TRUぬthematics 24−1 〔1988) TlIE I・旺…IGHT AT.TI田FIXED POINT ・. ..’一・「・ ..・・.
ON THE LINEARIZBD GROWTH CURVE IN THE SECOND PHASE
M)tosaburo MASUYAMA 〔胎ceW麺岬8・、・988・…ise輌・i・・6・i988)「. .1°Th・b・x・d it・鵬・n the・n・xt・P・g・i・・卿t・d f・・m・thg r・ference[5]with・・m・ modifications. 「 ’ . 二 L・t・汐一hirt2.・be・h・.belgh・.・f・h・Z・th i㎡iWa・・垣S. r寧唾・gica・ 、 age t in the second phase、〔the adolescent period). Then the coefficient of variation (CV) is about 4%. Hence the equations 〔1). and (2〕.hold [2i3]. Th・c・rre1・・i・rt・・eξfi・ig・・〔CC)・b・tyeep・i・nq’・bi・i・一゜・998・’°τ1ess[4L ・・th・t th・・e1・ti・n・(3)and(4〕・with C.r D・h・1d叩pr9麺t・1y・寧r・el・ti。・ (3〕 is verified graphically in. Fig. 1 G 2, based on the National Sam1)1e Survey as of 1983 [7L .』』』 ’ : . 』 By definition, X.(t) =》z. rt/ must be.・estimated旬the longitudina1..survξツ, bu…U・e the cr・ss−r・9?i・na・、・ne・’s?.・h・…he・・t・・d1妬迦γh卸C.’…11 biases of order O.ヱ rcm). − Assuming the strict cQnvgrgence of growth lines .(2),we set shrto)=Oat . the fixed point rら σノ,砲ich implies . . . . ’.(5) .h・rカ。ノ1−・ノ・わ・ 、 、.
2°Acc・rding t・th・S・rv・y[7],we・bt・in ’
『2.rカ ノ = 198.80 ± ヱ.94 rαηノ ‘ ・ .. 』 』 』、 O f・rthe rang・・3・ヱ7 ry・・ノ・f・b・y・・Th・CC b・tab・en・h(t)・aPd・h・rt.ノi・equa1 t・ −0.9949. And . ..』. . .、 、.、! 、.hT 「t。∼.=..189・9°,1 1・57・「・川、. t.t…. f°「the「ange 11∼ヱ7「卿ノ.°f gi「1s・The. CC bet鳩en%「麺d.h・似lis:equal to −0.9954● ・・ 、 These two estimates are too large con唖)ared with the regression esti皿ates ・ obtained by the pilot study, using the relation (4). The discrepancy seems to haVe arisen.. from the assumption of the.strict eonvergence of growth 1.ines..一・. ・ ・− Nbte:The obseIved.CV 6f height at the end’of’growth is abOut− ’3%∫whi:ch’is.very. small compared with the alnong−person SD of X=ZoθC,砲ere C denotes the 7576
M.MASUYAMA
THE ALMOST−ONE FARAM1正R H¥POTFESlS THE BIoLoGlcAL PRocEss xrt) I s PRIMITIVE, OR FUNDA岡EN「AL FOR THE LIFE. −t・ (IF NECESSARY, Xrt) I S TRANSFORMED IN ADVANCE IMO A NEW VARIA丁E, WH!CH IS DEN(汀ED旨AGAIN BY.xrt).) THE INDIVI肌、W駅1旭IUW OF聖「幻 FOR i = 1, 2⊃ … 診 」V IS SUFFICIENTLY SMA山WH‖…RE’nE VARIAB!し『Y IS REPRESE.g『ED BY A D!惰ENS!ONUESS MEASURE. ヒT・i、一駐距純P艀田TI拒(},. ’1−}−tE…i−ru lNDIVI㎜⊇噸bz lS A PdslTlvE DI“,ENSIONLESS・coNsTANT, THEN’ − yrt) BElNG ITS 窟…FERENCE P撤)CESS, M…HAME ・(1).. dAriノ’dei.r dY/dち . .. THIs MEANs THAT THE HEcHANIsH oF REGし‖_ATI.qN FOR THE VELOCITY IS IDFNT耳(礼一 Ii:Hls’戸RdPER TIME IS.usED, BY;㌔一’ . INTEGRATION,紺…OBTAIN c∼) Xi rt)=ai+birrt)・ IF N三SE丁 yrt) =X・(t), WE OBTAIN NATURAL ODND1TIONS α・ = O AND 1)・ = ヱ, THE DoT DENσTEs Tl・fE陸…AN wlTH REsPEcT コ TO W{… INDEX t. ONE (GENoTIC) FAcToR Is PREDoMINENT IN τH旺… PROCESS. . PARAME丁ERS CONTAINEI〕 IN τHE EXPLICIT .F(濃H OF.Xi rち, ARE M」丁U創]LY DEPENDENT. THE(紬ali.ATIdw COEFドICI白πρBE聴団 ・t’.Bb bi IS’EQUAL’Tσ一11”.−. THERE EXIST TWO CONSTANTS σAND・D SUCH THAT (4) αi=c一の〃 ORτ十‖…RE EXISTS A F!XED POI↑『「 r刀」 CJ ON ・’嘯gE LINEARI之ED cuRvE.(2).★ MY SP眺1舶, PO1肝s rbi・α∂艇 ON A・.FIXED・L!NE ?一σ一 Z為《ON・TFiE u−∂ PLANE. 嘘RM冊ua boNDITI〔別S, WE HAVE 』 THE RELATION O=D・ THE・STANnARD EEvl酊1㎝(F Xi !S A’”1 LINEAR FUNCT’1qN OF ヱrt,: (3) ・。ω一・[・α一8カyω], ・階iRE”6α州D eb DENO正.冊拒ST͡’@ ・
DEVIA丁10NS OF ai AND bi RESPECTIVELY. IN・REALITY川匪E F眠蹴…N’TAL・RELATIoNs(2),(3)AND(4)喘)ouLY・APPRoxlMA正LY、 SO THAT矩 SAY ’ALト⑩ST−(維PARA惰訂1ER’ AND ’QしいSJ−FlXED’ INSTEAD OF ’ONE PARA揃ER’ :AND’FIXED’ RESPECTIVELY, . ’ de @Accordi ng to circumstances, [6.]」・. ” ・ ” the LS estimate of r1)s C) may be a statistical artifact77 1 HEIGHIb eAT THE FIXED POIM