by a ball and tube joint. The motions of its long end
were registered by its pulling up a thread through a hole in a plate fixed
immediately below it. The thread was wound round a light pulley to which an
indicating pointer was attached. Another pendulum instrument was arranged
with eight indices, and was designed to give the direction of the horizontal
displacement. Dr. Wagener’s scheme also included a continuously registering
seismograph, in which the motions of a pendulum were to be recorded on a
drum, started into rotation by the earthquake, and travelling longitudinally as
it revolved, by means of a screw on its axis. This part of the plan does not

appear to be been put in practice.*
$ 37. Long Pendulum Seismograph.

A long pendulum seismograph provided with a pair of multiplying levers
by which two components of the horizontal motion were recorded on plates kept
in continuous motion by a elock, was designed and ereeted by the present writer
in 1879. Of this instrument, which was probably the earliest continuously
recording seismograph to be actually constructed and successfully used, a deserip-
tion will be found in Vol. I of the Transactions of the Seismological Society of
Japan. Separate smoked-glass plates were used to receive the traces of the two
pointers. This introduced certain practical difficulties, and the original arrange-
ment was also objectionable on account of the smallness of the plates. A
modified form has since been adopted, and is now in use in the writer’s obserya-
tory. The instrument as it now stands is shown in Plate VIII.

Fig. 24 gives a general view of the structure by which the point of suspen-
sion of the pendulum is supported. It consists of a very rigid wooden frame-
work, firmly founded on piles, and rising above the ground to a height of over
six metres, This is completely detached from the walls and roof of the building
in which it stands. The pendulum consists of a massive ring of cast-iron a,
weighing 25 kilogrammes, and hung in a horizontal plane from the top of
the framework, by three wires bbb. A serew and nut at the top allow the
height of the ring above the ground to be adjusted. To record the motion of
the earth relatively to the “steady point” of the pendulum (which is, of course,
its centre of percussion) two appliances are provided, which may be used either
alternatively or both together. In one of these the complete horizontal motion
is recorded by a single pointer: in the other, two rectangular components of the
horizontal motion are recorded by separate pointers, but on the same plate. The
first arrangement only is shown in fig. 24, and is also shown, on a larger scale,
in figs. 25 and 26. The second arrangement is shown in figs. 27 and 28.

The ring-bob of the pendulum (shown in section in figs. 25 and 27, and in
plan in figs. 26 and 28) carries an iron bar c across one of its diameters, in the

* Trans, of the Seismological Society of Japan, Vol. I, p. 54 and Vol. III, p. 107; also
Mittheilungen der deutschen Gesellschaft für Natur- und Völkerkunde Ostasien’s, 1878 and 1879.

centre of which a brass tube d (fig. 25) is fixed. The upper end of the multiply-
ing lever is a brass ball fitting easily but not loosely into this tube; it stands at
such a height in the tube as to be at the centre of percussion of the pendulum.
The fulcrum of the multiplying lever is a gimbal joint (see fig. 26) giving freedom
of rotation in any azimuth, but no freedom to rotate about a vertical axis. This
is carried by a stiff iron bar e projecting from a wooden upright f, which is
firmly nailed to a post g (fig. 24) stuck fast in the ground. The projecting bar
e is slotted where it is fixed to f by a serew-bolt, so that the position of the
gimbal joint may be adjusted horizontally, by turning e round or pushing it out
and in. Below the fulerum the multiplying lever consists of a light bamboo rod
h, terminated by a fork in which the marking index 7 is jointed on a horizontal
axis. The pointer is a piece of straw tipped with steel: it rests on the smoked-
glass plate j, and its pressure is adjustable by a small counterpoise k. The
glass plate is supported in the manner already shown in Plate II, and stands on
a board fixed to the top of the post g. It receives continuous motion from a
elock, by means of a roller, in precisely the same manner as the glass plate in
Plate II. The plate is kept continuously revolving, in expectation of an carth-
quake, and it is found that although the whole apparatus is remarkably free
from friction, the line traced on the plate does not widen to any very objec-
tionable extent, in the absence of earthquakes. The multiplying ratio is ten
to one.

The additional, or alternative, arrangement shown in figs. 27 and 28,
allows the same pendulum to serve as a two-component machine. A brass plate
/ is fixed to the cross-bar ¢ at the level of the centre of percussion of the pendu-
lum. In this two slots are cut, at right angles to each other (see fig. 28, where
a portion of the bar ¢ is removed), and in these slots the bent-up ends of two
short horizontal levers m, n, slide, m and n are pivotted about vertical axes at
o and p. The supports of these axes are fixed to the earth. Beyond the axes o
and p, to the right, are two light continuations of the levers m and n, arranged
parallel to each other. These (q and 7) record their motions on a large revolving
smoked-glass plate, a portion of which is shown at s. It will be obvious that a
horizontal motion of the earth parallel to n will simply cause n to slide in the
plate | without affecting the pointer r, but will cause m to revolve about the
axis o, and will therefore be recorded on a magnified scale by the pointer g.
One of the two levers mn is bent so as to pass below the other without touching
it. The upright ends by which they receive the motions of the bob are cylinders,
working easily, but without shaking, in the slots in the plate I. The long
pointers » and q are jointed horizontally near o and p, (see q, fig. 28) so that
they may follow inequalities of level in the revolving glass plate.

The inertia of the bob e is so great that there is no objection, on the score
of friction, to the simultaneous use of both arrangements for recording, and they
are constructed so as not to interfere with each other. ‘The recording levers are

so light that their influence, both static and kinetic, on the bob is negligible.

S 38. Influence of friction on the Pendulum and other Seismometers.

Before going further, it may be useful to enquire more particularly into the
effect of frictional resistance on a pendulum seismometer, in order that we may
estimate the amount of inaccuracy which is introduced when enough friction is
added to make static records practicable, by preventing the pendulum from
swinging. The theory is in fact applicable not only to the common pendulum,
but to ail seismometers whose stability would (in the absence of friction) cause
them to execute simple harmonic oscillations.

Let M be the effective inertia of the instrument 一 in other words the mass
referred to the steady point. Let F be the force per unit of that mass (due to
stability, and leaving friction out of account) which tends to restore the mass to its
position of equilibrium when the steady point is displaced through unit distance
from that position. Then, by the assumed condition, the total restoring force when
the displacement is r is equal to r MF. Also, let f be the frictional resistance
referred to the same point. This, which is made up of forces at the point or axis
of suspension, at the fulerum of the multiplying pointer, at the joint by which the
pointer is connected to the main mass, and at the marking pointer, will, in gene-
ral, be a constant force, sensibly independent of the velocity. Let the mass be dis-
placed so that the displacement of the steady point from its position of equilibrium
is ry, It will swing back, and over to the other side to a distance of, say, 7, from
the position of equilibrium. During this motion the energy dissipated through
friction is f(r, + 7). This must be equal to the loss of potential energy which is
「 に rMPdr, o ーー - =

2f
ore
and the same expression will give the decrease in amplitude of displacement
during every succeeding half-swing. The amplitudes diminish in arithmetical
a7

progression, successive displacements to the same side differing by — *

MF’

Hence r

From the above equation
1

(n—r)F
ge

duces whenever it acts. F\, if not directly calculable, is easily determined by
observing the period of free oscillation of the suspended mass,* and 7, — 7,

we have

for the acceleration which the frictional resistance pro-

may be at once determined by experiment in any actual instrument.
An earthquake in which the acceleration of the ground did not exceed this

‘ (r, — 7) F N 5 :
quantity, = would oceur without affecting the instrument, and would
: で 生き 前
* Since the complete period of free oscillation is — =
1

altogether fail to be recorded. The record of a sharper shock, given on a
moving plate, would be diminished by friction to an extent not easily calculable,
even if we assume a definite character (e. g. simple harmonic oscillation) for the
motion of the ground ; and the actual motions which oceur during an earthquake
are of too irregular a character to make any such assumption even approximately
correct.

Taking now the case of a common pendulum, let ! be the length of the
equivalent simple pendulum, or in other words the distance of the steady point
from the point of suspension. Then for any small displacement r, the restoring
7 Mg

if |

force (neglecting friction) is and the decrease of amplitude in a semi-

oscillation
2 が
人

Hence the acceleration due to friction, when it acts, is Ato art N

Observations with the long pendulum described in § 37 have shown that
7, — 7, is 0.03 em. in the ordinary condition of the instrument. Hence, since

0.03 7 g

2 x 540 ~ 36000
or 0.027 em. per sec. per sec. It is certain that a frictional resistance so small

the length 7 is 540 cm., the acceleration due to friction is

as this has no considerable influence on the records of even very minute
xurthquakes.

$ 39. Short Pendulum with considerable friction.

Mr. Gray (Phil. Mag., Sept. 1881, p. 203) has advised the use, for earth-
quake measurement, of a short (3 feet) pendulum provided, purposely, with
enough frictional resistance to bring it to rest after one half-swing when the
initial displacement is about that of the largest earthquake likely to occur, the
object being of course to avoid swinging. To fit such an instrument for use in
Japan we should provide for a maximum displacement of about 1 em., and, to
prevent swinging, 7, must be zero when the pendulum swings back after this
extent of displacement. Taking the length as 1 metre we should then have the

7
200

much in excess of even the maximum acceleration which occurs during many

acceleration due to friction = ‚or 4.9 em. per see. per see, This is a value

minor earthquakes, and these would therefore fail to be recorded. Probably
even so considerable an amount of friction as this would not introduce any very
large error into records of the greatest movements which the apparatus was fitted
to record, but it would certainly prevent many earthquakes from being recorded
at ali; it would give a much reduced record of greater motions; and it would
curtail such records as it gave, by causing the pendulum to end its registrations
long before the motion of the ground ceased—since the later part of an earthquake

consists always of slowly dying out undulations whose period is considerable and
whose amplitude is small. The same objections would apply, with scarcely
diminished force, to an instrument in which the frictional resistance was even
two or three times less than the value given above.

$ 40. Methods of making ‚a short pendulum astatie.

Various plans have been proposed for reducing the stability of a short
pendulum sufficiently to fit it for seismometrie work. Mr. Gray* has suggested
to fix on the bob a circular trough containing some liquid, which, when the
pendulum is defleeted, will accumulate on the inner side so as to bring the centre
of gravity of the system back to (or nearly to) the vertical line through the point
of suspension. Another plan is to use a ball rolling in a hollow curved surface
fixed to the bob, such as a sphere whose radius is longer than the pendulum.
“ Another method would be to attach a vertical spiral spring to a point in the
axis of the pendulum a little below the point of suspension and to a fixed point
above it, so that when the pendulum is deflected it would introduce a couple.”

The present writer} has suggested the following arrangement :—Let the
bob of the pendulum he jointed to the top of a short vertical lever which is
pivotted below the bob by a gimbal, or ball and socket joint, to a fixed support,
and let the connection of the lever with the bob be a ball and tube joint, so that
the top of the lever may be free to accompany the pendulum in all its (small)
displacements. Let a point on this lever anywhere between its pivot and the
bob of the pendulum be connected by a stretched spiral spring to a point (fixed
to the ground) vertically below and at some distance from the pivot of the lever.
Then when the pendulum is deflected to any side this spring will exert, through
the lever, a force tending to increase the deflection; and by a proper ar-
rangement of parts this foree may be made very nearly sufficient to hold the
pendulum deflected against gravity, and so to give a condition of nearly neutral
(feebly stable) equilibrium throughout a range of motion equal to that of the
largest earthquakes the apparatus is intended to record.

$ 41. Duplex Pendulum Seismometer.

An instrument resembling these in its action, though materially different in
its design, has been invented and successfully applied to the measurement of
earthquake motion, by the present writer, under the title of the Duplex Pendu-
lum. It consists of a combination of a common with an inverted pendulum.
The common pendulum is stable : the inverted pendulum with a rigid pivotted
supporting rod is unstable: by placing an inverted pendulum below a common
one and connecting the bobs so that any horizontal displacement must be common
to both, we may make the equilibrium of the jointed system neutral or as feebly
stable as may be desired. The following description is taken, with little alteration,

* Transactions of the Seismological Society of Japan, Vol. III, px 145.
† Ibid., p. 147,

a ee eee

from the writer’s original account of the instrument, given in the Transactions
of the Seismological Society of Japan, Vol. V, p. 89. The reference letters apply
to Plate IX, figs. 29 and 30, of the present memoir, which give two sectional
elevations of the instrument, taken at right angles to each other.

B, is the bob of the upper or common pendulum, consisting of a hollow
eylinder of lead hung (with freedom to oscillate in any azimuth) by two light
wooden rods aa from a cross piece b, which carries a steel pin ¢ whose point
rests in a conical cup of agate let into the upper fixed support S,. S, is a rigid
bracket standing out from the top of a post fixed in the earth. B,, also eylin-
drical, is the bob of the lower or inverted pendulum. It is fixed to a stout
circular rod d which is pivotted to a second fixed support or base S, by a some-
what peculiar joint. Two feet ee fixed to the rod of the inverted pendulum
stand respectively in a conical hole and a V-slot on the upper surface of a steel
plate f, on the lower surface of which there are another conical hole and V-slot
in a line at right angles to the line of those on the upper surface. Into the
lower hole and V a pair of inverted feet fixed to S, press up. The upper and
lower slots and holes are arranged so that their vertices are all in the same hori-
zontal plane. This mode of support also gives freedom to oscillate in any
azimuth, and it is emploved instead of the more simple method of pivotting a
single foot in a single conical hole, in order that there may be no freedom on the
part of the lower pendulum to rotate about a vertical axis. There is therefore
no objection to using a prolongation of the lower pendulum as the indicating
pointer. The bobs B, and B, are connected thus:—from a rigid brass bar
extending across the top of D, there depends a rigid vertical projecting piece
ending with a spherical ball which just fits in a cylindrical hole in a tube fixed
to B,. The pendulums therefore move freely together, this joint giving them
the necessary power of vertical sliding relatively to each other through a small
distance. 」

The spherieal ball on B, and the tube on B, are placed so that their point
of contact is at the centre of percussion of both pendulums. "This is the kinetic
condition which must be fulfilled in order that this point should be the “ steady
point” when a displacement of the earth occurs. The point of contact is, of
course, a short distance below the centre of gravity of B, and above that of B,.

If for brevity we call W, and W, the weights of the pendulums referred to
this point (that is, the actual weight of each multiplied by the distance of its
centre of support from its centre of gravity and divided by the distance of its
centre of support from the point of contact between the ball and tube, then the
static condition which will give neutral equilibrium in very small displacements
is that

W,l, = Wh,
where /, and /, are the lengths of the pendulums measured from their point of
contact to their respective points of support. In practice a small margin of
stability must be given by making W, 1, somewhat greater than W, /,.

The multiplying lever consists of a light wooden rod g rigidly fixed to B,
by a bridge h, and therefore forming a prolongation of d. At the top of this a
light arm of straw 7 is joimted (with freedom of motion about a horizontal axis),
and stands out horizontally with its end (steel-tipped) resting on a fixed or
revolving smoked-glass plate, of which a portion is shown at 7 in fig. 30.

When its equilibrium is nearly neutral the duplex pendulum forms an
exceedingly sensitive level. If the line joining the upper with the lower point
of support is not perfectly vertical a very large deflection of the pointer results.
It is easy to show mathematically that if the equilibrium were exactly neutral
when the line joining the supports was vertical, the apparatus would be unstable
should any deflection of this line from the vertieal take place, and that when
there is some small stability the effect of such a deflection is to cause a large
displacement of the bobs. The same characteristic is, however, shared by other
neutral or nearly neutral equilibrium seismometers, such as the rolling sphere or
the horizontal pendulum.

$ 42. Duplex Pendulum with a single bob.

The two bobs of the duplex pendulum may be united to form a single bob,
provided we make one of the supporting pieces, preferably that of the upper
pendulum, extensible. This idea has been carried out in the instrument shown
in Plate X, which was described to the Seismological Society of Japan on Feb.
15th, 1883, by the present writer.

Figs. 31 and 32, Plate X, give sectional elevations of the instrument, in
planes at right angles to each other. The bob is a cast-iron ring B, partially
supported by the stirrup-shaped piece a, at the bottom of which there is a steel
pin b standing in a cup of agate in the lower support s,. The weight of B is
does not all come on this feot-step: the remainder is borne by the spiral spring
e which is stretched vertically between the upper support s, and the middle of a
bar a which crosses the upper surface of the bob.

The apparatus is, in fact, a common and an inverted pendulum combined in
one, and its relation to the duplex pendulum is at once evident if we conceive of
the bob B as consisting of two parts: one borne by the tension of the spring,
the other by the upward thrust of the footstep. These correspond to the two
hobs B, and B, of the last article. The condition giving neutral equilibrium in
small displacements is that the thrust of the footstep shall bear to the pull of
the spring the same ratio as the length of the inverted pendulum bears to the
length of the spring. In other words, the weight of the bob must be shared
between its two supporting pieces in proportion to their respective lengths. In
practice, to give feebly stable equilibrium, the tension of the spring must be
somewhat increased. It is adjustable by the screw d and nut e. The initial
strain of the spring is so considerable that any actual displacement scarcely
increases the pull it exerts.

To register the motion, an upright multiplying lever f, carried by a gimbal

joint on the fixed bracket g, is connected to the bob at the centre of percussion
by a ball and tube joint. At its lower end a light horizontal pointer h is hinged,
and projects out horizontally, with its steel-tipped end resting on a revolving
smoked-glass plate, as in the instruments already described. The supports s,
and s, are brackets projecting from a pust fixed in the ground.

§ 43. Forbes’ Inverted Pendulum Seismometer,

The writer has only reeently become aware of how nearly, in its essential
idea, the instrument described in § 42 is a return to a very old seismometer, the
invention of the late Principal Forbes. The following account of it, as erected
for the observation of earthquakes at Comrie, is taken from the British
Association Report for 1841, p. 47 :—

“ The Inverted Pendulum Seismometer.—(1.) The smallest of the instru-
ments made on this principle has a pendulum thirty-nine inches long, and is
fixed into a brass socket at its lower end. The connection between the pendulum
and the socket consists of a strong elastic wire, which, by means of a pinching
serew, can be either raised or depressed in the socket, so as to increase or dimi-
nish the length and sensibility of the pendulum. There is a leaden ball near
the top of the pendulum from three to four pounds in weight: it has a hole
through its centre so as to allow the pendulum rod to pass freely through it, and
it can be fixed at any part of the rod by means of a pinching screw. At the
upper extremity of the pendulum there is a soft lead pencil, which rests on an
elastic wire contained in a brass tube. The pencil is thus pressed against a
white surface of paper, forming the segment of a sphere, having a radius of
thirty-nine inches. The paper is pasted on a piece ‘of copper beaten into the
proper shape. This copper segment rests on four upright iron rods which are
fixed into the base of the instrument. The base consists of four corresponding
flat iron bars, which cross in the middle, and support at that point the socket
above deseribed, to which the elastie wire of the pendulum is fixed............ The
instrument is fixed firmly to the floor of the room where it is set. By means of
three adjusting serews, which affect the socket, the upper extremity of the
pendulum is brought to the centre of the segment to be marked by it. Any
further description of this instrument is rendered unnecessary in consequence of
a paper by Professor Forbes, published lately in the Transactions of the Royal
Society of Edinburgh, where the mechanism and mathematical properties of it
are very clearly pointed out. (2.) The other instrument constructed on this
principle has a pendulum ten feet eight inches in length. The spherical segment,
on which the vibrations of its point are intended to be marked, is not, as in the
instrument just deseribed, supported on upright rods fixed to its base, hut is
suspended over the pendulum by a strong hold-fast of iron fixed into a wall. In
other respects, the mechanical construction of this instrument is much the same
as that of the former one.”

It has unfortunately been impossible to obtain access, in Japan, to a copy of

the paper in the Edinburgh Transactions, referred to above, but there seems no
reason to doubt that the combination of an inverted pendulum with a flexible
spring at its base was used to obtain an approach to neutral equilibrium. The
bending of the spring in this instrument plays the same part as the inclination of
the stretched spring in the seismometer of § 42, in somewhat more than counter-
balancing the overturning effect of gravity on the bob when the pendulum is
displaced.

Other appliances intended, primarily, to give information about the hori-
zontal movement of the ground during earthquakes will be mentioned in

Chapter VI.*

* See also the Appendix.

CHAPTER IV.

INSTRUMENTS FOR OBSERVING VERTICAL MOVEMENT,

$ 44. Loaded Spiral Spring.

In attempting to register the vertical component of earthquake movements
by the use of a mass whose inertia is to furnish a steady point, we at once meet
with the difficulty that gravity acts in the direction in which freedom of motion
is be retained. The weight must therefore be borne by some contrivance which,
by being extensible, or in some other way, will leave the mass freedom to
oscillate in a vertical line. The simplest example is afforded by a mass hung
by a long spiral spring from a fixed support. By limiting its freedom to one
line we may restrict this apparatus to the measurement of vertical movements, or
by leaving it free to oscillate horizontally, like a pendulum, as well as vertically,
we may obtain a universal seismometer, which, if furnished with suitable indices,
may simultaneously record three rectangular components of the motion. It will
generally be found convenient, however, to deal with the vertical component
separately, and therefore to deprive the instrument intended to record it of
any but vertical freedom.

To make the equilibrium of the hanging mass nearly enough neutral, the
spring would need to be of great length, since, when hung directly by a stretched
spring, the period of vertical oscillation of the mass will be only that of a
simple pendulum whose length is equal to the distance by which the spring is
elongated. This, in fact, makes a directly loaded spiral spring almost imprac-
ticable as a vertical seismometer.

$ 45. Horizontal Bar with flexible support and loaded end.

A committee of the British Association appointed in 1841 “for registering
shocks of earthquakes in Great Britain” describe in their report* a seismometer
for vertical motion, consisting of “a horizontal bar, fixed to a solid wall by means
of a strong flat watch-spring, and loaded at the opposite end. If the wall
suddenly rises or sinks, the loaded end of this horizontal rod remains, from its
vis inertia, nearly at rest, and thus can move any light substance (as paper or a
straw) brought against it by the vertical movement of the ground, and which
light substance is so adjusted as to stick wherever the rod leaves it.” The
report goes on to say that an instrument of this kind, set up at Comrie, gave on
one oceasion a record of vertical movement to the extent of half an inch. Nothing
is told as to the period of vertical oscillation of the bar, whose equilibrium was
probably too stable to make it act well as a seismometer.

* British Association Report for 1842, p. 94.

$ 46. Horizontal Bar with stretched supporting spring and loaded end.

Instead of supporting the horizontal bar by a flat spring at one end, we may
simply hinge it at one end to a fixed support, and hold it up by a stretched spiral
spring connecting another fixed support above it with some point in the bar,
preferably a point not far from the hinge. This arrangement has been used by
Mr. Gray, who has also added an ingenious plan by which the equilibrium may
be made neutral. *

The attachment of the supporting spring to a point near the hinge gives, of
itself, a system possessing much more feeble stability than would be possessed by
a mass hanging directly from a spring stretched to the same extent. Thus, in
the figure, let a heavy mass be fixed to the rod ABC at C, while the point B is
connected with a support D by a spiral spring, and the rod is hinged at A by

pressing against a fixed knife-edge. Let n be

D : - AU e ‘
- he ratio of lengths —— Ve have, in gene-
MU tl t f 1 gtl Wel 1 @
7 2 AB
= ral, for a mass freely executing simple harmo-
ニー 7
== 3 3 ‘ 5 M
= nic oscillations, the period < = 2 ェ 8
=, R
= も
= where J/ is the mass and R is the whole force
= 1 DT is tl IR is the whole fi
as

= which tends to restore the mass to its position
; of equilibrium when its displacement from
that position is unity. To apply this to the
present case, we may either refer the force to

=
re or refer the mass to B, the point of attachment
4° . 選 ( 0 ) of the spring. Both processes of course lead

to the same result. Taking the latter plan,

the end of the bar C’, where the mass is placed,

the mass referred to B is n?m, where m is the actual mass applied at C. The
restoring force for unit displacement (of the point of reference, B) is the force

: 5 2 : 8 nmg
required to produce unit extension of the spring, and is equal to 9 where I
is the actual total extension.

liane Pe
> I 72726 nd

Hence . ーー だ Nase = N)
nmg q

which is greater in the ratio }/” : 1 than the period in which a mass would
oscillate if directly hung to the spring, and stretching it to the same extent.
The arrangement has therefore the advantage of giving more nearly neutral
equilibrium than is attained by the use of a directly loaded spring of the same
length. On the other hand, the effective inertia of the system is n times less
than it would be if the spring were equally drawn out by a weight directly
applied to it.

* Transactions of the Seismological Society of Japan, Vol. III, p. 187.

The addition which Mr. Gray has made, with the view of making the
equilibrium still more nearly neutral, is a trough or tube containing a liquid,
which is connected to the horizontal bar in such a manner that when the bar is
depressed the liquid moves in the tube so as virtually to increase the load on the
bar, and when it rises the liquid moves so as virtually to decrease the load. The
apparatus isshown in sectional elevation in Plate VII, fig.35, which, with the follow-
ing description, is taken from the Philosophical Magazine for Sept. 1881, p. 209:—

“A vertical spring S is fixed at its upper end by means of a nut n, which
rests on the top of the frame J’, and serves to raise or lower the spring through a
short distance as a last adjustment for the position of the eross-arm A. The arm
A rests at one end on two sharp points, p, one resting in a conical hole and the
other in a V-slot; it is supported at B by the spring S, and is weighted at の
with a lead ring R. Over a pin at the point € a stirrup of thread is placed
which supports a small trough +. The trough ¢ is pivotted at a, has attached to
it the index © (which is hinged by means of a strip of tough paper at A, and
rests through a fine pin on the glass plate g), and is partly filled with mercury....
When the plane carrying the spring 4 is raised and lowered, the point a rises and
falls, but in consequence of the inertia and slow period of Z2 the point C remains
behind. In consequence of this the end of the trough ¢ falls and rises relatively
to a; and the mereury, running backwards and forwards, puts more or less foree
on the point C, and hence tends to keep this point stationary.”

An objection to this arrangement is that horizontal oscillations of the ground
produce a heaping up of the liquid at one and the other end of the tube alter-
nately, and so throw the bar into a state of forced vertical oscillation.

$ 47. Astatie Horizontal Bar, without liquid,

A simpler and equally effective method of reducing a loaded horizontal bar
to approximate astaticism, the invention of the present writer,* is as follows:—
Instead of changing the load upon the bar, when the bar rises or falls, we may
change the leverage at which the spring acts. The moment of the weight is
balanced, about the fulerum, by the moment of the upward pull of the spring.
In any small displacement the moment of the weight is sensibly constant. For
neutral equilibrium we must therefore make the moment of the pull of the spring
constant. ‘To do this we must cause the point at which the spring is attached to
the har to move towards the fulerum when the spring lengthens, and away from
the fulerum when the spring shortens. This is easily done by putting the point
of attachment B (in the sketch on p. 48) not in the horizontal line joining A with
C, but at a distance v vertically below that line. Let 4 be the horizontal
distance of the spring's line of action from the fulerum A. ‘Then if we suppose
the bar to be displaced downwards through any very small angle d@ from its
normal horizontal position, the point of attachment of the spring goes vertically

* Trans. of the Seismological Society of Japan, Vol, IIT, p. 140,

down through a distance hd@, and horizontally towards the fulerum through a
distance vd.

Now if we call / the total elongation of the spring in the normal position of
the lever, it follows from Hookes Law that the upward pull is increased (by the
supposed displacement) in the ratio

7 十 hdd .
nt
and the leverage at which this acts is diminished in the ratio
ヵ ー vdd
and ;

To make the system astatie, the produet of the pull into the leverage is to
be constant, being equal to the moment of the weight, and hence we must have
Ih = (l + hd) (h — vd).

From this, neglecting the term involving dé’, we have

大

ャ デラ っ
an equation which determines the distance v below the horizontal line of the bar,
at which the pull of the spring is to be applied, to give neutral equilibrium for
infinitesimally small displacements. In practice, v must be made somewhat
less than this, to leave a small margin of stability.

$ 48. Actual Vertical-Motion Seismograph.

An actual vertical seismograph designed on this plan, and erected in the
writer’s observatory, is shown on Plate XI, where fig. 34 gives a side elevation,
fie. 35 a front elevation, and fie. 36 a plan, The whole is supported on a post
P, stuck in the ground. The horizontal bar is a stout brass plate a loaded at
the outer edge with a cylinder of lead b. Close to the edge of a nearest to the
post a conical hole and V-slot are ent, as far apart from each other as the width
of the plate will alow; and into these a pair of steel points ce, fixed to the post,
press. These points determine the horizontal axis about which the plate is free
to rotate. A pair of springs dd are used to hold up the weight: the lower ends
are connected by the hanging bar e, at the centre of which there is a conical hole
against which the serew f in the plate a presses. ‘This part of the apparatus is
separately shown in fig. 37, which gives a transverse section through the plate
a, in the plane of the springs aud the hanging bar, The depth of the point
of f (the point, that is, where the pull of the springs acts) is adjustable by
turning the screw, and it is secured at any position by the jam-nut g. The
upper ends of the springs are fixed to a board A, whose position is determined
and whose height and plane are adjustable by three screws ij, in a manner
readily intelligible from the figures.

The steady line of the system (the instantancous axis with respect to a vertical
displacement of the axis of support) lies, of course, a little way outside the centre
of the bob 5b. At that line a pin % is fixed to the plate a, which gears into a slot

#
<

öl

in the cranked multiplying lever 7. "This lever is hinged about a fixed horizon-
tal axis at m, its bearings being carried by a bracket a standing out from one
side of the post. The longer part of the lever is of straw; and its lower end
carries a hinged marking pointer whose motions are recorded on the smoked-
glass plate 0. The multiplying ratio is 7 to 1.

In this instrument it is scarcely practicable to keep the glass plate revolving
continuously, Changes of temperature and other causes make the pull of the
springs so variable that the equilibrium position is rarely constant for any long
time, apart from earthquake disturbance. For the same reason static records are
useless; and the plan is adopted of starting the plate into rotation as carly as
possible in the disturbance, by making the earthquake close an electric circuit
which releases a eatch and allows a driving clock to come into action. The
arrangement will be deseribed in a later chapter.

The bob b might of course be pivotted at the instantancous axis of the plate,
instead of being fixed to the plate: but, as its diameter is small compared with
its distance from the axis of support, we should gain little, in the way of steadi-
ness, by this complication. The mass of the multiplying lever is very small;
otherwise we should place above its hinge m a counterpoise which would raise
its centre of gravity to the horizontal plane containing the hinge, in order to
prevent a horizontal displacement of the ground from producing any motion of
the system,

$ 49. Hydrometer Vertical-Motion Seismograph.

A proposed instrument is described by Mr. Gray under this title, in the
Philosophical Magazine for Sept. 1881. An enlarged hydrometer, with a narrow
projecting end, is immersed in a liquid, and weighted to have a slow period of
vertical vibration. It is constrained to move in a vertical line only, and the
projecting end is connected to a marking index.*

The same observer has also attempted to register the vertical motion during
earthquakes by using a metallic vessel with a corrugated bottom, filled with
liquid which can oscillate vertically on account of the flexibility of the bottom of
the vessel.

Both of these arrangements are unquestionably much less practicable than
the loaded horizontal bar, rendered nearly astatie either by Mr, Gray's or the
present writer’s method, and they may therefore be passed over with the briefest
mention.

* A buoy, free to sink and rise in a tub of water, had been previously used as a vertical-
motion seismometer by Dr. G. Wagener, (Trans, Seis. Soc. of Japan, Vol. I, p. 64.)

CHAPTER V.

RESULTS OF OBSERVATIONS.

§ 50. Stutie Records.

Previous to the first publication of the writers seismometrie observations *,
no observer had (so far as he is aware) succeeded in effecting a measurement of the
magnitude and direction of the ground’s motion, in conjunction with the time,
during any prolonged earthquake, Such measurements as had been made were,
almost without exception, of the static kind, into which the consideration of time
did not enter. ‘The information which statie records can yield is imperfect in a
very important particular. A knowledge of the direction and magnitude of
earthquake motions does not suffice to enable us to estimate the energy and
destructive power of the disturbance, since these things depend, most directly, on
the relation of displacement to time. And it may fairly be added, that no statie
record can determine conclusively even the amplitude of an earthquake vibration,
If the seismometer has much stability, its oscillations may become great enough
to exceed and altogether mask the true earthquake motion; unless, indeed, there
is much friction, in which case the records will be too small. On the other hand,
if the equilibrium of the seismometer is nearly neutral, it will be liable, during a
disturbance consisting of many successive waves, to work gradually away from
its original position, and so to give a record which is greatly in excess of the
amplitude of any single movement.

It would be scareely possible, and not at all profitable, to give a full account
of the early static measurements of earthquakes. The observations made by
Mr. Knipping in Tokio with Dr. @. Wagener’s pendulum seismometer have been
already alluded to ($ 36). These, and also the earlier observations of Dr. Verbeck
($ 31), were valuable as showing that the amplitude of motion in ordinary earth-
quakes is so exceedingly small that a considerable multiplication is indispensable
in the recording apparatus; and at the same time that it is needless to let the steady
mass be free to moye, or be in neutral equilibrium, throughout more than a very
small range of positions.

In an account of a somewhat destructive earthquake felt m Tokio and Yoko-
hama in February 1880+, Mr. Milne gives some static records, which he obtained
by means of a long pendulum, without multiplication. The eurves are very
irregular and not easily intelligible ; they show pretty clearly the comparatively

* Transactions of the Asiatic Society of Japan, Vol. IX, p. 40 (Dee. 14, 1880).
7 Transactions of the Seismological Society of Japan, Vol. I, part 2.

ee Se Bee ee FE ee as WA ee ale ae oe

wily し v4 a: . Lip cS ie et ale

BB}

rapid undulations of the earthquake superposed on larger displacements which
are presumably due to the swinging of the pendulum. Mr Milne also gives
results obtained with another long pendulum, which had two pointers writing
without multiplication on strips of smoked-glass, which were set in motion by the
disturbance, and were caused to move under the pointers for a time estimated at 3
seconds. During this time short portions of two undulating curves were traced ;
a but the amount of displacement shown was too small to admit of exact measure-
ment, and the time ineluded by the records was probably not more than a fiftieth
part of the whole duration of disturbance.

A remarkable series of examples of statie records has been furnished by the
observations of Father Faura at Manila during the great earthquakes of July

a 14th to 25th, 1880. These records, taken by a common pendulum writing in
: lycopodium dust without multiplication, will be found deseribed and figured in

the Proceedings of the Royal Society of London, Vol. XXXI, p. 460. Nothing
is said as to the scale of the figures, which are probably much less than full-size.*
In the description which accompanies them the amplitudes of horizontal motion
are stated only in terms of the angles of inclination made by the pendulum, but
7 its length is not given. Strangely enough, the writer appears to suppose that the
inclination of a vertical pendulum during an earthquake is a measure of the slope
of the surface of the ground, when that is horizontal before disturbance. The
c diagrams are very complicated, and it is impossible to make out, by inspection of
* them, how much of the recorded displacement is a true motion of the ground, and
how much is due to the swinging of the pendulum.

§ 51. Records on continuously moving plates.

The present writer was enabled, through the liberality of the President of り
the University of Tokio, to establish a seismometrical observatory there in 1880.
区 The instruments originally erected were the long pendulum with two multiplying
pointers (§ 37), and the horizontal pendulum, or rather pair of horizontal pendulums
ae (§ 23), also with multiplying pointers. In both cases, plates driven with a con-
tinuous motion by clockwork were used to receive the records, the writer having
been early convinced of the almost complete uselessness of static records, A
number of the other instruments described in Chapters II, IIT and IV have been
added, but most of the records hitherto obtained have been given by the horizontal
pendulum, in one form or another. From the establishment of the observatory to

the present time a very large number of earthquakes have been recorded, out of 4
2 which a few typical examples have been selected for deseription, and are given ü
- below. Plates XII to XX are fac-simile reproductions of photographs taken >

from the smoked-glass plates on which records were automatically inseribed.

I * A diagram of the Manila earthquake, which is the same as the fig. 2 of the Royal Society
* account, will be found in the Japan Weekly Mail of Aug. 14th, 1880. If this is (as it appears
to be) a fac-simile of the pendulum record, the Royal Society diagrams must have been reduced
in the proportion of about 1 to 24.

54 ;

$ 52. Eurliest Records.

The following account of the earliest earthquake record obtained by the
writer is taken, with little alteration, from a paper entitled “Notes on Some Recent
Earthquakes,” published in the Transactions of the Asiatic Society of Japan, Vol.
IX, p. 40. This was nothing more than a feeble shock, which did no damage
to buildings, and was just such as residents in Tokio expect to occur at intervals

of a few days.

EArtoqQuake or Noy. 3rp, 1880, 5u. 45 Mm. A.M.— RECORD GIVEN BY THE
HorizontaL PENDULUM SEISMOGRAPH.

One pointer registered East-West motion, the other North-South motion.
The glass plate of this instrument was revolving at the rate of one turn in thirty
seconds, and the record extended over three complete revolutions. In other
words, the earthquake lasted continuously during one and a half minutes of time,
During this time no fewer than 150 complete (double) oscillations of the earth’s
surface took place. The mean period of a complete oscillation was, as nearly as
possible, three-fifths of a second. The motion was almost wholly North-South ;
the other pointer showed a very small, though perceptible, amount of disturbance.
The records given by-both pointers began simultaneously, as far as can be judged,
but the East-West moyement soon ceased, while the North-South movement
lasted, as has been said, for 14 minutes, and consisted of more than 150 waves.

The earthquake did not begin suddenly. The waves began to appear so
gradually that it was impossible to say to which side the first deviation took
place. The amplitude increased, however, pretty rapidly, and reached a maximum
after about three complete waves. Allowing for the multiplication introduced by
the recording lever (6 to 1, in this case) the greatest displacement of a particle
‚on the earth’s surface was 0°29 mm. in a direction approximately North and
South, and 0:05 mm. in a direction approximately East and West. The two to-
gether gave a motion of about 0°30 mm. in the direction N. 15° W. and E. 15°S.

doth in amplitude and in period the successive waves were far from regular.
The disturbance did not consist of a series of simple harmonic displacements.

After about the third wave from the beginning of the disturbance the
amplitude of motion ceased to inerease. It then fluetuated considerably, some-
times becoming almost inappreciable, and again increasing to a value approaching
its first maximum. Before the earthquake ccased there were several maximums
and minimums in the amplitude of motion, but never a complete cessation until
the whole disturbance was over.

RECORD GIvEN BY THE Lona PENDULUM SEISMOGRAPH.

This record showed that the pendulum was set swinging by the shock through
a distance of about 1 mm: and as the swinging continued with gradually diminish-
ing amplitude during many revolutions of the plates, the records were much
obscured. It was, however, easy to see the comparatiyely rapid undulations

of the earthquake superposed on the long-period waves due to the swing of the
pendulum. The greatest actual displacement registered by the North-South
motion pointer was two-sevenths of a millimetre: the other pointer showed no
more than a trace of motion. These results agree very closely with those given
by the other instrument.

If we assume the greatest displacement to have taken place approximately
according to the simple harmonic law, then, taking the period as 0.6 second and
the total range of motion as 0.3 mm., we find for the greatest velocity 1.57 mm.
per second, and for the greatest rate of acceleration 16.4 mm. per second per
second, The value of y in the same units is 9797. Bodies connceted to the earth’s
surface sufliciently rigidly to move with it must therefore have experienced, in the
earthquake, a maximum horizontal force equal to about 45 of their own weight.

After describing (in the same paper) the records of four other small earth-
quakes which also occurred during the month of November, 1880, the writer
drew attention to the following as the most striking features of these carly
observations :—*(1) The very gradual beginning and ending of the disturbance,
In none of the observations did the maximum motion oceur until after several
complete oscillations had taken place. (2) The irregularity of the motion. The
suceessive undulations are widely different both in extent and in periodic time.
(3) The large number of undulations in a single earthquake, and the continuous
character of the shock. (4) The extreme minuteness of the motion of the
earth's surface.”

$ 53. Eurthquake of February 7th, 188 1*

Plate XII is the record of a small earthquake, given by the horizontal
pendulum seismograph (of the form deseribed in $ 23). On this oceasion the
displacement appears to have been confined to one direction, (approximately Past-
West): for this reason the tracings made by one pointer only are reproduced
in the Plate, The glass plate of the instrument made one revolution in 104
seconds, and the multiplying ratio was 6 to 1. The visible disturbance begins at
the point a, at the top of Plate XII. The innermost of the two cireles visible on
the Plate is that which the pointer was tracing out prior to the earthquake: on
it, at a, a feeble undulation will be seen to begin, and continue in the direction
of the arrow, which is drawn opposite to the direction in which the plate was
revolving. It is not until about twenty-five seconds, or nearly a whole quadrant
of the glass plate from the beginning of visible motion, that the displacements
become at all considerable. The greatest motion occurs at b (at the bottom of
the Plate). At that point the motion from one side to the other is 6 mm. on the
record, corresponding to I mm, of actual horizontal motion of the ground,

The undulations continue with varying amplitude for about 13 complete
revolutions of the plate. The latest motion which can be seen on the record

* See also Trans, of the Seismological Society of Japan, Vol, III, p. 115.

occurs at the point c on the outer cirele. The outer cirele is the line traced by
the pointer after the cessation of the disturbance.

It must not be supposed that the want of coincidence between the circles
traced before and after the earthquake is evidence of a permanent displacement
of the ground, since during a prolonged disturbance the accumulated effects of

frictional forces are almost sure to produce some resulting motion of the so-called °
steady line, especially when the equilibrium is nearly neutral,

This record exemplifies very well the characteristics mentioned at the end of
the preceding paragraph. The motions are at first so small that it is difficult to
point out the beginning definitely, and the undulations die out so gradually that
it is equally difficult to determine the end. Again, though the amplitudes are
far from regular, we cannot point to any one displacement as the principal shock.

At b, where the motion is greatest, the displacement (from West to East) is
6 mm. on the record, that is 1 mm. of actual earth motion. The motion there
is not very far from being simple harmonic, with a period of 1,4 sec. This gives
a maximum velocity of 2.3 mm, per sec. and a maximum acceleration of 9.6 mm.
per see. per sec., or say Tn of the acceleration due to gravity. In registering
earthquakes such as this, the necessity of avoiding friction in the seismometer
will be apparent (§ 38).

$ 54. Earthquake of March Sth, 1881*

Plate XIII is the record of this earthquake, which was one of unusual
violence compared with the minute disturbances to which residents in Tokio are
accustomed,

The record, which is in some respects the best the writer has obtained of
any earthquake, was given by the horizontal pendulum seismograph of § 23.
The multiplying ratio was again six to one. Of the two records on Plate XIII,
the inner is N-S motion, the outer E-W motion. The cross lines d and d’ show
the distance measured eireumferentially between the marking ends of the two
pointers. One revolution of the plate corresponds to about 80 seconds of time,
The beginning of motion is at a on the E-W circle (at the top of the Plate). At
a’, which is the corresponding point on the N-S circle, and for some time later, ;

scareely any displacement occurs. The earliest considerable N-S motion is at ,
which is synchronous with the E-W displacement が 。 Midway between a and b
there is a considerable E-W motion which has nothing to correspond with it on
the N-S cirele.

The most violent motions occur a little later than the points marked ヵ and
I’. The record extends over nearly two complete revolutions of the plate, up to

Bez た"

the point ¢ in the outer cirele, where it will be observed to leave off abruptly. a
This is because the writer, who had the good fortune to be present in the obser- 4
vatory during this earthquake, withdrew the plate at the point marked ce, in i

* See also Trans. of the Seismological Society of Japan, Vol. II, p. 121.

‘ 7 内 ar
al = RI Main 4 is

Earthquake af March 8 “er (To kto)
The. motion es magne fred ste times . P

The numbers denote yaa i ices of lime.

order to prevent confusion by the superposition of the later feeble undulations on
the earlier important ones. ‘Lhe point e', in the inner circle, corresponds to ce in
the outer. By the time the plate was removed the disturbance had lasted for
about 24 minutes, and nearly 200 waves had been registered. Feeble move-
ments were perceptible for some time longer.

The earlier and more interesting part of the record is copied in the accom-
panying figure (inserted here for convenience of reference), There the circular
path traced by each pointer has been cut up into four ares, each of which corres:
ponds to 20 seconds, and radial lines have been drawn to mark intervals of one
second. The four ares are consecutive. They begin at the point where the mo-
tion was first visible (the point a of Plate XIII). Each portion shows a
pair of records traced by the two pointers, and the lower one is turned round so
that contemporary points in the two lie on the same radius. Much care has been
taken to verify, by direct measurement of the glass plate, the coincidence in time
of the points here shown as coincident, since measurements made on the paper
photographs are liable to error on account of the unequal contraction of the sheet
in drying. The eirenlar path which was being traced by each pointer before the
surthquake is shown in the figure by a faint line.

Up to the tenth second the motion is almost wholly E-W. Then a large N-S
component comes in, and this continues to be present during the remainder of the
earthquake, except near the end of the record (as it appears on plate XIII),
where it will be seen that the N-S motion dies out sooner than the other.

Perhaps the most interesting feature of the record is the varying relation of
the W-E and N-S motions during the disturbance.

At first, for nine seconds, one occurs alone. From the tenth to the thirteenth
second there is synchronism between them; during that time displacements
toward FE. and S. are contemporaneous. A little later, about the thirtieth
second, there is an equally distinet synchronism ; but this time motions towards E.
and N. are contemporaneous :—there has been a change of phase-relation amount-
ing to just half a complete oscillation. In general, in other parts of the record
it is impossible to trace any distinct relation between the two components. For
instance, in the twenty-eighth second there is a notable movement from West to
East which is not associated with contemporary displacement of the other pointer.
The same remark applies to the forty-third second, and to the seventy-sixth.

Towards the end of the earthquake this independence of the two components
of motion was very clearly seen by the writer, who noticed that one pointer some-
times moved vigorously while the other was nearly at rest, whereas a few seconds
later the pointer formerly at rest took up the motion, and the movement of the
other almost ceased.

This change of phase-relation in the two components is evidence, of course,

that the direction of movement of the ground, was continually varying during the
disturbance. Ln some parts this variation oceurred so rapidly as to make the path
of a surface-partiele very far from rectilinear.

A remarkable instance of this occurs during the sixteenth second, when the
amplitude of motion was it its maximum. The W-E component was then al-
most exactly a quarter of a period on advance of the N-S component, and hence,
when compounded, the two records show the motion of the ground to have con-
sisted, at that time, of rotation in a very roughly circular path, The annexed

figure is the path determined by com-
N pounding the two records during three
seconds, (from the time 13'7 see. to 16°7
nn see, measured from the beginning of the
Ah ‘arthquake). The path described in this
E interval begins at p and ends at q.

The displacements plotted in this

| ュー sketch are taken without change of scale

from the record, and hence the scale of

the sketch is six times full size.

The greatest single horizontal dis-
placement (from one side to the other) is about 25 mm. on the record, which
corresponds to nearly 42 min, of actual motion. The greatest velocity of hori-
zontal movement of a surface particle was nearly 10 mm, per see. It is difficult
to find the greatest rate of acceleration from the record, on account of the very
irregular character of the motion: it appears to have been from 40 to 50 mm, per
see, per sec., or about 515 of 7. The period of the largest single complete oseil-
lation (that which occurred at 15 seconds from the beginning) was 1'1 seconds.

It may be added that this period was unusually long, which aceounts for the
fact that, notwithstanding the large amplitude of motion, this earthquake was not
of a destructive character. Although alarming enough, it did no damage except

cause some cracks in walls.
$ 55. Marthquakes of March Ist, 1882,

Two earthquakes which occurred on March Ist, 1882, of which the second
was unusually sharp, though not destructive, are shown in Plate XLV, which is
the record given by a single horizontal pendulum, of the form described in § 25.
The record gives only the West-East component of the horizontal motion : the
pointer which should have registered North-South motion was unfortunately out
of order, and failed to do more than show that a considerable amount of N-S
movement had also taken place.

Here the plate made one revolution in 85 secs., and the multiplying ratio
was 7 to 1. Of the two eireles which appear close together in the record, the
outer is the path traced before the first carthquake, and the inner is the path
traced after the second,

The earliest visible motion is a tremor of short-period waves, which
appears on the outer circle at a. Following the record in the direction in
which it was deposited (the direction shown by the arrow), we see these short-

59
period waves appearing very faintly at intervals, and these appear to comprise
the whole record of the first earthquake.

The second and relatively large shock, which oceurred about two hours
later, also begins in a tremor of short period waves, which become very promi-
nent at 5, although they may be detected for some time before that point is
reached. In a few seconds more these develop into large motions of much
longer period. The greatest displacement is at c, where a lurch from E. to W.
measures 35 mm. on the record, corresponding to 5 mm. of actual earth motion.
A little later an interval of comparative rest occurs; but violent motions begin
again at d. The disturbance continues irregularly, and with gradually diminish-
ing amplitude, for nearly 12 revolutions more. The last trace of visible motion
is at e on the inner cirele, The second record extends over fully 2 revolutions
in all, showing that the earthquake lasted for three minutes.

The short-period waves with which the earthquake begins are present, more
or less, during one minute. At first they exist alone. Then they may be traced
superposed on the great movements which constitute the principal shock. But
in the later parts of the record, during the gradual dying out of disturbance,
they do not appear: the motion then consists entirely of long period waves.

The short-period waves at the beginning and during the early part of the
earthquake have a period of only 0.17 see., and their amplitude is no more than
a very small fraction of a millimetre. The greatest back and forth motion, at ce,
oeeupies about 0.8 sec. As the disturbance ceases the average period of the
undulations appears to increase.

The record is, as has been said, incomplete, inasmuch as it gives only one
component of the horizontal displacement. Independent evidence showed, how-
ever, that the component given here was the principal part of the motion. or
a rough estimate of maximum velocity and acceleration in the East-West direction,
we may take the motion at ¢ as a simple harmonic displacement with period 0.8
see, and amplitude 2.6 mm. This gives a maximum velocity of 19.6 mm. per
si and a maximum acceleration of 154 mm. per see. per see., or I; of 7.

$ 56. Earthquake of March 11th, 1882*

Ten days later than the earthquake just deseribed, another oeeurred, of
suflieient violence to do a small amount of damage to buildings, by overturning
chimneys and cracking walls. A record of this shock, obtained by a pair of
horizontal pendulums of the eone-bob form (described in § 24), is given in Plate
XV, The inner pointer was placed somewhat obliquely to the radius of the
plate, which accounts for the skew character of its markings. It registered E-W
movement, and the outer pointer N-S movement. Previous to the occurrence of
the earthquake, the cireular path traced by the outer pointer had become so wide

* See also Trans, of the Seinmologien! Society of Japan, Vol, IV, p 73. In that paper the
multiplying ratio weed in obtaining this record was inadvertently given as 3 to 1, instead of
4tol,

— qa _ 2 Se ee ne

as partially to obscure the record. The multiplication was 4 to 1; and one
revolution of the plate took 88 seconds.

The earthquake began, as usual, with very small motions, at or near the
place marked a in the inner circle and a’ in the outer. (The circumferential
distance between the ends of the two pointers is shown by the simultaneously
inseribed lines b and 2"). After a few seconds, at ¢ and the contemporary point
c', the motion became suddenly violent, and, though the earthquake lasted for a
long time, the displacements which occurred at c and c’ were not exceeded or
even equalled during the remainder of the disturbance.

The record covers more than three revolutions, showing that the earthquake
continued for at least 44 minutes. The greatest single displacement (at c) from
E. to W. is 21 mm. on the record, or 5.25 mm. of actual motion, Contempora-
neous with this is the greatest S, to N. displacement (at ec’), which is 18 mm. on
the record, or 4.5 mm, of actual motion. Combining these, we find that the
greatest single movement of the ground was 7 mm., in the direction N.W.,
nearly, This is the return which followed the first great displacement, whose

direction was nearly S.E., and whose amplitude was somewhat less than that of

the return.

At this point the period was about 0.7 see. A simple harmonic oscillation
with this period, and 3.5 mm. amplitude, gives a maximum velocity of 31 mm.
per second, and a maximum acceleration of 280 mm. per sec. per sec., or z= of g.

The period of vibration lengthens towards the end of the disturbance; and
at the beginning there are traces of short-period waves preceding as well as
superposed on the principal movements.

$ 57. Earthquake of March 19th, 1882.

Plate XVI gives the record of another earthquake, which came in the same
month as the shocks described in the two preceding paragraphs, and was regis-
tered by the seismograph of § 24. This was a less severe, but very long
continued disturbance. The East-West record (the outer of the two) was some-
what obseured by a broadening of the eirele traced prior to the earthquake.
The disturbance extends over just three complete revolutions of the plate, and
the end overlaps the beginning in a way which makes it almost impossible to
distinguish the beginning or the end.

The multiplying ratio was 4 to 1; and one revolution of the plate occupied
75 seconds. The greatest displacement is about 2 mm, The mean period,
during the early part of the disturbance, is somewhat less than 1 see., but this

inereases very much towards the close.
S 58. Earthquake of August 18th, 1882,

The record of this somewhat sharp shock is shown in Plate XVIT, in
which the multiplying ratio is 4 to 1. The motion was principally E-W, and
the pointer registering that component happens to have been set somewhat

obliquely to the tangent of the circle it drew. The lines e and e’ serve to show
the relative positions of the pointers. The first visible disturbance is at a (on
the inner record). The record extends over about two revolutions. This is an
excellent example of the presence of short-period vibrations of small amplitude
in conjunction with larger and slower motions. The largest displacement, at /,
measures 4.25 1am. when reduced to the natural scale.

§ 59. Earthquake of Septe mber 29th, 1882,

A moderate earthquake of the above date illustrates well the arrival of
short-period vibrations in advance of the principal movements. Plate X VIII
shows the record, which consisted almost wholly of E-W motion. It began at
a and was traced in the direction of the arrow. From a to b nothing appears
bat a ripple of minute short-period waves. The greater motions, which begin
at b, have at first short-period waves mixed up with them; but these disappear
after a time, and the earthquake dies out in relatively long-period waves only.

In this ease the early part is shown with unusual clearness, owing to the
fact that previous to the shock the pointer was held by friction somewhat inside
of its position of equilibrium, and consequently the first shaking caused it to
ereep outwards from the centre.

The multiplying ratio was 6 to 1, and one revolution of the plate took 69
seconds.

§ 60. DBarthquake of April 23rd, 1883.

A sharp and Jong continued shock ‘(followed by another slighter shaking
about ten minutes later), which occurred at 10.35 p.m. on April 25rd, 1883, is
recorded on Plates XIX and XX. Plate XIX is the record given by the
horizontal pendulums of § 24, with a multiplication of 4. One revolution of the
plate took 54 seconds. The shaking lasted so long as to make it impossible to
follow all the motions, but the principal part of the disturbance is very clearly
shown. The earliest visible motion is at a on the E-W eirele. At b, nine
seconds later, the motions suddenly beeome large. The corresponding part of
the N-S record is }’. The record covers five revolutions of the plate, or 44
minutes of time.

—- Ts

Plate XX is a very remarkable record of the same earthquake, given by the
long pendulum instrument of § 37, with one pointer free to move in any azimuth,
and having a multiplication of 10. The plate revolved once every 72 seconds.
Nothing could show more clearly than this does the extraordinary irregularity of
earthquake motion. In fact, it is impossible to apply the terms used in speaking
of regular undulations to the spasmodie twists and wriggles which the surface of
the earth here executed,

In Plate XX the direction of motion of the plate under the pointer was
E.30 N. The principal movements are at the place marked ec, and these corres-
pond to the records given by the horizontal pendulums at b and 4’ on Plate XIX.
6

It is interesting to compare the records of the principal movements given by the
two instruments, and for this purpose the two components on Plate XIX have
been combined.

It must be bornerin mind that the diagram, as lithographed from a photo-
graph, bears to the original tracing on the plate the relation of a mirror image
to its object, and therefore the directions N.E.S.W. follow each other counter-
clockwise instead of clockwise.

Lines, nuinbered 1, 2, 3 ete., have been marked on the two components of
the horizontal pendulum record, to show corresponding instants of time; and
these lines are drawn (by means of a template) so that each is the path which the
corresponding pointer would have described if it had been caused to oscillate
while the plate stood at rest. Earthquake displacements are therefore to be
measured along them.

Tt will be seen at once that the phases of the two components do not agree,
in other words that the motion of the ground is not rectilinear. The actual path
of a surface particle, got by carefully compounding the displacements as deter-
mined by Plate XIX, from the points numbered 1 to 16, is shown in the upper
figure in the centre of Plate XX. To make it immediately comparable with the
long pendulum record it is, however, magnified two and a half times, so that it
bears to the earth’s motions the ratio of 10 to 1. The path begins at the origin
and is traced in the direction of the arrow heads: it ends at a point correspond-
ing to the points numbered 16 in Plate XIX. The time taken to trace this
path was 3°7 seconds.

The lower figure in the centre of Plate XX is a fac-simile of the record
given by the long pendulum during the same interval: it is inverted so as to
form a true fac-simile and not a mirror image. Its scale is also ten times
full size. The line marked on it E.30°N. is the direction of movement of the
plate under the pointer. Hence the lower diagram ought to agree with the
upper one, if we suppose the upper one to be stretched out at a uniform rate,
during the time of its being described, in the direction W. 30° S.—the direction
shown by the dotted line in the upper figure.

An inspection will show that the agreement between the two records is very
satisfactory, if we make allowance for slight movements on the part of the “steady
point”, which, in such a violent disturbance and during so considerable an inter-
val, are in fact almost inevitable. By compressing the lower figure along the
tangent to the circle we get a path very similar to that in the upper figure. The
two instruments which gave the records here compared are so different in all res-
pects that the agreement of the records, when thus examined, affords very strong
evidence of the substantial accuracy of both.

The greatest range of actual earth movement in this case was 5 mm, The
greatest vertical motion was less than 1 mm. The movements were so far from
cyclic that we cannot, except very loosely, speak of their period. For the prin-
cipal movements it was about 0.8 sec.

A record of the same earthquake, given by the duplex pendulum (§ 41) ona
stationary plate, with a multiplication of 3.5, is shown in the centre of Plate
XIX. Near the origin the motions are so numerous and so various in direction
that a considerable area of the jamp-black has simply been cleaned off, and it is
only the more prominent motions that are traceable. The prominent part of
the figure, however, agrees very closely with the path shown by the other two
instruments for the same portion of the motion, and so gives additional evidence
of the accuracy of the records.

§ 61. Summary of results.

The examples which have been given are fairly representative of a much
greater number of records, obtained during two and a half years. In the great
majority of cases, however, the motions have resembled the smaller rather than
the larger of the earthquakes here cited. In very many instances the recorded
displacements are too small to admit easily of reproduction.

The following conclusions may be taken as established, with respect to the
earthquakes which occur with much frequency in the Plain of Yedo.*

(1). In almost every instance the motion of the ground begins very gra-
dually, and at least some seconds pass before it reaches its maximum. This
makes it impossible to determine accurately the time of occurrence of an earth-
quake by means of a mechanical or electrical clock-stopping or time-taking
apparatus, since the time so determined will be dependent on the sensibility of
the apparatus to disturbance.

(2). An earthquake consists of many successive movements, and there is, al-
most always, no single larg+ one which stands out prominently from the rest.
As a rule there are many movements of nearly equal range, and sometimes several
maximums, with intervals of comparative rest between.

(3). The disturbance ends even more gradually than it begins. "The motion
of the ground dies out in a long series of undulations with irregularly diminishing
amplitudes.

(4). The range, the period, and the direction of movement are exceedingly
and irregularly variable during any one earthquake.

(5). The duration of disturbance of the ground is rarely less than one minute,
and often several minutes.

(6). Even in somewhat destructive earthquakes the greatest displacement of
a point on the surface of the soil is only a few millimetres, and there are very
many minor earthquakes in which it is less than one millimetre,

(7). In many cases the beginning of visible motion consists of a tremor of

* The writer's observatory is situated on one of the lowest parts of the great alluvial plain
which forms the site of the city of Tokio. It stands at a considerable distance from any elevated
ground. Water, in abundance, is reached about one metre below the surface. The walls of the
building which contains the seismographs are light wooden structures placed on the surface,
without any foundations to disturb the continuity of the ground inside and outside.

short-period waves of small amplitude, which are followed by the principal
movements. Often the early principal movements carry, superposed on them,
small short-period waves. These generally disappear before the end, and the
arthquake as it dies out consists of long-period waves alone. The first tremor
and the subsequent large movements sometimes occur in a manner stronely sug-
gestive of the idea that the first, or short-period movements, are normal waves and
the second, or principal movements, are transverse waves (§ 2).

(8). A progressive, though irregular, lengthening of period can sometimes
be detected towards the close of the disturbance.

(9). The period of the principal movements is usually from half a second to
a second ; but the short-period waves which occur at the beginning may have a
frequency of 5 to the second, or more. 1

(10), The vertical motion is generally much less than the horizontal.

SN 62. Complexity of the motion.

The long duration and great complexity of the movements which the sur-
face of the ground performs durme an ordinary earthquake in Tokio are no
doubt largely due to the (presumably) considerable distance of the observing
station from the origin of disturbance. The theoretical considerations which
have been adduced in Chapter I lead us to expect that (except as regards amplitude
of motion) earthquakes grow as they travel through heterogeneous media; and
observations of the same shock at different places show that this does happen.

In a paper descriptive of the earthquake of March 8th 1881,* which
supplied the earliest clear evidence of change in the direction of movement
during a disturbance, the present writer has suggested various explanations of
this change. These are: (1) the presence of normal and transverse waves toge-
ther. Even transverse waves alone may give non-rectilinear movements (§ 1),
whose horizontal components may also be curved provided the plane of the wave
is not vertical. (2) A possibly wide and not very distant origin, giving waves
with various azimuths, (3) The reflections, refractions, and diffractions which
oceur along the route. (4) The possible simultaneous, or nearly simultaneous,

- occurrence of two or more separate earthquakes. This last explanation is not so
extravagant as it may at first sight appear to be, if we consider that in a district
liable (as the Plain of Yedo is) to frequent earthquakes, there must often exist
states of critical equilibrium, ready to act as centres of earthquake disturbance.
The vibrations of a shock occurring at one point might therefore, on reaching
one of those places, start another convulsion, which would add its effects to those
of the first on neighbouring parts of the soil,

It appears, however, that changes in the direction of movement during an
earthquake are usual rather than exceptional; and they must therefore be
aseribed, in general, to the first three of the causes named above—probably to
the first and third in the great majority of cases.

* Trans. of the Seismological Society of Japan, Vol. III, p. 121.

The non-rectilinear character of the motion, which can be deduced (as has
been done above) by compounding two rectangular components of the horizontal
motion when these are separately recorded, is most directly shown by seismographs
possessing two degrees of horizontal freedom and giving records of which Plate
XX is an example. Even in very small earthquakes the indices of these
instruments frequently exhibit an immense number of movements in all
possible azimuths; so much so, that if the smoked-glass plate forming the
record receiver is stationary, the lamp-black is sometimes completely rubbed off
throughout a small area surrounding the point at which the index stands when at
rest. The writer has been close to the duplex pendulum seismograph (§ 41)
when a small (scarcely perceptible) earthquake began, and has watched the
marking pointer draw circles, loops, figures of 8, and numberless other curves.
It appears, in fact, that the tangled character of the movements, which is striking-
ly shown in Plate XX, is by no means confined to such considerable shocks as the
one there recorded, but is, in general, equally present in the most feeble disturb-
ances of which records have been obtained.

In the earthquake of § 60 the principal movements consist of wide loops,
nearly as broad as they are long. If we were to ascribe these to the simultaneous
arrival of two systems of rectilinear vibrations, we should have to conclude that
the principal movements in both systems reached the observing station together
—a thing in the highest degree improbable. Taken in conjunction with the
small amount of vertical movement, these (and similar features in other records)
appear to admit of but one rational explanation,—that they are transverse waves
whose direction of emergence is not very far from vertical.

CULAR EE iV

MISCELLANEOUS INSTRUMENTS,

§ 63. Fluid Pendulums.

In Mallet’s “Fourth Report on the Facts and Theory of Earthquake
Phenomena ”* an account will be found of a number of early seismometers, most
of which may be classed either as solid pendulums or as fluid pendulums, To
the first class belong the simple pendulum and the inverted pendulum of Forbes,
to which allusion has already been made (§ 43). Under the second head we
may include all arrangements in which a mass of fluid is caused to oscillate
relatively to the earth by the earth’s movements. These last are not well
adapted to the absolute measurement of earthquakes, and are mentioned here
chiefly for the sake of showing their inability to act as anything more than
seismoscopes. Like most other early seismometers, they seem to have been based
on the idea that an earthquake consists essentially of a single sudden impulse.

(1). Probably the oldest of these is that of Cacciatore, described thus by
Mallet :—“ It consists of a wooden circular dish about 10 in. diameter, placed
horizontally and filled with mercury to the brim-level of cight notches that face
the cardinal points and the bisecting rhumbs between, and are cut down through
the lip of the dish, equally in width and depth all round. Beneath each such
notch a small cup is placed, to receive such mercury as may be thrown out of
each notch by an oscillatory displacement of the main mass of mercury, due to a
general oscillation of the whole system. Either the volume or the weight of
mereury found in each cup is supposed to measure the value of the displacement,
and hence of the shock in its direction in azimuth.” (2). Another, suggested by
Babbage, is a bowl of molasses or other viscous liquid. (3). Another (Mallet)
is a cylindric tub with chalked or whitewashed sides, partially filled with liquid,
which is coloured so that it may show the height to which it is washed up
during the disturbance. (4). An improvement on any of these is a set of U-
shaped tubes, placed in different azimuths, and partially filled with mercury
(Palmieri and Mallet).

All these are liable to the objection already urged against a common short
pendulum,—that of having too great stability. This makes the amount of their
movements depend so greatly on the agreement or non-agreement of the liquid’s
period of oscillation with that of the earthquake as to deprive these of all value,
even as indications of the relative intensity of different shocks. If we avoid
accumulated oscillation by using a viscous liquid, (2), we introduce a friction
error of great but, generally, very unequal amount in different cases. The

* British Association Report for 1858, p・ 72 et seq.

futility of seismometers of the kind now under notice will be readily seen by
anyone who will take the trouble to imagine the motions which the liquid in a
bowl would undergo during any one of the complex disturbances which have
been described in the last chapter. From a knowledge of the height to which
the liquid had been washed up at all points round the inner rim it would be
impossible to determine anything definite as to the amplitude, direction, period,
or duration of the disturbance.

U-tubes (4), which (provided with indices to show the displacement of the
mercury in each) form an important part of Palmieri’s seismic apparatus (§ 69,
below), are equivalent as regards period to simple pendulums of half the whole
length of the oscillating column, if we neglect friction and suppose the bore to be
constant. A U-tube used as a seismometer differs, however, from a simple
pendulum in this important particular, that it is the inertia of the liquid con-
tained in the central or horizontal portion only that tends to keep the mass at
rest when the ground moves horizontally. Apart therefore from the oscillations
caused by its stability, the fluid does not behave as a steady mass. If, supposing
the bore constant, we call MT, the mass of fluid in the horizontal limb, and M,
the mass of fluid in the two vertical limbs taken together, then when an
acceleration a of the ground (to which the tube is fixed) takes place, the resis-
tance to acceleration is a M,, but the mass to be moved is MW, + M,. The fluid
will therefore suffer an acceleration in the same direction as a and equal to

70
A+M,

The U-tube seismometer might be rendered not altogether unsuitable for
absolute measurement by making it nearly straight, in the form either of an are
of large radius, or of a V with an angle of nearly 180° between the legs, or of
a straight line with both its ends bent up at a very small angle. In that case,
sensibly the whole of the mereury in it would act effectively in giving inertia ;
and, further, the equilibrium of the fluid might be much more nearly neutral
than is practicable with the U form. The period of free oscillation of the fluid

ーーー
27sin7

would then be 2 ェ \ ‚ # being the inclination, to the horizon, of the

tube at the ends of the Jiquid column.

$ 64. Standing Columns.

A simple form of rough seismometer recommended by Mallett is a group
of columns with plane bases normal to their lengths, which are set on a’ hori-

* Mallet (loc. cit, p. 81) falls into the same error as the Spanish writer referred to in $ 50,
when he says that the movement of the mercury in Palmieri’s instrument “ depends wholly upon
the U-tube being canted over more or less in its own plane, so as to throw the legs of the tube
out of plumb Probably in no case is there any sensible action of this kind: the movement,
of course, depends almost wholly on the inertia of the liquid. Even in considerable earthquake
the vertical motion is less than 1 mm., while the wave-length is generally some hundreds of
metres,—the slope is therefore quite insensible, と

† Loc. cit., also Admiralty Manual of Scientific Enquiry.

zontal surface attached to the earth. The ratio of height to width differs in the
columns forming the group, and any given earthquake is expected to throw over
certain of them and leave those with broader bases standing. Sand is scattered
round the base to prevent the overthrown columns from rolling, and the position
in which they are found is expected to show the direction of the impulse (that is,
if the base is circular). Mallet gives a formula for determining the maximum
velocity of earth movement from a knowledge of the height and diameter of the
most stable column which has been overthrown. But the formula involves the
assumption that the motion of the base begins suddenly and continues, with uni-
form velocity, until the column falls; or else, that having reached its maximum
value without disturbing the column, the motion suddenly ceases. Either of
these assumptions, it need scarcely be said, is wholly inapplicable to the case of
an earthquake. It would be very difficult to predict the behaviour of a stand-
ing column of given dimensions, if subjected to even the simplest of the disturb-
ances described in the last chapter; and it would be obviously impossible to
derive any definite knowledge of the character and amount and even the direction
of the motion from a knowledge of whether, and how, the column had fallen.
As a matter of fact, columns require to be very sensitive in order to be over-
thrown by even the sharpest shocks ordinarily experienced in Japan. Mr.
Milne* has given an interesting account of attempts to observe earthquakes by
their use. A column whose length was 10 times its diameter did not fall in any
earthquake during a year of use; and even when the columns were so long as to
require a very steady hand to set them up, they allowed many earthquakes to pass
unrecorded, Mr. Milne adds: “Unless I had practical experience with these
columns it would have seemed to me incredible that the smaller of them could
possibly have remained standing.” He suggests the use, in place of cylinders,
of columns shaped like an inverted bottle or a truncated cone. His observations
showed that columns of different sensibility sometimes fell in different directions,
apparently showing that the more stable had their plane of rocking changed be-
fore they fell. With sensitive columns, especially, it is difficult to attain such
perfect symmetry in all azimuths that a column will not fall more readily in some
directions than in others.

Mr Milne has also tried the plan of propping up small columns (such as
ordinary domestic pins or thin strips of glass) in a position as close as possible to
the vertical, leaving them free to fall, in an earthquake, away from the prop.
The results were not satisfactory.

S 65. Mallet’s Ball Seismometer.

In the “ Report” already referred to, Mallet gives a long description of an
instrument intended to act as a seismometer, in which balls are held in L-shaped
supports, which are fixed to the ground so as to move horizontally with it.

* Trans. of the Seismological Society of Japan, Vol. IIT, p. 46.

4 er

Four balls are used, facing the four cardinal points. When any one of the balls
is projected by an earthquake out of its L-support, it runs up an inclined plane
and then falls back again, and the interval between its leaving and returning to
the support is registered by means of a galvanie chronograph. The expectation
was, that the first arrival of a shock would displace the ball facing the origin of
disturbance, and hence, by breaking an electric contact, would register the instant of
arrival. Then the ball facing in the opposite direction would be carried forward
by its L-support, and so acquire a velocity equal to the maximum velocity of a
surface particle. When this maximum was passed, it would leave its support
and run up and back again along the inclined plane facing it. The interval of
its absence from its support would be a function of the velocity of displacement,
in calculating which the author ignores those movements of the support which
take place during the interval in question. The supports for the balls are fixed to
a platform held up by a spiral spring, whose compression was intended to give
some information as to the vertical component.

A simpler form consists of two pairs of balls set, in L-shaped supports, on
the tops of pillars, round the base of which damp sand is placed. The velocity
of projection of any ball is given by the horizontal distance from the foot of the
pillar to the point where it falls.

Apart from the errors which these methods of determination involve, a
knowledge of the maximum horizontal velocity of a particle on the earth’s sur-
face is not of great interest, and furnishes, of itself, a very incomplete account of
the motion.

$ 66. Pendulums intended to swing.

Attempts were made (before the introduction of the absolute measuring in-
struments deseribed in Chapters IT and III) by Mr. Milne,* by Mr. Gray,f and
also by the present writer, to determine roughly the period of an earthquake's
vibrations by seeing what length a pendulum (or other body capable of oseil-
lation) should have in order to be set into most violent motion by the disturbance,
In the author’s experiments a set of inverted pendulums were set up, consisting
of lead balls stuck on the ends of upright wires of various lengths and diameters:
Each was provided with a simple index, consisting of a thread which was at-
tached to the top of the ball, and passed through a hole in a fixed plate above,
on the surface of which it rested. The pendulum whose free period agreed most
closely with that of the earthquake waves would show that it had been most
violently disturbed by pulling its thread farthest through the hole. But the
period of the waves is too irregular to make the observation of much value.

$ 67. Perry and Ayrton’s proposed Seismograph,

A universal seismograph, whose action does not depend on the production of
a steady or approximately steady point, has been proposed by Messrs. Perry and

* Trans. of the Seismological Society of Japan, Vol. III, p. 28.
† Ibid, p. 20.

Ayrton in a paper “On a neglected Principle that may be employed in Earth-
quake Measurements.”* These writers conclude, as the result of a mathematical
examination of the dynamics of a body attached to the earth by stiff springs
when the earth oscillates periodically, that such a body has, in certain cases,
motions relative to the earth, which represent in miniature the motions of the
earth’s surface itself. This result is obtained by making the springs so stiff that
the free period of the body’s vibration is much shorter than the period of the
earthquake waves. By the introduction of friction it is possible to get an
approximation to accuracy with less stiff springs; and this is always desirable
when the earthquake vibrations are irregular. A proposed instrument is figured,
in which a mass is supported within a fixed case by five strong spiral springs,
and provided with three multiplying pointers, adapted to show three rectangular
components of its motion relative to the case. The pointers are to record their
displacements on a band of paper, drawn along by a clock which is set in motion
by the earthquake.

It appears that this suggestion has never been carried into effect. To sus-
pend a mass by springs which would render its period much less than that
of ordinary earthquake waves, and yet without practically preventing all
motion, would be a matter of great difficulty ; and the sudden changes which
take place in actual earthquake motions would be a serious obstacle to the intel-
ligibility of the records. Compared with the steady-point instruments of
Chapters II, III, and IV, this one presents many drawbacks, and few, if any,
advantages; and it is not remarkable that the principle on which it is based
should, even after attention was directed to it, have remained neglected.

§ 68. Electric Seismoscopes.

A simple method of detecting the existence of small earthquakes, which
has been very successfully used by Prof. Palmieri, consists in arranging oscillat-
ing pieces, such as short pendulums, in such a manner that a slight disturbance
will cause them to make an electric contact. ‘The current so established may be
used to work a magnetie indicator, to stop or start or take time from a clock, to
start a record-receiving plate, or, in other ways, to register the occurrence of the
earthquake and serve useful purposes in connection with its measurement. For
the detection of horizontal motion, Palmieri uses a simple pendulum, a few
decimetres long, to the bottom of whose bob a platinum wire is fixed which
stands over and dips into a hole or depression in a cup of mercury placed
beneath it. To make a hole in the mercury, an iron pin is fixed in the centre
of the cup, and the mercury is poured in so that its level is somewhat higher than
the top of the pin, but not so high as to make the liquid meet over the pin.
The pendulum hangs by a fine wire, forming part of the circuit; and the circuit
is closed when, by any shaking of the ground, the wire fixed to the under side of
the pendulum bob makes contact with the edge of the depression in the mercury.

“Phil. Mag., July 1879, p. 30; or Trans. of the Asiatie Society of Japan, Vol. V, p. 181.

For vertical movement, the pendulum bob is hung by a spiral spring just
above a continuous surface of mercury in a cup below, so that any vertical
vibration makes contact. Another form is a horizontal flat spring with a loaded
end, and a point on its under side which stands just clear of a mereury surface.

A convenient form of the horizontal-motion apparatus is shown in sectional
elevation and plan in figs. 38 and 39, Plate XXI. The pendulum a, consisting
of a fine wire and a lead bob with a platinum point below, is hung from the
top of the glass case b, The mercury is held in an iron cup e (shown in section),
the height of which is adjustable by the screw d. One of the terminals ee is
in electrical contact with the pendulum, the other with the cup of mercury,
The arrangement of connections preferred by the present writer is shown in fig.
40. A is the battery consisting of a single large “gravity” Daniell’s cell.
It is kept always in circuit with the electromagnet 2, whose resistance is
considerable. When the seismoscope C' acts the magnet is short-circuited, and
its armature is thereby released. The release of the armature can be used to
effect registration, start a record receiver, ete.

This arrangement may appear less simple than to put the seismoscope into a
single cireuit with the magnet and battery, the circuit being then normally open,
but closed by an earthquake. The closed-circuit plan has, however, several
advantages; amongst which perhaps the most important is this, that any failure
of the battery is at once detected by the release of the armature. Any number
of circuit-closing seismoscopes may, of course, be arranged as shunts to the same
electromagnet, so that the armature will be released if any one of them acts
during an earthquake.

To render Palmieri’s cireuit-closer additionally sensitive, Mr. Milne has
added a multiplying arrangement, which is shown in fig. 41, Plate XXI.* Any
deflection of the pendulum produces a magnified displacement of the lower end
of the lever /, which then makes contact with the mercury in the cup m.

S 69. Palmieri’s Seismic Apparatus. +

An apparatus designed by Prof. Palmieri for the registration of small earth-
quakes was placed by him in the observatory on Vesuvius in 1856, and has been
for some years in regular use in Japan, It consists chiefly of several circuit-closing
seismoscopes, with a clock to register the times of and intervals between distur-
bances, and a group of U-tubes in different azimuths, which are provided with
indices to record the displacement of the mercury and also with circuit-closing
contacts. The circuit-closers are arranged in two groups, for horizontal and
vertical motion respectively. These are connected to two electromagnets whose
armatures carry a red and black pencil respectively, to mark the oceurrence of
horizontal and vertical earthquake motions on a band of paper pulled along
below them by clockwork, This clockwork is started by whichever of the

* Trans. of the Seismological Society of Japan, Vol. IV, p. 98.
† Sismographes Electro-magnétiques de Louis Palmieri, Naples, 1878,

electromagnets is first affected, and simultaneously another clock is stopped.
The band of paper then continues to run for 48 hours at the rate of 1 mm. per
second, and so serves to receive notices of the occurrence of other disturbances
after the first. The time of the first contact is given by the clock which stops;
and the interval between that and subsequent contacts is determined by the
length of paper unrolled from the first to subsequent marks. An alternative
arrangement for receiving the marks caused by the attraction of the armatures
consists of a clock with a projecting arbor which carries three drums, one of
which rotates in 24 h., another in 1 h., and the third in 300 seconds. One of
the armatures carries three pencils which write on the three drums, and the
other carries one pencil which writes on the last drum only.

Tor the reasons which have been already sufficiently indicated, it cannot be
said that Palmieri’s apparatus can be trusted to measure even the relative
“ intensity ” of different earthquakes. As a recording seismoscope it acts admi-
rably, registering many slight earthquakes which would fail to affect a less

sensitive instrument.
$ 70. Time-takers.

3esides the apparatus of Palmieri, many arrangements have been contrived
by other observers with the view of determining the time of oceurrence of an
earthquake. Most of these act by stopping a clock, either by an electromagnet
in conjunction with a circuit-closing seismoscope, or by direct mechanical
action, as when a delicately supported mass is overthrown by the disturbance.
A time-taker used by Mr. Milne has the advantage over any of these, that it
records the time without stopping the clock. The apparatus is shown in fig. 42,
Plate XXI.* Cis a clock with a central seconds hand. The hour, minute, and
seconds hands (h, m, and s) stand forward from the face ; their extremities are
bent outwards and are tipped with pieces of cork smeared with oily ink. A light
wooden ring Z, covered, on the side facing the clock, with a varnished paper dial
which is graduated to correspond with the divisions of the clock face, is carried
by a truck on wheels, which can run forward and make the dial on Z touch the
clock hands. This happen at the time of an earthquake, when the electromagnet
M, which is in circuit with an electric seismoscope, allows the wheel P to be
released, and to be carried round through half a revolution by the weight W.
The rotation of P causes the dial to be advanced and withdrawn, by means of
the eonneeting-rod K. The clock is not sensibly affected by the contact, which
leaves marks on the dial R showing the position of the hands when it occurred.

From the account of earthquake motion given in Chapter V, it will be
obvious that the time recorded by this or any other time-taking instrument will
depend very largely on the delicacy of the seismoseope in connection with it.
Earthquake motions generally begin so feebly as to make the time of their oceur-

* Trans, of the Seismological Society of Japan, Vol. IV, p. 89.

~~ —s

rence, registered in this way, an exceedingly indefinite quantity. A valuable
addition to any time-taker is a contrivance by which the instant when time
is taken shall be marked on the continuously revolving plate of a recording scis-
mometer. We can then see, by examining the record, how long the earthquake
has been going on before the time-taking seismoscope acts, and obtain a much
more definite notion of the connection between the instant of time registered by

the clock and the principal motions in the disturbance,

$ 71. Rossi's Seismoscope and Microseismic Apparatus.*

In Rossi's registering seismoseope, a seconds pendulum with a heavy bob is
hung from a rigid frame, with freedom to swing in any azimuth, The bob is
tied by four silk threads to four uprights which support the pendulum. The
threads are made somewhat loose, and earry at the middle point of each a
metallic needle, whose weight causes the branches of the thread to make a very
obtuse angle with each other. Each needle stands just over a cup of mercury,
with which it makes contact when the pendulum swings, or rather when the
frame moves relatively to the hob. The vertical motion of the needle is } tan &
times the horizontal motion of the ground (in the plane of the thread), # being
the angle formed by the two halves of the thread at the needle. Fig. 43, Plate
XXI, gives a sketch of this ingenious method of multiplying the motion of a
pendulum, There a is the bob: the point of suspension is carried by four posts,
a part of one of which appears at b. ce is one of the four silk threads; d the
corresponding needle, which is partially supported by the delicate spiral spring
e; and f is the cup of mercury. The four cups are arranged to close four sepa-
rate circuits, each being provided with an electromagnet and pencil, which
records the closure of the cirenit on a strip of paper drawn by clockwork. Be-
sides acting as a seismoseope, the apparatus therefore shows the quadrant towards
which the pendulum moves each time contact occurs.

In the mieroseismoseope there are five pendulums of. different lengths—one
in the centre and four round it. Each of the latter is tied to the central pendu-
lum by a bent silk thread, which (as the former apparatus) causes a needle to
make contact with a cup of mercury whenever a relative displacement of the
connected bobs takes place. ‘The pendulums have different periods of oscillation,
and are therefore likely to acquire different amounts of swinging motion during
any disturbance. The apparatus has been successfully used in Italy to detect
very minute earthquakes since 1876.

a
§ 72. Other Microseismoscopes.

Rossi* in Italy and Milne} in Japan have used microphones to detect
minute disturbances of the ground. Another way of detecting the feeblest earth
tremors, if we know when to look for them, is to observe the image of a

* Telegraphic Journal, Vol LX, p. 460,
+ Trans, of the Seismological Society of Japan, Vol. III, p. 87.

fixed object, such as a star, in a basin of mercury. This plan was used by
Mallet in experiments on the speed of transit of artificial earthquake waves
through sand and rock (Brit. Assoc. Report for 1851). The following descrip-
tion of a registering seismoscope, capable of considerable sensibility, but some-
what liable to give false indications, is given by Milne*:—“If a light,
small sensitive compass needle be placed on a table, it will be found that a small
piece of iron like a nail may be pushed so near to it that the needle assumes a
position of extremely unstable equilibrium, If the table now receives the slight-
est tap or shake this condition is overcome, and the needle flies to the iron and
there remains. By making the support of the needle and the iron the poles of
an electric circuit it would be possible to register the time at which motion took

place with considerable accuracy.”
$ 75. Optical method of multiplying the displacement of a pendulum.

In the experiments referred to in the Introduction, Messrs. G. and H. Darwin
employed a very delicate means of measuring changes in the position of a pendu-
lum bob, which may also be employed in the measurement of very minute earth-
quakes. The method was to hang a light mirror by two fibres, one attached to the
bob and the other to an adjacent fixed support. When the bob moved the
plane of the mirror was changed, and the amount of change was read either by
reflecting a beam of light from a lamp upon a scale, or by observing the scale
reading reflected from the mirror into a telescope. The experiments of Messrs.
Darwin had for their object the detection of slow changes in the direction of the
vertieal, and for this reason they eliminated sudden tremors by hanging the
pendulum in a liquid. In the mieroseismie application of the same method, the
pendulum and mirror should be left as free as possible to respond to sudden
tremors of the ground; but slow displacements are to be altogether discarded.
Mr. Milne f has attempted by a similar method to register minute motions of the
soil. His pendulum bob was kept in contact with the ends of two wires laid in
fixed guiding pieces of glass tube, and placed at right angles to each other. The
other ends of the wires abutted against two mirrors, so as to turn them round if
the bob should become displaced, Some movement was found to have occurred
almost every time the instrument was examined ; but it is impossible to tell how
much of this was caused by earth tremors, how much by changes in the vertical,
and how much by movements of the pendulum’s supporting point, due to mois-
ture, unequal heating, and other causes. To arrange the apparatus in a man-
ner wwhielfavould eliminate all sources of error due to this last head would be
scarcely possible; and if, by continuous observation, actual tremors were distin-
guished, it would still be a matter of great difficulty to separate out those due to
such obvious causes as neighbouring vehicles and pedestrians from the natural
disturbances which it is the object of the apparatus to detect.

* TOC Cts, Po aie
† Ibid., p. 30.

に ネコ

$ 74. Measurement of earthquakes by reference to their Effects on Buildings.

Attempts have been made to infer the character of eartlıquake motions from
the traces which destructive shocks leave behind them, in fractured walls, pro-
jected fragments, overturned columns, ete. The two large volumes in which
Mallet has described the great Neapolitan earthquake of 1857* are in great part
filled with calculations of this kind. So far as these refer to the direction of the
principal movement (on the supposition that there was a principal movement
with a definite direction), they are not without value; but in applying his
measurements of projection and overthrow to the determination of the greatest
velocity of displacement, Mallet has throughout made an assumption which
entirely vitiates his results. ‘The assumption, already alluded to in speaking of
standing columns ($ 64), is that the body which suffers overthrow does so in
consequence of a sudden change of momentum with respect to the ground, and
that the momentum so acquired is left free to do its work, in causing the body to
fall, without further interference. In other words, it is assumed either that the
motion of the ground begins quite suddenly, and then continues (until the body
falls) with a uniform velocity—the velocity which is calculated. by Mallet’s
‘formulas; or else that, having communicated this velocity to the body, the
ground suddenly stops, and stays at rest until the body falls. It need scarcely be
pointed out, after the results given in Chapter V, how wide of the facts this
assumption is. And it is difficult to see how any observations of bodies projected

or overthrown can be of the smallest use as data from which to determine
earthquake motions, when these have the extraordinary complexity which
absolute measurements have shown them to possess in Japan.

Mallet’s treatment of fractured structures is even more unsatisfactory, and,
indeed, involves a distinet error. Taking the case of a column which is fractured
at a horizontal plane through the base by a horizontal impulse, he equates the
momentum multiplied by the height of the centre of gravity, with the moment
of resistance of the section to rupture. It is, howeyer, the greatest rate of
acceleration, and not the greatest velocity, which determines whether fracture
shall take place. If we assume the column to break without sensible bending,
so that the act of rupture (which is to be clearly distinguished from the over-
throw which may or may not follow it) occurs with sensible instantaneousness,
we should equate the moment of the resiftance to acceleration with the moment
of resistance to rupture. Calling MZ the mass above the fractured section, a the
acceleration, ん the height of the centre of gravity above the section, 7 the moment
of inertia of the section about a horizontal central axis perpendicular to the
direction of acceleration, b the shortest perpendicular distance of this axis from
the edge of the section, and f the modulus of rupture, we have

Mah = SI ; fl

or (7

b ー Min

* The First Principles of Observational Seismology, London, 1862.

CHAPTER Vink

SUMMARY AND COMPARISON OF INSTRUMENTS AND METHODS.

§ 75. Uselessness of all except Steady-point Seismometers.

Enough has been said to show that, of all actual seismometers, those only
are of value which aim at giving a steady line or a steady point during the
disturbancee—a steady line with respect to motion in one direction, or a steady
point with respect to motion in two directions, or in all. Cruder devices, such
as standing columns, fail because of the complexity of earthquake motion, They
cannot be trusted to yield even relative measurements of amplitude or of destruc-
tive power; and as to direction, it is impossible to speak of the direction of the
ground’s motion when we find displacements of nearly equal value occurring in
all possible azimuths, during the continuance of a single earthquake.

§ 76. Classification of Steady-point Seismometers.

We may classify these instruments with respect to the component or compo-
nents of motion they are designed to measure, as follows :—

I. Those possessing one degree of horizontal freedom, and therefore capable
of measuring one component of the earth’s horizontal motion. They are to be
used in pairs, giving two components, which are most conveniently taken at
right angles to each other. To this class belong the horizontal pendulum in its
various forms (SS 23-28); the common vertical pendulum (or any of the modifi-
cations of it given in Chapter IIT) when suspended, like the pendulum of a clock,
on an axis, not on a point; the rolling eylinder (§ 34), with or without a heavy
slab. A sphere rolling in a fixed V-groove might be added to the list, also a
flattened-out U-tube filled with mereury (§ 63).

IJ. Instruments possessing two degrees of horizontal freedom. This class
includes the common pendulum and its modifications (the duplex pendulum ete.)
when suspended by a point (SS 35 et seq.); the horizontal pendulum with
jointed supporting frame (§ 29); the rolling sphere, or spheres carrying a slab
($§ 31-33); a system of rolling eylinders at right angles to each other, with a
slab between ($ 34).

IIT, Instruments possessing vertical freedom only. The loaded horizontai
bar with flexible or extensible support, and its modifications (NN 45-48). Tt
would not be difficult, were it at all desirable, to design instruments which would
form two more classes, namely IV, instruments possessing vertical and also one
degree of horizontal freedom; and V, universal seismometers, having three
deerees of freedom.

SN 77. General principle of Steady-point Seismometers.

The steady line or steady point is obtained by pivotting a mass, or group of
masses, with appropriate freedom, and in nearly neutral equilibrium. The axis
about which the body or system spontaneously rotates in consequence of the
component of earth-displacement it is designed to measure, furnishes an approxi-
mate standard of rest.

It is only approximate ; first, because the earth’s motions are not indefinitely
small; second, because the stability of the system causes some motion to ensue
after every displacement; and third, because of friction at the joints and at the
index which is used to record the earth’s displacement with respect to the axis of
spontaneous rotation. The first error may be rendered inappreciable by placing
the axis of spontaneous rotation sufficiently far from the supports, compared with
the extent of motion to be measured. The degree to which we may reduce the
second and third sources of error depends on the method adopted for registering

the motions.

$ 78. Methods of registering the movements of the ground relative

>
to the Steady Point.

These are four in number 一

[1.] By an index which shows the greatest amplitude of relative displace-
ment, without showing its direction.

[2-] By indices which show separately the greatest displacements in two
or more directions.

[3.] By a writing pointer which draws a diagram of the motion on a
fixed plate,

[4.] By writing pointers which show the successive motions, or compo-
nents of them, in conjunction with the time, on a plate or drum
which is kept always moving uniformly.

[5.] By pointers writing on a plate or drum which moves uniformly
after being started into motion by the earthquake itself, but is
normally at rest.

In applying method [4] or [5] to a double-freedom instrument (of Class
IT), we may use either a single pointer to record the earth's complete horizontal
motions on the moving plate, or two pointers arranged so that each records only
one component. Plate VIII shows both plans, applied to a long pendulum

seismometer.

$ 79. Objections to Statice Records.

The statie records which are obtained by methods [1], [2], and [3] give,
of course, less complete information than [4] and [5], inasmuch as they do not
show the relation of displacement to time. The results obtained by method [3]
are of much greater interest than those which [1] and [2] can yield. But

——— ーー - au 人

besides possessing the defect just mentioned, all statie records are open to another
important objection. If the equilibrium of a seismometer is very nearly neutral,
the so-called steady-point is liable to become shifted during an earthquake to a
considerable distance from its initial position; and even during intervals of rest,
especially when these are long, a gradual creeping of the pointer occurs, due to
slight accidental displacements of the line or point of support, through changes
of temperature, warping, and other causes. Hence to use [1] or [2] we must
make the equilibrium tolerably stable. Another difficulty then presents itself,
namely the tendency which a stably hung body has to acquire, during a long
series of shakes, an oscillation whose amplitude may be comparable to or even
much greater than the motions of the point of support. Only one way remains
by which this objection may be overcome: the tendency to oscillate may be
removed by introducing a considerable frictional resistance ($ 39). This, again,
gives rise to error by making the recorded displacements too small, especially
in slight earthquakes. It is, however, the only method by which records
obtained by plans [1] and [2] can be prevented from being absolutely meaning-
less. In [3] the case is somewhat different. There the creeping of the steady
point, which occurs when the equilibrium is nearly neutral, is not so fatal an
objection as it is in [1] and [2]; for it is often possible in the record to distin-
guish, more or less completely, any motion of this kind, especially when the
principal earthquake motions are decidedly not rectilinear. The diagram in the
centre of Plate XIX may be referred to as an example of a static record which
has not been rendered. unintelligible by the nearly astatie character of the seismo-
graph which gave it.

$ 80. Records of displacement in conjunction with time.

Records which show each successive displacement in conjunction with the
time are far more valuable than statie records, both on aceount of the greater
fulness of the information they yield, and because they avoid to a great extent
the sources of doubtfulness or error to which statie methods of recording are
liable. When the movements are recorded on a continuously moving plate, the
equilibrium may safely be made much more nearly neutral than is practicable
with a statically recording instrument; and the friction, instead of being arti-
ficially inereased, may advantageously be reduced almost without limit. Any
moderate degree of progressive displacement of the so-called steady point, by the
accumulated effect of successive earth waves, leads to no serious confusion ; and
any long-period swinging that may occur is easily distinguished from the legiti-
mate portions of the record. A small amount of stability must be given to
prevent the former effect from becoming excessive; but the equilibrium may be so
nearly neutral that the slight friction which, in the most favourable conditions,
the pivots and the marking pointer must produce, will suffice to make immode-
rate swinging impossible. When registering on a continuously moving plate, a
well constructed and well adjusted seismograph should make one complete

oscillation in about 5 seconds, and its decrement of amplitude should not exceed
about 1 mm. per oscillation.

When the record is deposited on a plate or drum started by the earthquake
and stopped soon after it, the equilibrium may properly be made still less stable
than this; for in that case “creeping” of the pointer before or after the distur-
bance does no harm.

Methods [4] and [5] have each certain merits and defects, Experience
shows that when method [4] is used, and the equilibrium is nearly neutral, the
pointers ereep about through small distances almost continually, from causes
probably the same as those which prevented the mirror in Messrs. Darwin’s
experiments (referred to in the Introduction) from remaining at rest. The result
is that the circles which the pointers trace out on account of the continuous
rotation of the plate become gradually widened ; and hence when an earthquake
eeeurs its smaller movements are indistinctly recorded, and its beginning is
sometimes untraceable. Instances of this will be seen in several of the Plates.
The position of the pointers on the glass plate, and the plate itself, require to be
changed from time to time, when the lines become too wide. On the other hand,
method [5] may allow a part of the earthquake to pass unrecorded ; and it is
open to the further objection that there is no easy way of seeing that the starting
gear is in good order, whereas the apparatus used in method [4] shows at a
glance whether it is in a fit state to receive a record. As regards difficulty and
cost of construction, [4] requires a much more carefully made and costly piece
of clockwork, but avoids the circuit-closer, magnet, and battery which must be
combined with the simpler clock of method [5]. For use in an observatory where
the instruments can receive frequent inspection and skilled attention (a condition
essential to success in all seismie observation), the writer greatly prefers method [4].

Where the normal position of the index depends upon the pull exerted by
a spring, as in the vertical seismometers of SS 46 and 47, the creeping tendency
alluded to above is so great as to make [5] the only practicable method of
registration,

SN 81. Constructive details.

The principal points to be attended to in the design of an absolute seismo-
meter are. these :

The point or axis of support (points or axes, if there be a group of masses
instead of only one) must partake as exaetly as possible of the motion of the
earth at the place of observation. ‘This requires a rigid post or stone table (whose
dimensions must be greatly less than the shortest wave-lengths in the earth-
quakes to be measured), firmly imbedded in the earth; a steady clamp, giving
definite points of attachment (§ 23); and a rigid supporting frame to arry the
point or axis of support.

The distance from the axis of support to the steady line must greatly exceed
the greatest earth-motion to be measured, except where the apparatus is astatie,

EE = 2 ee En ee BE Eee EEE En

or nearly so, for large as well as for small displacements, and there is no sensible
rotation of the steady line, during displacement, about the axis of support.

The moment of frictional resistance, due to the pivots and pointer, must be
very small relatively to the moment of the effective inertia.

The displacement of the multiplying pointer, in the plane of the surface on
which it writes, must be a constant multiple of the displacement of the ground
with respect to the steady line. The pointer must, however, have freedom to
move, through a small distance at least, at right angles to the surface on which
it writes. To give it this freedom the writer has, after several trials, found no
plan so satisfactory as the use of a transverse joint, examples of which will be
found in all his seismographs.

The record-receiving plate or drum must have its axis of rotation definitely
attached to the earth, or, preferably, to a piece which is rigidly connected with
the axis of support of the seismometer whose displacement it is to record ; and
the motion of the record-receiver must be steady and continuous during an earth-
quake. These conditions are fulfilled in the writer’s apparatus by the indepen-
dent supporting frame which is used to carry the glass plate, without shake ; the
rolling-contact gearing, in place of toothed wheels, which, unless made with
extreme accuracy, would give unsteadiness to the motion; and the governor,
whose action is not sensibly spasmodic, and which is designed with a special view
to its being undisturbed by the motion of the ground, To fix the smoked-glass
plate directly to a slow-running arbor of the clock, though apparently a simpler
method of support, gives more chance of irregular motion, and a considerable
likelihood that the plate may shake.

A glass surface is preferable to paper for the reception of records, on account
of its smoothness; and a plate is in many ways more convenient than a drum.*

* To obtain permanent copies of a record traced on a smoked-glass plate, the plate is first to
he coated with photographer's varnish, ly pouring the liquid on it, and gently sloping it to allow
the varnish to spread uniformly. After itis dry, the “blue process’ of taking photographs may
be applied as follows. —

Dissolve one part of Ammonio-Citrate of Iron in eight parts of water, and make a separate
solution containing one part of Ferrieyanide of Potassium also dissolved in eight parts of water.
Immediately before taking the photographs mix four parts of the former solution with three
parts of the latter. Spread the mixture evenly over a sheet of paper (either with a glass rod or
with a flat Irush) and hang the paper up to dry in a dark room. When dry, expose it under the
glass plate, with the varnished side next to the paper, care being taken to bring the paper and the
plate into immediate contact. This is most easily done by laying the sensitized paper on a thick
cushion consisting of several layers of cloth, and then pressing the glass plate firmly against it
hy means of a second sheet of plate-glass which may be held down by clamps. In sunshine
the exposure should last from eight to twenty minutes.

After exposure, the paper is to be dipped into a tank of water and allowed to soak there for
about ten minutes. It must he thoroughly washed, so that every part of the unaffected chemicals
on the paper may be removed. It may be then dried in the open air.

The solution of Ammonio-Citrate of Iron should be kept in a dark place; and it is best to
prepare it not long before use. The sensitized paper may be prepared in large quantities and
laid by in a dark drawer; but if it is kept long the photographs are not so clear as those obtained
by the use of freshly prepared paper.

When the plate is to be set in motion by an earthquake, a light driving
clock is desirable, whose speed will quickly attain its steady value. The same
form of governor as is used in the continuously running clock may appropriately
be applied, and the same means of communicating motion to the plate.

Figs. 44 and 45, Plate XXI, give a section and plan of a driving clock
which is started by the release of the armature in fig. 40. The cord a passes
over a pulley on the ceiling, and carries a driving weight of about half a kilo.
b is a drum on which the cord is wound; its arbor projects on both sides and is
jointed, as in fig. 8, Plate III, to two spindles which revolve in slot-guides cc
and carry dises dd, which roll on two glass plates near their edges, and so cause
them to revolve. These form two record-receivers ; and a third is also driven by
the same clock, by means of an independent cord on b, which, when the clock
runs, drags along a carriage on a straight pair of rails, the carriage having a long
strip of smoked glass fixed to it. The axles and wheels of the carriage, and the
rails, are arranged on strict kinematical principles, to avoid the possibility of
any transverse shaking ; and the cord which drags the carriage is kept tight
enough to prevent longitudinal shaking.

$ 82. Requirements of a Seismological Observatory.

For the registration of ordinary earthquakes the equipment, to be fairly
complete, should comprise :—

[A]. A pair of single-freedom horizontal seismographs writing on a
plate which is kept always revolving.

[|B]. A pair of single-freedom horizontal seismographs writing on a plate
which is started into motion by an earthquake.

[©]. A double-freedom horizontal seismograph writing on a plate which
may either be kept always revolving, or he started into motion by an
earthquake,

[D]. A double-freedom horizontal seismograph writing on a fixed plate.

|]. A vertical motion (single-freedom) seismograph writing on a plate
which is started into motion by an earthquake.

[F |. For these one eontinnous-motion driving clock with a governor will
be needed—like that shown in Plate II. It may drive two plates,
if required, by having the arbor from which they! take their motion
prolonged outwards on both sides,

[G]. A second smaller clock, also with a governor, and with a magnetic
starting appendage, will be required for [B] and [EF].

[HH]. One or, better, several cirenit-closing seismoscopes ($$ 68 and 71)
should be put in connection with this clock.

[1]. The same eirenit-eloser should actuate a time-taker, to give a gene-
ral idea of the time of occurrence of the disturbance.

[J]. A useful addition is a set of electromagnetic time-tickers (in one cir-
cuit with a clock), which mark seconds simultaneously on all the

revolving plates. This apparatus is to be started by the same con-
tact which starts the plates of [B| and [E] and actuates [I]; and it
must be arranged to cease acting before one revolution of the plates
is completed, else the marks will become unintelligible.

The use of [J] is to allow the contemporary parts of all the records (except
those of [D]) to be readily identified, and to show at what part of the disturbance
the time-taking appliance has acted. It also provides, on the plates, a scale of
time which is convenient in determining the period, ete. of the waves.

For [A] it appears difficult to find any instrument superior to the horizontal
pendulum with light pivotted frame and pivotted bob ($ 24). The construction
shown in Plate II will be found very convenient; but the stand, there repre-
sented as made of wood, might with advantage be entirely of metal. This would

probably prevent, in part, a tendeney to “ creep”
1 Y 1 2 1

which the pointers are found to
possess. The instrument has the advantage of great compactness, facility of
transport, and an almost entire freedom from liability to get out of adjustment,
or out of order in any way.

[B] is desirable as a supplement to [A] in order that we may have a record
which is not obseured by the somewhat broad datum-line formed when [A] runs
for some time, either before or after the earthquake.*, The driving-clock for [B]
and [E] should of course be arranged to run for no more than a few minutes, so
that the record may not be obliterated by “creeping” after the disturbance
is over. [B] may suitably consist of a pair of horizontal pendulums with
flexible joints at one or both of the two points which determine the axis of
support of each (§ 27 or § 28). Their equilibrium may be more nearly neutral
than that of [A], and their frictional resistance should be reduced to the lowest
possible value.

[C] may conveniently consist of a long pendulum like that of Plate VIII.
By placing a weight on the multiplying lever, near the top, we may approximate
to the duplex pendulum of § 41, and in that case the length may be redueed.
The record given when a double-freedom horizontal seismograph writes with one
pointer on a revolving plate is not so completely intelligible as a pair of compo-
nent records, each showing motion in one azimuth, for it merges the tangential
part of the earthquake motion in the continuous motion of the plate itself; but
the single-pointer plan has the merit of showing, at a glance, the kind of motion
and the changes of direction which occur during the disturbance. These are,
of course, deducible from two-component records, but only by a somewhat

* An instance of the usefulness of an apparatus of this kind was furnished by a recent very
slight earthquake, the maximum amplitude of whose motions was so small that the wide datum-
lines on two continnously running plates obscured, almost completely, the records traced on
them. The disturbance, however, actuated an electric seismoscope and started a third plate, on
which a very clear record was given by a pair of horizontal pendulums and by the vertical-motion
seismograph of § 48. The greatest horizontal motion was about 0.8 mm., and the greatest
vertical motion about 0.1 mm.

tedious process, which it is practically difficult to apply when the motion of the
plate is slow and the period of the waves is short.*

The same steady point—namely, that of [C]—may be used to write a pair
of two-component records: this is easily done when the effective inertia is great,
as in the instrument of Plate VIII, where the arrangement in question is
adopted (figs. 27 and 28).

For a statie record, |D], the duplex pendulum will be found convenient.
There is, however, no reason why we should not take a statie record, as well as
a record on a moving plate, from the same long pendulum as gives [C], provided
its effective inertia is so great as to make the frietion of an additional marking
pointer scarcely noticeable. To do this a second transversely hinged pointer
might project from the vertical multiplying lever in fig. 25, Plate VIII, and
write on a fixed plate carried by a bracket above or alongside of the revolving
plate.

A double-freedom instrument might be made for [C] or [D] by combining
two single-freedom seismometers so that, instead of tracing separate records, their
motions should be compounded by a separate marking lever, pivotted with freedom
to move in any azimuth. It is scarcely likely, however, that this plan would
have any advantage as compared with the use of a double-freedom instrument of
the kind already described ; and it would of course involve a greater number of
parts. For [E] either of the horizontal bar instraments described in §§ 46-48
is suitable. The eireuit-elosers of Palmieri and Rossi appear perfectly satis-
factory for [H]; and Mr. Milne’s time-taker should serve well for [I].

The various seismographs may usefully be constructed with different ratios
of multiplication, appropriate to earthquakes of various amplitudes. One, pre-
ferably [C], should be capable of a considerable range of movement, and should
be provided with a supplementary pointer, writing without multiplication, to
serve for the registration of unusually large shocks, whose amplitude is so great
as to throw the ordinary indices out of gear.

$ 83. Determination of the Direction and Velocity of Transit.

The foregoing account of methods of earthquake measurement has referred
entirely to the determination of the character of the motions at a single observing
station. Another problem, of much interest, is the determination of the direction
and velocity of propagation of the earthquake waves. The direction, especially,
is important on account of the information it gives as to the position of the origin.
If we know, rigorously, the time of arrival of an earth-wave at three stations on
the surface, and assume its velocity and direction of propagation to be constant
between them, an obvious geometrical solution gives the direction in azimuth

* To facilitate the process of compounding the principal motions recorded by a two-compo-
nent instrament, the plate, before being withdrawn after an earthquake, should be turned round
by hand, and held successively in a number of positions, while lines, like those in Plate XIX,
are simultaneously drawn by displacing both pointers.

and the horizontal velocity of transit of the wave. Four sufficiently distant
stations furnish data for determining the epicentrum, or surface-point vertically
above the origin, the velocity being assumed constant and the wave-front
spherical, A fifth station, on the same assumptions, if properly chosen, gives
the additional information needful to determine the depth of the origin. Know-
ing the instants of arrival at five stations, we have five simultaneous equations for
the three coordinates of the origin, the velocity, and the instant of disturbance at
the origin,

But, in fact, even the first of these schemes, and still more the second and the
third, is generally incapable of sufficiently exact application to an earthquake as
a whole, on account of the ill defined beginning and long duration of the dis-
turbance. By the time it reaches a station distant from the origin, the shock is
very far from consisting of a single impulse whose time of arrival may be
recorded with precision, A glance at any of the diagrams described in Chapter
V will show how indefinite the time of arrival must be, whether we determine
it by the sensations of an observer or by using a mechanical or electrical seismo-
scope. The beginnings of motion are usually so gradual that the time shown
depends largely on the sensibility of the appliance used to detect disturbance ;
and in attempting to measure, by time-takers, the interval between the times of
arrival of an earthquake, as a whole, at two or more stations, we are liable to
errors which may even exceed the quan-

errors of perhaps ten seconds or more
tities to be measured. It is true that by increasing the distance between the
stations we make any given error relatively less important; but, on the other
hand, the character of the complex system of waves will then be more likely to
differ at different stations, and the errors of time measurement will thereby be
increased. It is only in violent earthquakes, where stations may be taken at
such distances from each other as to give intervals greatly exceeding the dura-
tion of the shaking at any one point, that trustworthy results can be obtained.

The same objection applies with equal force to the measurement of time-
intervals by telegraphing automatically the time of arrival at several stations,
and recording the signals on a chronograph.*

In a paper read before the Seismologial Society in February 1881,f the
present writer, after pointing out the impracticability of obtaining precise time-
intervals by reference to an earthquake as a whole, gave a scheme by which the
velocity and direction of transit might be determined provided we could identify,
at three or more stations, any one motion out of the group constituting an earth-
quake, It was proposed to put up three (or four) pairs of continuously recording
horizontal pendulum seismographs at three (or four) stations connected electrically

* Prof. W.S. Chaplin had designed and, in fact, prepared an apparatus to carry out this
method, just before the writer obtained his earliest records of earthquake motions on a moving
plate. The character which those records showed the motion to possess led Prof. Chaplin at
once to abandon his project.

7 Transactions, Vol. III, p. 111.

with each other. The plate at cach station was to be moved by an independent
clock of the usual kind, it heing needless that the plates should run at the same
speed. So far, the arrangement would give three (or four) pairs of independent
records of any earthquake. To complete the scheme, time must be simultaneous-
ly marked on all the plates during the disturbance. At one station a seismo-
scope was to be placed, which, on the occurrence of an earthquake, would start,
or throw into electrical connection with the station circuit, a clock which should
send time signals every second to all the stations, to be recorded by electro-
magnetic pointers on all the plates. The signal-sending clock should stop, or be
automatically cut out of the cireuit, before the plates had accomplished quite one
revolution from the beginning of the time signals.

The first signal, marked on all the plates, would serve to identify the same
instant of time in all the records, and the subsequent signals would determine
the velocity of the several plates. The scheme would work only if we could
recognise the same earth-wave on all. It would then be easy to measure with
much exactness the intervals hetween the instants at which the same phase of the
same wave appeared at all the stations.

The method gives so precise a means of determining small intervals of time
that we might use stations at no great distance from each other—say 500 metres,
which would probably give intervals of about one or two seconds. These,
however, would serve only to determine the azimuth of the epicentrum. For
its position, four much more distant stations might perhaps be used ; or, better,
two independent groups of three near stations, the groups being at a considerable
distance apart. Other plans might obviously be arranged by combining three
stations near each other with one distant station or more—all electrically
connected,

But the practicability of any of those plans depends on this fundamental con-
dition, that some one wave in the disturbance is to be identified at all the stations.
The writer's observations have shown (what, indeed, was to be anticipated) that
different seismographs recording the same earthquake at the same station give
closely accordant results. It remains to be seen how far the character of the
motion will prove constant when observations are made at stations whose dis-
tances from each other are comparable to or much greater than the wave-lengths
of the constituent vibrations. If the curved path in which a surface particle
moves be due to the simultaneous passage of two or more systems of independent
waves, travelling in different lines, no identification will be possible. But the
large loops of Plate XX show that, at least in some cases, the principal displace-
ments in two directions oceur together—a thing very unlikely to happen if they
are due to two independent systems of waves. There appears, therefore, a fair
prospect that, in favourable cases, the identification of some one prominent
motion at stations tolerably far apart may be possible; and in that case the
writer's method will serve to determine the speed and direction of transit.

S 84. Velocity of Transit of Artificial Earthquakes.

The method in question has been applied by Messrs. Milne and Gray to
measure the speed of transit of artificial disturbances.* The bearing of the
origin being known, two stations were of course sufficient, and at them horizontal
pendulum seismographs were set. The soil was hardened mud. The velocities
found were 438 feet per second for normal waves, and 357 feet per second for
transverse waves.

The earlier experiments of Mallet? gave the following velocities :—

Tisaing sero een 825 feet per second
In jointed granite ............ 1306 „ >
In solid granite ............. 1665 ,, 9

But from the known elasticity and density of solid rock, and from other
observations, it is probable that the last named velocity is much too low; and
that in a continuous mass of rock as high a velocity as 8000 or even 12000 feet
per second may be attained.

SN 85. Experimental Tests of Seismographs.t

To find experimentally the accuracy with which a seismograph registers
motions of any assigned form and period, let the seismograph to be tested, with
its receiving plate and driving clock, complete, be placed on a shaky table. Side
by side with the multiplying lever of the instrument let a second lever be
arranged, pivotted by an altogether independent fixed support, and connected to
the table at one point, so as to give the same ratio of multiplication as the actual
instrument. This lever is to be provided with a point which writes on the plate
close by the marking point of the seismograph. If the table is now shaken, the
component of earth-motion transverse to the levers is recorded by both pointers ;
and the agreement of the two records will show the accuracy of the seismograph.
To make the test as conclusive as possible the table should be forced to move in
such a manner as to give records resembling those which experience shows actual
earthquakes give. For a vertical-motion seismometer the same plan may be
followed, if we use, instead of a table, a spring board capable of vertical shaking.

In applying the test to a pair of horizontal pendulums with pivotted bobs, a
convenient plan is to place the two pendulums parallel or facing each other and
so that their marking points come on the same radius of the plate. Then let the
bob of one be held fixed by means of a bracket from a neighbouring wall: this
will ensure that its “steady line” is rigorously steady. The steadiness of the
other is then tested by shaking the table, and seeing how closely accordant the
two records are.

* Phil. Mag., Nov. 1881. Also Proc. Royal Society, Vol. XX XIII, p. 139. (Dec. 1881).
+ British Assoc. Reports for 1851 and 1852.
t See also Proc. Royal Society, Vol. XXXT, p. 444.

Plate XXII gives a number of examples of this test, as applied to the
horizontal pendulum of § 24. The curves marked aaa were given by the
pointer whose bob was fixed: they represent the true displacements of the table,
magnified 4 times. The curves bbb were given by the pointer whose bob was
free. The general agreement of the two records with each other shows that the
steady line of the free seismograph did in fact remain very nearly steady, during
movements of much variety. The pointers were set facing each other, and so
would have described ares of opposite curvatures on a stationary plate—a fact
which must be taken into account in comparing the records.

iri? て
HU

sd haft

Te Sree
は En HG
ees atin (eh
=
Pinay ek
» ose fg «

こき x are Life
+ Rat.
en Ar ri hi ‘uit ey
pe

wtil-fpinee Manta

Av sr lene Alk

“Theat tod ET
pl % Heard a wr til nay

APPENDIX.

ASTATIC SUSPENSION BY LINKWORK.

At a meeting of the Seismological Society of Japan, held while this memoir
has been passing through the press, Prof. C. D. West exhibited a model of a
new method of astatic suspension, applicable to seismometry. A massive bar
a (fig. 46, Plate XXIII) is hung by a system of links bhcc, like the so-called
parallel motion of Watt, whose effect is to give the bar freedom to move in an
approximately straight horizontal line, in the direction of its own length. A
small degree of stability is secured by placing weights dd at the bottom of the
links bb, or by making these slightly shorter than the lower linksec. To guard
against motion at right angles to the plane of the links a guide-bar ee is added,
which passes through holes in the upright supports. This, however, is liable to
introduce an objectionable amount of friction during an earthquake ; and, apart
from it, the numerous joints can scarcely be prevented from causing more fric-
tional resistance than is permissible in a good seismometer. The upper links,
which are ties, may be made of flexible cord or wire instead being of jointed, but
even then so many joints will remain as to constitute a serious defect.

More recently several methods of astatic suspension have occurred to the
present writer, one of which, in particular, seems well adapted to the measure-
ment of earthquake motions, especially when these are great.

It is obvious that the mass whose inertia is to give steadiness in a seismo-
meter may be hung in neutral equilibrium by any system of links which produces
an exactly or approximately straight-line motion, provided the links are placed
so as to make this line of motion horizontal. We might, for example, suspend
a mass by a pair of Peaucellier linkages, with freedom to move horizontally, and
thereby obtain absolute astaticism throughout the whole range of its motion; or,
by reducing the distance between the two pivot points in each linkage, we might
compel the mass to move in an are of very large radius, and thereby give it the
slight stability needful in a seismometer. But this plan would be open to the
same objection as the suspension by Watt’s linkage proposed by Mr. West,—the
multiplicity of joints would give rise to an intolerable amount of friction.

To avoid friction we should select a form of linkage with as few links as
possible, and these all ties, in order that we may easily substitute flexible cords
for rigid pieces with joints. No linkage satisfies these conditions better than the
approximate straight-line motion of Tchebicheff, illustrated in the sketch, AA
are fixed supports. BB are two equal links, which, when the apparatus takes
this form, may be flexible cords. They cross each other and are connected to
the end of a hanging bar ©. The vertical distance of the middle of の from the
line AA is equal to the distance 4 A, while the length of the hanging bar is

144. Then the middle point of the bar
moves in a line which is very approxi-

‘A mately straight and parallel to AA, pro-
vided its excursions lie within a range not
greater than the distance between the fixed
supports. These proportions will give
sensible astaticism when A A is horizontal ;
but by making the depth of the hanging
bar somewhat ‚greater, or by placing the
centre of gravity of the hanging bar below
the line of its attachment to the links BB,
we may give it any desired amount of
stability.

When BB are single cords the system is azimuthally unstable, but it is
easily prevented from rotating about a vertical axis by making each of the
suspending links B double, in two parts which form a V whose vertex is at the
end of €, and whose base is a line through A perpendicular to the plane of the

paper. Further, this prevents oscillation perpendicular to the plane of the paper,
and so leaves none but the desired freedom. To increase the steadiness we may
add a mass which should be as much as possible concentrated at the centre of の .
This may be pivotted about a horizontal axis perpendicular to C, through its
centre, and in that case the mass is equivalent to a particle concentrated there.
Figs. 47 and 48, Plate XXIII, show this arrangement in elevation and plan.
There the hanging bar is a light platfom on which a heavy lead weight is
pivotted about the axis ii on the points of two steel screws, which press up into
a conical hole and V-slot in a bar 7 to which the weight is rigidly attached.

Another plan is to use a pair of light suspended platforms, in line with
each other, and use them to carry a massive block by three sharp feet which
press into a hole and V-slot on one platform, and a V-slot parallel to these on
the cther. This arrangement is shown in figs. 49 and 50. :

But in both of these arrangements the friction at the pivots by which the
weight rests on the hanging platform is a disadvantage, which is all the more
felt because the platform tilts up through a considerable angle when displaced
from its mean position. For this reason the writer prefers the very simple form
shown in figs. 51 and 52. There the bob or heavy mass is a piece of lead
rigidly fixed to the hanging bar. The effect of this is that when a horizontal
displacement of the ground occurs, in the line of the bar, the centre of the bar
does not remain at rest, but moves through a small determinate distance in the
same direction as the ground.

Let M be the mass of the hanging piece (including the rigidly attached bob),
and let Mk? be its moment of inertia about its central transverse axis. When
the hanging piece is displaced, its motion is one of rotation about its instantane-
ous axis, which is always situated at the intersection of the suspending cords

(the point I in the sketch on p. 90). It is easy to show that, for any moderate
displacement, this axis moves in a sensibly horizontal direction, and through the
same distance as the centre of the bar. Hence the angular displacement of the
bar is, very nearly, proportional to its linear displacement, and is equal to the
latter divided by h, where h is the height of the instantaneous axis I above the
bar. As regards its resistance to rotation, the hanging piece is therefore equiva-
lent to a particle of mass M whose velocity-ratio relative to the centre of the bar

is sensibly constant, and equal to ;: Hence the extra inertia due to rotation is

Mi?
a” referred to the centre of the bar.
When a horizontal displacement of the supports occurs, we may consequently

consider the whole system as consisting of a particle M together with a connected
MR? 8 u 3 P
particle > of which only the first is effective in producing steadiness, although

both are constrained to share the same motion. The piece will therefore suffer
an acceleration in the same direction as the acceleration of the supports, and
bearing to it the ratio k* : kh? + だ だ . The centre of the bar will be displaced in
the direction of the displacement of the supports, and in the same proportion ;
and any measurements of earthquake motion which are taken with reference to
the centre of the bar as a datum-point must be multiplied by 1 + qe to
find the true displacement of the ground. When the bob is a lump of lead
whose dimensions are small compared with the length of the hanging bar, this
factor differs very little from unity.

Two light bars with dense rigidly attached bobs, and suspended by silk
threads so as to swing at right angles to each other, form an excellent two-
component seismograph, especially suitable for the measurement of large
earthquakes. The complete absence of joints makes the frictional resistance
exceedingly small: in this respect the method of suspension now under examina-
tion constrasts very favourably with most of the methods which have been
described in Chapters II and III. Moreover the construction, and also the
adjustment, is very simple. For large earthquakes the method of recording
shown in fig, 51 will probably be found suitable. A light pointer 7 is forked so
as to enclose the hanging bar, and is jointed to the bob at its centre. Its end m
rests on a smoked-glass plate, the pressure being regulated by a counterpoise n.
The plate should be set in rotation by an earthquake, by means of a seismoscope
which may conveniently be somewhat wanting in sensibility, in order that it may
act only when the motion becomes tolerably severe. The record will be less
than the true motion, in the ratio given above.

Fig. 52 shows an instrument of the same kind in plan, but with a multiply-
ing index p, which is pivotted to the ground on a vertical axis 0, and receives its
motion by having a bent-up end which gears into a slotted plate q fixed to the bob.

Fig. 53 is a sketch of a curious form of double-freedom seismometer, in
which a somewhat similar method of suspension is used. The base S, which is
fixed, is an equilateral triangle, at the corners of which there are three conical
cups forming sockets for three legs rrr of equal length. These legs press up
into three other sockets in the plate s, which are placed so as to form an equi-
lateral triangle of half the linear dimensions of the base. The height of s from
the base is 0.866 times the distance between the base sockets. The plate s carries
a massive bob w, slightly beneath it. The plate is then in nearly neutral (some-
what stable) equilibrium with respect to small motions in any azimuth, ‘ It is
necessary here to invert the system and use struts instead of ties, since a plate
hung by three crossed cords would be azimuthally unstable. Two of the legs
are made with loops to allow the three to cross each other. By adding a multi-
plying pointer to record the displacements of the base with respect to the centre
of the plate, we should obtain a compact form of double-freedom seismograph,
whose frictional resistance (though much greater than that of the single-freedom
instrument just described) could probably be kept within reasonable limits.

Scale '/>.

MVVU
ge a
NO )

ん

F
Sea
le |
2

Alm
T

Earthquake Measurement, Plate IV.

0a) ER)

1 me
RNNSSSSSSSSS Mt

Fig. 10.
Seale i, 2.

NN きつ

rn)

HS
WO

Earthquake Measurement, Plate VII.

NSS = R
DV

SS
ZY SY —
wm,
9

レノ /

Earthquake Measurement, Plate 7777

Scale */,

Fig.28. Scale /,

®
D

en mug N

3 《
ape
“2
np
¥
る
ree

WIR の

Earthquake Measurement, Plate IX.

Tro
~
Y

Fig. 2 9. Seale it: Fig. 30.

<i venk? だ お

Earthquake Measurement, Plate X.

‘a
ins)

ーー 一

ir)

A arn A nh
VAAN AAA

WVAWVVVVVVMMVVVWWW

=
=
=
=
=
=
>
:
:
>
=
て
=
2
ミ
ミ
Ss
=
:
ミ
oa
ミ
=
=
=
=

R tt,
フン イー L Ak
を N
ner ルル パソ / Wop |
1 vA wheal Wy Nv NAW Mid

Ir Ne 9
ト fig. 31. Seale "/,, Fig. 32.

teh ee ome 1)

>
N

3
»
を

=
:

2 ーー =
と ーーーーー ーー ーーー バー 一 =
a a > a =

ーーーー デ キー ーーーー bien ーー _

> wns

Earthquake Measurement, Plate XV.

a en“ u ET ee eS oe BB . = =. az
; F " ei h at. <=> =
. A aa
”

Measurement, Plate XVI.

Earthquake Measurement, Plate XVI.

“
S

eS eee SS en TS ae ee

XY >

Earthquake Measurement, Plate XVII.

A Earthquake Measurement, Plate XV;

ーー デー

ンジ ニニ ーー

Earthquake Measurement, Plate XIX

Earthquake Measurement, Plate XX.

2 a ャ ーー
ENDE

6 |
=m
|

li
I

td
WI

Earthquake Measurement, Plate

fa‘

WE 7 yr a 7

PC ee eS RED AS CAR) a On eee?
re ed ‘orate を ag - さか ト pay?

kp Measurement, Plate XXII.

4 % ae Me
ae a ial). 27 KR
ERS ER あう Tp dl a

ER a a Av

» Measurement, Plate XXIII.

7 も
まい
hh

お 6
ne -
at

I:
a

Karthquake Measurement, P.

er Pe 』、、 1 ig ‘
の y I zu
い \ 10%

ABHANDLUNGEN

DES

BMBORFIOTBATITGAKU.

(Universität zu Tokio.)

No.10.

PHYTOCHEMISCHE NOTIZ

UEBER EINIGE ん

JAPANISGHE PFLANZEN.

VON

J. F. EYKMAN.

PROFESSOR FÜR CHEMIE UND PHARMACIE AN DER MEDIOINISCHEN FACULTÄT,

HERAUSGEGEBEN VOM TOKIO DAIGAKU.
TOKIO.
2543 (1883.)

VAC RAT ASR IEA ASTLES

JAADIAAG OTAo?,T

AÄnT us Seis vii

と まい きじ jj

wo tei
& g8Ot a i AUL

DAADIAG OLIOT MOV FAHHDAORTZUER
OTAOT

(SBRT) Saat

ABHANDLUNGEN

DES

eer ree DATG Ak WU.

(Universitat zu Tokio.)

No. LO:

PHYTOCHEMISCHE NOTIZEN

UEBER EINIGE

FAPANDSEHE-PFRANZEN:.

VON

J. F. EYKMAN.

Proressor FÜR CHEMIE UND PHARMACIE AN DER MEDICINISCHEN FACULTÄT,

HERAUSGEGEBEN VOM TOKIO DAIGAKU.
TOKIO.
2543 (1883.)

EINLEITUNG.

Die Japanische Flora, so interessant in botanischer Hinsicht, bietet eine
Reihe von Pflanzen dar, welche auch in chemischer Hinsicht, toxicologisch wie
pharmacologisch, das höchste Interesse verdienen.

Nur eine kleine Anzahl ist bis jetzt Gegenstand solcher Forschungen
gewesen, und daher liegt hier noch ein ziemlieh unerschöpftes Feld zur
Bearbeitung vor.

Die Wichtigkeit phytochemischer Untersuchungen von einheimischen
Pflanzen erhellt leicht, wenn man die Schwierigkeiten ins Auge fasst, welche bei
hiesigen forensischen Untersuchungen auftreten und eine Folge sind von der
Unbekanntheit mit den giftigen Bestandteilen der einheimischen Flora. Wenn
auch die üblichen Methoden zur Ausmittelung der Gifte es gestatten, unbekannte
Gifte abzuscheiden und das physiologische Experiment an Tieren dann häufig
über ihre Giftigkeit Auskunft geben kann, so ist es doch für ihre weitere
Erkennung auf chemischem Wege erforderlich, dass man sowohl genauere
Kenntnisse über diese Gifte selbst besitzt als auch solche andere nicht giftige
Pflanzenbestandteilen kennt, welche zur Verwechslung mit bekannten Giften
Anlass geben könnten.

Auch von medieinischem Standpunkte sind mehrere Pflanzen oder ihre
Drogen wichtig, sei es dass sie wegen einer therapeutisch verwertbaren Wirkung
einen Platz im Arzneischatz verdienen, sei es dass sie mehr locale Bedeutung
besitzen, insofern sie als Ersatz von anerkannt wichtigen europöischen Pharmaca
dienen können.

Schon früher hatte ich begonnen einzelne einheimische Pflanzen von
obigem Staudpunkte aus zu untersuchen, und ich habe diesen Plan wieder
aufgenommen als mir mein Eintritt in die hiesige medicinische Facnltät dazu
weitere Gelegenheit bot. Es schien mir dabei zweckmiissig, meine Aufmerk-
samkeit einstweilen auf die wichtigeren Bestandteile zu beschränken, um so eine
grössere Zahl von Pflanzen oder Pflanzenteilen wenigstens vorläufig zu erledigen.
Meine Untersuchungen bieten daher weder ein abgeschlossenes Ganze noch
machen sie Anspruch auf etwaigen Erschöpfung der einzelnen Gegenstände.
Den Herren T. Shimoyama, K. Tamba und T. Niwa, die mir dabei abwechselnd,
jeder während eines Semesters, zur Seite standen bin ich für ihre Unterstützung
zu lebhaftem Dank verpflichtet. Denselben Herren verdanke ich auch sämmtliche
Angaben aus der mir nicht zugänglichen japanisch-chinesischen Litteratur.

Auch sei es mir hier gestattet, dankend die grosse Liberalität hervorzuheben,
womit seitens des Tokio-Daigaku sowohl die oft grossen Quantitäten des zur
Untersuchung dienenden Materials als das Laboratorium, auch während der
Ferien, zu meiner Verfügung gestellt wurde. Die beigefügten Tafeln sind von
Herrn K. Tsikasawa ausgeführt.

Leider stand mir über Phytochemie nur eine sehr beschränkte Litteratur
zu Gebote und hoffe ich, auch da, wo ich vielleicht nicht genügend die Arbeiten
Anderer über verwandte Gegenstände benutzte, nachsichtige Beurteilung zu
finden.

Tokio-Universität. J. F. EYKMAN.

VER

INHALT.

EINLEITUNG.
TL, -ANDROMEDA JAPONICA, THURS Se. cc cccomciccsnchpins erneuten
TEE SCOPOLEA 1) MPONTOAS MEAG Nee seen tosses
EET, MACE AS GORDAWAC RS BR een een nntea er
EV (.Camnmiontrome MAUS ER. 2.20 nee tie ov es een tee tar
Vi NEON AS DOMESTICA, IHONE ee
ME, (OnreAy Maponics RUB. 00a ee
VII SkIMmMIA Japonica THUNB. ..................... 2 この

41—46

47—52

I. ANDROMEDA JAPONICA THUNB.

Von den vielen in Japan und China einheimischen und zu der Gattuny
Andromeda gehoerenden Ericaceen ist die Andromeda Japonica eine hier
allgemein als giftig bekanute Stande. Als solche findet sie sich schon in den
ältesten Werken über Naturgeschichte Japan’s und China’s erwähnt.

Was sich in dieser Literatur über die Pflanze findet, lässt sich folgender-
massen kurz zusammenfassen.

Die länglichen, bitter und zusammenziehend schmeckenden Blätter sind
giftig und üben, von Pferden und Kühen gefressen, eine betäubende und
giftige Wirkung aus. Daher die Namen: Basuiboku 馬 酔 木 。 Umakuwasıu
ウゥ マク ハズ ②、 wie auch Shika kuwasu vy # 27»%() und Sishi kuwasu
YN IZ (4).

Ein Abzug der Blätter vertilgt Insecten und Würmer() und soll als
Waschmittel bei Ulcerationen und Scabies und auch als Gegengift gegen
Fugul®) gebraucht werden; die zerstossenen Blätter können nach Sohki (Chin.
Arzt) als Mittel gegen Schlangenbiss dienen, und der Geruch einer Abkochung
der Blätter verursacht beim Menschen Koptschmerz (7).

Die Pflanze wächst überall auf Bergen($), ist immergrün und wird daher
als Zierpflanze benutzt.

Eine Reihe von Namen sind in den verschiedenen Provinzen in Gebrauch.(®)

(1-3) Ba = Uma = Pferd, sui=betäubend, boku -- Baum, kuwasu =nicht essen, Shika ==
Hirsch.

(4) Thunberg, Flora Japonica: ,,sis kwas, i. e. Leo non edat vel leoni non condueit edere :
sis enim leonem significat. Item sishi gakure’’ (gakure = sich verstecken.)

(5) Honzokomoku keimo As MH #4 A PA

(6) Daher der Name Fugu Shiba (Fugu=Giftfisch, Shiba=Pflanze) ja] [% 48

(7) Yudoku Somoku Snsetsu {jf #/ MARR. Auch wird da erwähnt, dass hei Hirschen,
welche die Blütter fressen, die Hörner plötzlich abfüllen. In der Honzokomoku 本 並 綱目 vou
Lishishin 3 時 ¥3 wird noch angegeben, dass ein Abzug der Blätter das unreine Blut bei der Fran
nach der Geburt abtreibt.

(8) Thunberg und Oldham fanden sie bei Nagasaki, Bürger in Homamdake, Savatier in
Hakone und Sagami (Franchet und Savatier, Enumeratio plantaruım). Nach Sohki wächst sie in
China in den Wäldern von Kohto.

(9) In Honzokomoku keimo finden sich noch folgende Namen:

Yosebi 1er

Yoshimi 42?
Yoshimishiba ョ シミ シバ
Yomeba ョ メ バ
Dokushila ドク シバ
Kasukni カス タイ
Onasakamori ラナ サカ そ や リ
ザ Jenkishiba テア キシ レバ
Aseboshiba アセ ポレ シバ

2 :

Die am meisten üblichen sind: Asebo アセ ポ und Basuiboku.

E. Kämpfer in seinen Amönitates exoticae Fasc. V. p. 896. A° 1712
beschreibt diese Pflanze etwas ausführlicher unter den Namen Asjebo und
Asjemi. In Thunberg’s Flora Japonica p. 181 findet sich eine gute Beschreibung
nebst einer Abbildung der Pflanze (1)

Soweit ich in Erfahrung habe bringen kénnen, wurde diese Pflanze hier
schon mehrfach zum Gegenstande von Untersuchungen gemacht, und es hat
sich dabei herausgestellt, dass ihre Giftigkeit keinem Alkaloide zuzuschreiben
sei, dass die Blätter giftiger sind als das Holz und dass das Gift nicht durch
Bleiacetat gefällt wird, Diese Resultate wurden schon vor etwa sieben Jahren
im hiesigen Laboratorium erhalten. Unabhängig davon beschäftigte ich mich
kurze Zeit mit der Untersuchung des Holzes (Ende 1880-Anfang 1881) auf
Wunsch des Herrn Nagayo Sensai, Chefs des Central-Gesundheitsamts. Es
war ein Vergiftungsfall, der damals zu dieser Untersuchung Veranlassung gab.
Wiewohl es mir leicht gelang, eine amorphe, fast farblose Substanz von heftig
beizendem Geschmack abzuscheiden(2), so konnte ich doch, weil ich die
Untersuchung unterbrechen musste, damals nicht nachweisen, in wieweit diese
Snbstanz rein war, und in wieweit ihr als solcher die giftige Wirkung des Holzes

zuzuschreiben war.

Yosebo 3%#
Gomajakishiba ゴマ ヤキ シバ
Shiyari-shiyari シヤ リ シ ヤ リ
Hassasagi ハ サ サ ギ
Auch kommen in alten Liedern vor:
Asemi アセ ミ
Asemo アセ モ
Asebi アセ ビ
Chinesische Namen sind,
Shinboku #27
Tin-shu-kwa PH (tin = merkwürdig, shu = Kugel, kwa = Blume)
Bei-han-kwa 米飯 花 (Bei han = gekochter Reis, kwa = Blume)

Der Name Dodan, von Franchet & Savatier u. A. erwähnt für A. Japonica, ist der des
Enkyanthus Japonicus, wie mir Herr Prof. Ito Keisuke freundlichst mitteilte.

(1) Caulis arboreus, ramosus. tami terni vel plures subumbellati, striato-angulati,
glabri, purpurascentes, a casu foliorum nodulosi erecti, iterum ramulosi. Folia in ultimis
ramulis frequentia, alterna, petiolata, obovato-lanceolata, acuta, inferne atlenuata in petiolos a
medio ad apicem crenata, patentissima vel reflexa, nervosa, glabra, bipollicaria. Petioli
semiteretes, suleati, rubri, lineam longi. Flores in apicibus ramulorum racemosi. Racemi
alterni, laxi, rubri digitales. Pedicelli teretes, incrassati, erecti, lineam longi. Bracteae
subulatae, sparsae sub et in pedicellis. Perianthium ante florescentiam 5-gonum, acutum,
glabrum, altero Jatere rubrum aliero viride, ad basin fere 5-partitum : laciniae lanceolatae, vix
lineam longae. Corolla 1-petala, subeylindriea ore quinguefido, alba, calyce brevior, 5-striata.
Filamenta 10, receptaculo inserta, linearia, alla, calyce triplo breviora. Antherae ovatae, erectae,
intus gibbae, purpurascentes. Germen suyerum, 5-striatum, convexum, glabrum. Stylus
filiformis, viridis, brevissimus. Stygma simplex, oltusum viride purpureum. Capsula ovato-
globosa, 5 sulca angulis obtusis, glabra, 5 valvis, 5 locularis. Semina plurima, minuta.

(2) Sie wurde erhalten aus den mit Wasser oder verdiinntem Weingeist dargestellten
Extrakten durch wiederholte Ausziehung mit starkeın Alkohol und Versetzen der alkoholischen
concentrirten Lösung mit Aether. Die alkoholhaltige ätherische Lüsurg hinterliess beim
Verdunsten einen amorphen, in Wasser schwierig, in Alkohol und Essigsäure leicht loslichen
Rückstand.

Erst im Frühjahr 1882 konnte ich diese Untersuchung mit grösseren
Quantitäten Materials (Blätter) wieder aufnehmen.

Zur Darstellung des Giftes wurden ziemlich grosse Quantitäten der frischen
Blatter mit Wasser infundirt und das erhultene Infus auf dem Dampfbade zur
dünnen Sirupconsistenz eingedampft und danach filtrirt. Zuerst wurde etwa ein
halbes Liter dieses Infuses wiederholt mit offieinellem Aether ausgeschüttelt ;
der ätherische Rückstand nach dem Abilestilliren des Aethers löste sich nur'zum
Teil in Chloroform auf, und der Chloroform-Auszug hinterliess einen fast
farblosen Rückstand, welcher sich durch das physiologische Experiment auf
Kaninchen heftig giftig erwies(*). Eine andere Probe des Infuses mit
Chloroform geschüttelt zeigte, dass dieses Menstruum das Gift bei weitem reiner
aufnimmt, als Aether(f), weshalb ich zum Ausschütteln der übrigen, grösseren
Mengen des Infuses Chloroform benntzte.

Die durch Destillation concentrirte Chloroformlösung, welche srünlich
gefärbt war, wurde mit soviel Petrolenmäther versetzt, dass keine Ausscheidung
mehr stattfand(t). Dieselbe war nur wenig gefärbt und amorph; sie wurde in
alkoholhaltigem (käuflich officinellem) Aether aufgenommen und diese Lösung
wiederholt mit Wasser ausgeschüttelt. Die wässerigen Schichten wurden getrennt
und in flachen Schalen bei gelinder Wärme eingetrocknet. Sie schieden dabei
eine gelbe, ölartige Substanz ab, welche bei weiterem Eintrocknen in dünnen
Schichten glasklare, farblose oder schwach gelblich gefärbte Rückstände hinter-
liess, woran nicht die geringste Neigung zur Krystallisation zu erkennen war.
Auch durch partielle Fällung «der ätherischen oder chloroformösen Lösung
mittelst Petroleumäther erhielt ich nur amorphe Fällungen. Die meist farblosen
Anteile wurden gesondert aufbewahrt. Der Rest, sowie die in dem Aether etc.
zurückgebliebenen Portionen, welche nach dem Abdestiiliren der Lösungsmittel
zurückgewonnen wurden, bildeten eine gelbbräunliche, amorphe Masse, welche
in wässeriger Lösung auf Lakmus deutlich sauer reagirte. Weil die reineren

(*) Bei der häufig wiederholten Ausschüttelung mit Aether schieden sich aus dem Infus
Krystalle ab, welche gesammelt und abgewaschen fast völlig weiss waren und bei Verbrennung
bedeutende Mengen kohlensauren Kalk haltige Asche zurückliessen, wonach es scheint, dass in
den Blättern ein krystallisirendes Kalksalz vorhanden ist. Bei der weiteren Untersuchung
habe ich dies aber nicht weiter verfolgt. Auch wurde in den Blättern gefunden :

66.7 % Wasser,

1.83 % in Petroleumiither,

7.86 % in Alkohol lösliches Extrakt und
2.38 % viel Caleiumcarbonat haltige Asche.

(+) Aether nimmt ausserdem betriichtliche Mengen anderer Substanzen auf, welche mit
Bleiacetat einen starken gelben Niederschlag, mit Eisenchlorid Blaufiirbung erzeugen.

(t) Das wiisserige sirupöse Infus (mehrere Liter betragend) wurde in Portionen mit etwa
2-3 Liter Chloroform ansgeschtittelt。 dann das Chloroform abgehoben und abdestillirt und oft
zu erneuten Ausschiittelungen verwendet. Etwa 100 gr. in Chloroform lösliche Substanz von
schwach grünlichgelber Farbe wurden auf diese Weise erhalten, Die Ausfüllung mit Petroleum-
iither hatte den Zweck, die völlige Verjagung des Chloroforms durch lüngere Erwiirmung zu
umgehen und eventuell anwesende in Petroleumiither leicht lösliche Substanzen zu entfernen.
Durch Destillation der chloroformhaltigen Petroleumiitherlésung wurde der gelöste Anteil
zurückgewonnen und bei spiiteren Reinigungen verwendet,

Anteile keine oder nur eine weinrothe Färbung mit Lakmus erzeugten,
schien eine sanerreagirende Substanz beigemischt. Ich behandelte darum diese
Quantität wie folgt. Sie wurde, in Chloroform gelöst, mehrere Tage mit etwas
Bleihydroxyd unter öfterm Umschütteln stehen gelassen. Es bildete sich ein
bräunlich-gelber Bodensatz, während bedentende Mengen Blei in Lösung über-
singen. Von der filtrirten Lösnng wurde das Chloroform abdestillirt und über
den Rückstand schwach übererhitzter Wasserdampf geleitet(*) Was in der
tetorte zurückblieb, war eine ölige Flüssiekeit und eine wässerige. Die letztere
enthielt bedentende Mengen Blei, die ölige Flüssigkeit wurde nach dem
Erkalten gummiähnlich fest. Sie wurde von der wässerigen Flüssigkeit getrennt
und mit alkoholhaltigem (känflich officinellem) Aether behandelt. Hierin löste
sie sich zum grossen Teil, indem eine weissliche Bleiverbindung ungelöst
zurückblieb, welche daher wohl in Chloroform, nicht aber in Aether löslich war,
Die ätherische Lösung enthielt nur Spuren Blei, welche leicht durch einmalige
Ausschiittelung mit ein wenig Wasser nnd Durchleiten von etwas H,S entfernt
werden konnten. Die ätherische Lösung mit etwas Weingeist versetzt. lieferte,
mit Wasser ausgeschüttelt, Auszüge, welche bei der Eindunstung sich wie
früher verhielten und zu spröden, glasartigen, amorphen Massen eintrockneten.
Die am meisten farblosen Anteile wurden wieder gesondert aufgehoben und
stimmten in ihren Eigenschaften mit den früher erhaltenen Quantitäten
überein.

Die in Aether ımgelöst gebliebene Bleiverbindung wurde mit Chloroform
übergossen der Einwirkung von H, S ausgesetzt. Das bleifreie Filtrat lieferte
beim Verdampfen ebenfalls einen amorphen. gelben Rückstand.

Die etwas mehr gefärbten Rückstände der wässerigen Ausschiittelungen und
die in den ätherischen Menstrua gelöst zurückgebliebenen Quäntitäten wurden
nochmals in Chloroform gelöst, diese Lösung wurde dann mit Wasser und soviel
R,CO,- lösung geschüttelt, dass die Reaction sehr schwach alkalisch war. Aufbrau-
sen wurde dabei nicht beobachtet. Die wässerigen und chloroformösen Schichten,
welche beide brimlich gefärbt waren, wurden getrennt und die chloroformöse
Lösung mit etwas Tierkohle geschüttelt, wodurch sie sich klar filtriren liess.
Das Filtrat wurde partiell mit Petrolenmäther gefällt und die weniger gefärbten
Anteile gesondert getrocknet und mit absolutem Aether gekocht. Der filtrirte
Auszug liess beim Eindunsten einen schwach gelblichen Rückstand, welcher

(*) Es hatte dies den Zweck, leicht fliichtige Körper zu entfernen. Das wässerige Destillat,
worauf eine sehr geringe Menge eines fettigen Körpers zich zeigte, wurde mit Aether ausge-
schüttelt. Nach freiwilliger Verdunstung desselben blieb aber nur eine sehr geringe Menge
Riickstand, welcher, in Wasser verteilt und einem Kaninchen eingespritzt, keine deutlichen
Vergifinngssymptome zeigte, Mit concentrirter Salzsäure fiirbte es sich röthlich. Beim
Ueherleiten des Wasserdampfes entwickelte sich ein eigentümlicher, nicht sehr unangenehmer,
aber betiubender Geruch. Es kommt mir wahrscheinlich vor, dass schon beim Ueberleiten von
schwach übererhitztem Wasserdampfe Zersetzung eintritt unter Bildung eines betäubend
riechenden Körpers, weil auch bei mehrstiindigem Ueberleiten des Wasserdampfes dieser Geruch
in gleich intensiver Weise auftritt und das Nicht- Auftreten von Vergiftungssympiomen sowie die
röthliche Färbung mit I Cl dagegen spricht, dass die Substanz unverändert überdestillire.

うり

in wässeriger Lösung völlig neutral reagirte. Mit Salzsäure auf Pt-Blech erhitzt,
entstand Dunkelblaufärbung, nachher intensive Purpurfirbung, und unter
3ildung weisser Dämpfe blieben höchst geringe Spuren Asche zurück. Mit cone.
Salzsäure übergossen entstand nach einiger Zeit Dunkelblaufärbung.

Die wässerige K, CO, haltige Lösung wurde mit verdünnter Schwefelsäure
angesäuert-und mit Aether ansgeschüttelt; der abgehobene Aether, mit etwas
Wasser gewaschen und der freiwilligen Verdunstung tiberlassen, hinterliess wieder
einen amorphen (schwefelsäure-freien) Rückstand, der stark saure Reaction zeigte
und mit Salzsäure ebenfalls blau wurde. Hiernach scheint das reine Gift
entweder durch Oxydation oder unter dem Einflusse schwacher Basen als Zerset-
zungsprodukt einen sanerreagirenden Körper zu bilden ; vielleicht findet dies
auch schon durch Bleihydroxyd statt; dies würde erklären, wie beim Schütteln
der chloroformösen Lösung des Giftes mit Pb (O Hj* während mehrerer Tage
sich immer mehr von einer Bleiverbindung in dem Chloroform löst. Weil auch
verdünnte Säuren, sogar Oxalsäure, beim Erwärmen das Gift sichtbar zersetzen,
scheint mir bei der Darstellung und Reinigung jede Anwendung von chemisch
wirkenden Agentien unstatthaft und nur diejenige Methode zu empfehlen,
welche, wenn auch mit Verlust, die Erhaltung eines reinen Präparates
auf mehr mechanischem Wege ermöglicht.

Es wurden deshalb die erhaltenen reineren Portionen, nach dem Trocknen
zu Pulver zerrieben, wiederholt mit Petroleumäther und Benzol ausgezogen
und nochmals in alkoholhaltigem Aether gelöst. Nach dem Ausschütteln mit
Wasser wurden die resultirenden verdünnt alkoholischen Lösungen wieder bei
mässiger Wärme eingetrocknet und die spröde glasartige Substanz gepulvert.

Sie war schwierig in kaltem, besser in heissem Wasser und leicht in Weingeist,
Eisessig, Amylalkohol, Chloroform und alkoholhaltigem (käuflich ofücinellem)
Aether löslich. Petrolenmather und Benzol lösten dagegen sehr wenig. Die
wässerige Lösung wurde weder gefällt noch verändert durch Eisenchloril und
Silbernitrat und hatte einen bitteren, stark beizenden Geschmack.

Beim Verbrennen schmilzt sie, verbreitet scharfe Dämpfe und hinterlässt
keine Spur Asche. Mit concentrirter HCl] übergossen, zeigt sie, auf weisser
Unterlage betrachtet, nach einiger Zeit eine schöne reine Blanfärbnng; unter
Bildung einer bläulichgrauen Trübung wird die Lösung allmählich mehr röthlich
und zuletzt missfarbig.(*)

Wird die blaue Lösung auf dem Wasserbade erwärmt, so entsteht an
der Peripherie eine schöne violettlich rothe Farbe.

In alkoholischer Lösung mit concentrirter HCl erwärmt, entsteht eine
prachtvolle dunkle Purpurviolettfärbung. Mit verdünnter Salzsäure oder
Schwefelsäure erhitzt, trübte sich die wässerige Lösung und fürbte sich

(*) Neben dem mtechenden Geruch der Salzsiiurediimpfe liess sich ein eigentümlicher, etwa
an Spirii ulmaria erinnernder Geruch wahrnehmen, Auch beim Kochen der Substanz mit
verdiinnten Mineralstiuren war ebenfalls ein besonderer, etwas mehr rhamnusühnlicher Geruch
zu erkennen,

schön (carthaminähnlich) roth unter Ausscheidung eines bräunlichen harzigen
Stoffes. Concentrirte Schwefelsäure löste es mit bräunlichrother Farbe, die
beim Stehen an der Luft schöner, rosaroth wurde. Durch die Cyanprobe konnte
kein Stickstoff nachgewiesen werden.

Als dosis letalis für Kaninchen wurden pro K° Körpergewicht etwa 3 mgr.
bei subeutaner Injection gefunden.

einigen Stunden

Die äusseren Symptome, welche ich bei Kaninchen wahrnahm, sind etwa
folgende: bei ungefähr letaler Dosis füngt das Kaninchen nach 15-20 Minuten
an, unter Hervorbringung klagender oder schreiender J.aute zurückziehende
Bewegungen mit dem Kopfe zu machen. Diese werden allmählich stärker,
wobei dann die Bewegung des Kopfes nach vorwärts gepaart geht mit weiter
Oeffnung des Maules und der Nasenlöcher und schnalzenden Lauten. Oefters
wurde Defäcation oder Harnlassen beobachtet. Die Körperwärme nimmt deutlich
ab, und das Tier kann sich weniger gut bewegen. Es tritt Paralysis der
xxtremitaten auf, die Mucosa des Mundes und der Nasenhöhlen sind sehr
anämisch, und unter Ausfluss von Schleim tritt nach einigen Convulsionen,
anscheinend durch Erstickung, der Tod ein.

Hunde beginnen bald nach der Darreichung des Giftes zu erbrechen.⑪)

In Anklang mit dem japanischen Namen Asebodoku = Asebo-gift habe ich
für die Substanz den Namen Asebotoxin gewählt.

Zwei Elementaranalysen lieferten folgende Zahlen :

I. 0.2732 gr. bei 110°— 115° getrocknet gaben 0.2631 gr. Trocken-
substanz und nach Verbrennung 0.5815 gr. CO, und 0.1983 gr. H,O

II. 0.3288 gr. bei 110°—115° getrocknet lieferten 0.3181 gr. Trocken-
substanz und 0.7073 gr. CO, und 0.2078 gr. H,O.

Zur weiteren Bestätigung der Reinheit der Substanz wurde das noch
übrige Pulver mit absolutem Aether (über Na destillirt) übergossen und in

(7) Die Erscheinungen sind denjenigen analog, welche ein wiisseriges Infus der Blütter zeigt.
1 Cm’ eines solchen Infuses, correspondirend mit etwa 200 mgr. der jungen Blätter, tödtete ein
Kaninchen von etwa 1'14 K° Körpergewicht und stimmt deshalb in toxischer Wirkung mit
ungeführ 3—4 mgr. des Giftes überein.

zwei Portionen getrennt. Es zeigte sich dabei, dass das Asebotoxin in diesem
Menstrunm schwieriger als in nicht- absolutem Aether löslich ist. Sowohl von
dem in Aether gelösten als von dem darin ungelöst gebliebenen Teil wurden
nach dem Trocknen bei 110°—115° Elementaranalysen ausgeführt. Der in
Aether gelöste Teil löste sich nach dem Eintrocknen ebenso in absolutem
Aether schwierig auf.
Gefunden wurde:
III. In Aether ungelöster Anteil
0.3575 gr. lieferten 0.7929 gr. CO, und 0.2368 gr. H,O
lV. In Aether gelöster Teil
0.3898 gr. gaben 0.8651 gr. CO, und 0.2610 gr. H,O
Also in Procenten :

I II III IV Mittel
60.64 | 60.49 SE 60.48

—12: 60.28

|
7.53 | 7.29 | 7.36 7.44 | 7.405
| |

0 一 16| 32.19 32,17 32.15 32.04 32.115

Die Uebereinstimmung der Zahlen III und IV, auch in Vergleich mit I und
II, dürfte noch für die Reinheit des Asebotoxins sprechen. Es blieb mir nun
noch eine kleine Quantität von dem bei der partiellen Lösung in Aether
zurückgebliebenen Anteile, um einige weitere Reactionen anzustellen.

Die wässerige Lösung reducirte beim Kochen alkalische Kupferlösung(*)
und wurde nicht von Bleiacetat, Goldchlorid, Kupfersulfat, Quecksilberchlorid
und Eisenchlorid verändert noch gefällt. Durch Bleisnbacetat (Bleiacetat +
NH,) entstand eine flockige Trübung. Die Lösung in Alkohol reagirte auf
Lakmus neutral, die warm bereitete wässerige Lösung veränderte blaue Lak-
mustinktur nicht oder färbte sie kaum röthlich violett. Auch die Lösung in
Chloroform, mehrmals auf Lakmuspapier eingetrocknet, veränderte dieses nicht.

209 0°0-——

Ein Teil des wasserigen Infuses, woraus durch Chloroform das Asebotoain
erhalten war, wurde wiederholt mit oflieinellem Aether ausgeschiittelt. Bei dem
Abdestilliren des Aethers blieb, namentlich bei den, letzten Anteilen mehr
farblos, eine krystallinische Masse zurück, welche mit Bleiacetat einen chromat-
gelben, mit Quecksilberchlorid einen weissen und mit ammoniakaler Silber-
lösung einen starken schwarzen Niederschlag erzeugte; Eisenchlorid gab
Fällung und tiefe Grünblaufärbung. Die Rückstände reagirten sauer und
wurden mit Chloroform zur Entfernung des noch gelösten Asebotoxins ausge-

(*) Eine gleiche Quantität dieser wiisserigen Lösung, mit verdiinnter Schwefelsiiure gekocht,
lieferte nach dem Abfiltriren des braunen harzigen Zersetzungsproduktes und Neutralisation mit
Natronlauge eine anscheinend wenigstens gleichstarke Reduction, Da das abgeschiedene,
harzige, in Weingeist und Aether leicht lösliche Zersetzungsprodukt keine sichtbare Reduction
mit alkalischer Kupferlösung zeigte, scheint die gelöst bleibende Substanz eine stürkere Reduc-
tionsfühigkeit als das Asebotoxin zu besitzen.

zogen, dann mit Wasser gekocht und die Lösung mit essigsaurem Blei gefällt.
Nach dem Abfiltriren des orangegelben Präcipitates wurde aus dem Filtrate
das Plei entfernt. Nach Kindunsten der Flüssigkeit wurden gelblich gefärbte
Krystalle erhalten, die in grösseren Quantitäten aus dem Rest des wässerigen
Infuses gewonnen wurden. Dieser wurde warm mit Bleiacetat gefällt, aus dem
Filtrate das Blei grossenteils durch verdiinnte Schwefelsäure und zuletzt durch
H,S gefällt. Nach Eindampfen des Filtrates und längerem Stehen schied
sich eine Krystallmasse ab, welche durch Umkrystallisation aus Wasser und
verdünntem Alkohol etwas gereinigt wurde. Durch wiederholte Umkrystalli-
sation aus Wasser und verdünntem Alkohol konnte ich sie nicht völlig farblos
erhalten. Dies gelang aber folgendermassen: ein Teil wurde in 20 Teilen
absolutem Weingeist gelöst, wenn nöthig filtrirt und etwa 200 Teile absoluten
Aethers zugesetzt. Durch Schütteln dieser Lösung mit Tierkohle wurde die
Flüssigkeit völlig farblos und klar. Nach dem Abfiltriren wurden der Lösung
etwa 100 Teile Wasser zugesetzt und dann der Aether abdestillirt. Die
zurückbleibende verdünnt alkoholische Lösung lieferte nach dem Abkühlen
ganz farblose, glänzende, völlig gleichförmige Krystallnadeln. Sie wurden bei
gewöhnlicher Temperatur getrocknet und zur Entfernung von eventuell an-
hängenden Spnren Asebogenin mit Aether abgewaschen und nochmals aus
Wasser umkrystallisirt. Diese Substanz, Asebotin, hatte folgende Eigenschaften:

Farblose und glänzende Nadeln, welche sehr wenig in kaltem, gut aber
in kochendem Wasser löslich sind; Petroleumäther, Benzol, Chloroform und
absoluter Aether lösen es nicht oder sehr wenig.(*) Die Lösung in Wasser hat
einen rein bitterlichen Geschmack. Alkohol, auch absoluter, und Eisessig
lösen es sehr leicht. Es reagirt nentral, und seine wässerige Lösung wird von
den gewöhnlichen Metallsalzen, auch Bleiacetat, weder verändert noch gefällt.
Mit Bleisubacetat (Bleiacetat + NH。) aber giebt es eine starke weisse Fällung.
Für den Schmelzpunkt fand ich 147°5 (uncorr.), für das spec. Gew. 1.356 bei
15°. In kleinen Dosen, etwa 5 mgr., in etwas warmer Lösung Kaninchen
eingespritzt, zeigte es keine abnormen Erscheinungen.

In verdünnten Alkalien, etwas weniger leicht in Ammoniak, löst es sich
in grossen Mengen auf zu einer farblosen Lösung, worin verdünnte Säuren
wieder einen nach einiger Zeit krystallisirenden Niederschlag bewirken(t) An
der Luft färbt sich die alkalische Lösung nach längerer Zeit braun. Mit

(*) Einige Löslichkeitsbestimmungen bei gewöhnlicher Temperatur führten zu etwa folgen-
den Verhältnissen.

Wasser 1: 2000 & 2500
absoluter Aether i: 6000 & 6600
Chloroform 1: 25000

(+) 0.4758 Gr. des Asebotins lösten sich leicht auf in einer verdünnten Kalilésung, welche
0.105 gr. reine KOH enthielt. Brom verursacht in dieser alkalischen Lösung eine hellgelbe
Abscheidung, durch Ueberschuss übergeheid in eine in Chloroform und Aether sehr lösliche und
bei Verdampfung sich amorph abscheidende Substanz, welche allmählich eine hochrothe Farbe
annimmt.

Salpetersäure erwärmt und eingedampft blieb ein in Wasser mit gelber Farbo
löslicher Rückstand, der deutliche Quantitäten Oxalsänre enthielt.

Alkalische Rupferlösung wird in der Kochhitze durch das Asebotin nicht
redueirt, nach vorheriger Erhitzung mit verdünnten Mineralsäuren findet aber
starke Reduction statt : ammoniakale Silberlösung wird nicht reducirt.

In einer fenchten Atmosphäre von Ammoniak mehrere Tage unter eine
Glocke gestellt färben sich die mit Wasser befeuchteten Krystalle röthlich braun.

Bei 100° getrocknet verloren sie kein Krystallwasser.

Elementar-A nalysen :

I 0.3130 gr., bei 100° getrocknet, lieferten 0.1574 gr. H,O und
0.6548 gr. CO,
II 0.3533 gr., bei 100° getrocknet, lieferten 0.1775 gr. H,O und
0.7383 gr. CO,
oder in Procenten :

} Berechnet auf

| 1 | Il Mittel C., H., 0,.(*)
C=12| 67.05 | 56.99 | 57. | 66.7
H=1| 6,59 5.58 5.6 | 5.5
o=16! 37.36 | 37.8 | ara | are

Kocht man das Asebotin mit verdünnten Mineralsäuren, so wird die
Flüssigkeit bald trübe und scheidet sich eine krystallinische Substanz, Asebogenin,
aus, indem die Flüssigkeit nach Erkältung und Filtration starke Reaction
gegenüber alkalischer Kupferlösung zeigt.

1.427 Gr. Aselotin mit etwa 30 Cm? verdünnter Schwefelsänre (cr
1:7) gekocht, lieferten, nach dem Abwaschen und Trocknen bei 106°, 0.941
Gr. = 65.9% in absolutem Aether leicht lösliches AseLogenin. Filtrat und
Waschwasser wurden mit Barynmecarbonat unter Erwärmung behandelt, nach
Filtration die Flüssigkeit zu einem kleinen Volum eingeengt und mit starkem
Alkohol versetzt. Das Filtrat lieferte nach Verdunstung eine bedeutende
Quantität eines amorphen, hellgelblich gefärbten, ganz in ihrem imsseren
Ansehen Glucose- ähnlichen Substanz, die stark redncireud auf alkalische Kupfer-
lösung wirkte.

0,4462 gr. Asebotin lieferten, mit etwas grösserer Quantität verdiinnter
Schwefelsäure (1:15) gekocht, bis sich dus Asebogenin wieder abgeschieden
hatte, und nachher noch mehrere Stunden auf dem Wasserhade erhitzt, 0 280
gr. = 62.7%, ebenfalls in absolutem Aether ganz nnd leicht lösliches Ase-
bogenin. Das Filtrat wurde unter Erwärmung mit Na, CO, fast vollkommen
neutralisirt, die Flüssigkeit auf dem Wasserbade eingeenzt und aufein bestimmtes
Volum verdünnt; sie zeigte bei der Titration mit einer alkalischen Kupfer-

(*) Wegen der vollkommnen Uehereinstimmung der Resultate dieser beiden Annlysen
wurde eine dritte nicht ausgeführt. Wenn auch mehrere empirische l ormeln sich sus diesen
Zahlen berechnen lassen, welche mehr oder weniger annübernd stimmen, ist nur die obige
angeführt, weil sie am besten mit den Spaltungsprodukten durch Siiuren libereinstimunt.

lösung einen Reductionswerth von 36.2 Glucose. Weil ich keine grössere
Quantität des Asebotins mehr besass, musste ich die weitere quantitative
Bestimmung der Spaltungsprodukte unterlassen.

Das Asebogenin, welches bei andauerndem Kochen mit den verdünnten
Mineralsäuren sich grünlich gelb abschied, wurde leicht völlig farblos erhalten
durch mehrmalige Lösung in starkem Alkohol und Fällung mit der etwa
zehnfachen Menge Wasser.

Eigenschaften des Asebogenins.

Farblose, sehr feine Krystallnadeln, welche in kaltem, wie auch in kochendem
Wasser sehr wenig löslich sind. Auch in Chloroform ist es unlöslich, dagegen löst
es sich in Alkohol und Aether, auch absolutem, wie in Essigsäure, sehr leicht auf.

Es ist geruchlos, reagirt neutral und giebt mit Bleisubacetat (NH, + Blei-
acetat) starke weisse Fällung. Als Schmelzpunkt fand ich 162° 一 163° (uncorr.)
In Alkalien ist es leicht löslich.

0.2958 gr., bei 100° C getrocknet, lieferten 0.1439 sr. H,O und 0.6764 gr.

CO, oder in Procenten:
Berechnet auf

Cis H,; 0,
C=12| 62.37 62.43
H=1 5.40 5.20
O=16] 82.28 32.37

Eine zweite Analyse konnte ich wegen Mangels an Substanz nicht ausführen.

Durch seine Zusammensetzung, Schmelzpunkt etc. und namentlich schon
durch seine sehr geringe Löslichkeit in kochendem Wasser und leichte Löslich-
keit in absolutem Aether unterscheidet sich das Asebogenin scharf von dem
Asebotin.

Die Spaltung dieses letzteren konnte durch die Gleichung:

C„H„0,+H0=(C,H,0,+ C,H,0,

ausgedrückt werden. Nach dieser hätte ich erhalten müssen 68.1% Asebogenin
und 35.4% Glucose, während ich fand: 1° Mal 65.9%, 2° Mal 62.7% Asebogenin
und 36.2% Glucose. Das Deficit an Asebogenin, welches da am grössten war,
wo die grösste Quantität schwefelsäurehaltigen Wassers bei der Spaltung
gebraucht wurde, kann auf Rechnung der geringen Löslichkeit des Asebogenins
in kaltem Wasser gebracht werden.

Bei dem Vergleich der Eigenschaften des Asebotins mit anderen Bitterstoff-
glucosiden, kommt Phloridzin in vielen Hinsichten damit überein, unterscheidet
sich aber schon u. a. durch seine Phloridzein- Ammoniak- Reaction, wovon ich
mich durch Vergleich mit zwei Proben Phloridzin überzeugte.

Mehr weicht es in seinen Eigenschaften von dem in Ericaceen aufgefundenen
Arbutin(*) ab, welches aber nur um die Elemente des Wassers vom Asebotin zu
differiren scheint.

(*) Ein hier vorräthiges Muster Arbutin wurde in einer feuchten NH,—Atmosphiire blau
gefärbt, wasich in der mir zugänglichen Litteratur nicht erwähnt fand.

2C,H,0;, — 2 HO = C,, H, On
Arbutin Asebotin.

Auch ist das Asebogenin verschieden von dem Hydrokinon, welches Hesse
als Spaltungsprodukt des Arbutins fand.

Ebenso kann das Asebotin nicht mit dem Methylarbutin (Schmpt. 168°-169°)
identisch sein wegen seines niedrigeren Schmelzpunktes, der geringen Löslichkeit
in kaltem Wasser und der abweichenden Eigenschaften des Asebogenins (Schpt.
162°-163°) von dem Methylhydrokinon (Schpt. 53°)(*) Möglich aber ist es,
dass bei der Spaltung des Asebotins etwas Hydrokinon sich bildet nach der
Gleichung

C2, Hes O» + 4 H,0 = 2 C, Hie 0, + 2°C, H,O;
oder Chinhydron nach der Gleichung
0 H, 0, = C, BH, Os + Os。 Hy, O, + Hr 0,

worauf vielleicht die grünliche Färbung des Asebogenins hinweisen würde.

Ein Teil des aus dem wässerigen Infuse durch Bleiacetat gefällten
Niederschlags wurde, in Wasser verteilt, mit H, S zerlegt und heiss filtrirt.
Das Filtrat wurde eingeengt und die braune Extraktmasse mit starkem Alkohol
ausgezogen. Der Auszug wurde mit Aether versetzt, wobei sich eine braune
harzige Substanz ausschied. Die Lösung. die braun gefärbt war, lieferte nach
der Destillation einen braunen Rückstand, der mit essigsaurem Blei einen
orangegelben Niederschlag, mit Eisenchlorid dunkle Grünblaufärbung gab und
deutlich sauer reagirte.

Ich extrahirte diese braune Substanz dann mit Wasser, schüttelte die
Lösung wiederholt mit Aether aus und behandelte den Rückstand nach dem
Abdestilliren des Aethers mit Wasser und kohlensanrem Calcium, concentrirte
das Filtrat und füllte mit starkem Alkohol. Es entstand ein weisser Nieder-
schlag, welcher abfiltrirt, ausgewaschen und in Wasser gelöst wurde. Die mit
Salzsäure versetzte Lösung wurde wieder mit Aether ausgeschüttelt, welcher
dann einen Rückstand zurückliess, welcher zum Teil krystallisirte und nach
Reinigung durch Umkrystallisation aus Wasser zum Teil farblose, scharfe Krystalle
bildete. Dieselben hatten einen kühlenden sänerlichen Geschmack und gaben
mit Fe, Cl, dunkel grünlich blaue Färbung ohne Bildung eines Niederschlags.
Diese blane Lösung wurde auf Zusatz von Alkali intensiv violettroth, Bleiacetat

gab starken weissen Niederschlag, in Essigsänre löslich. In sanrer erwärmter

(*) Das von A. Michael synthetisch erhaltene Methylarbutin aus Acetochlorhydrose und
dem Kaliumsalze des Methylhydrokinons wird (Berl, Ber. 14. 2099) als leicht löslıch in Wasser
und Alkohol beac hrieben.

Fiir die bei 120° entwiisserte (wobei sie '/, Mol. H,O verlor) Substanz fünd 1. Michael, der
Theorie entsprechend, im Mittel 64.8% C und 6.6% H., während ich für das lufttrockne Asebotin
57% © und 5.6% H. fand, Zahlen, welche zu viel von einander differiren, um eine Identität beider
Körper wahrscheinlich zu machen,

Weil das synthetisch erhaltene Methylarbutin mit Ve, Cl, nicht blau wird, ist A. Michael der
Ansicht, dass diese Arbutin-Reaction dem Hydrokinonglucoside eigen ist. Da das Asebotin
diese Blaufürbung nicht zeigt, würde eine Identität mit dem Arbutin (Ilydrokinon-Glucosid)
auch biedurch ausgeschlossen sein.

Lösung entfarbt sie schnell KMn ©, und CrO,, beide unter CO, Entwickelung.
Sie reagirte in wässeriger Lösung stark sauer und war in Wasser gut, in Aether
ziemlich, in Alkohol leicht, in Chloroform nicht löslich. Alkalische Kupferlösung
wurde nicht, ammoniakale Siberlösung stark reducirt. Der Schmelzpunkt
(etwa 160°—170°) war schwierig zu bestimmen, weil sich dabei Gasbläschen
entwickelten. Beim Erhitzen zwischen zwei Uhrgläsern bildete sich ein
krystallinisches Sublimat, das sich Eisenchlorid gegenüber ganz wie die Säure
verhielt. Ob diese Substanz als solche vorkommt oder durch die Einwirkung
der Salzsäure als Zersetzungsprodukt entstand, muss ist unentschieden lassen.
Ich erhielt sie in zu kleiner Qnantität. um sie noch weiter zu untersuchen :
vielleicht ist sie aus einem durch Erwärmen mit HCl spaltbaren, ätherischen
Derivat einer (protocatechnsäureähnlichen) Säure entstanden (*)

——>-094.00—

Die harzige, sowie die gelbe, mit Bleiacetat orangegelben Niederschlag
gebende Substanz wurden aus einer Portion getrockneter Blätter in grossen
Qnantitäten erhalten. Die grob gepulverten Blätter wurden mit Weingeist
percolirt, der Auszug durch Destillation so weit concentrirt, dass sich eine grüne,
fettige Substanz abschied, dann colirt und bis zur Entraktconsistenz eingeengt.
Das Extrakt wurde mit sehr starkem Alkohol behandelt und dann etwa das
doppelte Volum Aether zugesetzt; dadurch schied sich eine rothbraune, dicke,
sirupöse Masse ab, welche gesondert wurde (siehe unten); von der Lösung wurde
der Aether und Alkohol durch eine geeignete Destillation zurückgewonnen. Als
Rückstand blieb dann eine rothbraune, sirupöse Masse, woraus sich allmählich
eine grosse Menge eines grangelben Bodensatzes abschied. Dieser wurde von dem
mehr flüssigen Teil getrennt und liess nach der partiellen Ausziehung mit absolutem
und verdiinntem Weingeist eine bedentende Quantität eines eitronengelben Pulvers
zurück. Durch wiederholte Reinigung und partielle Lösung oder Umkrystal-
lisation mittelst verdiinnten Alkohols konnte ich diese Substanz in zwei Teile
von verschiedener Zusammensetzung trennen. twa 15 gr. eines schon gelben
Pulvers, aus verdünntem, kochendem Alkohol in feinen Nadeln krystallisirend,
gab, nach dem Trocknen bei 120°—125° ©. analysirt, folgende Zahlen:

I 04558 gr. lieferten 0.1801 gr. H,O und 0.9148 gr. CO,
Tr Geo. YY ga eee BE RT) oe
LIT 2039975: pr, ONTDBR Ze me OOO

トル 39
oder in Proceuten:
i Il Ill Mittel
C=12| 5473 | 54.6 | 54.7 54.7
H=1 | -4.38 4.28 4.35 4.34
0O=16 | 40.89 | 41.12 | 40.95 | 40.96 に

(*) Möglich auch ist es, dass Chinasiiure, welche in anderen Ericaceen gefunden wurde.
vorhanden ist. Ein vorläufiger Versuch, durch Oxydation des alkohol-ätherischen Extraktes mit
Mn, und verd. H,SO,, Chinon zu erkennen, hatte keinen Erfolg. Es entwickelte sich viel CO,
und auch Ameisensiiure, ein Chinongeruch war jedoch nicht deutlich bemerkbar.

Eine andere Quantität, den mehr unlöslichen Anteil in Alkohol bildend
und durch längere Ausziehung der ursprünglich gelben Substanz mit Alkohol
zurückbleibend, lieferte nach mehrmahliger Umkrystallisation aus verdünntem
Alkohol von obigen abweichende Zahlen. Drei Proben als Resultat von drei
gesonderten Reinigungsprocessen wurden analysirt. Sie wurden bei 120—125°
getrocknet, wobei sie resp. 9.02%, 10.—% und 9.1% Wasser verloren. Auf
Trockensubstanz berechnet und nach Abzug von sehr geringen Spuren Asche
gaben

I 0.4114 er, 0.1260 gr. H,O und 0.8989 gr. CO,
IE Dane ON 22er und ———.,,_,, (*)
TI DEN BEN ey OHNE es il TE by oe
oder in Procenten :

I I Ul Mittel C,H, 0,(t)
C=12| 596 ーー | 594 | 595 | 60 一
H=1ı| | |
| |

0=16 | 37.2

3.4 | 3.41
| 37.11 36.67

|
3.37 | 8.89 | 3.33
— | sys |

Sowohl durch die procentische Zusammensetzung als durch ihre Eigen-
schaften stimmt diese Substanz sehr nahe mit dem Qnercetin überein. Sie bildet
gelbe, kleine Krystallnadeln, welche in kaltem Wasser fast unlöslich, schwer
löslich in Aether und leicht löslich in heissem, verdünntem Alkohol sind. In
Alkalien löst sich das Asebo-quercetin mit intensiv gelber Farbe auf, und diese
Lösung wird von verdünnter Schwefelsäure stark gelatinös gefällt. Die mit
Wasser verdünnte weingeistige Lösung gab mit Eisenchlorid Grünblaufärbung,
mit Bleiacetat orangerothen Niederschlag. Letzteres Reagens schlägt es voll-
kommen aus seinen Lösungen nieder. Conc. H, SO, und HC] fürben es hoch-
orangegelb. Mit ammoniakaler Silberlösung, wie auch beim Erwärmen mit
alkalischer Kupferlösung tritt starke Reduction ein.

Wird die mit verdünnter Schwefelsäure angesäuerte Lösung in Alkali
mit Natriumamalgam versetzt, so tritt beim schwachen Erwärmen die Steinsche
Paracarthaminreaction ein. Die entstandene rothe Farbe ging durch Alkalien
in grün über, und auch die alkalische Lösung des Asebo-quercetins gab sofort mit
Natriumalgam Grün fürbung.

Die andere Portion gelben Pulvers, für welche als mittlere Zusammensetzung
aus drei gut übereinstimmenden Analysen gefunden war:

54.7 %C

4.34 % H und

40.96 % O,
enthielt, mit den Zahlen für Asebo-Quercetin verglichen, weniger Kohlenstoff
(e* 5%) und mehr Wasserstoff (c* 1%) als letzteres und war auch in ihren
Eigenschaften etwas verschieden. Die Lösung in Alkali wurde nämlich von

(*) Das CO, konnte in Folge eines Versehens bei der Wiigung nicht bestimmt werden.
(+) Die Liebermann'sche Formel für Quercetin

verdünnter Schwefelsäure nicht gefällt, die Paracarthaminreaction trat schon
ohne Erwärmung intensiv ein, alkalische Kupferlösung wurde beim Kochen
weniger schnell redueirt, und auch die Löslichkeit in Wasser war grösser.

Die erhaltenen Zahlen stimmen aber ziemlich gut mit den für Quereitrin
angegebenen Formeln.

を C %H %0
(1) Cy Hy O, | 559 | 486 | 39.75

(2) Cys Hy 0。 ; 56.7 4.3 39.0

|
(1B) Gee les Ong poet 4.8 41.8

|
Asebo-Quercitrin | 54.7 4,34 40.96

Nur von der Liebermann’schen Formel.(3) weichen sie etwas erheblich ab.
nl. um 1.3% C und etwa 0.5% H. Dies liess sich aber erklären durch einen Gehalt
an Quercetin, weil die vollkommene Trennung von Quercetin und Quercitrin nicht
leicht gelingt. Durch weitere Umkrystallisation konnte ich auch noch eine
kleine Quantität mit den für Quercetin beschriebenen Eigenschaften absondern.
Leider ging mir aber eine grosse Quantität bei der Reinigung durch Zerspringen
(les Becherglases verloren, so dass ich die gereinigte Substanz nicht in genügender
Menge erhielt, um neue Elementaranalysen auszuführen. Dieselbe zeigte folgende
Bigenschaften :

Schwach hellgelbe, kleine Nadeln, welche in heissem Wasser ziemlich
gut, in Alkalien leicht mit intensivgelber Farbe löslich waren. In heissem,
verdünnten Weingeist löste es sich ebenfalls leicht auf. Die wässerige, wie die
verdünnt alkoholische Lösung gab mit Bleiacetat starken, orangegelben
Niederschlag, mit Eisenchlorid Blaugrünfärbung, welche durch Alkalien nicht
in violett überging. Ammoniakale Silberlösung wurde stark, alkalische Silber-
lösung beim Kochen /angsam reducirt. Die alkalische Lösung gab mit verdünnter
Schwefelsäure angesäuert seinen Niederschlag, und diese Lösung zeigte ohne
Erwärmen sehr schön die Paracarthaminreaction.

Auch die rothbraune Masse, welche aus dem weingeistigen Extrakte durch
starken Alkohol und Aether gefällt war, habe ich noch etwas näher untersucht.
Die Masse löste sich in wenig Wasser zu einem klaren dünnen Sirup auf, gab
aber auf Zusatz von viel Wasser einen starken hellröthlichbraunen Nieder-
schlag. Dieser wurde abfiltrirt, durch Lösen in wenig Weingeist und Fällung
mit Aether oder auch mit Wasser wiederholt gereinigt. Zur Entfernung eines
Gehalts aun Aschebestandteilen löste ich sie in Weingeist, setzte etwas Salzsäure
zu und fällte mit Wasser.

Diese Substanz, welche ich Asebofuscin nennen werde, trocknet im

(1) Aus Dragendorff’s Quali- und quantitative Analyse von Pflanzen und Pflanzenteilen.

(2) Hlasiwetz und Pflaundler, Kawalier und Rochleder.

(3) Die neueste Formel von Liebermann und Hamburger. Die Spaltung des Glucosides
in Quercetin und Tsodnloit findet nach folgender Gleichung statt: CH。。0。。 + 3H,0 =20,H ,0,—+
CHi。O・

feuchten Zustande auf dem Wasserbade zu einem dunkelschwarzbrausen Harz
zusammen. Bei gewöhnlicher Temperatur: getrocknet, bildet es aber ein hell-
bräunliches Pulver, welches auf dem Wasserbade nicht schmilzt. Es ist in
Wasser nicht oder äusserst wenig, in Alkohol leicht mit dunkelbrauner Farbe
löslich, nicht löslich dagezen in Aether und Chloroform, wie in kochendem Benzol,
ein wenig löslich in kochendem Amylalkohol. In Alkalien löst es sich mit brauner
Farbe. Beim Erwärmen mit alkalischer Kupferlösung, sowie mit ammoniakaler
Silberlösung ist keine deutlich sichtbare Reduction neben der dunklen Färbung
der Flüssigkeit zu beobachten. Die stark mit Wasser verdünnte weingeistige
Lösung wird von Eisenchlorid dunkler und grünlich braun gefärbt und gefällt.
Bleiacetat bildet in dieser verdünnt-alkoholischen Lösung, ebenso wie ammonia-
kale Chlorcaleiumlösung, hellbraungelben Niederschlag. Goldchlorid giebt
dunkelpurpurne Fallung. Mit Natriumamalgam entsteht ebensowenig in
saurer Roth-, als in alkalischer Lösung Grünfärbung. Wurde in die alkoholische
Lösung unter Erhitzung Salzsäuregas geleitet, so firbte sie sich schön
intensiv rothbraun. Wasser fällte daraus einen violetten Niederschlag ; dieser
wurde nochmals in Weingeist gelöst und mit Wasser ausgefällt. Ihre Lösung
in Alkohol ist schön dunkelweinroth. Diese Substanz, welcher der Name
Asebopurpurin beigelegt sei, löst sich beim Uebergiessen mit Kalilauge darin mit
prachtvoll grüner Farbe auf, welche nach einiger Zeit in braun übergeht. Die
stark mit Wasser verdünnte weingeistige Lösung gab mit Bleiacetat grünlichen
Niederschlag und färbte sich mit Fe, Cl; schmutzigbraun. Getrocknet bildet es
ein dunkelviolettliches Pulver.

Sowohl von dem Asebofuscin als dem Asebopurpurin wurden noch Elemen-
taranalysen gemacht.

Asebofuscin
I 0.3956 gr. der bis 120° getrockneten Substanz lieferten
0.1756 gr. H,O und 0.8625 gr. CO,
IL 0.4085 gr. der bis 120° getrockneten Substanz lieferten
0.1831 gr. H,O und 0.8978 gr. CO,

oder in Procenten:

I II Mittel
C=12; 59.6 | 59.9 69.7

1: 1| 4.93 | Pl 4.94
|

O=16 | 35,57 | 35.16 35.36

Asebopurpurin
I 0.3862 gr. der bis 145° getrockneten Substanz gaben
0.1615 gr. H,O und 0.9082 gr. CO,

II 0.3303 gr. der bis 160° getrockneten Substanz gaben
0.1436 gr. H,O und 0.7696 gr. CO,

oder in Procenten :

I i Mittel

Sr) 64.1 | 63.8 1 63.96

H= 1, 464 | 4.83 | 4.78

0=16 | 31.26 | 31.37 | 31.32
Weiter habe ich die Untersuchung der Andromeda Japonica Thunb.
nicht fortgesetzt; doch dürften die erhaltenen Resultate schon einen vorläufigen
Einblick in die Art ihrer Bestandteile geben. Auch habe ich es unterlassen
müssen, zu verfolgen, in welcher Form der Stickstoff in dieser Pflanze vorkommt.
Eigentümlich ist es, dass das Asebo-purpurin in feuchtem Zustande eine Farbe
besitzt, welche auch bei der Einwirkung von Natrium-amalgam auf Asebo-
quercetin und- quereitrin auftritt, und dass in allen diesen Fällen diese Farbe
durch Alkali in grün übergeht. Das Asebo-fuscin zeigt diese Farbereactionen
nicht. Weil auch das Asebotoxin beim Kochen mit verdünnten Säuren dieselbe
rothe Farbe erzeugt, welche bei der Einwirkung von Natriumamalgam auf
Quercetin in sauren Flüssigkeiten auftritt, (ohne dass diese Farbe nach dem
Alkalischmachen in grün überging) scheint mir dieses auf einen einfachen
genetischen Zusammenhang der verschiedenen Asebo-bestandteile hinzuweisen.

II. SCOPOLIA JAPONICA MAX.

Diese (wegen ihrer capsula bilocularis operculo circumscissa) den Hy-
oscyameen angehörende Solanee ist schon häufig Gegenstand von Untersuchungen
gewesen, weil sie in ihrer Wirkung der Atropa Belladonna ganz ähnlich zu sein
scheint. Ihre Wurzel wird auch gewöhnlich mit dem Namen “Japanische Bel-
ladonna” bezeichnet und scheint unter diesem Namen auch schon im europäischen
Handel angetroffen zu sein. In Japan werden die daraus dargestellten Galenica,
besonders das Extrakt. an der Stelle der Belladounapräparate therapeutisch
verwerthet.

Nach Dr. Martin enthält sie Solanin, nach Dr. Langgaard(*) verdankt sie aber
ihre Wirkung zwei Alkaloiden, welche er folgendermassen erhielt. Das erstere
wurde aus dem alkoholischen und mit Bleiacetat behandelten Auszuge dargestellt
und durch Ausschüttelung mit Chloroform der schwefelsäurehaltigen wässerigen
Lösung entzogen ; dieses hinterliess beim Verdampfen einen teilweise in zarten
Nadeln krystallisirten Rückstand, welcher mit einer gelben schmierigen Materie
verunreinigt war. Mit Wasser ausgekocht wurde die Flüssigkeit nach dem
Erkalten filtrirt und das Filtrat vorsichtig auf dem Wasserbade bei niedriger
Temperatur eingeengt und mit Ammoniak versetzt. Von dem dabei entstan-
denen weissen Niederschlage wurde abfiltrirt, der Niederschlag auf dem Filter
mit wenig Wasser ansgewaschen und in Alkohol aufgenommen, welcher nach
seiner Verdunstung farblose Krystalle zurückliess. Es wurde mit dem Namen
Rotoin belegt, nach dem japanischen Namen der Pflanze,, Roto”. Es zeigte das
Verhalten eines Alkaloides, wurde aber nur in geringer Menge erhalten, weshalb
eine genauere Prüfung nicht vorgenommen wurde. Einige Tropfen der wässeri-
gen Lösung des schwefelsauren Salzes, in das Auge eines Kaninchens gebracht,
riefen Pupilleerweiterung hervor. Das zweite Alkaloid, Scopolein genannt,
wurde in grösserer Menge in der Wurzel aufgefunden und durch weitere Aus-
schüttelung mittelst Chloroform der alkalisch gemachten Flüssigkeit entzogen.
Es bildete eine gelbliche, harzige Masse, die sich in vollkommen trocknem
Zustande zu einem gelblich grauen Pulver zerreiben liess und grosse Neigung
zusammenzukleben zeigte. Weder das Alkaloid noch seine Salze konnten aber
krystallisirt erhalten werden. Auf den tierischen Organismus zeigte es dem
Atropin ähnliche Wirkung. Auch in einer Notiz von Holmes (Pharm. Journ.
and trans. Apr. 1880) über die botanische Herkunft eines Specimens der Wurzel
von Scopolia Japonica findet sich, dass Prof. Flückiger der Ansicht sei, dass
darin Atropin vorzukommen scheint.

In der japanischen und chinesischen Literatur findet sich in Kurzem Folgen-
des: Die Pflanze wächst überall auf Bergen und in Thälern und erreicht eine
Höhe von über 1 Fuss. Die Blüthen sind violett oder gelb. Die perennirende,

(*) Mittheilungen der Deutschen Gesellschaft fiir Natur- und Völkerkunde Ost- Asiens,
Yokohama Dee, 1878,

im Frühjahr neue Sprossen treibende, bitter und scharf schmeckende Wurzel, wie
auch die Samen sind allgemein als giftig erwähnt und werden bei den folgenden
Krankheiten empfohlen :

chronische Katarrhe der Luftwege,

Dysenterie und wässeriger Stuhlgang,

Caries der Zähne, Schlundanschwellung,

Verhärtung der Brustwarzen, schlecht eiternde Furunkel, Prolapsus ani,

Syphilis, Hundsbiss, Krätze, Wechselfieber, Verletzungen mit

spitzen Gegenständen u. s. w.(*)

Die meist üblichen Namen sind: Roto 芝 und Hashiridokoro » v y

F = m. Andere Namen, dem Honzokomokn keimo entnommen, sind:
Onishirigusa ヲ ニ シリ グ サ
Oomirugusa ヲ ホ ミル グ サ
Yamasa ヤマ サ
Nanazugikuya ナ \ ツ
Koka FE Ri
Sogowo Hm
Gatoshi FRT-

Soweit mir bekannt, sind keine weiteren Mitteilungen über die Bestand-
teile dieser Pflanze gemacht worden. Die obigen, einander widersprechenden
und immerhin noch sehr dürftigen Angaben liessen es, mit Rücksicht auf die
hiesige therapeutische Verwerthung wünschenswert erscheinen, eine nähere
Untersuchung vorzunehmen.

Etwa 10 K° Pulver der frischen, im August von Herrn Tamba (in Nagano-
ken) eingesammelten Wurzel der wildwachsenden Pflanze stand uns dazu zur
Verfügung. Der grössere Teil(f) wurde mit c*. 85 procentigem Weingeist
einige Male perkolirt und von dem Perkolate der Weingeist bis auf einen Rück-
stand von etwa 4 Liter abdestillirt. Es schied sich dabei ein fettes Oel ab, die
Flüssigkeit reagirte sauer, war hellbräunlich gefärbt und zeigte starke Alkaloid-
reactionen, wie auch stark reducirende Eigenschaften gegenüber ammoniakaler
Silber- und alkalischer Kupferlösung.

Weil das abgeschiedene Oel nicht unbedeutende Mengen fettsaures Alkaloid
enthalten konnte, behandelte ich zu deren Zersetzung die Flüssigkeit mit

(*) Im Sanseizuye Z4 A findet sich, dass die Samen beim Menschen ein Gefühl verur-
sachen, als ob man den Teufel vor Augen sehe, und dass sie bei Wahnsinn angewendet werden.

(+) Die frische Wurzel eines cultivirten Exemplars enthielt 62.6% Wasser und 1.4% Asche,
die Blätter 84.1% Wasser und 2.15% Asche; während letztere nur Spuren Alkaloid enthielten,
wurde in 10 Gr. der Wurzel soviel Alkaloid gefunden als mit 8.6 CM? '/,, N. Mayers Lösung
übereinstimmt. In dem analysirten Pulver der wildwachsenden Wurzel wurde gefunden:
0.61% in Petroleumäther (Spt. 48°), 2.2% in absolutem Aether und 0.59% in absolutem Alkohol
lösliche Substanz und 7.18% Asche. In dieser, viel Carbonat enthaltenden Asche wurde von
einigen pharmaceutischen Schülern als Mittel aus zwei Bestimmungen, gefunden:

13.8% CO,, 3.9% SiO,, 3.4% Cl, 5.1% SO,, 29.4% K.O, 5.2% Na,O und viel P,O,, CaO
und MgO,

Bleioxyd(*) und liess damit mehrere Tage unter haufigem Umschiitteln stehen.
Die Flüssigkeit wurde dann zur weiteren Verjagung des Weingeistes noch etwas
eingeengt und sodann mit Chloroform wiederholt ausgeschiittelt, bis nichts
Bedeutendes mehr aufgenommen wurde. Die Chloroformlösungen wurden dann
nach dem Concentriren zwei Male mit schwefelsäurehaltigem Wasser ausgeschüt-
telt, worin das Alkaloid fast völlig überging. Die Chloroformlösung lieferte
aber nach dem Abdestilliren eine ziemliche Menge Krystalle, welchen eine
braune Materie beigemischt war. Sie wurden durch Abwaschung mit wenig
Chloroform und weiter durch Umkrystallisation aus kochendem Wasser und
verdünntem Alkohol gereinigt (Scopoletin).

Die schwefelsäurehaltige Lösung des Alkaloides, welche aus saurer Lösung
beim Schütteln mit Chloroform nur Spuren Alkaloid daran abgab, wurde mit
Kalicarbonat alkalisch ‘gemacht und dann mit Chloroform ausgeschüttelt.
Nach dem Abdestilliren des Chloroforms wurde das rückständige Alkaloid in
verdiinnter Schwefelsäure gelöst, die Lösung mit Tierkohle behandelt und
sodann nach dem Alkalischmachen mit Kalicarbonat wieder ausgeschüttelt.

Im Einklang mit den von Dr. Langgaard gemachten Erfahrungen wurde
also 1° eine krystallinische Substanz, aus saurer Lösung, und 2° ein Alkaloid, aus
alkalischer Lösung in Chloroform übergehend, gefunden. Weil ich aber mit
grösseren Quantitäten arbeitete, konnte ich das letztere Alkaloid fast völlig farblos
und grossenteils krystallisirt erhalten. Die aus saurer Lösung übergehende
Substanz erwies sich aber als stickstofffrei und scheint mir deshalb das von Dr.
Langgaard erhaltene Alkaloid aus saurer Lösung ein Gemisch gewesen zu sein.
Die obenerwähnte stickstoflose Substanz besitzt folgende Eigenschaften.

Scopoletin. Farblose, feine Nadeln oder grössere prismatische Krystalle,
welche meistens etwas grau- oder gelblich erhalten werden. Sie sind sehr
schwer in kaltem, etwas besser in kochendem Wasser, schwer auch in Aether,
besser in Chloroform und reichlich in warmem Weingeist und in Essigsäure
löslich, unlöslich oder fast unlöslich dagegen in kochendem Schwefelkohlen-
stofl oder Benzol. Von diesen Lösungen fluoresciren die wässerige, stärker
noch die alkoholische schön blau. Die wässerige, mit Säuren versetzte Lösung
zeigt violetten Stich, die alkalische Lösung dagegen besitzt gelbe Farbe mit
starker blauer Fluorescenz. Das Scopoletin löst sich leicht in Alkalien auf; die
wässerige Lösung reagirt schwach sauer auf Lakmus. Die gelbe Lösung in
Aetzalkali wird mit Natriumamalgam röthlich braun, welche Farbe beim schüt-
teln mit Luft weinroth wird.

Durch Tierkohle wird es, sogar aus weingeistiger Lösung, fast vollständig
fixirt, und es lässt sich dann durch kochenden Alkohol derselben sehr schwierig
wieder entziehen. Die Lösung in Ammoniak gab mit absolutem Alkohol
krystallinischen Niederschlag, welcher leicht in Wasser mit gelber Farbe und

{=}
(*) Hierzu wurde gebraucht das Priicipitat, entstanden durch Versetzen einer warmen
Bleiacetatlösung mit frischer verdünnter Kalkmilch. Nach mehrmaliger Auswaschung durch
Decanthation wurde der kalkhaltige Niederschlag verwendet.

blauer Fluorescenz löslich war; in dieser Lösung gab AuCl, kobaltblauen,
AgNO, bliulichen Niederschlag, Fe,Cl, Grünfärbung und grünlich gelben
Niederschlag. Auf dem Wasserbade verdampft hinterliess diese Lösung anschei-
nend unverändertes Scopoletin.

Die ammoniakale Lösung giebt mit Silbernitrat erwärmt langsame, auf Zusatz
von etwas Kalilauge starke Silberausscheidung. Esreducirt ebenfalls alkalische
Kupferlösung beim Erwärmen. Die wässerige mit verdünnter Schwefelsäure
angesäuerte Lösung entfärbt KMnO, leicht unter CO, Entwickelung.

In Eisessig gelöst entfärbt es Brom und giebt damit ein in Nadeln krystal-
lisirendes Substitut, welches in Wasser, Chloroform und Aether schwer, in
Alkohol leicht löslich ist.

Für den Schmelzpunkt fand ich 198° (uncorr.) Bei höherer Temperatur
verflüchtigt es sich und giebt ein krystallinisches Sublimat, das mit Ammoniak
wieder stark blau fluorescirt; es verbrennt zuletzt, ohne Asche zu hinterlassen.
Bei 100°-110° getrocknet verlor es kein Krystallwasser. Stickstoff konnte
qualitativ nicht nachgewiesen werden.

Elementaranalysen

I 0.2356 gr. der bei 100°-110° getrockneten Substanz
lieferten 0.5290 gr. CO, und 0.0881 gr. H,O

II 0.4334 gr. der bei 100°-110° getrockneten, nur aus
Wasser umkrystallisirten gelblichen Substanz lieferten
0.9683 gr. CO, und 0.1624 gr. H,O

oder in Procenten : Berechnet auf:
I II Mittel C,H,O; C,,H,O, C,H,0O, C,.H,,0,
CSS! | 61.2 60.98 | 61.1 61.5 62.5 60.— 60.67
は ドー 3 | 4,2 4.16 4.18 4.27 4.16 4.44 3.93
0 16 34.6 34,86 34.72 34.23 33.34 35.56 35.4.

— 0 —

Das Alkaloid, welches durch Tierkohle und durch Ausschüttelung seiner
mittelst K,CO, alkalisch gemachten Lösung mit Chloroform etwas gereinigt war,
schied sich bei freiwilliger Verdunstung der sehr concentrirten chloroformösen
Lösung an den Rändern völlig weiss und in feinen Nadeln krystallisirt und
übrigens als eine halb krystallisirte, schwach gelbliche Masse ab, welche aber
allmählich mehr und mehr krystallisirte. Die Krystalle waren jedoch wegen ihrer
Feinheit nicht mechanisch von der beigemischten amorphen Substanz zu trennen.

Das Alkaloid war in verdünnten Säuren leicht und völlig löslich ; diese
Lösungen gaben mit allgemeinen Alkaloidreagentien Fällungen(*) und wurden
durch Alkalien amorph, farblos gefällt.

Das Alkaloid schmolz noch unter 100° zu einer klaren, farblosen oder
schwach rothlich gefärbten Flüssigkeit. Mit concentrirter Schwefelsäure erwärmt

(*) Grenze der Fällbarkeit durch Mayer’ s Lösung C* 1: 15000.

gab es auf Zusatz von Kalinmdichromat oder Ammonmolybdänat starke Ent-
wickelung von Spiräa- und Benzaldehydgeruch.
Von diesem noch unreinen Alkaloide wurde eine Kohlenstoff- und Wasser-
stoffbestimmung gemacht.
0.3580 gr. der während mehrerer Stunden bei 105-110° getrockneten
und geschmolzenen Substanz (wonach sie aber immer noch in
Gewicht abnahm durch Verflüchtigung kleiner Quantitäten Alka-
loides) gaben :
0.8635 gr. CO, und 0.2342 gr. H,O oder

66.—% C und 7. 3% H.

Eine andere Portion wurde nochmals durch Lésen in verdiinnter Schwe-
felsäure, Behandeln mit Tierkohle und Ausschiitteln der mit K, CO, alkalisch
gemachten Lésung mit Chloroform gereinigt. Die durch Destillation concen-
trirte Chloroformlösung war auch jetzt völlig farblos, hinterliess jedoch bei
weiterer freiwilliger Verdunstung wieder einen nur zum Teil krystallisirten und
völlig weissen Rückstand, welcher sich in ein nicht- zusammenklebendes Pulver
verwandeln liess.

II 0.3359 gr. des bei 110° getrockneten und geschmolzenen Alkaloides
lieferten :
0.8358 gr. CO, und 0.2370 gr. H,O
III 0.2887 gr. des bei 105°—110° getrockneten Alkaloides ergaben
13.7 Cm? feuchten Stickstoff bei 18° und 758””

also in Procenten :
Tropein der Tropasäure

C,; Hos NOs
C— 72 | 67.9 70.58
Bu 7.84 7.95
N=14 5.4 4.84
wis | 18.86 | 16.61

Wenn auch diese Zahlen noch nicht sehr genau mit den fiir C,, H,, NO,
berechneten übereinstimmen, was wohl wegen der Unreinheit zu erwarten war,
scheinen sie mir nicht in Widerspruch mit der Vermutung, dass in der Scopolia
ein krystallisirtes Tropein vorhanden ist.

In kleineren Quantitäten erhielt ich es in völlig weissen Warzen von kleinen
Krystallnadeln ohne sichtbare Beimischung einer gelben amorphen Substanz.
Ihre wässerige Lösung gab mit Goldchlorid eine gelbe ölige Fällung, welche
in der Wärme sich ganz oder zum Teil löst und beim Abkühlen sich wieder
goldglänzend krystallinisch ausscheidet; unter dem Mikroskope zeigte es
Blattchen, welche nicht sehr regelmässig begrenzt waren. Pt Cl, giebt einen
zusammenklebenden Niederschlag; in stark verdünnten Lösungen tritt aber
keine Fällung ein. Weil die Wahrscheinlichkeit vorliegt, dass das erhaltene
Alkaloid, wofür der Name Scopolein beibehalten sei, welchen Dr. Langgaard

für das von ihm erhaltene, amorphe, graue Pulver wählte, aus mehreren
Tropeinen (vielleicht Hyoscyamin und Hyoscin) besteht, habe ich keine weiteren
Versuche vorgenommen zur Trennung ; die erhaltene Quantität reichte dazu nicht
aus. Dass aber ein den natürlichen Alkeinen der Solaneen nahe verwandtes
Alkaloid in der Scopolia vorhanden ist, konnte ich noch durch einen Spaltungs-
versuch mittelst Barytwasser weiter darthun. Es wurde dabei nämlich eine
Säure erhalten, welche in ihren Eigenschaften, soweit sie untersucht wurde,
völlig mit den der Atropasäure übereinstimmten. Etwa ein halbes Gramm des
Scopoleins wurde mit der doppelten Menge Barythydrat und der dreissigfachen
Menge Wasser in einer zugeschmolzenen Röhre während einiger Tage im
Wasserbade erhitzt. Beim Oeffnen der Röhre gab sich ein Spiräa- Geruch
zu erkennen ; der Inhalt wurde filtrirt, wobei eine sehr kleine Menge eines harzigen
Körpers zurückblieb. Beim Ansäuern des liltrates mit Salzsäure trat der
obenerwähnte Geruch stärker auf und schieden sich nach kurzer Zeit Krystalle
aus, welche durch einmalige Umkrystallisation aus heissem Wasser völlig farblos
und von vorzüglicher Reinheit erhalten wurden. Sie besassen folgende Eigen-
schaften:

Platte, glänzende, farblose Nadeln, welche in kaltem Wasser schwer, in
heissem leicht, reichlich in Alkohol, Essigsäure, Aether, Chloroform und
Schwefelkohlenstoff, ziemlich in warmem Benzol löslich sind. Auch sind sie
leicht in Alkalien löslich. Die wässerige Lösung reagirt sauer, verbreitet beim
Erwärmen den Geruch der Atropasäure (nach Spiräa & Benzoesäure) und giebt
mit essigsaurem Blei weissen Niederschlag. Mit ammoniakaler Silberlösung
erwärmt findet keine, auf Zusatz von etwas Kalilauge aber starke Reduction
statt. Auch alkalische AuCl, Lösung wird beim Erwärmen reducirt. Für den
Schmelzpunkt fand ich 104° © (nncorr.) und 105.°3 (corrig). Weiter erhitzt
entsteht ein krystallisirendes Sublimat und bleibt zuletzt keine Spur Asche
zurück. Diese Eigenschaften stimmen mit denen der Atropasäure (Schmpt.
106°) überein.

In concentrirter Schwefelsäure löste es sich farblos, und diese Lösung nahm
auf Zusatz von etwas Salpetersäure intensiv wein-bis braunrothe Färbung an.(*)

Das Filtrat der durch Salzsäure gefällten Atropasäure lieferte nach dem
Alkalischmachen mit Ammoniak, Ausschütteln mit Aether und freiwilliger
Verdunstung der ätherischen Lösung einen amorphen bräunlich gefärbten
Rückstand, der einen etwas anilinähnlichen Gernch besass und im Exsiccator
geringe Krystallisation zeigte. Seine Lösung in verdünnter Salzsäure gab mit
PtCl,, AuC], und Mayer’s Lösung Trübungen. Die Quantität dieses wahrschein-
lich unreinen Tropins war zur weiteren Untersuchung nicht ausreichend.

(*) Bald nachher hatte ich Gelegenheit diese Reaction mit aus Europa bezogener Atropa-
säure anzustellen. Sie verhielt sich völlig der aus Scopolein erhaltenen Säure gleich. Eine
Probe Tropasäure zeigte sich in dieser Beziehung von der Atropasiiure verschieden, indem sie sich
in conc. Schwefelsäure schwach röthlich, fast farblos löste, welche Lösung aber auf Zusatz von
Salpetersäure vollkommen farblos blieb,

Die ursprüngliche mit Bleioxyd behandelte und mit Chloroform ausge-
schüttelte Flüssigkeit gab nach dem längeren’ Stehen einen fast weissen krystal-
lisirten Absatz in ziemlich grosser Menge. Dieser wurde gesondert, wonach die
Flüssigkeit beim weiteren Eiudampfen noch eine neue Quantität lieferte. Sie
wurde mit kaltem Wasser ausgewaschen und durch Umkrystallisation ans
verdünntem Alkohol oder Wasser gereinigt.

Diese Substanz, Scopolin. welche sich als das Glucosid des schon beschrie-
beden Scopoletins herausstellte, besitzt folgende Eigenschaften.

Scopolin. Weisse, nadelförmige Krystalle, welche ziemlich in kaltem, leicht in
warmen Wasser wie in Weingeist löslich sind. Sie sind unlöslich in Aether und
Chloroform. Die wässerige Lösung reagirt neutral und reducirt auch beim
Kochen alkalische Kupferlösung nicht. Nach dem Kochen mit verdünnten
Mineralsäuren dagegen findet starke Reduction statt. Mit ammoniakaler Silber-
lösung färbt sich die Flüssigkeit beim Kochen gelb und giebt allmählich
Silberausscheidung, welche auf Zusatz von Rali sogleich stark eintritt. In concen-
trirter Salpetersäure löst es sich mit gelber Farbe. Die gelbe Lösung in conc.
Schwefelsäure fluorescirt blau. Schmelzpunkt 218° (uncorr.).

I 0.4385 gr., lufttrocken, lieferten nach dem Trocknen bei 105° 0,4100
gr. und beim Verbrennen 0.7608 gr. CO, und 0.1990 gr. H,O
II 0.3393 gr. bei 120° getrocknet gaben 0.3174 gr. Trockensubstanz und
0.5878 gr. CO, und 0.1530 gr. H,O
oder in Procenten, auf getrocknete Substanz berechnet:

の Berechnet auf.
I II Mittel ro
C=12 | 50.6 | 50.5 | 60.65 | 51.6
H= 1 54 | 5,86 | 5.88 ha
shel ads || aa, | 44,08

Bei der ersten Analyse wurde 6,5%, bei der zweiten 6.9% H,O, auf
Trockensubstanz berechnet, gefunden. Die Formel C,, H,, O,,. 2 H,O verlangt
6.45%. Stickstoff konnte weder qualitativ noch quantitativ nachgewiesen werden.

0.3388 gr. der bei 105° getrokneten Substanz lieferten nach der Methode
Dumas analysirt nur eine geringe Menge Gas, welche ich nicht weiter genan
bestimmte, weil sie höchstens 0.2% betrug, also innerhalb der Fehlergrenzen der
Analyse liegt.

0.8068 gr. des lufttrocknen Scopolins lieferten nach dem längeren Kochen
mit verdünnter Schwefelsäure und Filtration der völlig erkalteten Flüssigkeit und
Auswaschen mit Wasser 0.3033 gr. = 37.6% eines in Nadeln krystallisirenden
Zersetzungsproduktes.

Das Filtrat wurde mit Baryumcarbonat behandelt und nach der Neutralisa-
tion filtrirt. Das Filtrat, zum Trocknen verdampft, lieferte einen weisslichen
syrupösen Rückstand. Weil er eine Barytverbindung enthielt, warscheinlich des
gelöst gebliebenen Zersetzungsproduktes, habe ich auf die quantitative Bestim-
mung verzichtet; doch konnte ich durch wiederholte Ausziehung mit starkem

Alkohol 0.2561 gr. einer barytfreien, klaren und schwach gelblich gefärbten amor-
phen Substanz erhalten, welche das Reductionsvermögen der Glucose zeigte. Die
0,2561 gr. (bei 110° getrocknet), wurden in 50 Cm? Wasser gelöst und mit
Fehling’ scher Kupferlösung titrirt. Für 10 Cm? derselben wurden im Mittel
9.8 Cm? verbraucht, während für reine Glucose 9.76 Cm? verlangt werden.

Ein andermal wurden aus 1.2 gr. des lufttrocknen Scopolins durch Kochen
mit verdünnter Salzsäure 0.4571 gr. = 38.1% des unlöslichen Zersetzungs-
produktes erhalten.

Von einer gesondert gereinigten Quantität Scopolin, welche mehrere Tage
unter dem ICssiccator gestanden hatte, wurden 0.4135 gr. nochmals zur H,O
Bestimmung gebraucht. Durch Trocknen bis 155° verlor es 0.0275 gr. = 6.65%
H,O.

1.3035 gr. dieses unter dem Exsiccator getrockneten Scopolins gaben mit 20
Cm? (1:10) verdünnter Schwefelsäure gekocht 0.4894 gr. des krystallinischen
Zersetzungsproduktes. Zur Entfernung des noch gelösten Anteils wurde das
Filtrat dann wiederholt mit Chloroform ausgeschüttelt, welches nach dem Ab-
destilliren noch 0.0154 Gr. krystallisirten Rückstand liess. Es waren daher im
Ganzen 0.5048 gr. = 38.73% des krystallisirten Zersetzungsproduktes erhalten.
Das von dem Chloroform getrennte Filtrat wurde fast völlig neutralisirt, er-
wärmt und nach Abkühlung bis 100 Cm? verdünnt. Von dieser Flüssigkeit
wurden im Mittel 6.62 Cm? für 10 Cm? Fehling’s Lösung verbraucht, woraus
sich eine Quantität von 0.772 gr. = 59.2% Glucose berechnen lässt. Ein andermal
erhielt ich aus 0.6498 gr. unter dem Exsiccator getrockneten Scopolins durch
Kochen mit (1:10) verdünnter Schwefelsäure 35.3% des krystallisirten Zerse-
tzungsproduktes (Schmpt. 197.°5). Das Filtrat mit Ba CO, behandelt gab, bei
110° getrocknet, 0.419 gr. = 64.5% Rückstand, wovon 0.071 gr. beim Behan-
deln mit Aether-Alkohol in Lösung gingen und der Rest in 50 Cm? Wasser gelöst
einen Reductionswerth von 0.313 gr. Glucose zeigte.

Das krystallinische Zersetzungsprodukt, welches beim Kochen des Scopolins
mit verdünnten Mineralsäuren als völlig farblose Nadeln sich ausscheidet, besitzt
nach dem Trocknen gewöhnlich eine grauliche Farbe und konnte als identisch
mit dem schon beschriebenen Scopoletin erkannt werden. Es zeigte den
Schmelzpunkt 197? 一 198? (uncorr.) und weiter dieselben Reactionen wie oben
beschrieben. Auch gab eine Elementaranalyse Zahlen, die gleichfalls völlig mit
den für Scopoletin gefundenen übereinstimmten.

0.3213 gr. bei 100°—105° getrocknet, lieferten

0.7233 gr. CO, und 0.1170 gr. H,O oder 61.41% C und 4.05% H.,
indem für das Scopoletin früher gefunden war im Mittel 61.1% C und 4.18 H
(Seite 19).

Die Spaltung des Scopolins durch Säuren liess sich vorläufig am besten durch
die Gleichung

Coy Hy 03.2 H,0 = 2 Cy Hy, 0。 + Cre Hi O5
ausdrücken. Dieselbe verlangt 39.4% Scopoletin und 60.6% Glucose. Das Scopolin

hat keine dilatirende Wirkung auf die Pupille, das Scopolein dagegen wirkt stark
mydriatisch. 0.100 gr. Scopolin, einem Hunde subcutan eingespritzt, verursachten
mehreren Stunden Schlafsucht, ohne den Tod herbeizuführen, während ein Tropfen
einer Lösung des Scopoleins (1: 5000) beim Kaninchen, stärker beim Menschen
Dilatation der Pupille zeigte.

Die sirupöse Flüssigkeit, welche von dem auskrystallisirten Scopolin
abfiltrirt war, wurde mit K,CO.-Liésung alkalisch gemacht und wieder mit
Chloroform ausgeschüttelt. Das Chloroform hinterliess beim Abdestilliren nur
eine unbedeutende Quantität Alkaloid, in Form eines braun gefärbten amorphen
Rückstandes. Die ausgeschüttelte Flüssigkeit lieferte nach dem Ansäuern,
Eindampfen und langem Stehen noch eine ziemliche Menge des Scopolins.
Die davon getrennte sirupöse braune Masse, welche anscheinend sehr grosse
Mengen Glucose enthielt und noch starke Alkaloidreactionen zeigte, habe ich
ebensowenig, wie den braunen, wahrscheinlich tropinhaltigen, alkaloidischen
Rückstand, weiter untersucht.

Das Vorkommen eines den Alkeinen der Tropasäure nahestehenden
Alkaloides, wie auch eines Glucosids Scopolin mit seinem Zersetzungsprodukte,
dem Scopoletin, macht die Scopolia Japonica Thunb. zu einer höchst interessanten
Pflanze und hoffe ich mir gelegentlich grössere Quantitäte der beschriebenen
Bestandteile verschaffen zu können, um weitere Versuche über ihre Constitution
anzustellen. Ob auch in den übrigen therapeutisch verwertheten Solaneen
dieses Glucosid vorkommt, dürfte wohl angezweifelt werden auf Grund der
Häufigkeit, womit diese Pflanzen Gegenstand von Forschungen waren. Ande-
rerseits scheint mir das Vorkommen einer fluorescirenden Substanz in der Atropa
Belladonna die Möglichkeit des Vorhandenseins eines solchen oder ähnlichen
Glucosides nicht unwahrscheinlich zu machen und beabsichtige ich die Radix
Belladonnae in dieser Richtung nochmals zu pruefen.

III. MACLEYA CORDATA. R. Br.”

Diese als Giftpflanze bekannte Papaveracee(t) (Subtrib. Bocconieae) wächst
hier fast überall auf Hügeln und Bergen an uncultivirten Stellen. Sie erreicht
eine Höhe von bis c*2 M., ihre Blätter sind his 30 Ctm lang, und beim Verletzen
des Stengels, der Wurzel, Blattnerven und Früchte fliesst ein orangegefärbter
Milchsaft aus. Blüthezeit Juli. Die meist üblichen Namen sind:

Takenigusa 2 7% = 7 4, Tsiampangiku # ~ & vv ¥ 7 und Tachiobaku
タチ オォ バク. Der chinesische Name ist Hakurakukuai Aye 3a)

Auszüge der frischen, zerkleinerten Wurzel, Blätter und jungen Früchte
(1:10), durch Maceration mit schwefelsäurehaltigem Wasser, und Mischen mit
dem dreifachem Volum Weingeist angefertigt, wurden nach Filtration und Ver-
dampfen des Filtrates, zuletzt unter Zusatz von etwas Wasser, mit 25 Mayer’s
Lösung titrirt und die entstandenen Niederschläge durch Maceration mit
Weingeist in ein lösliches und ein unlösliches (Sanguinarin-) Doppeljodid
getrennt. Es zeigte sich dabei, dass der Alkaloidgehalt der Wurzel und
Blätter (c* 0.5-1. proc.) fast gleichgross ist, wie in Chelidonium majus(@). Auf
Trockensubstanz berechnet, enthielten die Früchte am meisten, die Wurzel
viel weniger und die Blätter am wenigsten Sanguinarin, während der Gehalt an
Alkaloid des in Alkohol löslichen Doppeljodids für Wurzel und Blätter fast
derselbe, für die Früchte etwas geringer war. Ziemlich grosse Mengen
der Wurzel wurden zur Erhaltung der beiden Alkaloide in Arbeit genom-

(*) Diese Untersuchung, welche hier in Auszug aufgenommen sei, weil sie sich der des
Chelidoniums anschliesst, wurde schon vor einigen Jahren von mir an dem Tokio Shiyakujio
ausgeführt (Beitrag z. Kenntn. der Papaveraceen-Alkaloide. Yokohama 1881). Eine andere
Untersuchung, ebendaselbst angefangen, die des Illieium religiosum Sieb., hoffe ich bald wieder
aufnehmen zu können und in mehr eingehender Weise fortzusetzen,
(+) Herba sinica, perennis; radice repente, foliis alternis, petiolatis, subrotundis, basi
cordatis, obsolete lobatis, grosse et obtuse dentatis, venosis, membranaceis, subtus glauco-albidis;
paniculae terminales, elongatae, laxae ramis erectis, paucis, ramulisque unibracteatis。 Calyx
diphyllus, foliolis coloratis, ovatis caducis. Corolla nulla. Stamina 24-48, hypogyna; filamenta,
filiformia, antherae biloculares, elongato-lineares, lutere longitudinaliter dehiscentes. Ovarium
sessile, ovatum compressum. Ovula juxta placentas intervalvulares oppositas utrinque 2-8,
horizontalia, anatropa. Stigma subsessile, bilobum, lobis oblongis. Capsula subcompressa,
unilocularis, bivalvis, tetra-hexasperma, valvis, a basi solutis, deciduis, replo annulari seminifero
persistente. Semina horizontalia, ovata, testa crustacea, basi superne strophiolata. Embryo in
basi albuminis carnosi minimus; cotyledonibus nunc tribus aequalibus, nunc duabus vel quatuor
inaequalibus, germinatione petiolatis, ovato-subrotundis. Radicula umbilico proxima, centrifuga.—
(Enpiicner, Systema Plantarum, 4817.)
(£) Andere Namen sind:
Sosoyaki y y ¥ #,Gurogi =% nm 2, Kenkuagusa 4 22 9 y#, Tsukeishi ウッ ケイ zy,
Takato タカ トウ , Urajiro ウラ ジロ , Okamedaoshi ラー カメ ダラ シ , Kajigusa カジ
y+, Sasayakigusa ササ ヤキ ダサ.
Die PHanze wird hier zum Fiirben von Bambus verwendet; daher der Name
Take-ni-gusa (Bambus- koch- Kraut).
Der Name Kara-Kachiba, von Franchet und Savatier (Enumeratio Plantarum)
erwiihnt, ist irrtiimlich der niichsten Seite im Honzo-zufu entnommen (Fol. 24
Vol. 21). Derselbe gehört der Ricinus communis an.

(2) Dragendorff, Chem. Werthbestimmung einiger starkwirkenden Drogen p. 100.

men. Auszüge mit verdünnter Schwefelsäure und Weingeist angefertigt, wur-
den nach dem Abdestilliren des Weingeistes mit Ammoniak übersättigt und
der Niederschlag nach dem Trocknen mit Aether ausgezogen bis dieser nicht
mehr gelb gefärbt wurde. Salzsäure fällte aus der ätherischen Lösung einen
scharlachrothen Niederschlag, welcher das salzsaure Salz eines, nach dem
Reinigen chemisch und physiologisch mit Sanguinarin übereinstimmenden, Alka-
loides bildete. Der in Aether ungelöst gebliebene Anteil wurde mit Weingeist
ausgezogen, der Weingeist abdestillirt, der Rückstand mit Essigsäure behandelt
bis zur schwach sauren Reaction und dann mit soviel Wasser versetzt, bis keine
Ausscheidung mehr stattfand. Aus dieser Lösung wurde ein Alkaloid erhalten,
welches nach völliger Reinigung folgende Eigenschaften zeigte.

Macleyin. In Wasser und Alkalien fast nnloslich: kalter Weingeist löst
fast nichts, heisser etwas mehr: diese Lösungen reagiren alkalisch. In Aether
sehr wenig löslich, besser in frisch gefälltem Zustande; schüttelt man nämlich
mit Ammoniak versetzte Macleyinsalzlösungen mit Aether aus, so geht das
Alkaloid leicht darin über, scheidet sich aber nach kurzer Zeit grossenteils
wieder aus, in Form von Kugeln oder mehr oder weniger gut ausgebildeten, an
den Enden abgestumpften oder abgerundeten Prismen. Benzol löst das Alkaloid
in der Kälte sehr wenig, in der Siedehitze besser.

Chloroform löst ziemlich gut, namentlich in der Wärme. Aus der concentrirten
Lösung fällt Aether das Macleyin wieder aus, bei langsamer Ausscheidung in
Formen wie Figur.

Das Alkaloid hat keinen Geschmack, seine Salze schmecken bitter, nachher
scharf und kühlend. Für den Schmelzpunkt fand ich 201° (uncorr.).
Bei 100° verliert es kein Krystallwasser.; die annähernden Löslichkeits-
bestimmungen bei gewöhnlicher Temperatur ergaben für
90% Weingeist 1 : 900
officinellen Aether 1 : 1000
Chloroform 1:15
Elementaranalysen :
0.31075 gr. lieferten 0.15325 gr. H,O und 0.7749 gr. CO,
u ze,
0.3107 ',, A 0,1505 TR REOLTART MIR
0.4528 ,, $ 16.69 Cm’ feuchten N, bei 22° © und 764" (corr.)
0.4691 ,, r Ina" N 7 +) 9, 28°.0 C und 765 ,, (corr.)

”

oder in Procenten:
I II II Mittel Berechnet auf
C=12| 67.98 | 67.64 | 67.94 | 67.85 | C,, 67.98 | C。 67.80 | CG, 69.81

H=1| 54| 546 | 5.88 | 5.44 || H, 5.38 | H, 5.65 | Hy .5.20
N=14) 4.19 | a4 | 一 416 || N Sy | N。 3.98] N 3.8

Oi!) 一 — | ee TO. Br, Os. 90 er ee

0.1396 gr. Platin = 17.4% (Pt = 194.5)

Die Formel (Cy H, NO,. HCl), PtCl, verlangt 17.43% Pt.

Das Macleyin zeigt folgende Reactionen :

Mit reiner, frisch destillirter Schwefelsäure färbt sich ein Krystall gelb, als-
bald ziehen sich aber Streifen von dem Alkaloide ab, welche die Schwefelsäure
violett fürben, während das Krystall fast schwarz erscheint. Wendet man zu
Pulver zerriebenes Alkaloid an, so färbt es sich zwar momentan gelb, die
Schwefelsäure wird aber sofort prachtvoll dunkelviolett; nach einiger Zeit,
schneller bei sanfter Erwärmnng, z. B. auf dem Wasserbade, geht die Farbe in
grasgrün über, schliesslich verändert sie sich in bräunlichgelb und die Flüssigkeit
entfärbt sich zuletzt fast völlig. In Berührung mit ZNO,—Dämpfen entwickelt
sich eine prachtvolle ultramarinblaue Farbe. Bewegt man einige Körnchen des
Alkaloides durch HNO, haltige Schwefelsäure, so entsteht eine prachtvoll blaue
oder bei etwas grösserem HNO, Gehalt dunkelgrüne Färbung, die bald in braun
übergeht.

HNO, von 1.2 spec. Gew. löst es ohne Färbung; mit gelber Farbe wird es
von HNO, von 1.34 spec. Gew gelöst.

Von eisenoxydhaltiger H,SO, wird es wie von reiner Schwefelsäure gelöst ;
nur ist die Nuance mehr blauviolett. Bei Erwärmung verwandelt sich die
Farbe ebenfalls in grün.

Die Lösung in Schwefelsäure wird auf Zusatz von Kalidichromat intensiver
blau gefärbt.

Molybdänsäurehaltige H,SO, giebt dunkelpurpurne, blaue und grüne
Färbung.

Eisenchlorid färbt nicht, auch wird ein Gemisch von Fe, C], und K, Fe, Cy,,
nicht redueirt. Mit den ungefärbten Säuren liefert das Macleyin farblose Salze,
welche sich alle, soweit sie untersucht wurden, in Krystallen darstellen liessen ;
die in Wasser leicht löslichen scheiden sich beim Verdampfen ihrer Lösungen
in der Wärme oft amorph aus, die schwer löslichen werden entweder amorph
oder in Form von Kugeln, Warzen, kleinen Prismen oder Nadeln gefällt. Beim
Eintrocknen schrumpfen manche, wie dies auch mit dem durch Alkalien gefällten
Alkaloid der Fall ist, zu einem bedeutend kleineren Volum zusammen. Die
Salze gelatiniren nicht. Ihre Lösungen werden allmählich, schneller beim Er-
wärmen, gelblich gefärbt und geben mit allgemeinen Alkaloidreagentien, auch
mit HgCl,, K, Cr, O,, KJ, CN.SK, Fällungen.

Hydrochlorat. Schiesst aus seiner wässerigen, neutral reagirenden Lösung
in 4-6 seitigen Prismen an. Löslichkeit in Wasser etwa 1: 140.

Hydrojodat. Jodkalinmlösung giebt in Macleyinsalzlösungen einen käsigen,
weissen Niederschlag, der nach einiger Zeit zusammenschrumpft. Unter dem
Mikroskope zeigt es sich im Form von Kugeln : es löst sich seh rwenig in kaltem
besser in heissem Wasser und Alkohol. Aus letzterer Lösung konnte das Salz in
kleinen Krystallen erhalten werden.

Sulphat (neutrales). Farblose Nadeln, in Wasser und Alkohol löslich,
neutral reagirend. Wird die wässerige Lösung bei 100° verdunstet, so trocknet
sie hornartig ein und löst sich dann sehr leicht in Wasser zu einer übersättigten
Lösung, welche auf Zusatz von etwas verdünnter Schwefelsäure weisse krystal-
linische Fällung giebt. Dasselbe wurde mit dem Hydrochlorat beobachtet.

Chromat. Kalidichromat erzeugt in Macleyinlösungen einen orangegelben
Niederschlag, welcher in heissem Wasser gut löslich ist und in Prismen erhalten
werden kann.

0.1914 gr. dieses bei 100° getrockneten Salzes lieferten 0.031 gr. Cr, 0。.
Die Formel (C,, H,, NO, H, Cr, O, verlangt 0.0316. gr.

Das Salz löst sich mit prachtvoll blauer Farbe in concentrirter Schwefel-
säure auf.

Platindoppelchlorid. Platinchlorid fällt Macleyinlésungen weisslichgelb.
Der Niederschlag löst sich beim Kochen ziemlich leicht auf. Beim Austrocknen
an der Luft schrumpft das Salz stark zusammen und löst sich dann sehr
wenig in Wasser.

0.8313 gr. des lufttrocknen Salzes verloren bei 100° 0.0255 gr. H,O = 3.07%

0.802 gr. dieses getrockneten Salzes lieferten 0.1396 gr. Platin = 17.4%.

0.2926 gr. des lufttrocknen Salzes verloren bei 110° getrocknet 0.0095
gr. = 3.24%

Die Formel (C,, H, NO,. HCl)? PtCl, + 2 H,O verlangt 17.43% Pt (auf
wasserfreies Salz berechnet) und 3.14% H,O

Quecksilberdoppeljodid. Jodkaliumquecksilberjodid giebt in Macleyin-
salzlésungen einen amorphen, käsigen, weissen Niederschlag, welcher beim Kochen
zu einer geschmolzenem Schwefel ähnlichen Masse schmilzt und in Weingeist gut
löslich ist.

Thiocyanat. Rhodankalium erzeugt in Macleyinsalzlésungen eine weisse
Fällnng, die beim Erwärmen leicht verschwindet und bei Abkühlung wieder
in Blättchen oder sternförmig gruppirten Nadeln ausscheidet.
gruppirten Nadeln und

Acetat. Krystallisirt in bündel- oder sternförmig
ist in Wasser und Alkohol gut löslich.

Saures Oxalat. Wird Macleyin mit Wasser und überschüssiger Oxalsäure
erwärmt, so scheidet nach Abkühlung das saure Salz in warzenförmig gruppirten
feinen Nadeln ans. Es löst sich schwer in kaltem, gut in heissen Wasser.

Saures Tartrat. Sternförmig gruppirte Nadeln, in kalten Wasser schwer

löslich.

Pikrat. Pikrinsaures Kali erzeugt in den Alkaloidsalzlésungen einen
gelben Niederschlag. Das Pikrat ist in kochendem Wasser sehr schwer löslich.
Mit conc. H, SO, färbt es sich prachtvoll blau.

Benzoat. Lange, weisse, seidenglänzende Nadeln, welche sehr schwer in
kaltem, etwas besser in heissem Wasser und gut in warmem Weingeist löslich
sind. Es schmilzt bei 166° (uncorr.) zu einer röthlichen Flüssigkeit.

Beim Vergleich des Macleyins mit den besser bekannten Papaveraceen-
Alkaloiden stimmen seine Eigenschaften mit dem von Hesse entdeckten Opium-
alkaloide Protopin auffallend überein, nl. in der characteristischen Kugel- und
Warzenform der Aetherausscheidungen, der Löslichkeit in verschiedenen
Menstrua, der Formel, der Zusammensetzung des Platindoppelsalzes u. s. w.,
Der Schmelzpunkt 201° (uncorr.), 205° (corr.) weicht nur wenig-von dem von
Hesse(*) für Protopin gefundenen 202° ab. Auch mit den Alkaloiden der übrigen
Papaveraceen, welche farblose Salze bilden, Chelidonin, Glaucopikrin, Sanguinaria-
porphyroxin u. s. w. stimmt das Macleyin in der violetten und grünen Farbe,
welche sie mit H,SO, erzeugen, überein. Das Chelidonin verhält sich weiter dem
Macleyin, und auch Protopin, Kryptopin, Lanthopin in der schwierigen Löslich-
keit in Alkohol und Aether, sowie des Hydrochlorats in Wasser, ähnlich, weicht
aber wesentlich von diesen Basen durch seinen niedrigeren Schmelzpunkt (130°
Will) und durch seine Zusammensetzung C,, H,, N, O, ab. Dieser hohe Stick-
stoffgehalt des Chelidonins scheint aber zweifelhaft, weil in allen anderen
analysirten Papaveraceen-Alkaloiden auf 1 Mol. (1 Aeq) nur ein Stickstoffatom
vorkommt, und ich hielt es daher nicht für überflüssig, eine weitere Prüfung
vorzunehmen.

(*) Beitr. z. Kenntniss d. Opiumbasen. Arch, d. Pharm. Jahrg. 61 Bd. CC. p. 117., Ann.
d. Chem. u. Pharm. VIII Supplementband. 3. Heft (Jan. 1872).

IV. CHELIDONIUM MAJUS.

In dem vorigen Aufsatze über Macleya cordata sprach ich die Vermuthung
aus, dass die bis jetzt für das Chelidonin adoptirte Formel C,, H,, N, O, mit
Hinsicht auf den Stickstoffgehalt nicht richtig sei, mich dabei stützend auf die
Analogie mit den übrigen besser untersuchten Papaveraceen- Alkaloiden.
Letztere enthalten alle auf ein Aequivalent der Base nur ein Stickstoffatom,
während das Chelidonin, welches dem Macleyin, Protopin u. s. w. nahe verwandt
scheint, drei Stickstoffatome enthalten soll. Will fand nämlich als Resultate
von Elementaranalysen(*)

C = 67.4 — 68.1%

Hl .6 7

N = 12.2%
und fiir den Platingehalt des Platindoppelsalzes 17.4-17.6%. Aus diesen Zahlen
berechnete er die Formel C,, H,, N, 0。, wofür später, mehr in Uebereinstimmung
mit dem Gesetze der paaren Atomzahlen, andere vorgeschlagen wurden, von
welchen die Formel C, H,, N, 0。 jetzt allgemein in chemische Werke
übergegangen ist.

Elementaranalysen, welche sowohl mit hier dagestelltem Chelidonin, wie
mit aus Deutschland bezogenem angestellt wurden, bestätigten diese Vermuthung
und zeigten, dass auch hier auf ein Aequivalent der Base nur ein Stickstoffatom
kommt.

Aus hier gesammelter und grob gepulverter Wurzel des Chelidonium
majus(t), wurde von Herrn Tamba das Chelidonin durch Auskochen mit
H,SO,- haltigem Wasser und Fällung der Auszüge mit Ammoniak dargestellt.
Der abgewaschene und getrocknete Niederschlag wurde mit schwefelsäure-
haltigem verdiinntem Weingeist ausgekocht und das Filtrat nach dem Verdünnen
mit Wasser durch Natronlauge gefällt. Der Niederschlag wurde getrocknet,
mit Aether zur Entfernung von Sanguinarin ausgewaschen und das ungelöst
gebliebene Alkaloid durch Lösen in H,SO, haltigem Weingeist und Freistellung
durch Natronlauge gereinigt; durch freiwillige Verdunstung der weingeistigen
Lösung wurden zuletzt etwa 2.5— 3 gr. ziemlich grosse und gut gebildete Krystalle
erhalten.

In concentrirter Schwefelsäure löste es sich anfangs farblos, bald bildeten
sich schwach bräunliche Streifen, welche nachher dunkelviolett wurden; bei

(*) Gmelin Handb. der organ. Chemie. IV. Bd. p. 1684. 一 C,, H, N, O, (Gerhardt)
Cy Hy N, 0。 (Gmelin). C, H,, N, 0。 Limpricht (Lehrb. 1197).

(t) Chin: Hakutsusai fı ff 2, oder Kusa no Oo = König der Kräuter > TE. Die
Blätter sollen nach Yamatohonzo, üusserlich in feingeriebenem Zustande applicirt, bei Syphilis
und Geschwulsten wirksam sein. Die Pflanze kommt hier wiewohl überall doch nur sparsam
vor und konnte deshalb nur eine kleine Quantität der Wurzel auf Chelidonin verarbeitet werden.—

Anwendung von etwas weniger concentrirter H,SO, schien die Farbung etwas
schöner einzutreten. Conc. H,SO, mit einer Spur HNO, färbt das Alkaloid
grün. In starker HNO, löst es sich mit gelber Farbe.

Als Schmelzpunkt fand ich 135°-136° C (uncorr.). Bei Temperaturen nahe
am Schmelzpunkt gelegen färbte das Chelidonin sich röthlich braun, beim
stärkeren Erhitzen purpurroth unter Bildung eines krystallinischen Sublimats
und Entwickelung von Dämpfen, welche Methylamin- und Benzoesiuregeruch
verbreiteten.

Bei 100° während mehrerer Stunden getrocknet, verlor. es nur Spuren
Wasser. Nach Will verliert das mit einer Molekel Wasser krystallisirte Chelidonin
schon bei 100° ©. 4.65—5.13% H,O.

Elementaranalysen.

I: 0.3762 gr. lieferten 13.5 Cm? feuchten N, bei 27° und 746 ™™.
II 0.4701 ,, ee AO es s S| get) Und aoe
III 0.3198 , bei 100°-110° getrocknet gaben 0.3158 gr. und
0.7470 gr. CO,, 0.1604 sr. H,O und 0.0014 gr. Asche.

IV 0.3376 ,, bei 100°-105° getrocknet gaben 0.3328 gr. und
0.7846 gr. CO,, 0.1514 gr. H,O und 0.0011 gr. Asche.

0.3896 gr. des Platindoppelsalzes, in lufttrocknem Zustande, ergaben bei
mehrstündigem Trocknen bei 100°-105° 0.0281 gr. und weiter bei 110° im
Ganzen 0.0310 gr. Gewichtsverlust und beim Glühen 0.1501 gr. Platin. Hieraus
berechnet sich für das Platindoppelsalz 3.6% H,O und 17.48% Pt. auf
getrocknetes Salz. Dieser Platingehalt stimmt völlig mit dem von Will
gefundenen 17.42-17.6% überein, weshalb ich auch später keine weiteren Be-
stimmungen ausgeführt habe. Weil die Analysen III und IV eine Differenz
von etwa 0.6% H. ergaben, liess sich keine Formel mit Sicherheit aufstellen,
wiewohl aus den Stickstoffbestimmungen schon hervorging, dass das Alkaloid
nicht 12%, sondern nur etwa 4% Stickstoff enthielt.

Nach längerer Zeit gelang es mir, eine Quantität Chelidonin (A), c* 10 Gr.,
aus Deutschland zu beziehen und haben wir damit noch weitere Analysen
angestellt. Um zu erfahren, ob dieses, ein weisses Pulver bildendes, Alkaloid
ein einheitlicher Körper sei, wurde es in mehrere Portionen getrennt und von
jeder Analysen ausgeführt.

Es wurde dazu eine Portion mit Alkohol gekocht, wobei der grössere Teil
zurückblieb (B). Mit der resultirenden alkoholischen Lösung wurde eine neue
Portion des Alkaloides unter Zusatz von verdünnter Salzsäure und etwas Wasser
gekocht. Hiedurch gingen grössere Quantitäten in Lösung und wurden daraus
nach Verdünnung mit Wasser durch Natronlauge gefällt. Es bildete sich ein
amorpher zusammenballender Niederschlag, der mit Aether ausgeschüttelt sich
vollständig darin löste. Die abgetrennte ätherische Lösung wurde nach der
Concentration durch Destilliren der freiwilligen Verdunstung überlassen und die
sich abscheidenden grösseren Krystalle gesammelt (C).

Der beim Erwärmen mit verdünnter alkoholischer Salzsäure ungelöst

zurückgebliebene Anteil wurde durch Behandlung mit Alkohol und Essigsäure
zum Teil in Lösung gebracht und daraus nach dem Verdünnen mit Wasser
mit Kalilauge gefällt. Das gefällte Alkaloid wurde aus Chloroform umkrystal-
lisirt und zur Entfernung von etwas gelbem Zersetzungs (Oxydations?) produkt
durch Digestion mit Alkohol gereinigt (D).

Was endlich nach der Behandlung mit Alkohol und Essigsäure zurück-
geblieben war und sich wie die übrigen Anteile qualitativ verhielt, wurde mit
verdünnter Natronlauge macerirt, längere Zeit mit Wasser abgewaschen und zur
Wasserbestimmung benutzt.

1.2150 gr. dieses lufttrocknen Alkaloides verloren nach einem Tage im
Exsiccator 0.0002 gr., nach weiteren 2 Tagen fand keine weitere Gewichts-
abnahme mehr statt, dann bei c* 135° während mehrerer Stunden erhitzt und
geschmolzen betrug die Gewichtsabnahme 0.0597 gr. und nach dem weiteren
Erhitzen bis 140 a 145° im Ganzen 0.0614 gr., also 5.05% auf wasserhaltige,
oder 5.23% auf getrocknete Substanz.

Elementar-Analysen. Einzelne wurden mit der entwässerten, andere mit
der zum Teil entwässerten, die übrigen mit der lufttrocknen Substanz ausgeführt.

A. urspriingliches Chelidonin,

V 0.2635 gr. verloren beim längeren Trocknen bei 100° 0.0009 gr.,
beim weiteren Trocknen bei 120°-130° noch 0.0118 gr. und
lieferten: 0.6122 gr. CO, und 0.1240 gr. H,O.

B. nach teilweiser Lösung in Weingeist

VI 0.3455 gr. bei 100° während 2 Stunden getrocknet verloren
0.0006 gr. und gaben 0.8062 gr. CO, und 0.1830 gr. H,O.

VII 0.3924 gr. bei 130° während etwa 1 Stunde getrocknet verloren
unter rothbräunlicher Färbung 0.0117 gr. und lieferten 0.9064 gr.
CO, und 0.1931 gr. H,O.

VIII 0.3500 gr. bei 130° während mehrerer Stunden erhitzt gaben
unter Bräunung 0.0168 gr. = 4.8% Verlust. Von dieser ge-
trockneten Substanz lieferten

0.2487 gr. nach dem Verbrennen
0.5980 gr. CO, und 0.1296 gr. H,O.

IX 0.2619 gr. ungetrocknet gaben 10.4 Cm? feuchten N, bei 23° und
758".

X 0.3713 gr. während 2 Stunden bei c* 125° getrocknet verloren
4.3% an Gewicht und lieferten 14.3 Cm? feuchten N, bei 25° und
7

C. aus Aether krystallisirt (siehe oben)

XI 0.3130 gr. bei 100° während 3 Stunden getrocknet verloren an

Gewicht 0.0016 gr. und ergaben 0.7285 gr. CO, und
0.1560 gr. H,O.

XII 0.2491 gr. bei 120°-130° verloren 0.0098 gr. und gaben 8.55 Cm?
feuchten N, bei 19° und 761.6™".
D. aus Chloroform krystallisirt (siehe oben)
XIII 0.3275 gr. ergaben 12.4 Cm? feuchten N, bei 22° und 758.3".
1.4534 gr. gaben während 2 Stunden bei c* 100° erwärmt
0.0100 gr., bei 135° geschmolzen noch 0.0496 und bei
160° im Ganzen 0.0719 gr. = 4.96% Gewichtsverlust.
Von diesem getrocknetem Alkaloide lieferten
XIV 0.3901 gr. beim Verbreunen 0.1894 gr. H,O(*)
Berechnen wir die Resnltate der Analysen auf lufttrocknes Alkaloid, indem
wir den Gewichtsverlust beim Trocknen als Wasser in Rechnung bringen, so
finden wir in Procenten :

| Ausgr. C. H. N.
1 Hier dargestelltes Chelidonin | 0.8747 rec wees 3,88
2 FR + 0.3680 Be Pe 義 > 3.91
Begs 52 0.3184 63.98 5.73
er 7 7 | 0,3367 63.59 5.16
5 A Chelidonin (Deutschland) 0.2635 63.4 5.76
6 B +i 2 | 0.8465 63.64 5.91
eee ” 9 0.3924 63.— 5.62
8, „ + 0.2612 62.44 6.04
eee ij | 809 | Mae | m+ 4.43
DD an Ar > DIET, oxincet me Aa deen er | 4.27
un . Fe | 0.8180 63.5 5.60 |
ck} ‘ r | 0.2491 m RE
Typ 5 5 | 0.8275 5 Di 4.24
un; 2 5 0.4094 a 5.66 |
Im Mittel ergiebt sich dann :
Gefunden Berechnet auf
C= 12 68.86 | Cy 68.84 | Cy 68.86 | Cy 64.7
H=1 5.68 | H, 6.82/H, 6.62) H,, 5.66
N= 14 4.09 N 3.92 N, 3.92 | N 3.77
O = 16 26.87 0。 26.92 0,。 26.60 0。 25.87

Aus der Uebereinstimmung der Analysen zeigte sich, dass das analysirte
Chelidonin ein einheitlicher Körper war und der Gewichtsverlust beim
Trocknen als Wasser betrachtet werden kann. Als Schmelzpunkt fand ich immer
135°—136° (uncorr.). Das bis 160° entwässerte Alkaloid wurde nach einigen

(*) Das CO, konnte nicht bestimmt werden, weil die Natronkalkröhre durch die starke
Ausdehnung ihres Inhalts zersprang.

Tagen auf seinen Schmelzpunkt untersucht, es fing an bei c* 110° zu
erweichen, war aber erst bei 132°—135° zu einer durchscheinenden Fliissigkeit
zusammengeschmolzen. Das Wasser wird nur sehr schwierig vollständig aus-
getrieben, sogar bei Temperaturen über dem Schmelzpunkt. Bei den angeführten
Elementaranalysen war der Gewichtsverlust bei 100°—105° ungefähr 0.2—0.7%,
bei e* 125° 3—4.4%. Bei völliger Entwässerung wurden gefunden 5.05 und
4.96%, während für die Formel C,,H,,NO;. H,O 5.04%, für C,,H,,NO,. H,O.
4.85% Wasser verlangt werden.
Für das Platindoppelsalz : Pt 王 194.5| Pt—19

I (C,H, NO,. HCl)? Pt Cl, wird verlangt 17.3 % | 17.5 %

1. ( Cpe NO; BCS PO 3 17.9 % | 18.1 %

BI (C,H, NO HCl yy PtCl, ~~ > + 17.43% | 17.62%

während von Will 17.4—17.6, von mir 17.48%, gefunden wurde. Die zweite
Formel wiirde hiernach nichtin Betracht kommen. Leider besass ich nicht
weitere Substanz noch auch Verbrennungsröhren genug, um Analysen des
Platindoppelsalzes auszuführen.

V. NANDINA DOMESTICA THUNB.” |

Diese Berberidee ist in China(2) wie in Japan einheimisch und wird als
Zierpflanze häufig cultivirt.(3) Sie erreicht eine Höhe von bis etwa 2 Meter.
Den Blättern werden, wiewohl nicht giftige, doch emetische Eigenschaften zuge-
schrieben, und alle Teile dieser Pflanze finden in der Medicin Verwendung.
Blätter wie Wurzelrinde werden in Form eines wässerigen Extraktes gebraucht.
Sie sollen die Muskulatur stärken, geistige Thätigkeit erhöhen, heitere Gemüts-
stimmung herbeiführen, Schlaflosigkeit verursachen. und das Leben verlängern,
weiter rheumatische Affectionen heilen, sogar Gicht, auch Diarrhöe und Sper-
matorrhöe, die Gesichtsfarbe verschönern und Haare, selbst graugewordene,
verstärken und schwärzen.

Auf Durchschnitt zeigt die Wurzel gelbe Farbe, von welcher ich vermutete,
dass sie mit einem Berberingehalt zusammenhänge. Doch auch die weissen Teile
der Wurzelrinde wie die Blätter schmecken ziemlich bitter und schien wahrscheinliga
noch ein zweites Alkaloid vorhanden, wie auch in auderen berberinhaltigen
Pflanzen farblose Alkaloidez. B. Oxyacanthin, Hydrastin aufgefunden sind. Eine
vorläufige Untersuchung bestätigte dies und gelang es leicht, ein fast farbloses in
Aether und Benzol lösliches Alkaloid abzuscheiden. Zur Erhaltung von etwas
grösseren Quantitäten wurden mehrere Kilo’s der Wurzelrinde(4) von Herrn
Shimoyama in Arbeit genommen. Aus einer Portion wurde ein wässeriges Ex-
trakt dargestellt (die Auszüge färbten sich beim Eindampfen schwarzbraun)
und dieses wiederholt mit Weingeist ausgekocht. Nach dem Abdestilliren des
Weingeistes wurde der Rückstand mit Wasser und Ammoniak versetzt, wobei
sich ein harziger Niederschlag ausschied ; dieser wurde mit Aether wiederholt
ausgeschüttelt. Die abgetrennten ätherischen Lösungen lieferten nach der
Destillation einen braunen amorphen Rückstand, welcher mit essigsäurehaltigem
Wasser ausgezogen wurde. Nach Fällung mit Ammoniak wurde das immer
noch gefärbte Alkaloid gereinigt durch Lösung in der erforderlichen Menge
verdünnter Essigsäure, Zusatz von etwas Bleiacetat und Durchleiten von H,;S. ;

(1). Jap. Nanten, 南天 der meist übliche Name.
Andere Namen sind:
Natsuten ナッ テン と , Nansoku Hg, Nasjoku 南 煽
Nansoboku 南 草木 , Rantensiku 南天 竹
Nansjokuso #a{§##, Nantensjoku 南天 煽
Nantensjokusidsu 7#jA, Sannio =#, Uhanio Bett
Koku Uhansjo #ékff, Tensiku 天 竹 , Tensikusishi 天 符 枝 子
Uso 鳥 草 , Taishun KF, Yotoso 揚 桐 草

(2). Aufdem Berge Kosan #1, auch in der Provinz Toshiwa.

(3) Wegen der schönen Verteilung der Blätter, der rothen Farbe, welche sie nach der Bliithe
annehmen, und des reichen, schönen, rothen Fruchtstands. Sie kommt auch mit weissen und (
hellvioletten Früchten vor. Sie wird benutzt als Zusatz zu Geschenken, aus Esswaren bestehend,
und hat dann die Bedeutung, dass man sich nicht vor der Giftigkeit der Speisen zu fürchten hat,
weil, falls dieselben auch giftig wären, die Nandina die giftigen Eigeuschaften wegnehmen würde.

(4) Sie enthielt 37.65% H,O, 1.17% Stickstoff und 3,37% Asche.

die farblose Lösung wurde dann partiell mit Ammoniak gefällt. Eine andere
Portion der Wurzelrinde wurde in getrocknetem und grobgepulvertem Zustande
mit Weingeist ausgezogen und nach dem Abdestilliren des Weingeistes dem
Riickstande das Alkaloid durch Zusatz von Ammoniak und Ausschüttelung mit
Aether entzogen. Der Aether hinterliess einen amorphen braunen Rückstand,
worin sich nach längerer Zeit eine sehr geringe Krystallbildung zeigte. Diese
Krystalle wurden so gut wie möglich aus dem übrigens braunen und amorphen
Anteil entfernt, doch waren sie in zu kleiner Quantität anwesend, um weitere
Reinigung und Untersuchung vorzunehmen. Das amorphe Alkaloid wurde
wie früher mit Hülfe von Bleiacetat und H,S gereinigt. Nur durch wiederholte
Reinigung und mit ziemlichem Verlust konnten einige Gramm völlig weiss
erhalten werden, die uns gestatteten es noch etwas näher zu untersuchen. In
Uebereinstimmung mit dem japanischen Namen Nanten und dem daraus
hergeleiteten Namen Nandina sei dem Alkaloide der Name Nandinin beigelegt.

Das Nandinin bildet ein amorphes, weisses Pulver, das namentlich in
feuchtem Zustande, wie in Lösung Neigung zeigt, sich bräunlich zu färben, und
dann nur durch wiederholte Reinigung und mit ziemlichem Verlust wieder weiss
zu erhalten ist(f)

Es ist leicht in Weingeist, Aether, Benzol und Chloroform, wie in verdünnten
Säuren löslich. Auf keine Weise konnten krystallisirte Salze dargestellt
werden, weder durch freiwillige Verdunstung der neutralen Lösungen in
Sänren, auch unter dem Exsiccator, noch durch Vermischen der ätherischen
Lösung des Nandinins mit ätherischen Lösungen der Säuren u. s.w. Unter dem
Mikroskope liess sich niemals eine Krystallisation beobachten. Auch beim
längeren Verweilen einer Portion des Alkaloides über einem Uhrglas mit Salzsäure
in einem verschlossenem Raum zeigte sich keine Krystallisation.

Die Lösung in verdünnter Säure wird von allgemeinen Alkaloidreagentien
eefällt. =

Alkalien geben weissen Niederschlag, in grossem Ueberschuss des Fällungs-
mittels löslich,

Hydrargyrichlorid weissen Niederschlag, in Salzsäure löslich,

Kaliumdichromat gelben Niederschlag,

Tannin weissen Niederschlag, in Ueberschuss, wie auch in verdünnter
Essigsäure löslich. Salzsäure erzeugt in der Lösung in Tannin wieder
einen Niederschlag, welcher von Neuem auf Zusatz von Tannin oder
Essigsäure verschwindet, durch Salzsäure aber wieder hervortritt,

Mayers Reagens weissen Niederschlag, in Alkohol löslich,

Kaliumeadmiumjodid, Kaliumwismuthjodid füllen ebenfalls weiss.

Platinchlorid gelblich weissen Niederschlag,

Pikrinsäure gelben Niederschlag,

(t) Herr Shimoyama fand, dass Frösche bei subeutaner Injection von 1--8 mgr. des
Nandinins getödtet werden.

Conc. Schwefelsäure löste mit violettlichrother Farbe. Auf Zusatz einer
Spur Salpetersinre färbt sich die Lösung prachtvoll blau. Ebenso
erzeugen andere Oxydationsmittel, wie Chlor- oder Bromwasser,
Kalichromat, Ammonmolybdänat, auch Ferrichlorid in dieser schwefel-
sauren Lösung grünblaue bis blaue Färbung. Auch mit Selensäure
oder Tellursäure wird diese Lösung parchtvoll grün, später indigblau
Das Platindoppelsalz wird ebenfalls von H,SO, prachtvoll blau gefärbt.

Chlor- oler Bromwasser färben grün,

Salpetersäure grasgrün, nachher braun.

Blementaranalysen.

T 0.2990 er. des lufttrocknen Alkaloides, bei LOO— 105° getrocknet
eben 0.2943 er. 'Trockensubstanz und nach dem Verbrennen
0.1566 gr. H,O und 0.7473 gr. CO,

If 03133 gr. lieferten nach dem Trocknen bei 105°, 0.3091 gr. und
beim Verbrennen 0.1658 gr. H,O und 0.7877 gr. CO,

III 0.3072 er. des bei 100° getrockneten Alkaloides lieferten 11.9 cm?
feuchten N, bei 11° und 707.8”".

oder in Procenten :

N ul II Mittel NO
a A 69.5 ーーー 69.35 70.17
H=1 5.96 5.94 ーーー 5.95 5.84
Ey eee eta 4.81
0) 三 16| — ーー | 一 一 20.42 19.68

1.004 er. des Iufttrocknen Platindoppelsalzes verloren nach einigen Tagen im
Exsiccator 0.0128 gr. und nachher bis zu constantem Gewicht bei 105° getrocknet
noch 0.0398 gr. und lieferten beim Glühen 0.1716 gr. = 18.23% Platin, auf
getrocknetes Salz berechnet.

Das Salz (C,, H,, NO, .. HCl)? Pt Cl, verlangt 18.35% Pt. (Pt = 194.5)

Nach obiger Formel würde das Nandinin mit Hydroberberin C, N, NO,
homolog sein.

Das Berberin liess sich aus den mit Wasser oder Weingeist dargestellten
Anszügen nicht in befriedigender Weise durch Stehenlassen mit Salzänre erhalten.
Es wurde aber in genügender Menge abgeschieden, um seine Anwesenheit fest-
zustellen, als die mit Ammoniak versetzte und mit Aether wiederholt aus-
geschüttelte Flüssigkeit nach dem Alkalischmachen mit Chloroform ausgeschüt-
telt wurde. Aus dem Rückstande konnte durch Versetzen mit verdünnter
Schwefelsäure, Umkrystalisation der ausgeschiedenen Krystalle aus Alkohol und
weitere Reinizung durch Fällung der wässerigen Lösnng mit verdünnter
Schwefelsäure eine kleine Menge reiner, gelber Krystalle erhalten werden, die
ganz das Aussehen von Berberinsulphat zeigten und auch in ihren Reaktionen
damit übereinstimmten.

Vi. ORIXA JAPONICA THUNB.

Ueber diese einheimische, auf hohen Bergen vorkommende Rutacee findet sich
in der japanischen Litteratur Folgendes. Das Wurzel- und Stammholz und die
Blätter dieses etwa 10 Fuss hohen, im Frühling blühenden Baumes werden in
Japan, wie in China als Heilmittel bei Typhus, Frost- und Wechselfieber, Malaria,
Speichelfluss benutzt.(*) Sie sollen auch Geschwulste in der Halsgegend beseiti-
gen und bei Insectenstichen und Schlangenbiss Verwendung finden. Der chine-
sische Name Siyousan 75 | ist der übliche für die Wurzeldroge, Siyokuschizn
5) 7 der für die Pflanze.

Japanische Namen sind Rokusagi コタ クタ サキ, Nogusa 7 7 % Heminotiya
~ 2) #+¥, Haneboku 2 # 7, Tomome +} € », Tiyabisiyaku # 7 & » r X
u.s. w. Die Wurzel soll verfälscht werden mit der des Clerodendron tricho-
tomum (Kusagi クタ サギ) oder der Hydrangea hirta Sieb. (Yama asisai ヤマ
? » +A). Was im Handel Kaisiyu-Siyousan ( 海 州 常山 ) heisst, ist die Wurzel
von Clerodendron trichotomum.

Wurzel- und Stammholz wurden einer vorliufigen Untersuchung unter-
worfen, besonders zur Feststellung, ob die gelbe Farbe, welche beide besitzen,
auch hier mit einem Gehalt an Berberin zusammenhänge. Herr 'Tamba ex-
trahirte dazu eine ziemliche Menge des groben Pulvers durch Perkoliren mit
Weingeist. destillirte den Weingeist vom Perkolate ab und behandelte den Riick-
stand mit Wasser. Es schied sich dabei ein gelbbraunes Harz ab, welches
abfiltrirt wurde. Das eingeengte Filtrat lieferte nach längerem Stehen mit
Salzsäure eine ziemliche Menge gelbe Krystalle. Diese wurden gesammelt und
durch Umkrystallisation aus Alkohol- oder Wasser oder durch Fällung der wäs-
serigen Lösung mit verdiinnter Salzsänre gereinigt.

Mit salzsaurem Berberin verglichen zeigten sich die gereinigten Krystalle
qualitativ völlig damit identisch.

Das Harz, welches wiederholte Male durch Lösen in Alkohol und Fällung
durch viel Wasser unter Zusatz von etwas Salzsänre gereinigt wurde, bildete ein
graulich gelbes Pulver. Is ist geschmack- und geruchlos, unlöslich in Wasser,
leicht löslich in Alkohol und Alkalien, unlöslich in Aether.

(*) Verschiedene Recepte werden empfohlen, Das Holz wie die Blätter werden z. B. mit
Glycyrrhizae radix zusammen gediimpft an der Sonne getrocknet und während einer Nacht in Sake
macerirt, nach dem Coliren wieder an der Sonne grtrocknet, dann gerüstet und zerrtossen. Auch
wird Siyusan mit Essig zubereitet und bositzt dns Praeparat dann eine brechenerregende Neben-
wirkung. Auch wird eine Abkochung (15 :250) mit Reiswasser bei Fieberanfall auf einmal
genommen, um Erbrechen zu erregen. Ohkin erwähnt in seiner Gesundheitslehre, dass er 40
Jahre lang ausgezeichnete Wirkung dabei gesehen hat.

VII. SKIMMIA JAPONICA THUNB.

Diese Rutacee ist eine zu den einheimischen Giftpflanzen gehörende Stande,
welche überall in Japan vorkommt. Der japanische Name ist Miyama の た
kimi = ミヤ マシ キミ (chin. In-wn B3E)0). Der Flora Japonica v. Siebold ist Fol-
gendes entnommen: Die Pflanze(2) kommt überall in den Wäldern an schattigen
Orten auf Bergen vor (Fundort Nagasaki auf dem Berge Kawara, c* 600 Meter);
Höhe 3-4 Fuss mit nach dem Boden geneigten Zweigen, in cultivirtem Zustande
erreicht sie eine gréssere Höhe. Die Blüthen erscheinen im März-April und
verbreiten, namentlich Abends, einen angenehmen an Daphne odora erinnernden
Geruch. Die Früchte sind roth oder bei einer cultivirten Varietät weiss und
reifen im October. Die Blätter schmecken aromatisch und scharf. Die Pflanze
wird allgemein in Gärten und bei Tempeln cultivirt und wegen ihrer schönen
Inflorescenz, des angenehmen Geruchs ihrer Blüthen und schönen, rothen
Früchte als Zierpflanze angewendet. Die Japaner und Chinesen zählen sie
zu den Giftpflanzen (sikimi=böse Frucht).

Durch die freundlichen Bemühungen des Herrn Shimoyama gelang es mir,
eine grosse Quantität der im Frühjahr (Jan.) in der Provinz Boshin in schattigen
Thälern des Berges Kiyosumi yama gesammelten Pflanzen zu erhalten.

Aus den frischen Blättern wurden durch Dampfdestillation etwa 50-70 gr.
des ätherischen Oels dargestellt. Dieses war klar, fast farblos, sehr schwach
gelblich gefärbt und von eigenartigem Geruche, etwas an das ätherische Oel
von Citrus bigaradiae und Juniperusarten erinnerend. Es hatte ein spec. Gew.

(1) Andere Namen sind:

Sinsan, it. Portugalnoki (Thunberg. Flora Jap. p. 62), Haharagusa (Honzohiyuhumebiko
wakunsjo), Yamarincho (Honzokoi)

(2) Frutex tri-quadripedalis (nunquam, ut voluit Kaempferus, arbor vasta) ramis in planta
spontanea plerumque deflexis, in culta erectis vel patentibus strictis alternis teretibus, cortice glabro
glanduloso-verrueuloso sordide e cinereo flavescente vel fuscescente valde aromatico vestitis, novel-
lis virentibus. Folia alterna, quatuor ad sex e quavis gemma, per tres annos persistentia, cujusvis
anni fasciculato- approximata et ab iis anni praecedentis remota, petiolata petiolis semiteretibus
erassiusculis glabris basi articulatis semipollicaribus; lamina obovata-oblonga vel oblonga,
utrinque attennata, basi non articulata, acuta, integerrima, coriacea, penninervia, nervo medio
subtus valde prominente, lateralibus in planta viva vix conspicuis, utrinque glabra et laete viridis
vel rarius subtus verruculis prominulis adspersa et fuscescens, 4-6 pollices longa, 15-20 lineas
lata. Stipulae nullae. Gemmae perulatae; Perulae imbricatae, lanceolatae, acutae, folia-
ceae, post vernationem longe a se invicem remotae, eo modo ut rami hornotini magis inter
perulas distantes quam inter folia approximata et fasciculata extendantur indeque a basi ad duas
tertias longitudinis aphylli et perularum delapsarum cicatricibus tantum sint notati. Flores ex
apice ramorum, in paniculam bi-vel tripollicarem densam thyrsoideam et subpyramidatam
dispositi, polygami, suaveolentes, vernales, Rachis paniculae stricte erecta, teres, glabra, albida,
rami primarii alterni, patentes, singuli Bractea lanceolata vel lineari-lanceolata acuta coriacea
tandem decidua suffulti; secundarii subdichotomi, bracteolati; Pedunculi uniflori, florem
subaquaentes vel es longiores, teretes, sursum clavato- incrassati. Calyx hypogynus, gamosepalus,
brevis, urceolatus, quadri- quinquefidus, persistens ; limbi laciniae ovato semiorbiculares, acutius-
culae vel ohtusne, integerrimae, erecto- patentes, glabrae, albae, aestivatione imbricatae. Corolla
hypogyna, tetra- vel pentapetala, decidua; Petala cum calycis laciniis alternantia, sessilia ovato-

von 0.8633 bei 20°, war löslich in Weingeist und Essigsäure mit schwacher
Opalescenz, verpuffte mit Jol und wurde von Schwefelsäure réthlichbraun
gefärbt. Mit Natrumbisulfitlésung öfters geschüttelt, schied sich erst nach
einigen Wochen ein wenig einer butterähnlichen Masse ab. Mit Kali geschüt-
telt fand keine merkbare Volumabnahme statt ; beide Schichten wurden bräunlich
gefärbt, auf Zusatz von ammoniakaler Silberlösung fand aber nur schwache
Reduction statt.

Im Soleil- Ventzke’s Polarisations Apparat zeigte es p. 1 Dem. eine Rechts-
drehung von +7°.45. Der wiederholten fractionirten Destillation unterworfen
destillirte von den flüchtigeren Anteilen der grössere "Peil zwischen 170°-173°
(uncorr.), von dem Reste ging der grössere Teil zwischen 225° und 235° (uncorr.)
über.

Die niedrigst siedenden Fractionen waren farblos, wurden von Schwefelsäure
orangeroth, von trocknem Salzsäuregas violettlichrothbrann gefärbt, an der Luft
verdickten sie sich allmählich.

0.2655 gr. des Destillates 170°-178° gaben bei Verbrennung 0.2769 gr. H,O
und 0.8436 gr. CO,

oder in Procenten :
Versuch 0。 Hy,
C=12| 86.7 88.2

H = H 11.6 | 11.8
0. 一

Die höher siedenden Fractionen waren schwach grün- bis gelblich gefärbt.
Von dem Destillate 225°-235° lieferten
0.3872 gr. 0.3754 gr. H,O und 1,1018 gr. CO,

oblonga vel oblonga, acutiuscula, integerrima, erassiuscula, superne parum concava et sulco
longitudinali exarata, alba, subtus convexiusenla, alba marginem versus roseo-suffusa, erassius-
cula, inter se aequalia, calyce triplo longiora, aestivatione imbricata. Stamina tot quot petala et
cum his alternantia, hypogyna, extra torum aflixa, in floribus foemineis abortiva, ovario breviora ;
Filamenta inter se libera, subulata, erecta, glabra, alba, petalis parum longiora; Antherae dorso
parum supra basin aflıxae, ovatac, obtusae, basi cordatae, quadriloculares et antice longitudinaliter
quadrivalyes: Torus in floribus masculis quadrilobus, lobis brevibus transverse ellipticis carnosis
cum staminibus alternantibus, in hermaphroditis et foemineis annulus, brevissimus basin ovarii
arcte cingens, virescens. Ovarium superum, liberum, toro cinetum, ovato- globosum, glabrum,
plerumque quadriloeulare : locula uniovulata, ovulo ex angulo centrali pendulo anatropo ovata.
Rudimentum ovarii in flore masculo subconicum, breviter apieulatum. Stylus simplex, erectus,
quadrisulcatus, crassus, ovario et staminibus lrevior, glaher., Stigma capitato- incrassatam, quadri-
vel quinquelobum, lobis abbreviatis convexis superne papillosis et suleo tenui exaratis, glabrum,
virens. Drupa supera, globosa vel obsolete quadri- vel quinqueloba, magnitudine pisi, glabra,
eoceineo- rubra; sarcocarpium carnosum glandulis immersis obsitum, tandem siccum, septis
teniibus membranaceis ; endocarpinm in coceos quatuor vel quinque apice affixos ceterum a
earcocarpio solutos cartilagineos siccos indchiscentes trigonos utrinque altenuatos dorso convesos
lateribus planos glabros monospermos mutatum. Semen nnieum in quovis cocco, ex nngulo
contrali pendulum, ellipticum dorso convexum, Jateribus planis angulis acutis, glabrum, Testa
membranacea, in angulo centrali raphe notata lineari ab hilo elliptico in chalazam usque,
verticem seminis oecupantem, producta, ‘Tunica interior tenuissima, a testa vix separanda, fusca,
Albumen carnosum, nequabile., Embryo orthotropus, radicula brevissima hilum spectante,
cotyledonibus magnis carnosis late ellipticis utrinque rotundatis plane sibi impositis et bine
simul param concavis, plumula inconspicua, (v. Sieb. Flor. Jap.)

oder in Procenten.

Versuch Cu H,,0

C12 Wis | 18,9
|
I

|
H= yy 10.7 10.55
|

(0) == is 10,55

Die in dem Destillationsgefässe zurückbleibenden (oberhalb 250° siedenden)
Anteile wurden beim Erkalten fest, waren in Weingeist, Wasser, Petroleumäther
und Essigsäure nicht oder kaum, leicht dagegen in Chloroform löslich. Die
Resultate dieser Analysen, wiewohl nur je einmal ausgeführt, zeigen an, dass
ausser einem Terpen (Skimmen) vielleicht eine Kamferart vorliegt. Das in Ol.
rutae vorkommende Methylnonylketon verlangt 77.65% C, 12.94% H und
9.41% 0. Ein Terpenhydrat C,, H,, O würde 77.9% C, 11.7% H und 10.4% O
verlangen. Der Kohlenwasserstoff kann weder Nonylhydrür C, H,,noch Nonylen
GC, H,, sein. Hrsteres enthält berechnet 15.6% H, letzteres 14.3% H. Fine
Dampfdichtebestimmung konnte ist bis jetzt nicht ausführen.

Aus einer gesonderten Portion des Holzes mit Rinde wurde ein weingeistiges
Extrakt dargestellt, und es gelang mir leicht, durch vorläufige Versuche einen
crystallinischen Körper daraus abzuscheiden, welchen ich in grösserer Menge
dargestellt und auch noch etwas weiter untersucht habe.

Schon durch Erwiirmen des Extraktes mit einer geringen Menge Wasser
lässt es sich in eine dunkelschwarzgrüne harzige und eine obere sirupöse
braune Masse trennen. Letztere liefert nach längerem Stehen in geeigneter
Concentration einen erheblichen krystallinischen Bodensatz, welcher durch Um-
krystallisation aus verdünntem Weingeist und Wasser mit Hülfe von Tierkohle
völlig weisse Krystallnadeln liefert.

Skimmin. Weisse Krystallnadeln, wenig in kaltem, leicht in heissem
Wasser und Weingeist löslich, sehr wenigin Aether und Chloroform, leicht in
Alkalien mit schön blauer Fluorescenz löslich. Die wässerige und alkoholische
Lösung fluoreseiren nicht und schmecken bitter. Schmelzpunkt 210° (uncorr.).
Es scheint nicht giftig zu sein: 0.100 gr., in c* 5-6 Cm* warmem Wasser mit 2
Tropfen Natronlange einem Hunde subentan injieirt, verursachten keine lue-
sonderen Symptome ausser etwas Lustlosigkeit.

Mit Bleisubacetat (Bleiacetat + NH,) entsteht weisse Tällung.

Die wässerige Lösung reagirt neutral, reducirt in der Siedehitze alkalische
Kupferlösung nicht und wird von den gewöhnlichen Metallsalzen, auch
Bleiacetat, Ferrosulfat, Bisenchlorid und Goldchlorid, weder gefällt noch
verändert. Sowohl von dem im Essiecator getrockneten als dem bei höherer
Temperatur entwässerten Skimmin führte ich Elementaranalysen aus. Im
Exsiccator verlor das lufttrockne Skimmin 0.5% und weiter bei 130°-135°
3.96-4.2 proc. Wasser.

I 0.2906 gr. des entwässerten Skimmins lieferten

0.1319 gr. H,O und 0.5857 gr. CO,

II 0.3150 gr. des entwässerten Skimmins lieferten
0.1328 gr. H,O und 0.6287 gr. CO,
III 0.3534 gr. des während mehrerer Tage im Exsiccator getreckneten
Skimmins gaben 0.1597 gr. H,O und 0.6860 er. CO,
1V 0.3274 gr. des im Essiccator getrockneten Skimmins gaben 0.1422 er.
H,O und 0.6382 gr. CO,

oder in Procenten, auf wasserfreie Substanz berechnet.(*)

r 日 Berechnet auf
I II Ill IV Mittel OHO.
C=12| 550 | 544 ! 55.8 55.5 | 55.— | 55.5
|
H= 1 BO | ae) AS Ae are 4.9
| ! | |
O=16| 40 40.9 | 39.9 | 39.9 | 40.2 | 39.5

Beim Kochen mit verdünnten Mineralsänren scheidet sich ein in Wasser
unlöslicher krystallinischer Körper (Skimmetin) ab, während das Filtrat starke
Reduction gegenüber alkalischer Kupferlösung zeigt.

5.938 gr. mit 35 Cm? verdünnter Schwefelsäure (1:25) gekocht, bis Lösung
stattfand, und in geschlossenem Kolben dann mehrere (6) Stunden
auf dem Wasserbade erhitzt gaben nach dem völligen Erkalten
abfiltrirt und ausgewaschen 2.783 gr. = 46.8% Skimmetin.

10 Cm? des Filtrates (100 Cm?) bis 50 Cm? verdünnt wurden mit Fehling’s
Lösung titrirt. Im Mittel wurden für 10 Cm? Kupferlösung 8.2 Cm? verbraucht,
woraus sich 51.35% Glucose auf lufttrocknes Skimmin berechnet.

SO Cm’ des Filtrates wurden mit Baryumcarbonat behandelt und lieferten
nach dem Trocknen bei 110° 2.527 gr. = 53.2% einer braungelben sirupösen
Substanz, welche in Wasser (c* 20 Cm?) bis auf 0.039 gr. = 0.8% löslich war.
Bringen wir diese 0.8% als Skimmetin in Rechnung, so ergiebt sich aus 100
Teilen Skimmin 47.6 Teile Skimmetin und 52.4 (titrirt 51.35%) Glucose. Das
lufttrockne Glucosid scheint daher bei seiner Spaltung kein weiteres Wasser
aufzunehmen.

Das glucoseähnliche Spaltungsprodukt wurde in Wasser gelöst und die
Lösung mit Tierkohle entfärbt; die filtrirte Lösung enthielt, durch Wägung
bestimmt 7.16%, durch "Titration 7.2% Glucose. 1 dm dieser Lösung im Soleil-
Ventzke’s Polarisator zeigte eine Drehung von + 5.2(t), woraus sich ein spec.
Drehungsvermögen berechnet von

52.10 .... 251 31

en 10a

also etwa die Hälfte von dem der Dextrose (58.7 bei 20°, Tollens) (möglich
durch Glucosangehalt).

= + 24.5⑪).

Skimmetin. Parblose Krystallnadeln, welche in kaltem Wasser fast völlig
unlöslich, in kochendem etwas besser löslich sind; auch löslich in Weingeist,
(*) III und IV sind mit einer gesondert dargestellten und gereinigten Probe angestellt,

(t) bei Petroleumlicht,
($) d = 1.026 (spec. Gew. einer 7.2% Zuckerlösung).

Aether, Chloroform und Hisessig. In verdiinnten Alkalien lést es sich leicht,
ohne die alkalische Reaction abzustumpfen. Die wässerige, weingeistige und
alkalische Lösung fluorescirt schön blau. Auch cone. Schwefelsäure löst es
mit intensiver blauer Fluorescenz. Es reducirt auch beim Kochen alkalische
Kupferlésung nicht. Die wässerige Lösung wird von Bleiacetat gefällt, der Nie-
derschlag ist in Weingeist löslich. Auch wird die warm gesättigte wässerige
Lösung von Ferrichlorid blau gefärbt, von AuCl, rosa, später violett und blan.
Für die Löslichkeit in Wasser fand ich bei 11° 0.022%, bei 23° 0.03% und für
den Schmelzpunkt 223° (uncorr.) Bei stärkerer Erhitzung entsteht ein krystal-
linisches Sublimat, und es verfliichtigt, ole eine Spur Asche zu hinterlassen,
Cumaringeruch konnte ich bei der Verflüchtigung nicht bemerken.
Elementaranalysen von der bei 110° vetrockneten Substanz, wobei sie

kein Krystallwasser verlor.

0.3120 er. lieferten 0.1068 er. H,O und 0.7632 er. CO,

03143... , 0.1011, E50. „0.7738,

Von einer anderen Probe, ans einer gesondert dargestellen Portion Skimmin

„

erhalten und zuletzt aus der alkalischen Lösung mit Salzsäure gefüllt, lieferten
0.2943 gr. 0.1000 er. H,O und 0.7155 er. CO,.

oder in Procenten :

: Berechnet anf
l II II Mittel C。 H,O,
C=12 66.7 | 67.1 66.3 66.7 66.67
H= 1 3.8 3.57 3.78 | 3.71 3.7
OSS 15 29.5 29.33 29.92 | 29.59 | 29,63

Die Spaltung des Skimmins kann durch die Gleichung,
05870: BO 105040, «¢, 5.0,
Skimmin Glucoseart Skimmetin

ausgedrückt werden.

Das wasserhaltige Glucosid wiirde nach dieser Gleichung 47.3% Skimmetin
und 52.7% Glucose liefern, während ich fand 47.6% Skimmetin und 52.42%
Glucose.

Die Formel C,, H,, O, . H,O verlangt 5.26% Wasser, während ich fand für
das unter dem Exsiceator getrocknete Skimmin 3.92-4.2%. Im Exsiccator
verlor es bei einer Bestimmung 0.5%, also im Ganzen 4.42-4.7%. Warscheinlich
war das analysirte lufttrockne Skimmin schon etwas verwittert.

Das Skimmin und Skimmetin zeigen mit dem Scopolin und Scopoletin grosse
Uebereinstimmung. Sie unterscheiden sich von einander z. B. durch Folgendes.

Schmpt. Mit Ammoniak

Skimmin | schwieriger löslich in kaltem Wasser | 210° | gelbe Lösung und grünblauer Reflex

Scopoletin 5) stark Pe ee 198° | 61.% C und 4.2% H .

Bei dem Vergleich mit anderen Glucosiden zeigen sie sich nahe mit Aesculin
und Daphnin verwandt, und, was das Skimmetin anbetrifft, scheint es mit
Umbelliferon identisch. Es ist dies um so interessanter als aus Daphnearten—
deren Blüthe denselben Geruch wie Skimmia Japonica zeigen —ılas Umbelliferon
durch trockne Destillation des Daphnins oder des Extraktes entsteht. Das
Daplhnin C,, W,, O, und das damit isomere Aesculin liefern bei ihrer Spaltung
durch Säuren Oxyumbelliferon, C, H,O, (Daphnetin und Aeseuletin).

In der nächsten Tabelle habe ich zum Vergleich mit Scopolin und Skimmin
einige Eigenschaften dieser Körper übersichtlich zusammengestellt, so weit die
mir zugänliche Litteratur reichte.

Das Skimmetin, wofür genau die procentische Zusammensetzung gefunden
wurde, welche der Formel des Umbelliferons

CH : CH. CO

GA, 0 一 一

OH
auch mit dem des von Tiemann und Reimer für Umbelliferon festgestellten
Schmelzpunkt (223°— 224°) übereinstimmt, dürfte wohl anch wegen der völlig
analogen Kluorescenzerscheinungen als identisch mit Umbelliferon angesehen
werden und in dem Skimmin deshalb das Glucosid dieses inneren Anhydrides der
Umbellasäure vorliegen.

entspricht und dessen Schmelzpunkt (223° uncorr.)

Auch das Seopoletin scheint mir nach seinen Eigenschaften sehr nahe mit
den Zimmtsäurederivaten Aesculin, Kaffeesäure, Pernlasänren, Umbelliferon ete.
zusammenzuhangen. Eine Identität mit Aesenletin (Schpt. über 270°) und
Daphnetin (Schpt 253.—256°) kann nicht vorliegen wegen der Differenzen der
Schmelzpunkte. Auch enthält das Scopoletin mehr H. Vielleicht dass bei
näherer Untersuchung sich die Formel C,, H, O, (=Methylisenletin) für das
Seopoletin bewährt und dann «die Spaltung «des Glucosides nach folgender
Gleichung stattfindet.

20,120, =3G, Hs u, +36, 8, 0, oder
GEH 0, 08,0, CeO}

Es würde dabei 48.3% Glucose entstehen müssen, womit die letzt angeführte
Bestimmung durch Titration (Seite 24.— 0.313 gr. Glucose ans 0.6498 gr.
Scopolin = 48.1%) in Einklang steht.

Das von Tiemann und W. Will dargestellte Monomethyläseuletin
6, H, O, OCH, hat den Schmelzpunkt 184°. und wird von le, Cl, nicht gefärbt, das

(OW),
Dioxy 8. Methyleumarin 0, H, O ——— | von MH. V. Peehmann und
) (CH,):CH. CO
©, Duisberg schmilzt bei 235°, die Lösung in cone. Schwefelsäure Huoreseirt nicht.

Würde sich das Scopoletin als Cumarinderivat herausstellen, so wären in der

Scopolia Japonica Körper mit den beiden Kohlenstoffkernen

eg nnd Ci 2.70.08

で

vertreten, ersterer als Tropasäure in dem Alkaloide (Scopolein), letzterer in
dem (Scopoletin), und Scopolin.

Was die anderen Bestandteile der Skimmia Japonica anbetrifft, so kann
ich darüber vorläufig nur noch Folgendes mitteilen: Das Skimmetin wurde
auch als solches aus der Pflanze mit dem Schmelzpunkte 222°-223° (uncorr.)
abgeschieden, war aber schwierig von mehreren amorphen, zum Teil harzigen
Körper zu trennen. Weiter erhielt ich einen weissen krystallisirten Körper,
welcher bei c* 244° (nncorr.) schmolz, in Wasser nicht, sehr wenig in Weingeit,
besser in Petroleumiither, Aether, leichter in Chloroform und auch in Essigsäure
löslich. In Alkalien löst es sich nicht und zeigt keine Fluorescenzerscheinnngen.
Die Lösung in Essigsäure gibt mit H,SO, Braunfärbung. Das Gift wurde bis
Jetzt als eine bräunliche, amorphe Substanz erhalten. Sie ist wenig in kaltem,
ziemlich in kochendem Wasser, leicht in Weingeist, Aether und Chloroform
löslich. Einige Milligramm. in | Cm? Wasser gelöst töten resche unter fast

völliger Lähmung.

Wasserfrei.

Alkalische
| Schmelz- Lösung in Lösung in
Pr: | P nn Fe, Cl, AuCl;. H, SO,
TOC. roc. roc. punkt. bene Wasser. Alkalien,
Cc H ‘One|
| Aesculin. 、 160° reducirt. nicht. grünblau, beim | starke blaue ーー gelb mit blauer
Erwärmen Gold- | Fluorescenz, noch Fluorescenz
ausscheidung. in einer Lösung
Cys Hy, Oo. 2 H, O 529| 47| 424 | Son bate 002000
Daphnin. | 200° redueirt blau. (?) ーー 一 — — goldgelb.
nach
liingerem
Kochen.
|
1 (Gy, OM 上 0) 50.6 5.4 | 44.- 217° nicht. nichts, nichts farblos wird gelb und löst | schwach gelb.
Scopolin ET : beim Erwärmen | beim Kochen fluorescirt nicht. |sich fast farblos &
(Gi。 Hy O1) braungelb. Goldausschei- mit schwach blau-
dung. em Reflex,
i in ©); Hy, 0。. H,O 55.— 4,8 | 40.2 nicht. nichts, nichts, arblos, ist sich farblos farblos,
Skimmin Gi His 0。. H, 55 210° ich il ich farbl 1 ich farbl farbl
beim Erwärmen |kann mit AuCl,} fluorescirt nicht. | mit schwach
braungelb. gekocht a violettblauem Re-
ohne Färbung flex.
| und Goldabschei-
dung.
|
ae über 270° rile riinblaue Lö-|in der warmen| kalt gesättigt | schwach gelblich | intensi lb mit
Besoin chins ° | 2 SM Fite mit Alkali gesiittigten Lö-| farblos mit | mit schwach et Fluo:
CH: CHC.O | Oxyumbelliferon. roth, durch Säure ‚sung Rothbraun- | schwachem blauer Fluores- | rescenz。
C,H, De ala Diane fiirbung, bald Au) blauen Reflex. | cenz.
OH ) C, H, 0。 60.6 8.86 | 86.04 Ausscheidung.
Dioxycumarin.
Daphnetin. 253°—256°| reducirt. | griin mit Alkali = gelb. gelb. orange.
roth.
Scopoletin ©, Hn 0; 61.1 4.2] 35.7 198° reducirt. |dunkel bläulich- | blau, beim Ko- | farblos mit blauer | gelb mit blauer goldgelb, mit
C,H, O grüner Nieder- chen grüner und | Fluorescenz, Fluorescenz. Weingeist ver-
(Gi H; O,) schlag. grauschwarzer diinnt grünlich-
flockiger Nieder- gelb und mit
schlag. starkem blauen
Reflex,
Skimmetin ©, H;, 0; 66.7 3.7 | 29.6 9939 nicht, bläulich, beim Erhitzen | farblos mit blauer | farblos mit stark goldgelb, mit
rosa, später Fluorescenz, blauer I luores- | Weingeist ver-
violett und blau. cenz. dünnt fast farblos
mit starker blauer
Fluorescenz.
Umbelliferon C, H; 0; 66.67 8.7 | 29.68 | 2239-2249) 一 一 ーーー —- farblos mit blauer | blaue Fluorescenz ーーー
Fluorescenz.
CH:CH. CO
eg)
OH
Monomethylasculetin 184° ene ECU
CH: CH. CO
C,H, O |
2 Cy H, 0,| 625| 4.17 | 38.88
Dioxy B Methylenmarin | 0550 intensiv grüne fluorescirt nicht. gelb,
C. (CH,): CH. CO Färbung.
De ||
OH
OH

Feuchte

Atmosphäre,

goldgelb,

farblos,
nach mehreren
Tagen bläu bis
grün,

farblos,
nach mehreren
Tagen schün
grünblau,

fleischfarbig, wie
Mangansulphid.

goldgelb.

farblos, nachher
schwach braun,

ANHANG.

Andromeda Japonica Th.

Nachdem ich die Untersuchung der Andromeda Japonica beendet hatte
erschienen aus dem Laboratorinm von Prof. Dragendorff ausführliche Mittei-
lungen von R. Thal under den Titel: Erneute Untersuchungen über die
Zusammensetzung und Spaltungsprodukte des Ericolins und seine Verbreitung
in der Familie der Ericaceen nebst einem Anhang über die Leditannsäure, die
Callutannsäure und das Pinipikrin. (Pharm. Zeitschr. f. Russland Jg. XXII,
1883, No 14-18).

Wiewohl keine Andromeda’s in den Kreis seiner Untersuchungen aufgenom-
men wurden, scheinen mir dieselben sich so nahe an den von mir mit der A. Japo-
nica angestellten anzuschliessen, dass ich es erwünscht erachte, hier ein kurzes
Referat über diese Untersuchungen von Thal einzuschalten.

T. stellte das zu seine Versuchen dienende Ericolin, aus 300 Ib. zerhacktem
Ledumkranutes dar durch Fällung des wässerigen Infuses mit Bleiacetat und
Bleiessig. Das Filtrat wurde eingedampft, nochmals filtrirt und mit H,S entbleit.
Nach dem Viltriren wurde bis zur Extraktconsistenz eingeengt. Das Extrakt
wurde wiederholt mit einem Gemische von wasserfreiem Aether und Alkohol
(1:2) ausgekocht, der Aetheralkohol abdestillirt und der Rückstand wiederholt
diesem Auszichungsverfabren unterworfen, bis er sich in dem genannten Lösungs-
mittel völlig löste.

Das so erhaltene Präparat wurde einige Stunden lang bei 95°-100° C
getrocknet nnd dann über H,SO, im Vacuum gehalten. Es besass nun Latrakt-
consistenz und enthielt 0.36% Asche. Als solches der Elementaranalyse unter-
worfen, während der Feuchtigkeitsgehalt in besonleren Proben bei 95°-100? C
bestimmt wurde, ergab es im Mittel aus 3 ziemlich übereinstimmenden
Anlysen und anf trockne, aschefreie Substanz berenchet 71.82% C, 6.38% H
und 21.8% 0. Mit einem zweiten Präparate erhielt er aber sehr abweichende
Zahlen 59.59% C, 7.02% il und 3339% O als Mittel von 3 besser üübereinstim-
menden Analysen. 'T. wurde hiednrch veranlasst sein Präparat nochmals durch
wiederholte Lösung in Aether- Alkohol zu reinigen, bis die Lösung nach 48
stiindigem Stehen nichts mehr ausschied. Das so gereinigte Ericolin wurde
dann während 12 Stunden bei 95° C getrocknet und eine Woche über Schwefel-
sänre gehalten. Es enthielt noch 0.32% Asche. Ms wurdein fenchtem Zustande
analysirt und der Trockenverlust, im Mittel 36.20% betragend und bei 959-100?
in gesonderten Portionen bestimmt, als Wasser in Rechnung gebracht. Er erhielt
als Mittel aus 4 Bestimmungen:

82.46% C 5.89% H und 11.65% O.

Thal beschreibt sein Ericolin als geruchlos, braungelb, klebend hygro-

scopisch, stark bitter schmeckend, in Aether und Alkohol leicht, in reinem

(*) N. Tydschr, v. Pharm. in Nederland Jg. 1882 No, 11, Jg. 1888 No. 3 u. No. 8 wie
auch New-Remedies Vol. XI, No. 10, Vol. XII, No. 3 u. #8.

Aether sehr schwer löslich. Beim Erhitzen schon unter 100° erleidet es eine
teilweise Zersetzung unter Entwickelung von Ericinol. Es ist wenig in Benzol,
noch weniger in Chloroform, leicht in Aether-Alkohol löslich und giebt bei
der Spaltung durch Säuren Zucker und Ericinol. Letzteres wird sofort nach
der Abscheidung teilweise oxydirt, teilweise in eine Hydroericinol genannte
Substanz verwandelt.

Auf Grund von mehreren weiteren Elementaranalysen kommt T. zu dem
Schlusse, dass dieses Ericolin bei der successiven Behandlung mit Benzin,
Chloroform, Aether-Alkohol und Wasser schon eine teilweise Zersetzung
erleidet unter Freiwerden von Ericinol, wie dies aus folgenden Data ersichtlich

Asche Verlust beim C% H% 0%

Trocknen

I Reines Ericolin (Thal) ...... 0.32 | 36.20 | 82.46 | 5.89 | 11.66
II Benzinrückstand ,, 0.— | 2440| 78.93] 8.18] 18.64
III Chloroform ,, A 0,09 11.70 59.08 | 8.85 82.57
IV Aether-Alkohol ,, ,, ’ 0.65 7.28) 5428| 7.12 38.65
V Ericolin (Schwarz. u. Rochleder) 10,65 | —— 51.71| 7.19 | 41.10
VI Pinipikrin (Thal)(*) aus H* Sabine ONT 13.43 | 57.46 | 8.33 | 34,21
VIL Pinipikrin (Kawalier) |

a aus Pinus silvestris —-- au 55.61 | 7.60 | 36.8

3 aus Thuya orientalis ニー | <— | ET...

Vergleichen wir diese Zahlen mit denen, welche ich fand fiir
PSOE LORAIN sist ort hints ary ne eae aka .— — 60.48 7.4 32.12

so zeigen sie am meisten Aehnlichkeit, mit den Analysen IIL bis VII sind
aber sehr abweichend von. den Analysen I u. II welche mit Substanz angestellt
wurden, welche beim Trocknen einen sehr grossen Gewichtsverlust (24.4-36.2%)
zeigten.

Es war einleuchtend, dass der um so höher gefundene Kohlenstoffgehalt, je
bedeutender der Verlust beim Trocknen, zum Teil daraus erklärt werden
musste, dass die Substanz ohne weiteres Trocknen analysirt und dass der in
einer gesonderten Probe bei 95°-100° bestimmte Gewichtsverlust als Wasser
in Rechnung getragen wurde, während dabei nicht nur Wasser, sondern auch
Erieinol verflüchtigt. 7’hal redueirt dem entsprechend auch die Resulte seiner
Analysen (1) auf:

80% C, 7.69% H und 12.31% O
Zahlen, welche er berechnete aus der Formel C,, H,, O,, welche bei Spaltung
besser als die anfangs aufgestellte und aus den Analysen (I) berechnete Formel
C。。 H,, O, zu verwerten war. Leider standen mir nur die Nos. 14. 15. 17 u.

(*) Dieses Pinipikrin besass dieselben Eigenschaften und lieferte dieselben Spaltungs-
produkte wie Ericolin.

18 der Pharm. Zeitschr. f. Russland Jg. 1883 zu Gebote und gelang es mir
nicht das fehlende Heft anderweitig einzusehen, so dass ich auch die darin zu
erwartenden Spaltungsversuche des Ericolins unerwähnt lassen muss.

Es fragt sich nun, ob auch für den giftigen Bestandteil der Andromeda
Japonica Th. eine derartig leichte Zersetzbarkeit durch blosse Behandlung mit
Lösungsmitteln wie Benzin, Chloroform etc. zutreffend ist, wie von Thal für das,
wie mir scheint verwandte, Ericolin aus seinen Analysen abgeleitet wird. Es
wäre möglich, dass das von mir, nach vielen solchen Behandlungen mit mehreren
Menstrua, erhaltene Asebotoxin als Produkt einer teilweisen Spaltung des
wirklichen Bestandteils aufzufassen sei und wären dann weitere Versuche und
Analysen, wie sie von Thal mit dem Ericolin angestellt wurden, auch für das
Andromedagift erwünscht. Leider scheint T. keine physiologischen Versuche
über die toxische Wirkung angestellt zu haben, und darauf bezügliche frühere,
Angaben konnte ich in der mir zur Verfügung stehenden Litteratur nicht finden,
so dass ich auch in dieser Hinsicht keine Vergleiche zwischen dem Ericolin und
Asebotoxin anstellen kann.

Nachdem ich wieder in Besitz einiger Verbrennungsröhren, gekommen war
habe ich während des Drucks dieser Bogen noch einige Versuche und Analysen
ausgeführt mit Präparaten, welche ich durch weitere Reinigung aus Resten,
von früheren Reinigungsversuchen herstammend, erhielt.

A. Geruchloses, aschefreies Präparat, ein weisses Pulver bildend, in geschmol-
zenem Zustande schwach bräunlichgelb gefärbt. Mit conc. Salzsäure entsteht
allmählich schöne Blaufärbung unter Verbreitung des früher erwähnten Geruches
etc. Die wässerige Lösung, auch die in der Kochhitze gesättigte, reagirt völlig
neutral und wird von Fe, Cl, nicht gefärbt ; mit verdünnter Schwefelsäure gekocht
trübt sie sich weisslich, bald nimmt die Flüssigkeit rothe Farbe an, mit carmin-
rothem Reflex, es scheidet sich eine anfangs wenig gefärbte ölige Flüssigkeit
ab, die bei andauerndem Kochen grünlich, nachher braun wird und sich dann
harzig abscheidet. Das Filtrat zeigte auch hier beim Kochen mit alkalischer
Kupferlösung eine wenigstens gleichstarke Reduction wie die nicht vorher mit
Säuren gekochte Lösung. In der Kälte zeigt es mit alkalischer Kupferlösung
keine oder höchst geringe Reduction. 8 mgr., einem Kaninchen von 2.95 K°.
Körpergewicht (= 2.7 mgr. p K°) in wässeriger Lösung subcutan injicirt,
töteten innerhalb 2 Stunden unter den früher beschriebenen Symptomen. Das
Präparat zeigte sich deshalb dem früher erhaltenen völlig gleich.

0.3668 gr. lieferten nach einstündigem Trocknen bei 105°-110° 0.3523 gr.

Rückstand und 0.7974 gr. CO, und 0.2433 gr. H,O.

oder 61.73% C
7.67% H
30.6 % 0.

Ein ähnliches Präparat, in wässeriger Lösung neutral reagirend und mit
Fe, Cl, schwache, zweifelhafte Färbung gebend, lieferte nach einstündigem
Trocknen bei 110°

60.96% C, 7.56% H und 31.48% O.(*)

B. Ein fast reines Präparat., von dem vorigen darin abweichend, dass es in
geschmolzenem Zustande mehr röthlichbraune Farbe zeigte, in wässeriger Lösung
sauer reagirte und mit Fe, Cl, violettröthliche Färbung erzeugte.

0.4167 gr. lieferten während 1 Stunde bei c* 110° getrocknet 0.4120 gr.

Rückstand und
0.2942 gr. H,O und 0.9442 gr. CO,
oder 62.5% C. 7.91% H und 29.59% O.

C. ähnliches Präparat wie B, in geschmolzenem Zustande noch mehr röthlich
gefärbt. Die wässerige Lösung, welche es beim Verdampfen zurückliess, war
anfangs völlig farblos, färbte sich jedoch während des Verdunstens röthlich.

Die wässerige Lösung reagirte stark sauer und gab mit Fe, Cl, ziemlich
intensive Violettrothfärbung.

0.3715 gr., während einer Stunde bei 110° getrocknet, gaben 0.3425 gr.
Rückstand und

0.7913 gr. CO, und 0.2336 gr. H,O oder
63% C 7.6% H und 29.4% O.

D. Dieses Präparat war erhalten durch Fällung einer Lösung eines sauer-
reagirenden Präparates in Chloroform mit Petroleumäther, bis sich nichts mehr
ausschied. Die Petroleumätherlösung wurde dann concentrirt und einmal mit
Wasser ausgeschüttelt. Die abgehobene Petroleumätherschicht wurde verdampft
und auf dem Wasserbade getrocknet. Der Rückstand war gelb gefärbt, klebrig
zähe, löste sich in kochendem Wasser nicht völlig auf und verbreitete beim
Trocknen in hohem Masse den betäubenden Geruch.(t)

Die wiisserige Lösung reagirte ziemlich sauer und gab mit Fe, Cl, ziemlich
intensive Violettrothfärbung.

0.4019 gr. während 14 Stunden bei 105°-110° getrocknet gaben 0.3657 gr.

Rückstand und
0.8904 gr. CO, und 0.2617 gr. H,O oder
66.% C, 7.94% H und 25.66% O.

Durch subeutane Injection von 0.0002 gr. der Präparate A. B. C und D,
in 1 Cm? Wasser gelöst, bei Fröschen wurde nun versucht, die relative Giftwir-
kung und Reinheit festzustellen. Wiewohl die Präparate A. B. und C keine
grossen Unterschiede zeigten, war doch im Allgemeinen die stärkste Giftwirkung

(*) In diesen Präparaten findet sich die Bestätigung, dass das reine Absebotoxin in wässeriger
Lösung neutral reagirt, was ich auch früher schon fand, doch bei dem damals analysirten
Präparate weiter festzustellen vernachlässigte. Bloss den bei dem Ausziehen mit absolutem
Aether zurückbleibenden Anteil habe ich auf seine neutrale Reaction controlirt.

(+) Es ist wohl diese flüchtige Substanz, welche den im Yudoku somoku susetsu erwähnten
Kopfschmerz verursacht.

bei A und der Reihenfolge nach etwas schwächer bei B und C.

Das Präparat D. zeigte eine deutlich geringere Wirkung. Dass die Prä-
parate A. B und C, wiewohl die letzteren durch ihre saure Reaction und ihr
Verhalten gegenüber Eisenchlorid sich unrein erwiesen,(*) dennoch nur wenig
Unterschied in toxischer Wirkung zeigten, scheint mir dadurch erklärlich, dass die
Unreinigkeiten auch für sich nicht unwirksam sind oder vielleicht nur zu
einzelnen Procenten in B und C vorhanden sind, sodass kleine Schwankungen in
den Dosen bei Fröschen von verschiedenem Gewicht, 12-15 gr., und Individualität
keine constanten Unterschiede in der relativen Giftwirkung zeigen.

Nach den Analysen zu urteilen würde eine grössere Unreinheit einem
höheren Kohlenstoffgehalt entsprechen, doch sind Schwankungen, wie für die
reinen Präparate gefunden wurden, von 60.3-61.7% C und 7.3-7.7% H, auch
derart zu interpretiren, dass die analysirten Präparate, welche vorher schon
unter dem Ersiccator oder durch Schmelzung auf dem Wasserbade teilweise
getrocknet waren, nicht bis zum constanten Gewicht, sondern mit Rücksicht
auf die Verbreitung eines besonderen Geruches nur während etwas 1 Stunde bei
C* 110° getrocknet wurden.

Weil der Gewichtsverlust dabei nur 3-4% betrug, war die Verfliichtigung
des eigentümlich riechenden Stoffes (Ericinol?) jedenfalls keine erhebliche und
glaube ich daher, dass die angeführten Zahlen einen ziemlich genauen Ausdruck
für die Zusammensetzung des Asebotoxins geben. Die von Thal weiter unter-
suchten, von Rochleder und Schwarz entdeckten, Körper Callutannsäure, Ledi-
tannsänre und Ledixanthin scheinen mir die Analoga zu den von mir abgeschie-
denen Körper, Aseboquercetin etc. zu sein. Aus den Angaben von T. geht
nicht hervor, in wie weit jene Körper die Stein’sche Paracarthaminreaction geben.

Scopolica Japonica Max. Längere Zeit, nachdem ich die Untersuchung
dieser Pflanze abgeschlossen hatte, wurde mir eine kleine Probe gelber Krystalle
übergeben mit der Bezeichnung Solanin aus Scopolia Japonica und welche
schon vor mehreren Jahren dargestellt zu sein schien. Bei der Untersuchung
stellte sich dieses Präparat als etwas unreines Scopoletin heraus. Es löste sich
schwierig in Wasser, leicht in Ammoniak mit gelber Farbe und blauer Fluorescenz
etc. und scheint somit das Scopoletin die Substanz zu sein, welche von Dr. Martin
als Solanin aufgefasst und auch dieselbe, welche von Dr. Langgaard mit dem
Namen Rotoin bezeichnet wurde.

(*) In ihren sonstigen Eigenschaften verhielten sie sich fast gleich. Alle 4 Priiparate
gaben gesüttigt wiisserige Lösungen, welche sich mit Bleiessig (officinellem) klar mischen liessen
und mit Mayer's Reagens milchartig weisse Trübungen gaben, welche suf Zusatz von etwa
gleichem Volum Wasser wieder verschwanden. Auch reducirten sie ein Gemisch von Eisen-
chlorid und Ferricyankalium. Die Reaktion mit Salzsiiure trat bei B und C etwas weniger rein
ein als bei A, ; bei D zeigte sich nur eine bräunliche Fürbung.

VII

VIII

TAFELN.

Andromeda Japonica Thunb.

Scopolia Japonica Max.

Scopolia Japonica Maz., Wurzel und Querdurchschnitt
Macleya cordata i Br:

Nandina domestica Thunb. Exemplar in Frucht.
Nandina domestica Thunb. Blüthestand

の zzzg Japonica Thunb. mit Droge.

Skinmia Japonica Thunb.

木

ANDROMEDA JAPONICA THUNB.

m
L DER SN(

OPOLIA JAPONICA.

1
4

WURZE

ae a: En:
BE Mimi.

| !
1}
」
」
|
] 5
| 4
at
( 還
も
74
』
-
*
- H
. F.
é
・

ae
nal

a
0
= eh (の

+ & am
MACLEYA CORDATA. R. BR.

i ee
=

NANDINA DOMESTICA THUNB.

ORIXA JAPONICA THUNB.

|
.
|
i

*
SKIMMIA JAPONICA THUNB

| に)

yt"

In Betreff der „Abhandlungen des To-
kio Daigaku ” war bisher beabsichtigt,
den Abhandlungen jeder einzelnen
Facultiit laufende Nummern zu geben.
Von jetzt an, mit No. 10 beginnend,
sollen aber die “ Abhandlungen des
Tokio Daigaku” einerlei von welcher
Facultiit sie stammen, in der Ordnung
ihrer Veröffentlichung mit laufenden
Nummern versehen werden, Die bisher
veröffentlichten Abhandlungen (1 bis
9) gehören sämmtlich zu einer Facultät,
der naturwissenschaftlichen, und be-
halten daher ihre bisherigen Nummern
wie unten angegeben.

Abhandlungen des

Abhandlungen der
Tokio Daigaku.

naturwissenschaft-
lichen Facultiit

Bd I Heft—1 wird künftig No.1
gezählt als
sap! ik iow
ULL 3 wre
No. 4 : a
ess, の
eG 6
ae ik r |
goes 3 A
ao i, a,

The plan hitherto followed with the
Memoirs of the University has been to
issue and number them as Memoirs of
some one of the Departments. But
hereafter the Memoirs of all the
Departments will be issued as “Memoirs
of Tokio Daigaku” and numbered as
such in the order of their publication
beginning with number 10. Those of
the Science Department heretofore
issued will be numbered in the new
series as shown below.

Sci. Dept. Tokio Daigaku

Vol I pt—1 willbe No.1
Vol II „ » 2
Vol III pt—1 1 Nat:
No. 4 ” » 4
Fa) 2 aa
3 5 al
m Pr Peer (
Pai: = ie.

9 7 wae

meee? @ OEE] MP Oo ee’ 2” S 1A ©, — — 4a???
=> a? DP) pP um, » pumy» リン II p> %
Du mp DDB m) mw) 2 N, DDD >)» om
ERLITTEN SL we) > 5
>»? sm»)»p m Dad) DEB» » Dur, S$ の
» 22 mm) Ey» WWD )) BD Wumm)» >Z 7 2, D> DYE)
> Pwr) m pw» PPD y 2). > on 2
>») aD) ıD DEB) PPD» 1, PF う あ PVP 5725)
> >>> Ebb > PEED »>> PY map) y are? > m) > 7 ee
p Dp wm) >» m YRprs) Day) “yy \@ »D 7 4 > Lip ) - )
> J) > IDI SD PH n > Ww?» 》)) BP m») By)
> DP WH DD We» dD YYW)» HD PD DD)
) 22 DD DD »mw DP か 必 2p> の BD? D> Pw wy
DP TID >. ua P YD) ww» か “DD WD yyy yy»
> > ヵ IVD 2» DD Y» Y)DY)) YW)? BD yw) WH
> >? D2» VR PP» B)DD»)) DW ZIAD D> „mm!
> DD mm )> PWD DBD) BHP >») Bw? BD YD DDB)
y DP IID DD» DD YP» PID »» 024 < は 5 >» > Pp
> Dp mph») Pip mM) U m)) ) BERD A) . ye
DD D> PB BY: DM N a > Zu B.
> DDR DD)» ID D)52)) 95
> IR HR I»: )) N > I» ea yy >) Sa ーf
) DIB» ) ) >» ) 2) 名 りう ゅ JIVDOO >» ノ ) か
‚> WM) wy)?» Wi) > >>» ) im Pp DP) yar
>.» »)» DD DD ) WD), D2» DY} “yy? \) » »y )
»> Py) y» >»), >». I) 222 まず )) ) DD I i> » PP yy y
2) 21) »» DAW?) DADDY? PD) Epp DB PS >» aw
> > つり MID) » J) WB) / 2 DID) BP) ラノ 5 VD YP» yy
か つゆ mm Be win A ID) D>ZmmDB» クル pj
yD») LHP ) ツ BD m »)Dı» Don >>> ノリ
yw ) yw 2, DD クウ PV) DY?)
) ゆ DP VY AY DID Wh DYDD YAH) PPP HVS
> フツ OD” DY) DDP nn BD WWI リア Dy yD:
2 20 dei » 20> WWM? DID» 2WW 9 m) ) Da 2
ちあ) DDI VDRO の の) リラ YY 27-29’ DD DD»
» DD D mw DIA YIP Wy oe pam AMD >), m y»
> WD DP DD) DW) Min» WII ») » yo
DPT) DYDD » IrBD)B BY Hay » I 。 PP LW? YD yyy)
DP WER Dy) » Dela» Dim» WPD WM DD yyy”
PD TSP PD» rm» BD Yow? YW WPD WW) 9
2) 9 WP? 2 》 m 220 |) BWW D WB WB Wy Wy»
> PPD) P29) » DIE) » YP ) PD » m m Ty)
©» io» 5 nr me) Dwi» っ WE ’BO) る Hp)
>» MD. DB 2, PIP)? うら DD) » 5 ww)
>) Sm)Dy m WWD)? 2 Diem» Bw PD» » Jren)n)
» I> BYP) RD Win» PD Dim) >a? IE >) > Dis»)
PIP OP? KS Dumm I» V) 3 LD? AD ノ を >»)
PD De ) D>)» 9 Sy ).D > YY) m we)
> うけ ゅ Pdr wD > "MIND »>»> yy a ‚a pnt > >)» ‘> Wo
FB So wm 2 > > „2 mie > ・ » >) { wD mm )
2» em DD PED pp DD ウッ ) I DD DD DAW)
PWD DW »Dn >> WM)» )» PS m) DPD”)
>»? DD >= D> 必須 > » AW D> wm wm)
D> Du @ P 9)? wr ) 1) py» | BD
2 2 PED 2) BD) DP BD vd) DW
» ID? m 5% »)»>»») DP DAW D> DWP
> DP »? D> YD BWYPwD» D)) >> D> DED wi
> DD 2 2 の > yd ın yy) DE $e 0 »5 |
> Dep m) DID Bd? DDr) »] PD 3 . yom 4

SMITHSONIAN INSTITUTION LIBRARIES

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Memoirs of the Science Department, University of Tokio, Japan"

See other formats