A　study　of　microseisms　observed　in　Nagoya　and　its　vicinity

ABSTRACT

　Microseismic　observation　was　carried　out　at　Nagoya　University　during　the　International　Geophysical　Year　1957-58　and　later　periods.The　seismograph　was　deviced　for　this　observation.The　general　nature　of　microseisms　as　well　as　microseismic　storms　accompanying　typhoons　and　cyclones　was　studied.The　comparative　observation　of　microseisms　was　also　performed　at　various　stations　of　the　Nobi　Plain,centering　around　Nagoya.
　The　microseisms　dominant　all　the　year　round　at　Nagoya　University　have　about　3　sec　period　and　3　to　8　μ　amplitude.The　amplitude　in　the　winter　season　is　nearly　twice　as　much　as　that　in　the　summer　normal　period.When　microseismic　storms　culminate,the　period　comes　to　present　the　second　peak　at　5-7　sec　in　the　case　of　cyclone,or　8-9　sec　in　the　case　of　typhoon,in　addition　to　the　first　peak　at　about　3　sec.As　to　amplitude,its　distribution　becomes　so　uniform　that　any　dominant　amplitudes　are　hardly　pointed　out.The　result　of　data　of　several　typhoons　shows　that　many　of　the　observed　variations　of　microseismic　amplitude　and　period　may　be　explained　simultaneously　by　regarding　the　shore　line　near　the　observation　point　as　a　source　area　of　microseisms.The　results　of　comparative　observations　of　microseisms　are　summarized　below.
　The　predominant　period　of　microseisms　in　Nagoya　and　its　vicinity　is　3　to　4　sec　which　is　understood　as　the　foundation　characteristics　representing　a　generalized　geologic　structure　of　the　area.Amplitude　varies　considerably　with　location　of　observation　point,ranging　in　value　from　0.2　to　8　when　the　value　at　Nagoya　University　is　taken　as　unity,value　of　amplitude　is　related　to　the　thickness　of　the　soft　foundation　represented　by　Quaternary　beds.

INTRODUCTION

　Microseismic　observations　have　been　carried　out　as　a　study　in　the　International　Geophysical　Year　program　which　was　commenced　in　July　1957.Part　of　the　results　has　been　already　reported　[8],revealing　value　of　such　observations.Pertaining　to　the　mechanism　of　generation　of　microseisms　under　abnormal　meteorological　conditions,a　tentative　conclusion　has　been　reached.However,microseismic　motions　are　largely　influenced　by　regional　characteristics　of　the　land,and　this　fact　adds　importance　to　local　observations.Ever　since　the　IGY　program　was　started,geophysicists　of　Nagoya　University　have　been　performing　continuous　observation　of　microseisms　and　the　result　has　been　partly　reported　[3].
　The　writers　of　the　present　paper　studied　on（i）general　nature　of　microseisms　as　observed　in　the　Nagoya　district,and（ii）characteristics　of　microseismic　storm　caused　by　meteorological　abnormality　on　the　basis　of　the　one-point　observation　data　obtained　during　the　IGY　period.The　result　of　the　study　disclosed　that　the　data　were　not　sufficient　for　understanding　the　general　nature　of　microseisms　in　this　district.Hence,the　writers　performed　comparative　observations　of　microseisms　in　Nagoya　City　and　the　surrounding　areas.The　result　is　reported　herein.PartI　deals　chiefly　with　the　result　of　one-point　observations　by　Nagoya　University,and　PartII　with　the　result　of　comparative　observations.In　this　paper　the　writers　attempted　to　clarify　the　mechanism　of　generation　and　propagation　of　microseisms.
　The　seismograph　used　in　this　study　was　specially　devised　by　the　senior　writer　for　observation　of　microseisms.Constants　are　listed　in　Table　1.Characteristic　curves　are　shown　in　Fig.1.Fig.2　is　a　photograph　of　the　general　view　of　the　seismograph.

PART　I.CONTINUOUS　OBSERVATION　OF　MICROSEISMS　IN　NAGOYA

1.General　nature　of　microseisms　in　the　district
　Except　for　the　observations　performed　by　us　during　the　IGY　period,no　record　of　systematic　observation　of　microseisms　in　the　region　around　Nagoya　is　known.Thereupon,in　the　first　place　we　had　to　acquire　general　knowledge　of　amplitude　and　period　of　microseisms,as　well　as　their　seasonal　variations,in　the　area　to　be　studied.We　examined　the　records（two　horizontal　components　1　of　1958　for　the　most　part,and　read　the　amplitudes　following　the　IGY　microseismic　observation　regulations.One　point　of　amplitude　for　each　day　was　picked　up　and　a　monthly　running　average　of　vicissitude　was　calculated.The　result　is　given　in　Fig.3　which　may　be　regarded　as　representing　seasonal　changes　in　the　amplitude　of　microseisms.It　is　known　from　the　figure　that　the　amplitude　variation　with　direction　is　not　conspicuous,and　seasonal　or　any　regular　changes　are　almost　unrecognizable.A　similar　fact　is　found　in　the　records　of　observations　in　other　districts.However,it　can　be　pointed　out,if　necessary,that　the　N-S　component　is　slightly　larger　than　the　E-W　component　in　August　and　September.
　The　average　amplitude　for　the　year　is　about　10　μ　which　tells　that　Nagoya　is　one　of　the　areas　of　most　remarkable　microseisms,like　Tokyo,Sakata　and　Tori　shima.Considering　the　fact　that　Tokyo　and　Sakata（Yamagata　Prefecture）are　situated　in　the　geologically　young　sedimentary　plains,the　large　amplitudes　of　microseisms　in　Nagoya　are　not　contradictory　to　the　geology　of　the　neighborhood　of　our　observation　station,consisting　of　Pleistocene　sediments.
　The　amplitude　of　microseisms　is　believed　to　vary　greatly　with　season,attaining　its　maximum　generally　in　winter　and　becoming　smaller　in　summer.Our　case　is　not　exceptional.The　amplitude　as　small　as　7　p　in　June　to　July　becomes　twice　as　much,15　μ,in　December　to　January.The　abrupt　rise　in　the　amplitude　during　the　period　from　August　to　September　may　be　explained　by　the　fact　that　meteorological　abnormality（especially　typhoon）is　most　frequent　during　this　period,giving　rise　to　microseismic　storms.For　comparison,the　data　related　to　the　E-W　components　in　the　records　of　amplitude　in　Tokyo　during　the　same　period　are　plotted　in　Fig.3.The　figure　reveals　the　same　tendency　as　mentioned　above.According　to　the　annual　records,the　maximum　amplitude　is　about　100　μ,seldom　exceeding　this　value　even　in　a　fairly　strong　microseismic　storm.
　The　foregoing　paragraphs　have　dealt　with　the　amplitude.Now,as　to　the　period,annual　changes　are　very　small,being　about　3　sec　throughout　the　year,as　was　already　reported　partially.This　tendency　is　one　of　the　marked　differences　between　Nagoya　and　other　districts.
　Therefore,we　made　somewhat　detailed　researches　on　the　distribution　of　amplitude　and　period.For　the　sake　of　convenience,we　divided　one　year　into　two　seasons,summer（April-September）and　winter（October-March）,referring　to　the　past　results.For　each　season　we　obtained　the　frequency　distribution　by　the　following　method:In　the　summer　season,we　selected　seven　periods　of　two　days　each　during　which　no　microseismic　storm　or　similar　abnormal　phenomenon　took　place（such　period　will　be　called　here　a　normal　time　of　microseisms）,and　microseismic　waves　recorded　in　about　20minutes　of　each　period　were　picked　up.Prior　to　this,we　had　ascertained　that　an　almost　stationary　frequency　distribution　would　be　obtained　by　reading　the　microseismic　waves　for　every　five　minutes　separately　and　then　reading　continuously　for　about　20minutes.Thus,about　3000　waves　in　the　normal　times　of　the　summer　season　were　examined　and　the　frequency　distributions　of　microseismic　amplitude　and　period　were　obtained,as　illustrated　in　Figs.4（a）and　5（a）.As　to　the　winter　season,on　the　other　hand,there　was　no　calm　period　that　lasted　as　long　as　two　days,so　we　had　to　set　the　duration　of　normal　time　at　one　day　each.Reading　of　about　3000　waves　was　performed　likewise.The　result　is　illustrated　in　Figs.4（b）and　5（b）.The　figures　can　be　regarded　as　representing　the　general　characteristics　of　amplitude　and　period　in　normal　times.
　In　these　figures,it　is　noticed　that　the　amplitude　and　the　period　are　symmetrically　distributed　on　both　sides　of　the　dominant　value　in　each　period.In　other　words,both　amplitude　and　period　have　only　one　dominant　value,without　the　second　and　the　third　dominant　values　as　found　in　the　observation　data　of　Tokyo　and　Kyoto　[8].
　Comparison　between　Figs.4（a）and　4（b）reveals　that　the　dominant　amplitude　in　the　summer　normal　time　is　2.5　to　3　μ,I,whereas　that　in　the　winter　season　attains　to　5　μ　nearly　twice　as　much　the　former,as　in　the　case　of　Fig.3.The　upper　limit　of　the　amplitude　is　7　to　8　μ　in　the　former,while　in　the　latter　the　value　becomes　as　large　as　13　μ　Thus,the　amplitude　distribution　is　clearly　different　between　the　summer　and　the　winter　seasons.As　to　the　period,however,the　distribution　shows　good　concordance　between　the　two　seasons　even　in　detailed　points,including　the　values　of　the　dominant　period（2.9　sec　in　summer,2.9　sec　in　winter）,the　upper　limit（4.9　sec　in　summer,5.1　sec　in　winter）,and　the　range（1.3　to　4.9　sec　in　summer,1.3　to　5.1　sec　in　winter）.The　values　larger　than　4　sec　in　winter　are　exceptional,but　these　may　be　ascribed　to　the　duration　of　a　calm　time　which　was　about　one　day　shorter　than　that　of　the　summer　normal　period.These　facts　will　lead　to　a　conclusion　that　the　distribution　of　the　period　during　calm　times　of　microseisms　is　generally　constant　throughout　the　year.
　What　would,then,be　the　cause　of　such　marked　difference　in　amplitude　between　times　as　mentioned　above?　Here　let　us　take　a　look　at　Figs.6（a）and　6（b）which　were　prepared　from　the　data　of　one　day（June　18,1958）when　a　weak　cyclone　was　situated　near　Japan　and　a　cold　front　had　just　passed　through　the　country.With　regard　to　the　period,Fig.6　bears　a　close　resemblance　to　Figs.4（b）and　5（b）,rather　than　to　Figs.4（a）and　5（a）.As　to　the　amplitude,the　resemblance　is　closer　to　the　winter　tendency　than　to　summer,although　Fig.6（b）has　somewhat　larger　values.This　relationship　is　applicable　to　other　less　abnormal　meteorological　conditions.（For　example,refer　to　Section　2　which　discusses　the　relationship　between　cyclones　and　microseisms.）
　Thus,it　may　be　summarized　that　the　influence　of　meteorological　abnormality　upon　microseisms　is　manifested　more　conspicuously　by　the　increase　of　amplitude　than　by　the　elongation　of　period.This　result　agrees　well　with　those　obtained　by　other　investigators　[7,10,11].The　markedly　large　values　of　dominant　and　maximum　amplitudes　in　the　winter　normal　times　are　explained　by　that　the　variation　of　meteorological　abnormality　is　far　greater　in　winter　than　in　summer,hence　the　influence　appears　in　the　amplitude　of　microseisms　during　the　periods　which　we　regarded　as　the　winter　normal　times.
　On　the　basis　of　the　above-mentioned　results,we　can　conclude　that　the　microseisms　dominant　all　the　year　round　at　the　observation　point　in　the　compounds　of　Nagoya　University　have　about　3　sec　period　and　3　to　8　it　amplitude.
2.Microseismic　storms　accompanying　abnormal　meteorological　conditions
　The　foregoing　section　has　discussed　the　results　of　observations　of　the　days　when　meteorological　changes　were　relatively　small,that　is,calm　periods　of　microseisms.It　is　known,however,that　amplitude　and　period　of　microseisms　present　distinct　changes−microseismic　storms—with　approaching　meteorological　abnormality　such　as　typhoon　and　cyclone.Accordingly,if　the　mode　of　changes　of　microseisms　is　studied,in　relation　to　the　kind,scope,character,and　course　of　progress　of　the　meteorological　abnormality,we　may　find　a　clue　useful　for　understanding　the　mechanism　of　generation　of　microseisms.Many　of　the　past　research　works　on　microseisms　are　supposed　to　have　been　performed　from　this　point　of　view.We　shall　proceed　with　our　discussion　from　the　same　view-point.The　data　herein　dealt　with　were　selected　from　among　the　readable　records　of　the　times　of　abnormal　meteorological　conditions　during　the　years　of　1958　through　1960.For　convenience　of　explanation,the　data　were　divided　into　two　portions,one　is　for　the　microseismic　storms　that　are　most　conspicuously　manifested　in　typhoons,and　the　other　is　for　those　accompanying　cyclones.
　I.Microseisms　accompanying　typhoons.—Typhoons　used　in　our　study　are　four（one　in　1958;three　in　1960）,and　their　scope　and　path　are　illustrated　in　Fig.7.Record　reading　was　made　at　intervals　of　12hours,as　represented　in　Fig.7　by　the　black　circles　on　the　path　of　each　typhoon.A　few　examples　of　the　frequency　distribution　thus　obtained　are　given　in　Figs.8　and　9.These　are　the　results　pertaining　to　T.5811　and　T.6014.For　comparison,the　data　of　Fig.4（a）and　Fig.5（a）—distributions　in　the　summer　normal　times—are　plotted　in　the　distribution　diagram　of　T.5811.
　Examining　the　frequency　distribution　of　amplitude　and　period,in　reference　to　the　path　of　typhoons　in　Fig.7,we　find　that　the　distribution　bears　a　close　resemblance　to　that　in　the　summer　normal　times,if　the　location　of　the　typhoon　is　far　away　from　Japan（the　distance　between　the　center　of　the　typhoon　and　the　observation　point　is　greater　than　1600km）.That　is　to　say,any　influence　of　the　typhoon　is　not　recognized　yet.As　the　typhoon　approaches,however,the　amplitude　shows　a　gradual　increase,and　two　maxima　that　are　barely　noticeable　at　900km　distance　expand　markedly　at　400km　distance.Especially,the　rise　of　the　upper　limit　of　amplitude　is　much　greater　than　the　increase　of　dominant　amplitudes.When　microseismic　storms　culminate,the　period　comes　to　present　the　second　peak　at　5-7　sec,in　addition　to　the　first　peak　at　about　3　sec.As　to　the　amplitude,its　distribution　becomes　so　uniform　that　any　dominant　amplitudes　are　hardly　pointed　out.For　easier　understanding　of　this　status,we　prepared　Figs.10（a）and（b）showing　the　maximum　and　the　mean　maximum　values　of　amplitude　and　period　which　were　picked　out　of　12　hour　each　distributions,and　their　relations　to　the　time.The　maximum　amplitude　or　period　represents　the　maximum　values　in　the　read　records,and　the　mean　maximum　is　the　average　of　several　values　before　and　after　the　largest　amplitude　or　period.
　Ever　since　the　theory　of　「stationary　waves」　was　advocated　by　Longuet-Higgins　[5].a　most　prevalent　idea　has　been　that　the　relationship　between　meteorological　abnormality　and　generation　of　microseisms　should　be　ascribed　directly　to　a　change　in　pressure　exerted　on　the　sea　bottom　by　the　waves　caused　by　abnormal　meteorological　conditions.At　present,discussion　of　the　generation　of　microseisms　seems　to　be　focused　rather　on　the　source　of　stationary　waves.With　regard　to　this　subject,there　are　the　following　three　views:
（1）The　central　area　of　a　typhoon　is　where　the　stationary　waves　will　be　generated.
（2）Waves　generated　in　the　central　area　of　a　typhoon　approach　the　coast,and　there　the　stationary　waves　will　be　generated.
（3）Both（1）and（2）are　the　cause　of　stationary　waves.
　No　matter　which　of　the　above　three　views　is　accepted,the　main　point　of　argument　is　laid　on　the　relationship　between　the　location　of　a　typhoon　center　and　the　direction　of　propagation　of　microseisms,and　on　the　time　of　changes　in　the　microseismic　amplitude　and　period　correlated　with　the　time　of　the　records　made　by　the　sea　wave　meters　on　the　coast.Santo　[11,12,13],Okano　[9,10]　and　others　think,on　the　basis　of　the　IGY　data,that　the　view　of（2）is　most　appropriate　as　far　as　the　microseisms　observed　in　Japan　are　concerned.Keeping　this　view　in　mind,let　us　examine　Figs.10（a）and（b）.In　both　amplitude　and　period　the　maximum　value（or　the　mean　maximum　value）increased　roughly　in　proportion　to　the　approach　of　a　typhoon.Increase　in　amplitude　is　more　conspicuous　and　rapid　than　increase　in　period,so　that　even　after　the　amplitude　reached　its　maximum（20th,15h）and　began　to　decrease　the　period　is　still　increasing.Similar　relations　were　reported　by　Dinger　[1]　and　others.However,Okano　[9]　states　that　a　reverse　relation　is　possible.That　is,depending　on　the　typhoon’s　path,the　degree　of　its　development,and　the　location　of　the　observation　point,the　time　when　the　maximum　amplitude　can　be　attained　is　later　than　the　time　when　the　period　reaches　its　maximum.Comparison　of　Fig.10　with　Fig.7　reveals　that　the　time　when　the　microseismic　storm　culminated　is　roughly　coincident　with　the　time　when　the　typhoon　neared　the　observation　point.This　fact　may　be　explained　as　follows:Although　the　extension　of　the　central　area　of　a　typhoon　is　difficult　to　estimate　as　it　is　intricately　related　to　the　central　pressure,the　barometric　gradient,the　travel　time　of　wind,etc.,it　is　believed　to　be　250km　in　radius　when　the　central　pressure　is　over　985mb　and　about　50km　if　the　central　pressure　exceeds　1000mb.Using　these　values,the　size　of　the　disturbance　source　of　typhoon　T　6014　is　assumed　to　be　about　250km,and　that　of　T.6012　is　about　100km.This　indicates　that　when　these　typhoons　were　nearest　to　Japan,the　Pacific　coast（T.6014）or　the　Japan　Sea　coast（T.6011,T.6012）adjacent　to　the　observation　point　was　already　within　the　source　of　disturbance.As　a　result,the　time　when　the　microseismic　storm　reached　its　maximum　would　coincide　with　the　time　the　typhoon　was　nearest　to　the　land.On　the　other　hand,the　intensity　of　microseismic　storms　is　far　greater　in　T.6014　than　in　T.6011,despite　that　the　size　of　disturbance　source　is　nearly　the　same　in　the　two　typhoons.The　path　of　typhoon　T.6011　runs　from　the　southern　tip　of　Shikoku,passes　through　the　Kii　Channel,and　advances　to　the　Japan　Sea,whereas　that　of　typhoon　T.6014　progresses　eastward　along　the　Pacific　coast.Judging　from　these　different　paths,it　may　be　said　that　the　microseisms　observed　in　Nagoya　are　more　conspicuous　when　the　source　of　disturbance　is　located　on　the　Pacific　side　than　when　it　is　found　on　the　Japan　Sea　side.And　yet,it　is　not　known　whether　this　fact　is　attributable　to　the　difference　in　the　distance　from　the　source　of　microseisms　or　to　the　difference　in　the　coastal　topography　as　suggested　by　Okano　[9]　who　believes　that　the　terrain　of　the　Pacific　coast　is　in　favor　of　generation　of　microseisms.In　the　cases　of　T.6011　and　T.6012　which　took　nearly　identical　paths,the　latter　is　of　a　much　larger　scale,and　this　difference　is　clearly　manifested　in　the　amplitude　of　microseisms.
　The　foregoing　paragraphs　have　discussed　chiefly　on　the　amplitude　of　micro-seisms.As　to　the　period,however　its　changes　with　approaching　typhoons　are　generally　very　slight.As　seen　in　Fig.10（b）the　maximum　period　increases　from　6　sec　to　8-9　sec　as　the　typhoon　approaches,but　the　dominant　period　is　still　around　3　sec　which　agrees　quite　well　with　the　dominant　period　in　the　normal　times　of　microseisms.When　the　typhoon　is　of　a　fairly　large　scale　and　is　situated　near　the　observation　point（as　in　the　case　of　T.6014）,another　dominant　period　is　seen　at　about　5　sec　but　it　does　not　show　any　particular　increase　with　approach　of　the　typhoon.
　So　far,the　relationship　between　typhoon　and　microseismic　storm　has　been　studied　using　data　of　several　typhoons.The　result　suggests　that　many　of　the　observed　variations　of　microseismic　amplitude　and　period　may　be　explained　simultaneously　by　regarding　the　shore　line　near　the　observation　point　as　the　source　area　of　microseisms,taking　into　account　the　typhoon’s　scope,location　and　distance　from　the　observation　point.
　II.Cyclones　and　micros　isms.Cyclone　is　another　abnormality　of　meteorological　conditions.Now,let　us　examine　microseismic　storms　that　would　appear　with　passing　cyclones.Path,central　pressure　and　duration　of　the　cyclones　used　in　our　study　are　given　in　Fig.11.The　cyclones　are　denoted　as　LA,LB,LC.....LF,in　order　of　generation.The　five　cases,LA　through　LE,are　when　the　meteorological　conditions　near　Japan　were　relatively　simple　and　the　respective　cyclones　were　the　only　abnormalities,but　LF　is　a　case　of　a　rather　complex　situation　where　several　cyclones　existed　on　both　Pacific　and　Japan　Sea　sides.By　the　same　method　as　before,the　dominant　amplitude　and　period　and　the　maximum　amplitude　and　period　were　obtained　from　their　frequency　distributions.The　results　are　illustrated　in　Fig.12.As　compared　with　the　data　resulting　from　typhoons,the　data　in　Fig.12　indicate　that　changes　are　generally　smaller　in　both　amplitude　and　period,especially　in　period.This　must　be　due　to　that　cyclones　are　much　smaller　in　scale　of　meteorological　abnormality　than　typhoons.
　Figs.11　and　12　tell　that　the　view　on　the　microseisms　accompanying　typhoons　as　mentioned　in　the　foregoing　section　can　be　applied　to　LA,LB,LD　and　LE.In　other　words,by　taking　into　account　the　cyclone’s　location,scope　and　its　distance　from　the　observation　point,we　can　explain　rise　and　fall　of　the　observed　microseismic　storms　On　the　other　hand,the　above　view　does　not　seem　applicable　to　LC　and　LF.For　example,in　the　case　of　LC　in　Fig.11,even　when　the　center　of　the　cyclone　was　nearest　to　Japan（Nov.2,6　h）it　was　still　at　a　distance　as　600km　north　of　the　Japan　Sea　coast,and　yet　the　microseismic　storms　attained　the　culmination　within　several　hours.The　central　pressure　of　this　cyclone　was　998mb,consequently　the　extension　of　the　disturbance　source　must　have　been　less　than　150km　in　radius,and　the　velocity　of　waves　coming　from　the　source　less　than　25km/h.At　this　velocity,it　would　take　at　least　24hours　for　the　waves　to　reach　the　coast　of　the　Japan　Sea.Thus,the　source　of　microseisms　cannot　be　situated　near　the　coast.In　this　case,to　regard　the　cyclone’s　central　area　as　the　source　of　microseisms　would　lit　better.As　to　the　case　of　LF,Fig.12,we　cannot　take　it　as　a　single　abnorinality.In　fact,there　were　complex　changes　in　amplitude.
　At　any　rate,the　maximum　amplitude　of　microseisms　resulting　from　cyclones　would　seldom　exceed　30　μ.The　maximum　period,7　sec,is　somewhat　smaller　than　the　value　of　8-9　sec　in　the　case　of　typhoons.It　is　interesting　to　note　that　in　each　of　LA　through　LF　the　dominant　period　is　always　about　3　sec　which　is　similar　to　that　of　the　normal　times　of　microseisms.
3.Discussion
　Since　Longuet-Higgins　his　theory　of　stationary　waves　has　been　regarded　as　most　relevant　for　explaining　the　generation　mechanism　of　microseisms.One　of　the　major　reasons　for　supporting　his　theory　is　that　the　theoretically　obtained　one　to-two　relation　of　the　microseismic　period　to　the　sea　wave　period　would　explain　quite　well　the　observed　facts.According　to　this　theory,when　the　dominant　period　observed　in　Nagoya　is　3　sec　the　dominant　period　of　sea　waves　should　be　6　sec.The　location　of　the　source　of　microseisms　present　another　problem.In　section　2,we　tentatively　called　it　a　coastal　area,but　in　our　case　it　is　still　unknown　whether　the　source　is　situated　within　Ise　Bay　or　somewhere　in　the　outer　sea.
　In　May　1962,a　pressure　type　sea-wavemeter　was　installed　in　Ise　Bay　by　the　Ministry　of　Construction,which　enabled　simultaneous　observations　of　microseisms　from　August　to　September　at　a　point　near　Nagoya　Harbor　in　the　inner　part　of　the　bay　and　at　the　Institute　of　Earth　Sciences,Nagoya　University.Fig.13　shows　an　example　of　distribution　of　microseismic　period　as　observed　at　these　two　points　and　of　period　of　sea　waves.Sea　waves　inside　the　bay　are　very　weak　all　the　year　round,with　amplitude　seldom　exceeding　50cm　even　under　considerably　abnormal　conditions.Hence,we　are　unable　to　make　a　comparative　study　of　microseismic　waves　and　sea　waves　for　a　long　duration.The　example　given　in　Fig.13　is　the　case　when　typhoon　T.6212　passed　through　the　southern　coast　of　the　Pacific　side　and　the　amplitude　of　sea　waves　attained　to　several　tens　of　centimeters.The　period　of　sea　waves　is　dominantly　12-13　sec,much　larger　than　6　sec.A　similar　pattern　of　sea　wave　period　was　observed　in　another　typhoon,T.6214,of　the　same　year.Also,the　amplitude　of　sea　waves　is　generally　small　inside　the　bay.These　facts　suggest　that　at　least　the　sea　waves　inside　the　bay　do　not　contribute　much　to　the　microseismic　activity.The　difference　in　microseisms　between　the　Institute　of　Earth　Sciences　of　Nagoya　University　and　Nagoya　Harbor　will　be　discussed　in　Part　II.Oceanographic　information　tells　us　that　the　mean　period　of　sea　waves　around　Japan　is　8-10　sec,which　is　supposed　to　be　the　value　at　normal　times.The　period　of　sea　waves　under　abnormal　meteorological　conditions　is　dominantly　8-15　sec,with　maximum　period　attaining　to　15-20　sec,considering　from　the　wind　velocity（20-50　knots）in　the　central　area　of　abnormality.
　In　the　results　of　our　microseismic　observations,however,the　dominant　period　is　still　around　3　sec,although　the　maximum　period　shows　a　distinct　increase　with　the　approach　of　the　abnormality.Thus,as　far　as　our　results　are　concerned,the　one　to-two　ratio　relation,i.e.,the　1/2　period　theory　may　not　be　applicable.
　According　to　Ewing　et　al.[2],in　discussion　of　microseisms　it　is　desirable　to　consider　the　following　factors　separately:（a）nature　of　the　source,（b）mode　of　propagation　in　oceanic　paths,and（c）mode　of　propagation　in　continental　paths.In　our　case　assuming　a　source　of　microseisms　near　the　coast,factor（b）can　be　excluded.That　is,we　are　to　consider　the　nature　of　the　source　and　the　mode　of　propagation　of　microseisms　in　the　inland　area.
　Denoting　the　spectrum　of　sea　waves　in　the　source　of　microseisms　as　Y_1（ω）,the　land　characteristics　of　the　path　of　propagation　as　Y_2（ω）,and　the　characteristics　of　the　seismograph　as　Y_3（ω）,the　spectrum　of　the　microseisms　we　observe　is　expressed　as　Y（ω）＝　Y_1（2　ω）・Y_2（ω）・Y_3（ω）.In　Y_1　＝　Y_1（2ω）the　theory　of　Longuet-Higgins　was　applied　to　the　relation　between　the　period　of　sea　waves　and　that　of　microseisms.Even　if　Y_3（ω）was　ignored,we　would　get　Y（ω）＝　Y_1（2ω）・Y_2（ω）,not　Y（ω）＝　Y_1（2ω）.That　is　to　say,unless　the　land　characteristics　Y_2（ω）cover　a　wide　range,the　relation　of　the　period　of　observed　microseisms　to　the　period　of　sea　waves　is　not　necessarily　1:2.In　view　of　the　propagating　direction　and　the　result　of　orbit　analysis,microseisms　are　regarded　as　the　Rayleigh　type　surface　waves.If　so,Y_2（ω）should　have　the　band　pass　characteristics　related　to　geological　structure,and　could　never　be　flat,Accordingly,there　can　be　cases　where　the　characteristics　of　Y_2（ω）surpass　those　of　Y_1（2ω）,so　that　the　pattern　of　the　observed　microseismic　period　is　reflecting　Y_2（ω）for　the　most　part.
　If　the　predominance　of　3　sec　microseisms　in　our　case　was　thus　interpreted,the　above-mentioned　results　of　observations　of　the　period　of　sea　waves　and　microseisms　would　indicate　merely　that　the　influence　of　sea　waves　inside　Ise　Bay　on　microseisms　is　negligible,and　would　not　offer　any　counter-evidence　against　the　1/2　period　theory.
　In　order　to　separate　the　influence　of　the　land　characteristics　on　amplitude　and　period　of　microseisms　from　the　source　characteristics　in　the　disturbance　source,the　observation　results　of　only　one　point　and　of　a　given　duration　of　IGY　are　far　from　sufficient.Therefore,we　later　carried　out　comparative　observations　of　microseisms　from　1961　to　1962,and　achieved　our　object　to　a　certain　extent.The　results　are　reported　in　PartII.

PART　II.COMPARATIVE　OBSERVATIONS　OF　MICROSEISMS　IN　NAGOYA　AND　ITS　VICINITY

1.Location,duration　and　method　of　observation
　From　the　results　of　our　observations　of　microseisms　during　the　IGY　and　later　periods,we　learned　that　microseisms　of　about　3　sec　in　period　and　5μ　in　amplitude　are　dominant　all　the　year　round,and　understood　the　nature　of　microseismic　storms　under　abnormal　meteorological　conditions.Since　these　are　the　results　of　one-point　observations　at　the　Institute　of　Earth　Sciences,Nagoya　University,during　the　above-mentioned　periods,we　cannot　assert　that　they　are　representing　the　characteristics　of　the　entire　area　of　the　Nobi　Plain.On　the　other　hand,the　land　characteristics　of　the　site　of　observation　may　account　for　the　easy　occurrence　of　3　sec　microseisms,as　has　been　briefly　mentioned　in　PartI.However,in　this　respect　also,we　should　not　make　a　hasty　conclusion.Thereupon,we　decided　to　carry　out　comparative　observations　of　microseisms.Of　the　two　horizontal　component　seismographs,one　was　installed　at　the　Institute　of　Earth　Sciences　to　he　used　as　a　key　station.and　the　other　was　used　in　mobile　observations　at　various　points　of　the　Nobi　Plain,centering　around　Nagoya.Some　typical　examples　of　the　records　of　microseisms　obtained　are　shown　in　Fig.14　together　with　those　used　in　PartI.Comparative　observations　were　performed　from　1961　to　1962　at　the　stations　listed　in　Table　2.Duration　of　observation　at　one　station　was　limited　between　2　and　4　weeks,which　is　supposedly　long　enough　for　our　purpose.As　stated　before,one　of　the　two　horizontal　component　seismographs　having　the　same　characteristics　was　used　in　fixed　point　observation　and　the　other　was　used　in　mobile　observation.For　the　sake　of　convenience　the　components　were　unified　in　a　N-S　component,since　it　is　known　that　the　variation　of　microseismic　amplitude　with　direction　is　not　very　great　in　this　district　as　already　mentioned　in　Part　I.Distribution　of　observation　stations　and　geological　outline　of　the　Nobi　Plain　are　shown　in　Fig.15.
　In　analyzing　the　data　the　same　method　as　the　one　stated　in　Part　I　was　employed.The　only　difference　is　that　the　values　at　the　fixed　station　of　the　Institute　of　Earth　Sciences,Nagoya　University,were　used　as　the　standard　for　comparison　between　the　values　obtained　at　other　stations.
2.Result　of　observation　and　discussion
　Fig.16　shows　some　of　the　amplitude　and　period　distributions　obtained　at　various　stations　by　the　same　method　as　in　Part　I.The　distribution　at　the　key　station　during　the　same　period　is　also　given　in　the　figure.These　records,except　that　of　Minato^1,are　generally　of　calm　times　of　microseisms.
　So　far　as　our　observations　go,the　period　of　microseisms　is　dominantly　3　to　4　sec,seldom　at　higher　or　lower　values.Of　the　above-mentioned　nine　observation　stations,the　dominant　period　at　Nagoya　University,Ichinomiya　and　Gifu　was　rather　small,being　2.7　to　3.3　sec,but　all　the　other　stations　recorded　the　dominant　period　at　nearly　4　sec.Putting　aside　the　4　sec　period　at　Seto,we　can　say　that　4　sec　period　is　dominant　generally　in　the　area　near　the　Nagoya　Harbor（Minato）,whereas　3　sec　is　the　dominant　period　in　the　area　farther　inland　from　Ise　Bay.Nevertheless,variation　of　microseismic　period　with　geographic　distribution　of　observation　stations　is　very　small　and　cannot　furnish　information　useful　for　reasonable　interpretation　of　its　relation　to　the　ground　characteristics.
　On　the　other　hand,variation　of　amplitude　with　location　is　fairly　conspicuous,the　value　occasionally　attaining　to　ten　times　as　much　that　of　the　standard　value　at　the　key　station　in　Nagoya　University.Referring　to　Fig.16　the　dominant　periods　and　amplitudes　at　the　respective　observation　stations　are　summarized　in　Table3.
　As　the　table　indicates,variation　of　characteristics　with　location　of　obser-vation　station　is　more　conspicuous　in　the　amplitude　than　in　the　period.At　stations　in　the　southern　part　of　the　district,namely,Tobishima,Kuwana,Shiratori,Ishin　and　Minato,the　values　of　dominant　amplitude　are　very　large,three　to　eight　times　as　much　as　that　at　the　key　station,whereas　the　value　at　Ichinomiya　is　roughly　the　same　as　that　of　the　key　station,being　about　1.0　to　1.2　times　the　latter.The　value　becomes　smaller　at　Seto（0.66-0.8　times）,and　the　smallest　at　Gifu（0.2-0.4　times）.In　other　words,amplitude　is　greater　in　the　south　and　decreases　northward.Considering　that　the　main　source　of　microseisms　recorded　in　this　district　is　located　near　the　Pacific　coast　in　the　south,the　above　values　may　suggest　a　correlationship　between　the　increasing　distance　from　the　source　and　the　decreasing　amplitude　of　microseisms.On　the　other　hand,in　Fig.15　showing　the　location　and　geology　of　the　observation　stations　we　know　that　almost　all　stations　in　the　southern　area　stand　on　the　Holocene　beds,while　Gifu　and　Seto　are　in　the　area　of　Paleozoic　strata　or　igneous　rocks.Thus,a　correlationship　can　be　assumed　also　between　the　hardness　of　foundation　and　the　amplitude　of　microseisms.In　order　to　find　out　which　correlationship　is　more　effective,we　prepared　a　diagram,Fig.17,showing　the　relation　between　distance　and　relative　amplitude,taking　two　directions,i.e.,NE-SW　line　connecting　Seto,Nagoya　University　and　Tobishima,and　N-S　line　connecting　Gifu,Ichinomiya　and　Tobishima.In　the　NE-SW　direction,the　amplitude　attains　a　peak　at　Tobishima,southernmost　station,and　decreases　rapidly　northward,in　the　order　of　Ishin,Minato,Shiratori　and　Nagoya　University.Between　Nagoya　and　Seto,the　decrease　in　amplitude　is　very　small.This　is　applicable　to　the　N-S　line,too,where　the　amplitude　shows　a　rapid　change　between　Tobishima　and　Ichinomiya　while　the　change　is　very　small　between　Ichinomiya　and　Gifu.
　If　these　facts　should　be　interpreted　as　attenuation　of　surface　waves　according　to　distance,as　has　been　generally　believed,the　source　of　the　microseisms　must　be　situated　very　close　to　the　observation　point（such　as　the　innermost　part　of　Ise　Bay）,and　must　be　a　very　local　one,too.However,as　already　pointed　out　in　Part　I,the　result　of　our　comparative　observations　of　microseisms　and　sea　waves　showed　that　the　sea　waves　in　the　innermost　part　of　the　bay　cannot　become　a　source　of　microseisms.And,if　the　decrease　in　the　amplitude　is　explained　as　surface　waves’　attenuation　with　distance,the　period　must　become　longer　with　increasing　distance　from　the　source.Our　observations　revealed,however,a　reverse　relation,that　is,the　period　becomes　shorter　inlandward.According　to　Okano　[8],an　area　where　microseisms　can　be　generated　is　limited　roughly　in　the　near-coast　sea　of　rather　dense　isobathic　contours.
　It　is　difficult　to　know　an　exact　location　of　the　source　of　the　microseisms　we　observed,but　the　afore-said　facts　suggest　at　least　that　the　source　is　somewhere　near　the　coast　at　a　distance　somewhat　larger　than　that　between　the　observation　stations.
　Thus,in　both　Tobishima-Nagoya　University-Seto　line　and　Tobishima-Ichinomiya-Gifu　line,the　correlationship　between　the　amplitude　characteristics　and　the　distance　from　the　source　of　microseisms　becomes　secondary,so　that　the　effect　of　the　nature　of　the　medium　of　the　path　of　propagation　of　microseisms,i.e.,ground　characteristics,is　considered　the　primary　one.Shida’s　study　[14]　seems　to　follow　this　viewpoint.
3.Geologic　structure　and　microseisms
　In　an　attempt　to　clarify　the　characteristics　of　the　microseisms　in　this　district,the　observed　facts　so　far　described　must　be　fully　elucidated.That　is,we　must　find　an　explanation　that　would　simultaneously　satisfy　the　following　facts:3　to　4　sec　periods　are　predominant,and　microseisms　of　4　sec　period　are　abundant　in　the　southern　area;the　amplitude　varies　greatly　with　location　of　observation　station,and　the　value　becomes　larger　southward.Now,let　us　study　these　problems.
　As　has　been　mentioned　in　the　last　section　of　PartI,the　spectrum　of　the　microseisms　we　are　to　observe　is　expressed　as　Y（ω）＝　Y_1（2ω）・　Y_2（ω）・Y_3（ω）.As　correction　was　made　for　the　seismograph’s　characteristics,Y_3（ω）,the　spectrum　can　be　represented　by　Y（ω）＝　Y_1（2ω）・Y_2（ω）.Of　the　values　to　be　used,direct　data　of　the　spectrum　of　sea　waves　denoted　as　Y_1（2ω）are　very　few,but　refer-ring　to　the　analytical　results　of　sea　waves　at　various　parts　of　the　country　during　the　IGY　period　we　get　a　comparatively　flat　distribution　of　spectrum,although　8-12　sec　periods　are　generally　dominant.Thus,Y_1（2ω）is　considered　to　cover　a　fairly　broad　area.There　are　cases　where　dominant　period　moves　toward　a　longer　period　with　the　approach　of　meteorological　abnormality,and　occasionally　exceeds　20　sec.However,in　the　result　of　our　observation　of　period　of　microseisms,the　microseisms　of　3　to　4　sec　were　dominant　with　or　without　existence　of　abnormal　meteorological　conditions.This　suggests　that　Y_2（ω）is　closely　related　to　the　period.
　Pertaining　to　the　underground　structure　of　Nagoya　and　vicinity,there　is　a　study　by　Iida　and　Aoki　[4]　based　on　near　earthquakes.According　to　this　study,the　following　two　models　are　equally　possible:
　表：Model　I・Model　II
　Considering　from　the　location　of　observation　stations　whose　data　were　used　in　the　analysis,the　above　models　are　regarded　as　representing　the　under-ground　structure　of　the　area　around　Nagoya　City.On　these　models,the　amplitude　versus　period　characteristics　of　Rayleigh　waves　can　be　calculated,but　a　method　of　calculation　for　a　three-layered　structure　has　not　been　established　yet.Hence,we　carried　out　calculation　by　assuming　the　structure　to　be　two-layered.In　calculation　we　adopted　the　idea　of　Tazime　[15]　i.e.,the　dominant　period　of　surface　waves　can　be　obtained　by　the　1/4　wave　length　law.If　the　first　and　the　second　layers　were　grouped　together　as　one　layer,the　resultant　period　would　become　about　8　sec　which　is　too　large　as　compared　with　the　known　dominant　value　of　3　to　4　sec.Therefore,we　used　the　first　and　the　second　layers　only,ignoring　the　third　layer,for　the　reasons　that　the　velocity　contrast　between　the　first　and　the　second　layers　is　considerably　large　and　the　boundary　between　the　second　and　the　third　layers　is　at　a　great　depth　of　9.3km　from　the　ground　surface.Thus,we　obtained　the　period　of　2.6　sec　in　both　models.
　From　the　result　of　our　calculation　on　the　Rayleigh　waves　for　a　two-layered　structure,the　amplitude　versus　period　characteristics,Y_2（ω）,and　the　dispersion　curves　of　the　cases　closest　to　the　above-mentioned　two　models　were　selected　and　diagramed　in　Fig.18.As　is　seen　in　the　diagram　of　calculated　models,both　model　I　and　model　II　are　closer　to　model　[A_2]　than　to　model　[A_1].In　either　of　[A_1]　and　[A_2],dominance　of　the　amplitude　is　found　at　about　3　sec　of　period.Especially　model　[A_2]　seems　to　present　the　characteristics　of　a　fairly　distinct　band　pass　centering　around　3　sec.
　Both　model　I　and　model　II　have　some　defects　which　prevent　us　to　build　a　decisive　conclusion,so　any　further　discussion　on　Y_2（ω）would　be　meaningless.However,the　calculated　result　may　present　an　interpretation　of　the　fact　that　3-4　sec　microseisms　are　apt　to　dominate　in　the　Nagoya　district.
　Next,we　must　think　about　the　difference　in　microseismic　amplitude　between　observation　stations.This　problem　reminds　us　of　Shida’s　study　[14].Through　the　multi-points　observations　in　the　Shonai　district,Shida　pointed　out　a　close　relationship　between　the　amplitude　of　microseisms　and　the　strength　of　foundation　which　will　be　known　from　observations　of　ground　vibrations.This　relationship　suggests　that　the　subsurface　weak　beds　would　affect　microseismic　waves　more　strongly　than　expected,despite　the　fact　that　the　wave　length　of　microseisms　is　much　longer　than　that　of　ground　vibrations.In　order　to　ascertain　this,observation　of　microseisms　must　be　performed　in　an　area　of　known　underground　structure.
　Details　of　underground　structure　of　the　whole　area　of　Nobi　Plain　are　not　clarified　as　yet,but　fortunately　the　results　of　numerous　borings　down　to　the　depth　of　several　hundred　meters　are　available　for　a　line　connecting　our　observation　stations　between　Nagoya　University　and　Kuwana,via　Tobishima.Based　on　these　boring　records,Matsuzawa　[6]　prepared　a　geologic　cross-section.According　to　the　outline（Fig.19）of　this　cross-section,with　an　exception　of　the　observation　station　at　Nagoya　University　which　is　situated　near　the　base　of　Diluvium,all　other　stations　such　as　Shiratori,Minato,Ishin　and　Tobishima　stand　on　the　Alluvial　beds,and　the　Quaternary　beds（Alluvium　plus　Diluvium）become　thicker　almost　linearly　in　the　above-mentioned　order.Only　Kuwana　stands　on　the　Tertiary　foundation　which　is　in　fault　contact　with　the　Quaternary　beds.
　By　comparing　the　crosssection　with　the　distribution　map　of　relative　amplitude　of　microseisms　in　Fig.17（a）,a　very　close　relationship　is　found　between　the　two.That　is　to　say,the　increase　in　the　amplitude　of　microseisms　is　roughly　proportional　to　the　thickness　of　the　Quaternary　strata.In　other　words,assuming　the　direction　of　propagation　of　microseisms　at　nearly　right　angle　to　the　cross-section,the　difference　in　the　amplitude　between　Nagoya　University　and　other　stations　south　of　Shiratori　is　largely　attributed　to　the　existence　of　the　Quaternary　soft　beds.
　For　quantitative　examination　of　this　relationship,the　amplitude　characteristics　of　the　Rayleigh　waves　in　the　three-layered　structure　must　be　known,as　in　the　case　mentioned　before.However,such　a　task　is　very　difficult　at　the　present　stage　of　our　knowledge.Thereupon,considering　that　the　characteristics　of　Rayleigh　waves　are　most　strongly　affected　by　the　interlayer　velocity　of　S　waves　and　that　our　microseismic　observations　are　being　made　on　the　horizontal　components,we　attempted　to　detect　the　effect　of　soft　beds　on　the　amplitude.As　to　the　values　of　S　waves　the　results　by　Iida　and　Aoki　[4]　were　referred　to,as　before.In　addition,as　the　uppermost　layer　a　layer　of　small　velocity　was　assumed.Thus,we　obtained　models　[B]　and　[C].The　result　is　given　in　Fig.20.The　figure　shows　that,if　the　thickness　of　the　uppermost　layer　is　definite,amplification　of　amplitude　becomes　larger　with　decreasing　velocity　of　S　waves　in　this　layer,and　if　the　velocity　of　S　waves,V,in　the　layer　is　definite,the　amplitude　increases　with　the　layer’s　thickness.At　the　same　time,the　dominant　period　increases,more　so　in　[B].The　result　may　not　be　allowed　to　substitute　the　amplitude/　period　characteristics　of　surface　waves,but　it　becomes　known　at　least　qualitatively　that　the　surface　amplitude　increases　markedly　with　increasing　thickness　of　soft　beds,and　the　period　also　increases　but　not　so　remarkably　as　amplitude　does.This　result　is　not　contradictory　to　the　inclination　revealed　in　the　result　of　our　observations　of　microseisms　in　the　Nagoya　University-Tobishima　line.For　discussion　of　the　same　subject　on　other　observation　stations,we　do　not　have　sufficient　information　of　under-ground　structure,so　a　hasty　conclusion　should　not　be　reached.Nevertheless,it　may　be　reasonable　to　apply　the　same　theory　to　other　stations.This　theory,however,cannot　offer　a　satisfactory　explanation　of　the　amplitude　at　Kuwana　and　the　period　at　Seto.Despite　the　fact　that　Kuwana　is　situated　on　the　Tertiary　foundation,the　amplitude　is　fairly　larger　than　at　Nagoya　University.But,considering　it　from　a　different　angle,we　may　be　able　to　say　that　the　amplitude,which　rapidly　increased　with　increasing　thickness　of　the　Quaternary　beds　in　the　Nagoya　University-Tobishima　line　showed　a　sudden　decrease　in　the　Tobishima-Kuwana　line,is　reflecting　a　large　fault　supposedly　existing　between　Kuwana　and　Tobishima.With　regard　to　the　period　at　Seto　we　cannot　say　anything　decisive　until　more　precise　observations　are　made　in　the　area　between　Nagoya　and　Seto.

CONCLUSIONS

　As　the　result　of　our　researches　so　far　reported,we　have　learned　the　general　characteristics　of　microseisms　observed　in　Nagoya　and　the　surrounding　areas　during　the　International　Geophysical　Year　1957-1958　and　later　periods.The　results　are　summarized　as　follows.
　The　microseisms　dominant　all　the　year　round　at　Nagoya　University　have　about　3　sec　period　and　3　to　8　μ　amplitude.The　amplitude　in　the　winter　season　is　nearly　twice　as　much　as　that　in　the　summer　normal　time.The　result　of　data　of　several　typhoons　shows　that　many　of　the　observed　variations　of　microseismic　amplitude　and　period　will　be　explained　simultaneously　by　regarding　the　shore　line　near　the　observation　point　as　a　source　area　of　microseisms.
　From　the　analysis　of　comparative　observations　the　predominant　period　of　microseisms　in　Nagoya　and　its　vicinity　is　found　to　be　3　to　4　sec　which　is　understood　as　the　foundation　characteristics　representing　a　generalized　geologic　structure　of　the　area.Amplitude　varies　considerably　with　location　of　observation　point,ranging　in　value　from　0.2　to　8　when　the　value　at　Nagoya　University　is　taken　as　unity,value　of　amplitude　is　related　to　the　thickness　of　the　soft　foundation　represented　by　Quaternary　beds.By　advancing　the　conclusion　in　the　paragraph　above　mentioned,we　may　be　able　to　know　the　hardness　of　foundation　or　the　extent　of　soft　foundation　through　multi-points　observation　of　microseisms.

ACKNOWLEDGEMENTS

　The　writers　wish　to　express　their　heartiest　thanks　to　the　staffs　of　schools　and　institutions　concerned,who　gave　them　a　ready　consent　to　the　use　of　the　facilities.To　Mr.Kitajima,chief　of　the　Construction　Section,Ise　Bay　Harbor　Construction　Department,who　permitted　the　writers　to　utilize　the　records　of　sea　waves,our　thanks　are　also　due.The　writers　are　indebted　to　Messrs.Yoshio　Osada,Toshimi　Kitano　and　Akio　Kato,for　their　cooperation　in　record　reading　and　other　troublesome　works.

REFERENCES

[1]　DINGER,J.E.,Discussion　of「Can　sea　waves　cause　microseisms.」　Symposium　on　microseisms:Natl.Acad.Sci.,Natl.Research　Council.,Publ.No.306,87-92（1953）.
[2]　EWING,W.M.,WENCESLAS,S.J.,and　FRANK,P.,Elastic　waves　in　layered　media:McGraw-Hill　Book　Co.,New　York（1957）.
[3]　IIDA,K.,and　WADA,T.,Microseisms　observed　at　Nagoya　University.Preliminary　report:Read　at　the　Seismological　Society　of　Japan（1957）.
[4]　—,and　AOKI,H.,Phases　in　the　preliminary　tremors　of　seismograms　and　crustal　structure　in　Nagoya　district（in　Japanese）:Zisin（J.Siesm.Soc.Japan）,2nd　Ser.,12,75-81（1959）.
[5]　LONGUET-HIGGINS,M.S.,A　theory　of　the　origin　of　microseisms:Phil.Trans.Roy.Soc.Am.,A　243,1-35（1950）.
[6]　MATSUZAWA,I.,The　geologic　structure　of　the　northern　coastal　region　of　Ise　Bay（in　Japanese）:Ministry　of　Construction,pp.4（1962）.
[7]　MATUZAWA,T.,ASADA,T.,DEN,N.,and　TAKANO,K.,Tripartite　observations　of　microseisms:Japan.J.Geophys.,3,1-74（1962）.
[8]　National　Committee　for　the　International　Geophysical　Year,Report　of　microseismic　and　sea　wave　observations　in　Japan　during　the　International　Geophysical　Year　1957-1958:Science　Council　of　Japan,Tokyo（1959）.
[9]　OKANO,K.,Microseisms　observed　at　Abuyama　Seismological　Observatory（in　Japanese）:Zisin（J.Seism.Soc.Japan）,2nd　Ser.,2,182-190（1959）.
[10]　—,Direction　of　approach　of　microseisms　observed　by　vector　seismographs（in　Japanese）:Zisin（J.Seism.Soc.Japan）,2nd　Ser.,13,37-42（1960）.
[11]　SANTO,T.,Investigations　into　microseisms　using　the　observational　data　of　many　stations　in　Japan（Part　1）.On　the　origin　of　microseisms:Bull.Earthquake　Research　Inst.,Tokyo　Univ.,37,307-325（1959）.
[12]　—,Investigations　into　microseisms　by　the　observational　data　of　many　stations（Part　2）.Further　considerations　on　the　origin　of　microseisms:Bull.Earthquake　Research　Inst.,Tokyo　Univ.,37,483-494（1959）.
[13]　—.The　observation　of　microseisms　at　a　wave　gauge　station.On　the　origin　of　microseisms（Part　3）:Bull.Earthquake　Research　Inst.,Tokyo　Univ.,38,241-254（1960）.
[14]　SHIDA,I.,Relationship　between　the　microseisms　observed　in　Shonai　district　and　its　ground　condition（Part　2）（in　Japanese）:Publ.Ground　Invest.Group　Shonai　District,1-10（1960）.
[15]　TAZIME,K.,Relation　between　dispersive　surface　waves　and　the　nature　of　the　ground（in　Japanese）:Geophys.Exploration,9,58-62（1956）.