fulltext (2)
fulltext[1]
![fulltext[1]](https://img.taocdn.com/s3/m/01bfacf6ba0d4a7302763a57.png)
Mathematical Modeling of the Hot Strip Rolling of Microalloyed Nb,Multiply-Alloyed Cr-Mo,and Plain C-Mn Steels FULVIO SICILIANO,Jr.and JOHN J.JONASIndustrial mill logs from seven different hot strip mills(HSMs)were analyzed in order to calculatethe mean flow stresses(MFSs)developed in each stand.The schedules were typical of the processingof microalloyed Nb,multiply-alloyed Cr-Mo,and plain C-Mn steels.The calculations,based on theSims analysis,take into account work roll flattening,redundant strain,and the forward slip ratio.Themeasured stresses are then compared to the predictions of a model based on an improved MisakaMFS equation,in which solute effects,strain accumulation,and the kinetics of static recrystallization(SRX)and metadynamic recrystallization(MDRX)are fully accounted for.Good agreement betweenthe measured and predicted MFSs is obtained over the whole range of rolling temperatures.Theevolution of grain size and the fractional softening are also predicted by the model during all stagesof strip rolling.Special attention was paid to the Nb steels,in which the occurrence of Nb(C,N)precipitation strongly influences the rolling behavior,preventing softening between passes.The presentstudy leads to the conclusion that Mn addition retards the strain-induced precipitation of Nb;bycontrast,Si addition has an accelerating effect.The critical strain for the onset of dynamic recrystalliza-tion(DRX)in Nb steels is derived,and it is shown that the critical strain/peak strain ratio decreaseswith increasing Nb content;furthermore,Mn and Si have marginal but opposite effects.It is demon-strated that DRX followed by MDRX occurs under most conditions of hot strip rolling;during theinitial passes,it is due to high strains,low strain rates,and high temperatures,and,in the final passes,it is a consequence of strain accumulation.I.INTRODUCTION been achieved.The goal of the present research was,there-fore,to characterize strip rolling in terms of softening mecha-T HE metallurgical features of plate rolling are now nism,strain accumulation,grain size,and precipitation.Also,largely understood.[1–10]The long interpass times allow com-the rolling behavior of microalloyed Nb,plain C-Mn,andsome Cr-Mo grades will be considered and compared. plete static recrystallization(SRX)to take place if the steelAnother important point that arose from this study is that is being rolled above the interpass recrystallization stopthe levels of Mn and Si present in the steel appear to influence temperature(T nr),i.e.,in the absence of carbonitride precipi-the precipitation behavior during strip rolling.The occur-tation.Plate rolling below T nr leads to austenite pancaking,rence of precipitation changes the rolling load and must, because strain-induced precipitation prevents any furthertherefore,be accurately predicted.SRX from occurring.By contrast,during the very shortHodgson[15]has carried out an excellent analysis of the interpass times involved in rod rolling,neither SRX nordifferent types of mathematical models and has supplied a precipitation can take place,[11–14]and,in contrast to platelist of the advantages associated with the application of a rolling,the strain accumulation that takes place leads toparticular model to a given practical situation.According dynamic recrystallization(DRX)followed by metadynamicto this author,the main advantages are(1)a reduction in recrystallization(MDRX).Thus,the metallurgical character-the number of mill trials,(2)an evaluation of hardware istics of plate rolling,on the one hand,and of rod rolling,modifications,(3)the prediction of variables that cannot on the other hand,are now fairly clear.be measured,(4)an estimation of the effect of interactions, From the point of view of interpass time,strip rolling(5)the potential for performance enhancement,and(6) falls between the two processes discussed previously.Duringinexpensive research costs.the initial passes,when the interpass times are still relativelyThe applicability of any model must be intensively tested long,the metallurgical behavior is similar to plate rolling.in advance,preferably with industrial data.That is the As the interpass intervals become shorter,the characteristicsapproach that was employed here.After an improved accu-approach those of rod rolling.Because the classification of racy in predicting the operational parameters has been dem-strip rolling as being like plate rolling,on the one hand,or onstrated,the model can then be applied to production. like rod rolling,on the other,depends on chemistry and The analysis of mean flow stress(MFS)behavior as a rolling schedule as well as pass number,a relatively simple function of inverse absolute temperature permits identifica-and generally accepted analysis of this process has not yet tion of the main microstructural changes taking place duringrolling.These include SRX,DRX followed by MDRX,strainaccumulation,and phase transformation.Figure1illustratesthese phenomena schematically in the form of an MFS vs FULVIO SICILIANO,Jr.,Research Associate,and JOHN J.JONAS,1/T curve for a hypothetical five-pass schedule.Beginning Professor,are with the Department of Metallurgical Engineering,McGillat the first pass(on the left-hand,or high-temperature side), University,Montreal,PQ,Canada H3A2B2.Manuscript submitted March22,1999.there is a low-slope region,within which SRX occurs.The(a )Fig.1—Schematic representation of the evolution of MFS as a function of the inverse absolute temperature.Each characteristic slope is associated with a distinct metallurgical phenomenon.high temperature permits full softening to take place during the interpass interval.After pass 2,the lower temperature does not permit full softening,leading to strain accumula-tion.This accumulation then leads to the onset of DRX (as long as there is no precipitation),which is followed by MDRX between passes 3and 4.The analysis of MFS curves as described previously was first proposed by Boratto et al.[1]for determination of the (b )three critical temperatures of steel rolling (Ar3,Ar1,and Tnr ).[2,3,4]This technique has also been used to deduce that DRX occurs in seamless tube rolling [16,17]as well as in hot strip mills (HSMs).[18,19]Sarmento and Evans [19]came to similar conclusions in their analysis of industrial data from two HSMs.The main types of controlled rolling considered here are listed in the following paragraphs.Recrystallization-Controlled Rolling .In the schedules considered subsequently,as long as strip rolling is carried out above T nr ,SRX is considered to take place.Conversely,below T nr ,there is strain accumulation.Conventional-Controlled Rolling .Here,finishing is employed to flatten or “pancake”the austenite grains at temperatures below T nr .In this case,the particle pinning (c )resulting from the precipitation of Nb(C,N)retards or even Fig.2—Temperature-time diagrams comparing three rolling approaches:prevents the occurrence of recrystallization.In the absence (a )recrystallization controlled rolling,(b )conventional controlled rolling,of precipitation,as long as rolling is carried out below T nr and (c )dynamic recrystallization controlled rolling.for static recrystallization,the strain accumulation that takes place can trigger DRX followed by MDRX,leading to rapid softening between passes.In terms of mathematical model-ing,if there is no precipitation and the accumulated strain exceeds the critical strain,DRX is initiated,often causing caused by DRX when high strain rates are employed and large strains are applied (this corresponds to single-peak full and fast softening.This is usually associated with unpre-dictable load drops in the final passes.[20]behavior in the stress-strain curve).Circumstantial evidence for the occurrence of DRX in seamless tube rolling [16,17]and Dynamic Recrystallization-Controlled Rolling .This type of process consists of inducing DRX in one or more passes HSMs [18,19,21,22]can be found in the literature.Figure 2illustrates schematically the three different rolling during the rolling schedule.This can be done either by applying large single strains to the material or via strain paradigms described previously for a hypothetical five-pass schedule.The conditions associated with each type of rolling accumulation.Both methods allow the total strain to exceed the critical strain for the initiation of DRX.Some of the are displayed in Table I.Knowledge of the rolling parameters and process limitations associated with each method makesbenefits of this approach involve the intense grain refinementTable I.Mechanistic Conditions Pertaining to the Three Controlled Rolling TechniquesRole of Relation betweenType of T Range with Strain-Induced Precipitation andProcess Respect to T nr Precipitation RecrystallizationRCR above absence required SRX before precipitationCCR below presence required precipitation before SRX or DRX DRCR below absence required no SRX;DRX before precipitation possible the design of rolling schedules to fit the needs and B.Analysis of Mill Log Dataconstraints of each case.Logged data were collected from the following fiveHSMs:Dofasco seven-stand HSM(Hamilton,Canada), II EXPERIMENTAL PROCEDURE Algoma six-stand–2.69-m-wide HSM(Sault Ste.Marie,Canada),Sumitomo seven-stand HSM(Kashima,Japan), The present work combines the use of models developedSumitomo seven-stand HSM(Wakayama,Japan),and BHP from hot torsion test data and from the analysis of industrialsix-stand HSM(Port Kembla,Australia).Data from two mill logs.other HSMs were taken from the literature.[19]One is fromthe Usiminas six-stand HSM in Brazil,and the other is an A.Materials unidentified seven-stand HSM referred to here as“Davy.”Some1300logs were made available for analysis,of Various materials were tested and were divided into threewhich about300were examined in detail.The data that groups:microalloyed Nb(group A),multiply alloyed Cr-were employed consisted of interstand distances,work roll Mo(group B),and plain C-Mn(group C)steels.Tables IIdiameters,and pyrometer locations.For each strip,the fol-and III list the grades studied here.In group A(Table II),lowing data were used:chemical composition,strip width, some steels have low Si contents,with the same base compo-strip thicknesses before and after all passes(H and h,respec-sition(e.g.,AD5and AD6and AD7and AD8).Other typestively),work roll rotational speeds,roll forces,and tempera-have different Mn contents,with the same base compositiontures(mean values)for each pass,according to a model or (e.g.,AD9and AD10and AD2,AD3,and AD4).Theseto entry and exit temperatures.grades are used to study the influences of Mn and Si on theThe previous parameters were then employed to calculate precipitation of Nb(C,N)and on the critical strain/peakthe true strains,strain rates,interpass times,and MFSs, strain ratio.Table III gives the chemical compositions ofaccording to the Sims formulation.[4,23]The corrections for some of the multiply alloyed Cr-Mo and plain C-Mn gradesroll flattening,[24,25]redundant strain,and forward slip studied here.The steels used for the torsion tests are markedwith“TT.”between roll and strip[25]were taken into account in theseTable II.Chemical Compositions of the Niobium Steels Investigated(Group A)Steel Plant C Mn Si Nb Ti N*Al P S AA1TT Algoma0.050.350.0100.035—0.0040.0430.0080.006 AA2Algoma0.050.700.1000.053—0.0050.0450.0070.004 AA3Algoma0.060.700.1100.058—0.0050.0450.0080.004 AS1TT Sumitomo0.07 1.120.0500.0230.0160.0000.0290.0190.002 AS2TT Sumitomo0.09 1.330.0600.0360.0160.0030.0190.0170.020 AB BHP0.11 1.050.0100.031—0.0030.0400.0120.012 AD1Dofasco0.060.650.2250.020—0.0040.0350.0080.005 AD2Dofasco0.140.650.2250.020—0.0040.0350.0080.005 AD3Dofasco0.120.850.2250.020—0.0040.0350.0080.005 AD4Dofasco0.12 1.000.2250.020—0.0040.0350.0080.005 AD5Dofasco0.060.650.1150.030—0.0040.0350.0080.005 AD6Dofasco0.060.650.0100.030—0.0040.0350.0080.005 AD7Dofasco0.060.650.1150.045—0.0040.0350.0080.005 AD8Dofasco0.060.650.0100.045—0.0040.0350.0080.005 AD9Dofasco0.060.450.0100.008—0.0040.0350.0080.005 AD10Dofasco0.060.650.0100.008—0.0040.0350.0080.005 AD11Dofasco0.06 1.250.3250.0750.0240.0040.0350.0080.005 AD12Dofasco0.06 1.250.3250.080—0.0040.0350.0080.005 AM1“Davy”0.060.700.0700.050—0.0040.0510.0180.013 AM2“Davy”0.050.450.0200.020—0.0050.0470.0140.008 AU Usiminas0.110.540.0020.018—0.0040.0570.0140.008 *For the Dofasco grades,a mean value of40ppm N is listed here.The actual values are taken into account in the spreadsheets.Table III.Chemical Compositions of the Multiply-Alloyed and Plain C-Mn SteelsGroup Steel Plant C Mn Si Nb Ti Cr Mo V Ni N Al P S B BCM Sumitomo 0.280.520.220——0.830.15——0.0050.0260.0160.004BCMV Sumitomo 0.410.630.280—0.015 1.380.600.270.020.0060.0460.0130.001BCMVN Sumitomo 0.470.660.1700.0160.0160.980.970.120.460.0040.0420.0160.004CCA TT Algoma 0.030.270.010——————0.0040.0420.0080.010CS1TT Sumitomo 0.10 1.080.060——————0.0030.0200.0170.003CS2Sumitomo 0.450.760.210——————0.0050.0040.0170.004CD Dofasco 0.060.270.000——————0.0040.0350.0070.005CM “Davy”0.030.240.020——————0.0040.0420.0090.006CUUsiminas0.050.240.002——————0.0040.0300.0160.010calculations and were organized using MICROSOFT [1];here,the MFS (M )is a function of the strain,strain rate,temperature,and carbon content in weight percent.EXCEL*spreadsheet software.Some typical spreadsheet*MICROSOFT and EXCEL are trademarks of the Microsoft Corpora-M ϭexp 0.126Ϫ1.75C ϩ0.594C 2[1]tion,Redmond,WA.inputs and outputs are shown in Table IV .The same spread-sheet was used for the microalloyed Nb,multiply alloyed ϩ2851ϩ2968C Ϫ1120C 2T0.21˙0.13Cr-Mo,and plain C-Mn grades;this is because it does not consider microstructure but only mechanical parameters.According to the present method,the Sims formulation is employed to calculate and plot the MFS vs 1000/T for several bars.Because strip rolling reductions are applied at III.THE SUBMODELS APPLIED TO HOTSTRIP ROLLING various strains and strain rates,all the derived MFSs are“normalized”to ϭ0.4and ˙ϭ5s Ϫ1.[27,28]An example of this approach for grade D5is illustrated A.The MFS Modelin Figure 3.It can be seen that Misaka’s equation overpre-dicts the MFSs obtained from the mill logs for this 0.03pct Improvement of the Misaka equationMisaka’s equation [26]has often been employed to specify Nb steel.On the other hand,it underpredicts the MFSs for some higher Nb grades.[28]The trend of the Misaka equation the MFS for C-Mn steels during hot strip rolling.It will be used here as the basis for a modified equation that takes fits the mill log values reasonably well in the region where the slope is quite shallow (i.e.,at low 1/T or high T values).into account the effects of different alloying elements,such as Mn,Nb,and Ti.The Misaka equation is displayed in Eq.This indicates that full SRX is occurring between the passesTable IV .Example of the Spreadsheet Calculations Carried Out Using the Mill DataInputs:Data from mill logsRoll Radius Roll Speed Width Gage TemperatureRoll Force Pass (mm)(rpm)(mm)(mm)(ЊC)(Tonne)————30.60——F139433.9126417.339872157F239154.5126410.799512223F338179.212647.429152116F4365119.01264 5.109071691F5363147.11264 3.908961357F6376167.21264 3.148841264F7378172.012642.618721627Outputs Hitchcock Forward Nominal Total Strain Rate Interpass 1000/T Sims MFS Pass R Ј(mm)Slip Factor Strain Strain*(s Ϫ1)Time (s)(K Ϫ1)(MPa)F1403 1.100.660.7512.9 3.480.79116F2412 1.090.550.6125.2 2.140.82151F3416 1.080.430.4841.3 1.480.84179F4402 1.080.430.4772.7 1.010.85151F5428 1.060.310.3494.20.790.86178F6472 1.050.250.271150.630.86185F75611.040.210.23131—0.87250*Includes the redundant strain.effects of the other elements are considered to be linear over their concentration intervals.**Comparison of Tables II and III indicates that the Eq.[2](group A)steels (Table II)have only half the average Si levels (about 0.11)compared to the Eq.[4](group B)steels of Table III (about 0.22pct Si).At the time this analysis was carried out,no account was taken of the Si levels on the MFS and,therefore,on the rolling load.In retrospect,however,this would have been desirable,as it appears from a study of the previous relations that the introduction of such a term may have permitted the use of a single set of coefficients.B.Modeling of Grain Size and Fractional SofteningDuring a particular pass of a rolling schedule,the sum ofFig.3—Comparison of Misaka’s equation and the present equation (MFS*)the retained and applied strains will determine which soften-based on chemical composition and fitted to the mill (Sims)data in the ing mechanism (SRX or DRX ϩMDRX)will operate.SRX region.Here,the mill data for grade AD5(0.03pct Nb)are corrected Depending on the type of softening,different equations are to a constant strain of 0.4and a constant strain rate of 5s Ϫ1.then employed to specify the grain size and fractional soften-ing.In this section,a method is described that can be used to follow the microstructural evolution during multipass roll-in this region.A solution-strengthening factor can,therefore,ing.For now,no attention is paid to the exact onset of DRX,be added that allows Misaka’s equation to fit all grades.For and the critical strain is simply taken as a fixed fraction of compositions AS1,AS2,and AB,for example,a correction the peak strain.The example given subsequently involves for solute strengthening due to the high Mn content (1.12,modeling microstructural evolution in a plain C-Mn grade,1.33,and 1.08,respectively),allows Misaka’s equation to and the critical strain is considered to be 0.8p .The onset fit the mill values.[28]Another correction is still required for of DRX in Nb steels is discussed in the next section,IIIC.the occurrence of strain accumulation and DRX,which leads During rolling,the temperature decreases continuously.to departures from Misaka-type behavior at high values of Thus,the temperature adopted for each interpass interval 1/T (i.e.,low T ).For this purpose,the model by Yada and is taken as the average of the prior and subsequent pass Senuma [29,30,31]was adapted and tested.[25,27,29]This pro-temperatures.This assumption is employed because the pres-duced the following final MFS equation:ent equations were derived for isothermal conditions and the interpass times are short enough to allow the use of a MFS ϩNbϭ(MFS Misaka (0.768ϩ0.51Nb ϩ0.137Mn [2]single temperature value.In the present work,the recrystallization and grain-size ϩ4.217Ti))ϫ(1ϪX dyn )ϩK ss X dynrelations described subsequently were incorporated into the Here,X dyn is the softening attributable to DRX,ss is theMICROSOFT EXCEL spreadsheet software,as shown in steady-state stress,and K ϭ1.14is a parameter that converts more detail in Section IV .The submodels involved here flow stress to MFS.were assembled by following a method developed for rod Equation [2]is valid over the following concentration rolling.[11]The recrystallization kinetics equations used here ranges:0.020to 0.080pct Nb,0.35to 1.33pct Mn,and 0are adaptations of the Avrami–Johnson–Mehl–Kolmogorov to 0.024pct Ti.This approach has been used successfully equation,with parameters that were selected to fit the mill to calculate the MFSs in plain C-Mn [25,27]as well as in Cr-data.Mo steels,[27,32]and the respective equations are displayed 1.Softening between passesin Eqs.[3]and [4],respectivelyA softening model uses parameters such as strain,strain MFS ϩC-Mn ϭ(MFS Misaka (0.768ϩ0.137Mn))[3]rate,initial grain size,and temperature to decide upon the mechanism and calculate the extent of softening.In an HSM,ϫ(1ϪX dyn )ϩK ss X dynthe softening between passes can be calculated with the aid of an MFS equation and temperature corrections.In the The Mn concentrations studied ranged from 0.27to 1.08present work,the microstructural evolution equations were pct.For the multiply alloyed steels,the following relation tested directly using the spreadsheet for the three groups of is applicable:[27,32]steels.The t 0.5equation selected for each group is shown in MFS ϩCr-Moϭ(MFS Misaka (0.835ϩ0.51Nb ϩ0.098Mn Table V .Note that Eq.[6]is the only one available for the DRX ϩ0.128Cr 0.8ϩ0.144Mo 0.3ϩ0.175V [4]kinetics of Nb steels.A similar expression is employed for ϩ0.01Ni))ϫ(1ϪX dyn )ϩK ss X dynthe multiply alloyed grades,as derived in a recent study.[32]For the C-Mn steels,several equations are available;how-Equation [4]is considered to apply to the following composi-ever,the one developed by Hodgson et al.[33,36]provided the tion ranges:0.52to 0.66pct Mn,0to 0.08pct Nb,0.83tobest fit to the mill logs.1.38pct Cr,0to 0.46pct Ni,0.15to 0.97pct Mo,and 0to 0.27pct V .Due to the high maximum concentrations of Cr2.Strain accumulation between passesPartial recrystallization between passes results in retained and Mo,the solution effects approached “saturation;”this is why exponents are employed in this expression.Thestrain,which must be added to the strain applied in theTable V .Equations Describing the Softening KineticsGroup Type EquationReferenceASRX t SRX 0.5ϭ(Ϫ5.24ϩ550[Nb])ϫ10Ϫ18(Ϫ4.0ϩ77[Nb])d 20exp (330,000/R T )[5]33DRXt MDRX 0.5ϭ4.42ϫ10Ϫ7˙(Ϫ0.59)exp (153,000/R T )[6]34,35B SRX32t 0.5ϭ1.57ϫ10Ϫ14иd 20иϪ2.9exp 271,000R T [7]DRX32t 0.5ϭ1.84ϫͫ˙иexp 330,000R TͬϪ0.86exp271,000R T[8]C SRX32,36t SRX0.5ϭ2.3ϫ10Ϫ15Ϫ2.5d 20exp230,000R T [9]DRX14t MDRX 0.5ϭ0.4˙иexp 300,000R TϪ0.8exp240,000R T[10]subsequent stand.The accumulated strain in pass i (i Ͼ1)is small,d 0i ϩ1will be close to the original grain size,d 0i ;in the latter case,the grains only change their shapes because then becomes [37]of the applied strain.[11]a i ϭi ϩK acc (1ϪX i Ϫ1)i Ϫ1[11]b.Grain growth after recrystallizationwhere X is the fractional softening and K acc is a constant.After complete recrystallization,the microstructure is sub-The value of K acc was reported in the literature as falling jected to grain growth;this is driven by the decrease in free between 0.5and 1.[11,37]The parameter K acc can be related energy associated with the grain boundaries.For the group to the rate of recovery.High rates of recovery result in less A and B steels,one single equation (Eq.[19])was employed accumulated strain.This is clear from the work of Gibbs et to describe grain growth:al.,[37]where longer interpass times led to K acc ϭ0.5and shorter interpass times (less recovery)to K acc ϭ1.In the present hot strip model,the K acc constant is assumed to be d 4.5ϭd 4.50ϩ4.1ϫ1023ϫt ip[19]1and the accumulated strain is used in all the calculations.The latter is considered to represent the average strain pres-ϫexp (Ϫ435,000/R T )ent within the material.3.Grain-size evolutionAlthough the previous equation was derived for Nb steels,a.Recrystallized grain sizeit is used here to describe grain growth in the multiply This calculation simply takes into account the initial grain alloyed steels as well.This is due to the lack of an equation size,strain,and/or strain rate.The grain size after SRX is for the multiply alloyed grades,and because the alloying well known to be strongly dependent on the prior strain and elements present in the group B grades make it more rea-only depends on the strain rate to a minor extent.On the other sonable to adopt an equation derived for Nb steels than hand,the grain sizes after MDRX are strongly dependent on one derived for C-Mn grades.The grain-growth kinetics the strain rate.[14,15,34–36,38–41]The equations used here are in Nb and multiply alloyed steels are,of course,expected listed in Table VI.to be slower than in plain C-Mn compositions due to solute The grain size pertaining to the entrance of a given pass drag.For the group C (C-Mn)steels,however,a group of is considered to be the “initial”grain size (d 0).Note that equations is available to describe grain growth.The the equations used to model the recrystallized grain sizes method used here was proposed by Hodgson et al.[11,36,44]for the steels pertaining to groups A and B are the same;and is a “pragmatic”one,based on both laboratory [44]and this is due to the lack of appropriate equations for the multi-industrial [11]observations.ply alloyed steels.It should be noted that there is a large difference between In the case of incomplete recrystallization,the initial grain the rate of growth during the first second of the interpass size for the following pass (d 0i ϩ1)can be calculated using interval and the remaining time.This difference may arise the following relation,[11,33,43]which specifies a kind of because of the large driving force for grain growth present “average”derived from the freshly formed and original in the initially fine-grained structure.Another possible effect grain sizes:concerns the presence of “deformation”vacancies immedi-ately after rolling,which could accelerate growth.[45]This d 0i ϩ1ϭd rex i ϫX 4/3i ϩd 0i (1ϪX i )2[18]transition was handled by adopting different grain-growth exponents for each stage [11]and by using the following With this formulation,when X i is close to 1,the initial grainsize for the following pass is d rex i .On the other hand,if X iequations.Table VI.Equations Describing the Recrystallized Grain SizeGroup Type Equation Ref.A SRX d SRXϭ1.1иd0.670иϪ0.67[12]42DRX34d MDRXϭ1370ϫϪ0.13expϪ45,000R T[13]B SRX d SRXϭ1.1иd0.670иϪ0.67[14]41DRX35d MDRXϭ1370ϫϪ0.13expϪ45,000R T[15]C SRX33,36d SRXϭ343иd0.40иϪ0.5expϪ45,000R T[16]DRX33,36d MDRXϭ2.6ϫ104и˙иexp300,000R TϪ0.23[17]SRX,t ipϽ1s[20]Eq.[3],based on the Misaka equation.The observed andpredicted MFSs are compared in Table VIII,where the differ-ences between the three sets of MFSs are also shown(in d2ϭd2SRXϩ4.0ϫ107(t ipϪ4.32t0.5)expϪ113,000R Tpercentage).The“basic”version of the prediction spreadsheet MDRX,t ipϽ1s[21]described previously,allowing for strain accumulation andDRXϩMDRX,is considered to be fully applicable to the d2ϭd2MDRXϩ1.2ϫ107(t ipϪ2.65t0.5)expϪ113,000R Tgroup C(C-Mn)grades.However,for Nb steels,the criticalstrain for the initiation of DRX must be accurately known,and agreement regarding this quantity is lacking in the litera-SRX,t ipϾ1s[22]ture.The actual precipitation start time during hot strip roll-ing is also an unknown quantity.In the next two sections d7ϭd7SRXϩ1.5ϫ1027(t ipϪ4.32t0.5)expϪ400,000R T(C and D),some improved methods for the estimation ofthese two parameters will,therefore,be proposed anddescribed for the Nb grades.MDRX,t ipϾ1s[23]d7ϭd7MDRXϩ8.2ϫ1025(t ipϪ2.65t0.5)expϪ400,000R T C.Critical Strain for the Initiation of DRX4.Design of the microstructural-predictionThe critical strain for the onset of DRX is an important spreadsheet parameter employed in the mathematical modeling of micro-The previous equations describing the microstructuralstructural evolution and of rolling load.Knowledge of the events can now be organized into a spreadsheet.Basically,critical strain for the initiation of DRX is a requirement for the spreadsheet parameters displayed in Table VII are usedprediction of the operating softening mechanisms in hot as input data to simulate the microstructural changes taking working processes.place during hot rolling,in a pass-by-pass analysis.TheFor the present purpose,it is useful to express the critical starting grain size(after roughing and before strip rolling)strain for the initiation of DRX(c)as a function of the is adopted as100m for the C-Mn grades and as80mpeak strain(p),as determined from a stress-strain curve. for the others.The subsequent grain sizes are calculated This is because several equations are available to specify after recrystallization and grain growth,and the results con-the peak strain as a function of initial grain size,temperature, stitute the input data for the next pass.Both the accumulated and strain rate for Nb steels.Thec/p ratio often lies strain as well as the redundant strain are employed through-between0.67and0.86[46]and is generally considered to be out the calculations.Some typical inputs and outputs of the0.8for plain C-Mn steels.Previous workers have reportedvalues for Nb steels as low as0.65.[34]The effects of Nb, microstructural-evolution spreadsheet are given in Table VII.The following example uses data from the Dofasco HSM,Mn,and Si were taken into account in a recent investiga-grade CD.tion,[47]and a fit was found for thec/p ratio.The method In C-Mn steels,mechanisms such as carbonitride precipi-used was to test several possible values ofc/p and then tation followed by strain accumulation do not take place,soto select the ratio that provides the best fit to the Sims MFS only SRX as well as DRXϩMDRX effects are considered curve.This procedure,thus,specifies the conditions underwhich DRX will occur.Over100mill logs were tested in here.The observed MFS is calculated using the Sims formu-lation(from the mill data).Then,predictions are made using this way using the grades listed in Table I.The resulting。
MySQL数据库考试试题与答案
专业：移动通信科目：MySQL数据库一、单项选择题1．以下聚合函数求数据总和的是( )A．MAXB．SUMC．COUNTD．AVG答案：B2．可以用( )来声明游标A．CREATE CURSORB．ALTER CURSORC．SET CURSORD．DECLARE CURSOR答案：D3．SELECT语句的完整语法较复杂，但至少包括的部分是( )A．仅SELECTB．SELECT，FROMC．SELECT，GROUPD．SELECT，INTO答案：B4．SQL语句中的条件用以下哪一项来表达( )A．THENB．WHILEC．WHERED．IF答案：C5．使用CREATE TABLE语句的( )子句，在创建基本表时可以启用全文本搜索A．FULLTEXTB．ENGINEC．FROMD．WHRER答案：A6．以下能够删除一列的是( )A．alter table emp remove addcolumnB．alter table emp drop column addcolumnC．alter table emp delete column addcolumnD．alter table emp delete addcolumn答案：B7．若要撤销数据库中已经存在的表S，可用（）。
A．DELETE TABLE SB．DELETE SC．DROP SD．DROP TABLE S答案：D8．查找表结构用以下哪一项( )A．FINDB．SELETEC．ALTERD．DESC答案：D9．要得到最后一句SELECT查询到的总行数，可以使用的函数是( ) A．FOUND_ROWSB．LAST_ROWSC．ROW_COUNTD．LAST_INSERT_ID答案：A10．在视图上不能完成的操作是( )A．查询B．在视图上定义新的视图C．更新视图D．在视图上定义新的表答案：D11．UNIQUE惟一索引的作用是( )A．保证各行在该索引上的值都不得重复B．保证各行在该索引上的值不得为NULLC．保证参加惟一索引的各列，不得再参加其他的索引D．保证惟一索引不能被删除答案：A12．用于将事务处理写到数据库的命令是( )A．insertB．rollbackC．commitD．savepoint答案：C13．查找条件为：姓名不是NULL的记录( )A．WHERE NAME ! NULLB．WHERE NAME NOT NULLC．WHERE NAME IS NOT NULLD．WHERE NAME!=NULL答案：C14．主键的建立有( )种方法A．一B．四C．二D．三答案：D15．在视图上不能完成的操作是( )A．更新视图数据B．在视图上定义新的基本表C．在视图上定义新的视图D．查询答案：B16．在SQL语言中，子查询是（）。
mysql索引类型normal,unique,fulltext
mysql索引类型normal,unique,fulltext问题1：mysql索引类型normal，unique，full text的区别是什么？normal：表⽰普通索引unique：表⽰唯⼀的，不允许重复的索引，如果该字段信息保证不会重复例如⾝份证号⽤作索引时，可设置为uniquefull textl: 表⽰全⽂搜索的索引。
FULLTEXT ⽤于搜索很长⼀篇⽂章的时候，效果最好。
⽤在⽐较短的⽂本，如果就⼀两⾏字的，普通的INDEX 也可以。
总结，索引的类别由建⽴索引的字段内容特性来决定，通常normal最常见。
问题2：在实际操作过程中，应该选取表中哪些字段作为索引？为了使索引的使⽤效率更⾼，在创建索引时，必须考虑在哪些字段上创建索引和创建什么类型的索引,有7⼤原则：1．选择唯⼀性索引2．为经常需要排序、分组和联合操作的字段建⽴索引3．为常作为查询条件的字段建⽴索引4．限制索引的数⽬5．尽量使⽤数据量少的索引6．尽量使⽤前缀来索引7．删除不再使⽤或者很少使⽤的索引⼀、 MySQL: 索引以B树格式保存 Memory存储引擎可以选择Hash或BTree索引，Hash索引只能⽤于=或<=>的等式⽐较。
1、普通索引：create index on Tablename(列的列表) alter table TableName add index (列的列表) create table TableName([...], index [IndexName] (列的列表) 2、唯⼀性索引：create unique index alter ... add unique 主键：⼀种唯⼀性索引，必须指定为primary key 3、全⽂索引：从3.23.23版开始⽀持全⽂索引和全⽂检索，FULLTEXT， 可以在char、varchar或text类型的列上创建。
4、单列索引、多列索引： 多个单列索引与单个多列索引的查询效果不同，因为： 执⾏查询时，MySQL只能使⽤⼀个索引，会从多个索引中选择⼀个限制最为严格的索引。
fulltext(3)
As a result of the fast development of new materials, urged by the demand of more durable and resistant materials, fibre reinforced plastic composites (FRP) were introduced as replacement for the reinforcing steel. Carbon FRP has been used since the seventies in the space industry, where the requirement for lightweight, high-tensile strength, corrosion and high temperature resistance is essential. Recently, in Brazil and many other countries, FILl) composites (mainly carbon) have been used in strengthening slabs (of bridges and parking floors), beams (of
Materials and Structures/Mat~riaux et Constructions,Vol.35, Januanj-February 2002, pp 50-58
Strength of short concrete columns confined with CFRP sheets
I
1. I N T R O D U C T I O N
External confinement has been successfully applied to concrete columns in order to increase their strength and ductility or to recover them from eventual deterioration originated from different sources. In early constructions, concrete-filled steel tubes were used for some of the reasons mentioned above. The main disadvantages of this technique are the high cost for maintenance against corrosion, the low fire resistance and the heavy weight of the tubes. 1359-5997/02 9 R.ILEM 50
fulltext
International Journal of Wireless Information Networks, Vol.4, No.4,19971068-9605/97/1000-0249$12.50/01997Plenum Publishing Corporation249GSM Point-to-Point Short Message ServiceYi-Bing Lin 1GSM short message service provides a connectionless transfer of messages with low-capacity and low-time performance. This article provides a tutorial on GSM point-to-point short messaging. The short message protocol hierarchy is introduced and the mobile originated /terminated messaging procedures are described.KEY WORDS:GSM; short message service; cellular system.1. INTRODUCTIONGSM short message service provides a connection-less transfer of messages with low-capacity and low-time performance. Each message can transmit up to 140octets and operates like a paging service with the added capa-bility that messages can pass in both directions and con-rmation can be provided to indicate that the sent mes-sage has been received. The short messages are trans-ported on the GSM signaling links. Thus, the messages can be received while the mobile users are in conversa-tions.Two types of the short message services have been de®ned:Cell broadcast service periodically delivers shortmessages to all subscribers in a given area.Point-to-point service enables short messages to a speci® c user; this GSM feature can be considered as an enhanced two-way paging service.This article will focus on the point-to-poin t service.An example of the GSM Short Message Service network architecture is illustrated in Fig.1. A complete list of abbreviations used in this article is given in Table I. The short message is ®rst delivered from the message sender (either a paging input device or a GSM mobile station) to a1Departmentof Computer Science and Information Engineering,National Chiao Tung University, Hsinchu, Taiwan, R.O.C.; e-mail:liny@ .tw.Fig.1.GSM short message service network architecture.Short Message Service Center (SM-SC). The SM-SC is connected to the GSM network through a speci® c GSM Mobile Switching Center (MSC) called the Short Message Service Gateway MSC (SMS GMSC). The SM-SC may connect to several GSM networks, and to several SMS GMSCs in a GSM network. Following the GSM roam-ing protocol, the SMS GMSC locates the current MSC of the message receiver and forwards the message to that MSC. The MSC sends the paging request to the Base Sta-tion Subsystem s, and the base stations page the destination Mobile Station (MS). The MS used for short message ser-vices must contain special software to enable the messages to be decoded and stored. Typically, messages will beLin 250Table I.Acronym ListCM-Sub: Connection Management SublayerGMSC: Gateway MSCHLR: Home Location RegisterIMSI: International Mobile Subscriber IdentityIWMSC: Interworking MSCLA: Location AreaMAP: Mobile Application PartMM: Mobility ManagementMO: Mobile OriginatedMS: Mobile StationMSC: Mobile Switching CenterMSISDN: Mobile Station ISDN NumberMT: Mobile TerminatedRR: Radio ResourceRPDU: Relay Protocol Data UnitSM-AL: Short Message Application LayerSMC: Short Message ControlSM-CP: Short Message Control ProtocolSMI: Short Message Identi®erSMR: Short Message RelaySM-RL: Short Message Relay LayerSM-RP: Short Message Relay ProtocolSMS: Short Message ServiceSM-SC: Short Message Service CenterSM-TL: Short Message Transfer LayerSS7: Signaling System Number7TCAP: Transaction Capabilities Application PartTPDU: Transfer Protocol Data UnitTP-UD: Transfer Protocol User DataVLR: Visitor Location Registerstored on the Subscriber Identity Module of the MS, which can be displayed on the standard screen of the MS.An MS may send or replay a short message. The message is delivered to a Short Message Service Inter-working MSC(IWMSC) and then to the SM-SC (the details will be elaborated later).2. SMS PROTOCOL HIERARCHYFigure2illustrates the SMS protocol hierarchy for mobile originated messaging [6]. (The architecture for mobile terminated messaging is similar, with the excep-tion that the IWMSC is replaced by the GMSC.)In this hierarchy, the protocol between the IWMSC and the SM-SC below the transfer layer is not speci®ed in GSM.Some example protocols are described in ref.2. Other potential protocols include Telocator Alphanu-meric Input Protocol [7] and Telocator Data Protocol [10]. The protocol between the MSC and IWMSC is GSM MAP[4],which utilizes the services provided by Signaling System Number7Transaction Capabilities Application Part(SS7TCAP) [9,4].The layers below the CM-sublayer are the Mobil-ity Management(MM) Sublayer and the Radio Resource (RR)management sublayer.At the RR-sublayer,the short message service is supported by a control channel (such as SDCCH or SACCH). The upper layer protocols are described below.2.1. The Short Message Transfer LayerThe Short Message Transfer Layer(SM-TL) pro-vides services to transfer Short Message Application Layer(SM-AL) short messages and the corresponding delivery reports [3]. These service primitives (between SM-TL and SM-AL) generate a reference number called Short Message Identi®er(SMI) for every short message associated with the primitives. This SMI at the MS is not carried between the peer entity at the SM-SC. That is, a short message may have different SMIs at the MS and the SM-SC sides.The SM-TP consists of four types of Transfer Pro-tocol Data Units(TPDUs):Fig.2.SMS MS-MSC protocol hierarchy (mobile origination).GSM Point-to-Point Short Message Service251SMS-SUBMIT conveys a short message (UserData or TP-UD) from the MS to the SM-SC. ThisTPDU optionally speci®es the validity period thatthe short message can be buffered in the SM-SCif it cannot be delivered to the recipient immedi-ately.SMS-DELIVER conveys a short message fromthe SM-SC to the MS. This TPDU includes theservice center time stamp used by the SM-SC toinform the recipient MS about the arrival time ofthe short message at the SM-TL of the SM-SC.A Boolean parameter More-Message-To-Send isused to indicate if one or more messages are wait-ing in the SM-SC to be delivered to the recipientMS.SMS-STATUS-REPORT conveys a report fromthe SM-SC to the MS to describe the status of theprevious sent mobile originated short message. Ifthe previous short message is not delivered suc-cessfully, this TPDU may report permanent errors(such as validity period expiration or incompati-ble destination) or temporary errors (such as con-gestion). This TPDU is optionally initiated by theSMS-SUBMIT.SMS-COMMAND conveys a command from theMS to the SC. The command can be a query aboutthe previous submitted short message, cancella-tion of the status report, or deletion of the sub-mitted message.The TPDUs (except for SMS-STATUS-REPORT) have a protocol identi®er that identi®es the layer protocol above SM-TL.For an unsuccessful delivery due to the temporary absence of the MS,messages-waiting information may be stored in the HLR/VLR of the recipient MS. The TPDUs are of either priority of nonpriority. A nonpri-ority TPDU will not be attempted if the MS is absent.2.2. The Short Message Relay LayerThe Short Message Relay Layer(SM-RL)pro-vides services to transfer TPDUs and the correspond-ing delivery reports for SM-TL. These service primitives (between SM-RL and SM-TL) generate SM-RL SMI for every short message associated with the primitives. Sim-ilar to the SM-TL SMI, the SM-RL SMI at the MS is not carried between the peer entity at the SM-SC. For every short message, the SM-RL SMI is mapped to and from the SM-TL SMI.At this layer, the Short Message Relay(SMR) entity at the MS communicates with the peer SMR at the MSC by using the Short Message Relay Protocol(SM-RP). SM-RP is the protocol for the networking func-tions between MS and SM-SC, which interworks with TCAP/MAP in the MSC.SM-RP consists of the following Relay Protocol Data Unit(RPDU) types:RP-DATA is invoked by the SM-RL-DATA ser-vice primitive, which passes the TPDU and neces-sary control information from the MS to the net-work (or from the network to the MS).RP-DATAcontains the originating address (MS or SM-SC),the terminating address (SM-SC or MS), and theuser data containing TPDU.RP-SM-MEMORY-AVAILABLE is invoked bythe SM-RL-MEMORY-AVAILABLE primitive,which passes the necessary control informationfrom the MS to the network to indicate that theMS has memory available to receive one or moreshort messages.RP-ACK is invoked by the SM-RL-REPORTprimitive to acknowledge the corresponding RP-DATA or RP-SM-MEMORY-AVAILABLE.RP-ERROR is invoked by the SM-RL-REPORTprimitive to report error of a corresponding RP-DATA. An error may occur because the messageis too short to contain a complete message typeinformation (and the message should be ignored[5]), the message reference is unknown, the mes-sage type is unknown, or the message content issemantically incorrect.2.3. The Connection Management SublayerThe Connection Management Sublayer(CM-Sub) for SMS provides services to support the SM-RL. In this layer, the Short Message Control(SMC) entity at the MS communicates with the peer SMC at the MSC by using the Short Message Control Protocol(SM-CP). The MS has two SMC entities, one handles MS originated (MO) short message service, and the other handles MS termi-nated (MT) short message service. Note that the SMC entities cannot simultaneously perform messaging in the same direction.The SM-CP consists of the following protocol ele-ments:CP-DATA is invoked by the SM-CP serviceLin 252Fig.3.Mobile originated short messaging (I).primitives MNSMS-DATA or MNSMS-ESTab-lish, which delivers RPDU between the MS andthe MSC.CP-ACK acknowledges the corresponding CP-DATA. The CP-ACK does not contain any spe-ci® c information element.CP-ERROR is invoked by the SM-CP ser-vice primitives MNSMS-ABORT or MNSMS-ERROR, which provides the error cause of themessaging procure.Note that before any CP message is delivered, an MM-connection must be established. The SMC uses the MNSMS-ESTablish primitive to establish an MM-con-nection and then transfer a RPDU on that MM-connec-tion. (Note that the primitives with pre®x MNSMS are between SMC and SMR as indicated in Fig. 3.The primitives between SMC and MM are pre®xed with MMSMS.) The SMC uses the MNSMS-DATA primi-tive to transfer an RPDU on an already established MM-connection. When the short message delivery is com-pleted, the MM-connection is released by the SMC using the MNSMS-RELease service primitive.In abnormal cases (e.g., RR-connection failure or SM-CP protocol error), the SMC releases an MM-connection by using the MNSMS-ABORT service primitive. When an error occurs (the causes are similar to that reported by RP-ERROR), the SMC releases an MM-connection by using the MNSMS-ERROR service primitive.3. MOBILE ORIGINATED MESSAGINGIn mobile originated(MO)messaging,the MS sends a short message to the SM-SC. Depending on the destination of the message, the SM-SC may deliver the message to another MS or to a normal pager through the traditional paging system [8]. Figure3illustrates the message ¯ow for MO messaging. The logical message path is MS>originating MSC>IWMSC>SM-SC (note that the originating MSC can be the IWMSC). The procedure is described in the following steps: Step1.The SM-TL entity issues an SMS-SUBMIT TPDU to the SM-RL by the SM-RL-DATA(Request) primitive.Step2.The SMR entity creates an RP-DATA(MO) RPDU and invokes the MNSMS-ESTablish(Request) primitive to transfer the RPDU. The SMR sets the timer TR1M and expects to receive the RP-ACK before the timer expires. If the timer expires, the CM-connection (and thus the MM-connection) is aborted, and a report indication is passed to SM-TL.Step3.The SMC rst establishes the MM-connec-tion as illustrated in Fig.4.Step3a.The MS sends a CM SERVICE RE-QUEST message to the MSC. The service type of this message is `short message service.’Fig.4.MM-connection establishment for mobile originated short mes-saging.GSM Point-to-Point Short Message Service253 Step3b.The SMS coordinating process in theMSC (Co SMS MSC) is invoked, which sends theMAP PROCESS ACCESS REQUEST messageto the VLR. This message (and the messages withpre®x MAP) is GSM MAP message that is deliv-ered through the SS7network by using the SS7TCAP protocol.Step3c.The VLR checks if the MS is legal toreceive the short message service. If so, it acknowl-edges the MSC request.Step3d.The MSC forwards the acceptanceto the MS by sending the message CM SER-VICE ACCEPT.Step4.After the connection to the MSC is estab-lished, the SMS creates the CP-DATA containing the RPDU, and sends the message to the MSC. It sets a timer TC1*and expects to receive the result before the timer expires. The value of the timer may vary with the length of CP-DATA and the type of the radio channel (SDCCH or SACCH) used for the transmission. If TC1*expires, the CP-DATA message may be retransmitted for at most three times. If the attempt is not successful, then the error is reported to SM-RL, and the CM-connection is released.Step 5.The MSC acknowledges the CP-DATA message by returning a CP-ACK. When the SMC of the MS receives the CP-ACK, the TC1*timer is reset.Step 6.The SMS-SUBMIT request is forwarded from the SMR entity to the SM-TL of the MSC by the MNSMS-ESTablish(Indication) and the SM-RL-DATA-(Indication) primitives. The SMR sets the timer TR2N and expects to receive the SMS-STATUS-REPORT (Step13) before the timer expires. If the timer expires, the SMR requests the SMC to abort the CM-connection (and thus the MM-connection). It also sends a report indication to the SM-TL. (Note that the transfer layer protocol only involves the MS and the SM-SC.) Step7.(See Fig. 5.)After the MSC receives the TPDU,the mobile originated short message ser-vice process in the MSC(MOSM MSC)is invoked. The MOSM MSC sends the MAP SEND INFO FOR MO SMS message to the VLR to request the subscriber-related information (e.g., the MSISDN of the MS). This message is used to check if the request violates supple-mentary services invoked or any restrictions imposed.Step8.The VLR either acknowledges the requestpositively or provides the following error causes: Tele-service is not provisioned, call barred, unexpected datavalue, or data missing.Step9.If the VLR response is positive, then theMSC sends the GSM MAP message MAP FORWARDFig.5.Mobile originated short messaging (II). SHORT MESSAGE to the IWMSC (the gateway tothe SM-SC). Note that if the MSC is the IWMSC, thenthis message is saved. The message includes the SM-SC address (provided by the MSC; the address is an E.164number [1]), the sender’s MSISDN, and the short mes-sage (i.e., the TPDU).Steps10±12.The short message is delivered to theSM-SC,and the SM-SC returns a delivery report to the IWMSC. This report is included in the GSM MAPmessage MAP FORWARD SHORT MESSAGE ack sent from the IWMSC to the MSC.Steps13and14.(See Fig.6.) The SM-TL entitysends SMS-STATUS-REPORT to the SMR entity at SM-RL, and the TR2N timer is stopped. The SMR gener-ates RP-ACK and sends it to SMC at CM-Sub. The SMR also releases the CM-connection (and thus the MM-con-nection).Step15.The SMC entity generates the CP-DATA and sends it to the MS. The TC1*is set as in Step4.Step16.When the SMC entity of the MS receivesthe CP-DATA, it forwards the RP-ACK to the SMR en-tity at the SM-RL. It also sends the CP-ACK to the MSC.Step17.The SMR stops the TR1M timer, forwards SMS-STATUS-REPORT to SM-TL, and invokes the MNSMS-RELease primitive to release the CM-connec-tion.Step18.The CM-connection and the MM-connec-tion are released. If the short message transfer is not suc-cessful, the status report provides the error cause to the MS. The error can be unknown service center address, service center congestion, invalid short message entity address, subscriber not service center subscriber, or pro-tocol error.Note that the MO messaging procedure is similar tothat for GSM call origination [4], with one major differ-Lin 254Fig.6.Mobile originated short messaging (III).ence. MO messaging needs IWMSC to interwork to SM-SC. On the other hand, in GSM call origination, the orig-inating MSC processes the call request just like a typicalcentral of®ce in the public switched telephone network.4. MOBILE TERMINATED MESSAGINGIn mobile terminated(MT)messaging,an MSreceives a short message from the SM-SC. The messagesender can be another MS in the GSM network or aninput device in the traditional paging system. The logi-cal message path is SM-SC>GMSC>terminatingMSC>MS where the GMSC can be the terminatingMSC. Figure7illustrates the message ¯ow for MT mes-saging. Note that in this procedure, the GMSC needs toquery the HLR to locate the terminating MSC. The pro-cedure is described in the following steps:Steps1and2.The SMS GMSC receives a shortmessage(RP user data)from the SM-SC.The SMSGMSC requests the routing information by sending themessage MAP SEND ROUTING INFO FOR SM tothe HLR. The message includes the MSISDN of the re-cipient MS.Step 3.The HLR uses the received MSISDN toretrieve the routing information of the MS. The HLRreturns MAP SEND ROUTING INFO FOR SM ackto the SMS GMSC. If the operation fails, then an erroris indicated in the message. The error causes are simi-lar to the ones listed in Step8of the MO messaging.If the operation is successful, then the message includesthe IMSI and the serving MSC address of the recipientMS.Step4.The SMS GMSC delivers the short mes-sage to the MSC by sending the message MAP FOR-WARD SHORT MESSAGE.This message is de-scribed in Step9of MO messaging, with one majorexception. In MT messaging, the message should spec-ify the More-Messages-To-Send parameter to indicate ifthere are more messages to send. This parameter is notFig.7.Mobile terminated short messaging (I).GSM Point-to-Point Short Message Service 255Fig.8.Mobile terminated short messaging (II).used in MO messaging. Also, the destination address is IMSI in MT messaging.Step 5.The MSC sends MAP SEND INFO FOR MT SMS to the VLR to obtain the subscriber-related information.Step 6.When the VLR receives the message MAP SEND INFO FOR MT SMS , a micro Check Indica-tion is invoked to check the data value of the message.If the tests are passed, the VLR initiates the paging pro-cedure. If the VLR knows the Location Area (LA) of the recipient MS, then MAP PAGE is sent to page the MS in the LA. Otherwise,MAP SEARCH FOR SUB-SCRIBER is sent to page all LAs in the MSC. For sim-plicity, assume that MAP PAGE is sent.Steps 7±9.The MSC performs the paging opera-tion. If the operation is successful, then the MSC sends the MAP PROCESS ACCESS REQUEST message to the VLR.Step 10.The VLR sends MAP SEND INFO FOR MT SMS ack to the MSC, and the MSC is al-lowed to initiate forwarding of the short message to the MS.Steps 11±20.(See Fig.8.)The SMR entity of the MSC generates RP-DATA from the TPDU SMS-DELIVER , and relays it to the MS. These steps are sim-ilar to Steps 1±6and 13±19in MO messaging.Step 21.(See Fig.9)After the MSC receivesFig.9.Mobile terminated short messaging (III).the SMS-STATUS-REPORT ,the result is included in MAP FORWARD SHORT MESSAGE ack to ac-knowledge the message sent in Step 4.Step 22.If the result sent from MSC indicates that the recipient MS is absent,the SMS GMSC may invoke the procedure MAP REPORT SM DE-LIVER STATUS and informs the HLR the status for further processing. The SMS GMSC may take actions for other failures /errors. In any case, it forwards the result to the SM-SC.The MT messaging procedure is similar to GSM call termination [4], with one major difference. In MT messaging, when the HLR is queried by the GMSC, the destination MSC number stored in the HLR is sent back to the GMSC directly. In GSM call termination, the HLR always queries the VLR to obtain the MSC number and the related routing information. Note that the MSC num-ber stored in the VLR is more up to date than that in the HLR. Since a voice call is circuit-switch-oriented andLin 256the voice trunk must be set up from the GMSC to theterminating MSC, accurate routing information must beobtained from the VLR to avoid unnecessary trunk reser-vation due to misrouting. On the other hand, MT messag-ing is circuit-switch-oriented, where trunk setup is notneeded. Thus, the MSC number is obtained directly fromthe HLR to simplify the routable address query proce-dure. If the MSC number is out of date (which is not verylikely), the SMS protocol considers the MS is absent(and Step22is executed).5. CONCLUSIONSThis article provided a tutorial on GSM point-to-point short messaging. The short message protocol hier-archy was introduced and the mobile originated/ter-minated messaging procedures were described.Otherprocedures, such as the memory available noti®cationprocedure, follow the same rules described in the mobileoriginated/terminated messaging and are not covered inthis article. The reader is referred to ref.6. Inter MSChandover may occur during the short message delivery.The details can be found in ref.6. The short messageinformation ¯ow is described at Annex3in ref.3. Arecipient MS can reply a short message as speci®ed atAnnex4in ref.3.ACKNOWLEDGMENTSThe author would like to thank Kaveh Pahlavanand the three reviewers for their valuable comments andassistance in improving the quality of this paper. Thiswork was supported in part by National Science Coun-cil Contract No. NSC86-2213-E-009-074.REFERENCES1. CCITT.Numbering plan for the ISDN era. Technical Report Rec-ommendation E.164(COM II-45-E), ITU-T,1991.2. ETSI/TC.Technical Report Recommendation GSM03.47, ETSI,1993.3. ETSI/TC. Technical Realization of the Short Message ServicePoint-to-Point, Version4.6.0. Technical Report RecommendationGSM03.40, ETSI,1993.4. ETSI/TC.Mobile Application Part(MAP)Speci®cation,Ver-sion4.8.0. Technical Report Recommendation GSM09.02, ETSI,1994.5. ETSI/TC.Mobile Radio Interface Signaling Layer3(Phase2),Version:4.2.0. Technical Report Recommendation GSM04.07,ETSI,1994.6. ETSI/TC.Point-to-Point (PP) Short Message Service (SMS) Sup-port on Mobile Radio Interface, Version:4.5.0. Technical ReportRecommendation GSM04.11, ETSI,1994.7. Glenarye Electronics, IXO TAP Protocol. Technical Report GLP-3000-180, Glenarye Electronics,1991[see also ftp://mirror.lcs./telecom-archiv es/technical/ixo.tap.protocol].8. Y.-B.Lin,Paging systems:A tutorial.Technical Report PCS-NCTU-96-24, Department CSIE, National Chiao Tung University,1996[see also .tw].9. Y.-B.Lin and S. K. DeVries, PCS network signaling using SS7.IEEE Personal Communication s Magazine,Vol.2,No.3,pp.44±55,1995[see also .tw].10. PCIA,Telocator data protocol. Technical Report ttp://www.mot.com/MIMS/MSPG/pcia protocols/tdp,Personal Communica-tions Industry Association, July1995.Yi-Bing Lin received his BSEE degree from National ChengKung University in1983, and his Ph.D. degree in Computer Sciencefrom the University of Washington in1990. From1990to1995, hewas with the Applied Research Area at Bell Communications Research(Bellcore), Morristown, New Jersey. In1995, he was appointed Profes-sor in the Department of Computer Science and Information Engineer-ing (CSIE), National Chiao Tung University (NCTU). In1996, he wasappointed Deputy Director of Microelectronics and Information Sys-tems Research Center, NCTU. In1997, he become Chairman of theCSIE, NCTU. His current research interests include design and analy-sis of personal communications services network, mobile computing,distributed simulation, and performance modeling.Dr. Lin is a subject area editor of the Journal of Parallel andDistributed Computing, an associate editor of the International Jour-nal in Computer Simulation, an associate editor of IEEE Networks,an associate editor of SIMULATION magazine, an area editor of ACMMobile Computing and Communication Review, a columnist of ACMSimulation Digest, a member of the editorial board of InternationalJournal of Communications Systems, a member of the editorial boardof ACM/Baltzer Wireless Networks, a member of the editorial board ofComputer Simulation Modeling and Analysis, Program Chair for the8th Workshop on Distributed and Parallel Simulation, General Chairfor the9th Workshop on Distributed and Parallel Simulation, ProgramChair for the2nd International Mobile Computing Conference, thePublicity Chair of ACM Sigmobile, Guest Editor for the ACM/BaltzerMONET special issue on Personal Communications, Guest Editor forIEEE Transactions on Computers special issue on Mobile Computing,and Guest Editor for IEEE Network special issue on PCS NetworkManagement. Dr. Lin is a senior member of IEEE.。
文献检索
10、下面提供E-mail检索服务的是( A )。
A. Bigfoot B. yahoo C. Altavist D. Anywho
11、检索工具的功能主要体现在( C )。
A.附录部分B.目次表格C.索引部分D.使用说明
12、计算机文献检索中，每一种文献特征对应于计算机数据库中的( C )。
A.量大分布广，离散程度高B.组织形式多样化
C．检索方式灵活D．规范化程度较高
22、在CBMdisc检索中，查找期刊《肿瘤》上的文献，其正确表达式为( D )。
A.肿瘤in SO B. SO=肿瘤
C.肿瘤in TA D. TA=肿瘤
23、文献检索中，主题语言是一类检索语言，以下哪一个不是检索语言（A）。
2、在CIN20年
C. 15年D. C. 15年
18、下列说法中哪项是错误的?( A )
A. SCI是由美国国立医学图书馆编辑出版
B. SCI收录有期刊论文、会议摘要、综述以及部分期刊上的书评
C. SCI是一个综合性的大型科技文献检索工具刊
D. SCI为双月刊，每两个月出版一期
19、在计算机文献检索中，如果检索课题所涉及的内容比较专深，往往采用( B )的检索方法。
29、中国生物医学文献数据库的英文缩写是：( C）
A. GBMdisc B. ZGBMdisc
C. CBMdisc D. CGBMdisc
30、医学主题词表（MeSH）有哪二个主要部分（B）
A.主题部分、著者部分B.字顺表、树状结构表部分
C.述评部分、主题部分D.注释部分、树状结构表部分
31、下面哪个检索式的作用与短语检索"digiatal camera"等价：( A )。
fulltext(2)
Wanagel, J., Sass, S.L., and Batterman, B.W., Phys. Stat. Sol. (a), 10, 49-57 (1972)ቤተ መጻሕፍቲ ባይዱ 72Wan2: Wanagel, J., Sass, S.L., and Batterman, B.W., Phys. Stat. Sol. (a), 11, K97-K100 (1972). 72Wesl: Westlake, D.G. and Mueller, M.H., Proc. Int. Meeting on Hydrogen in Metals, Jiilich, 1972, 66-75; abstract available in Ber. Bun. Gesellschaft fiir Phys. Chem., 76(8), 757 (1972). 72Wes2: Westlake, D.G., Ockers, S.T., Mueller, M.H., and An~terson, K.D., Met. Trans., 3, 1709-1710 (1972). 73Asal: Asano, H. and Hirabayashi, M., Phys. Stat. Sol. (a), 15, 267-279 (1973). 73Asa2: Asano, H. and Hirabayashi, M., Phys. Stat. Sol. (a), 16, 67-72 (1973). 73Can: Canelli, R. and Mazzolai, F.M., Appl. Phys., 1, 111-118 (1973). 73Cha: Chang, H.Y. and Wert, C.A., Acta Met., 21, 1233-1242 (1973). 73Wes: Westlake, D.G. and Ockers, S.T., Met. Trans., 4, 1355-1357 (1973). 74Aro: Arons, R.R., Bohn, H.G. and Liitgemeier, H., J. Phys. Chem. Solids, 35, 207-214 (1974). 74Cam: Cambini, M., Serreels, R., and Gevers, R., Phys. Star. Sol. (a), 21, K57-K59 (1974). 74Kle: Kleppa, O.J., Dantzer, P., and Melnichak, M.E., J. Chem. Phys., 61, 4048-4058 (1974). 74'lhk: Takano, S. and Suzuki, T., Acta Met., 22, 265-274 (1974). 75Fuk: Fukai, Y. and Kazama, S., Scripta Met., 9, 1073-1076 (1975). 75Wes: Westlake, D.G. and Ockers, S.T., Met. Trans., 6A, 399-402 (1975). 76Asal: Asano, H., Abe, Y., and Hirabayashi, M., Acta Met., 24, 95-99 (1976). 76Asa2: Asano, H., Abe, Y., and Hirabayashi, M., J. Phys. Soc. Jpn., 41, 974-976 (1976). 76Ent: Entin, I.R., Somenkov, V.A., and Shil'shtein, S.Sh., Fiz. Tver. Tela., 18, 2965-2970 (1976) in Russian; TR: Soy. Phys.-Solid State, 18, 1729-1734 (1976).
fulltext翻译
J Appl Electrochem (2009) 39:577-582DOI 10.1007/s10800-008-9695-zORIGINAL PAPERElectrochemical treatment of pharmaceutical azo dye amaranthfrom waste waterRajeev Jain . Nidhi Sharma . Keisham RadhapyariReceived: 25 February 2008/Accepted: 13 0ctober 2008/Published online: I November 2008@ Springer Science+Business Media B.V. 2008Abstract The electrochemical behavior of pharmaceuti-cal azo dye amaranth has been investigated in distilledwater and Britton-Robinson buffer. One well-definedirreversible cathodic peak is observed. This may beattributed to the reduction of the -N=N- group. Calculationof the number of electrons transferred in the reductionprocess has been performed and a reduction mechanismproposed. Results indicate that the electrode process isdiffusion controlled. The cathodic peak in the case of controlled potential electrolysis is found to reduce substantially with a decrease in color and absorbance. The reaction has first order kinetics with k value 5.75 x10-2 abs min-l. The effciency of different electrode materials (platinum and steel) for decolorisation is compared. Chemical oxygen demand (COD) decreases substan-tially from 2,680 t0 96 ppm at platinum and t0 142 ppm atsteel. This translates t0 97% COD removal at platinum and95% at steel.Keywords Electrochemical treatment . AmaranthAzo dye . Industrial effluents . CV, CODI IntroductionAzo dyes continue to be a source of pollution fromindustrial processes which employ dyes to color paper,R. Jain . N. Sharma . K. RadhapyariSchool of Studies in Chemistry, Jiwaji UniversityGwalior 474011, Indiae-mail: raj eevj ain54 @ yahoo.co.inplastics, foodstuffs, pharmaceutical products, and naturaland artificial fibers [1, 2]. It is reported that approximately5 tonnes of dye discharge from dye and coloration industriesevery year [3]. The release of such compounds into theenvironment is of great concern due to their toxicity,mutagenicity, carcinogenicity, and bio-transformationproducts [4-6]. Hence. much research has focused onmethods of azo dye destruction. Many treatment processeshave been investigated extensively to treat wastewaters suchas chemical precipitation [7], adsorption [8], biologicaltreatment [9], photocatalytic degradation [10, 11], electro-catalytic oxidation [12], ozonation [13], Fentons' reaction[14], and electrochemical methods [15-19]. Electrochemi-cal techniques are an attractive methodology for thetreatment of dye wastewaters. This technique has significantadvantages viz., wide application, simple equipment, easyoperation, lower temperature requirements. and no sludgeformation [20-24].Amaranth (Fig.1) [trisodium salt of l-(4-sulpho-l-naphthylazo)-2-naphthol-3, 6-disulphonic acid] is an acidicmonoazo dye used as food and pharmaceutical colorant.Only a few analytical methods, such as square waveadsorptive stripping voltammetry (SWAdSV) and spec-trophotometry are available for determination of amaranthin soft drink samples [25]. Degradation of amaranth fromenvironmental samples has been studied on activated car-bon fiber (ACF) electrodes [26,27]. The amaranth azo dyehas electroactive groups. However. its electrochemicalbehavior and treatment have not been investigated.Therefore, cyclic voltammetric and differential pulsepolarographic studies have been undertaken in the presentwork for understanding the electrochemical behavior ofamaranth. Results have been analyzed employing the cri-terion of complete decolorisation of the dye-containingsolutions.Springer578J Appl Electrochem (2009) 39:577-582Na07SOH S03NaFig. 1 Structure of amaranth2 Experimental2.1 InstrumentationO,NaCyclic voltammetric (CV) studies were carried out on an EGand G potentiostat (Princeton Applied Research) integratedwith applied electrochemistry software. The working elec-trode potential was cycled between -1.2 and +1.2 V atdifferent sweep rates (50-2.000 mV s-1). The electrochemical cell consisted of three electrodes (in close proximity) immersed in the solution to be electrolyzed. The voltammetric behavior was studied using platinum as working electrode, SCE as reference, and platinum wire as counter electrode. Controlled potential electrolysis was carried out using a cyclic voltammograph coupled to a digital electronic 2000 0mnigraph x-y/t recorder. The working electrodes used for controlled potential electrolysis (CPE) were platinum foil (3 x 3 cm2)and steel foil (4.5 x 3.5 cm2), Ag/AgCI as reference elec-trode and platinum wire as counter electrode.Differential pulse polarographic (DPP) measurementswere carried out on an Elico pulse polarograph model CL 90 connected with Polarocord recorder model LR-108.Triple distilled mercury was used for the DME. The capillary had a flow rate of 3.02 mg s-l together with a drop time of 3 s. The pH measurements were made on a Hach digital EC-40 Benchtop pH/ISE meter. The absorption spectra of samples were recorded using an Elico SL 159 UV-Visible spectrophotometer. The chemical oxygen demand (COD) was determined using the open reflux method using COD digester apparatus (Spectralab 2015-S).The synthetic azo dye amaranth was obtained from Aldrich USA. All chemicals used were AR grade. Readymade silica gel G plate for TLC, having fine coating on alumina sheet. was obtained from E-Merck.2.2 Reagents and materialsStock solution of amaranth (2 x l0-3 M) was prepared in doubly distilled water. In order to evaluate the effect of varying pH, BR buffers in the pH range 2.5-12.0 were prepared as per a literature method[,. ]. The supportingSpringerelectrolyte was l.0 M KCI.For the COD experiments reagents were prepared in accordance with standard methods (APHA. 1995) [29].2.3 ProcedureThe CV and DPP studies were carried out by mixing l.0 mL of potassium chloride, 1.0 mL stock solution, and 8.0 mL of appropriate BR buffer/distilled water. Solutions of different concentrations were prepared. Dissolved oxygen was removed from the solution by passing nitrogen gas for about 15 min. The polarograms and cyclic voltammograms were then recorded. The redox behavior was studied at varying pH (2.5-12.0), concentrations and sweep rates (100-2,000mV s-l). Controlled potential electrolysis of the dye solution was performed at slightly more negative potential than the peak potential of the respective peak. Absorbance of the olution was measured at different time intervals at 520 nm.The value of the rate constant k was calculated from the l091(absorbance) vs time plots. The number of electrons transferred was calculated from the decrease in current with time during electrolysis. Controlled potential coulometry was also carried out at different pH values. The progress of electrolysis was monitored by recording cyclic voltammograms at regular intervals of time. The end products of electrolysis were identified by TLC. For COD studies. The experiments were carried out as per standard methods.2.4 CoulometryFor the coulometric determination of number of electrons "n" consumed in the reduction. a solution of depolarizer,potassium chloride, and buffer/distilled water was mixed in the same ratio as that for CV and DPP studies. The solution was de-aerated by passing nitrogen gas for 15 min and cyclic voltammograms were recorded at slightly more negative potential than the peak potential. With the progress of electrolysis the color of the solution gradually faded and finally a colorless solution was obtained. The number of electrons "n" transferred at the platinum electrodewas determined from the formula Q = nFN.3 Results and discussion3.1 Cyclic voltammetric (CV) studiesThe cyclic voltammogram of amaranth in distilled water (2 x 10-4 M) exhibits reduction peak at -0.872 V and corresponding oxidation peak at -0.779 V at scan rate of 100 mV s-l. The cathodic peak can be safely assigned to the reduction of the azo (-N=N-) group. The separation between cathodic and anodic peak potentials is more than 60 mV. indicating the irreversible nature of the electrode process [30]. As the scan rate (v) is increased. the reduction peak potential shows negative shift and the oxidation peakpotential shows positive shift. At higher scan rate of1,000 mV s-1 the reduction peak appears at -1.0 V andthe oxidation peak at -0.697 V. The peak potential separation (AEp) increases gradually from 0.093 t0 0.306 V asthe scan rate is increased from 100 to 1,000 mV s-l suggesting the irreversible nature of the electrode process.The plot of ip,c vs 1/2 in the 6.5 pH solution is a straight line passing through the origin indicating the diffusion controlled nature of the electrode process (Fig. 2 (I)) This proportionality may be attributed to the fact that a steeper concentration gradient is established thereby increasing the rate of diffusion at faster scan rates. Reduced species will diffuse away from the electrode faster as the scan rate is increased [31,32]. A similar plot for the dye solution in neutral aqueous medium and at pH 8.8 is,once again, a linear behavior, however, not passing through the origin. This indicates adsorption effects contributing to diffusion current. In such a case an adsorbed species may undergo electron transfer.3.2 Effect of pHWell-defined cathodic peaks in the acidic pH range are obtained with both platinum and glassy carbon electrodes.With increase in pH the cathodic peak shifts negatively and the anodic peak positively with increasing pH, which indicates that proton transfer occurs as a step consecutive to an irreversible electrode process [33]. The plot of Ep.c vs pH is linear up to pH 6.5 as also above 6.5.However. the two linear segments have different slopes.The value of pH 6.5, is in accord with the pKa value.Above pH 6.5, the Ep,c becomes practically independent of pH. This indicates the reduction of unprotonated species34]. These results are presented in Table .3.3 Differential pulse polarographic studiesA single four-electron irreversible reduction peak is observed in the pH range 2.5 t0 12.0 at the mercury electrode. This may be attributed to reduction of the azo group(-N=N-). A shift in Ep,c towards more cathodic potentialwith pH, along with a break at pH 6.5, is observed. Beyondthis there is near constancy in Ep,c (Fig.3). This suggestsparticipation of protons in the rate determinatiAnalysisofthepeakEd.e.vs log刍）- O;,at4m, lo[rt]Xtand shifting of the peak potential towards more negativevalue with concentration suggests the irreversible nature ofthe electrode process [35,36]. The results of DPP studies arethus in close agreement with those of the CV studies.3.4 Controlled potential coulometric studiesBy using controlled potential coulometry, the number ofelectrons transferred "n" at platinum electrode werecalculated and this lies in the range 2土0.2 (Table ..). Thevalue of "n" lies in the range 4 + 0.2 at the steel electrode.Controlled potential electrolysis of amaranth (2 x 10-4 M)at platinum and steel electrodes was carried out at -1.20 V.Bench scale electrochemical treatment was carried out.Electrolysis results in complete disappearance of colorwhich is fairly faster in case of platinum where the solutionis decolorized within 10 min of electrolysis leading to thecomplete disappearance of the cathodic peak current. Withsteel. the colored solution took 80 min for complete colorremoval. The data are summarized in Table '_ . A compara-tive overlay of the dye solution of pH 8.8 before and afterelectrochemical treatment is presented as Fig. i.3.5 Spectral studiesUV-visible spectra of amaranth (2 x 10-4 M) in distilledwater were recorded at Amax = 532.5 nm. The progress of 有错controlled potential electrolysis was monitored by record-ing spectral changes at different time intervals. AtAnax = 532.5 nm, absorbance decreases systematically 有错with the progress of electrolysis. The kinetic measurementswere conducted at steel electrodes. The observed rateconstant, k= 5.75 x 10-2 a s min-l was determinedfrom the first-order kinetics plot of log absorbance vs time3.6 COD removalThe electrolyzed solution shows a substantial decrease inCOD from an initial value of 2,680 ppm to a final value of96 ppm at platinum and t0 142 ppm at steel.3.7 Proposed mechanismOn the basis of the coulometry, controlled potential elec-trolysis, chromatographic and spectral analysis, tworeduction mechanisms are proposed. One is the two elec-tron reduction to hydrazo compound at platinum (Fig. ).A four electron reduction at steel (Fig. ) may be favoredby the presence of the strong electron releasing substituent,the -OH group [37].In Fig.6, amaranth undergoes 2e- reduction in acidic aswell as alkaline medium. The reduction process proceedsby the protonation of the polarized molecule resulting in the formation of (B). In the second step, which is slow and rate determining, (B) accepts 2e- and a proton and results in the formation of stable hydrazo (-NH-NH-) form (C).In Fig.7, 4e- reduction takes place at the steel foil electrode. The I and II steps of reduction proceed in the same way as described in Fig. . In step III. (c) the hydrazo moiety undergoes a further two electron reduction resulting into two products (D) and (E).4 ConclusionThe newly developed electrochemical method gives satis- factory and promising results. The electrochemical reduction of amaranth under the experimental conditions described in this work is an irreversible process controlled by diffusion. Both platinum and steel electrodes exhibit great stability and resistance to redox and acidic/basic environments showing no deactivation. During the elec- trochemical degradation process, the COD decreases by approximately 97u/o at the platinum electrode and by 95% at the steel foil electrode with complete color removal. The electrochemical treatment developed achieves higher de- coloration and COD removal than that reported previously through the ACF electrode treatment method [26] which gives only 60% COD removal. Hence, the present elec- trochemical procedure is a better alternative approach for wastewater treatment resulting in significant lowering of toxicity.。
- 1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。
Environmental Monitoring and Assessment(2005)102:285–308DOI:10.1007/s10661-005-6028-0c Springer2005 FLORISTICS,PERFORMANCE AND PROGNOSIS OF HISTORICAL TREES IN THE URBAN FOREST OF GUANGZHOU CITY(CHINA)C.Y.JIMDepartment of Geography,The University of Hong Kong,Pokfulam Road,Hong Kong(e-mail:hragjcy@hkucc.hku.hk)(Received10July2003;accepted9April2004)Abstract.Outstanding historical trees embedded in cities constitute pertinent environmental as-sets,yet they are widely threatened in third-world cities.Inadequate understanding of this valuable natural-cum-cultural heritage hinders proper conservation.A case study of Guangzhou in south China evaluatedfloristic composition,age profile and biomass structure of historical trees,assessed their performance in major habitats(institutional,park and roadside),and established a prognosis for future growth and management.The348historical trees examined belonged to only25species,vis-`a-vis 254trees in the entire urban forest,dominated byfive species and native members.Roadside had more trees,followed by institutional and park,with merely the most common four species shared by all habitats.The limited commonality reflected tree-performance differentiation by habitats exerting selection pressure on species.The institutional growth-regime was more conducive to nurturing high-caliber specimens,whereas park is less capable.Individual species achievement by habitats,derived from tree-count ranking and relative-abundance indices,could inform species choice and tree conser-vation.Few trees exceeded300years of age in the millennium-old city,echoing a history of intense tree–city conflicts.Potential life-span,trunk and crown diameters indicated ample opportunities for further expansion of biomass and landscape impacts,which would be straitjacketed by the tightening urban fabric.Keywords:urban trees,urban forest,historical trees,tree habitat,tree performance,tree protection, Guangzhou,China1.IntroductionIn the course of a city’s development,pre-urbanization natural features such as individual or groves of outstanding trees could sometimes be preserved in situ and subsequently became embedded within built-up areas(Jim1989;Nowak,1993). Some trees could survive the trying transition from a natural forest habitat to a stressful,fragmented and isolated urban habitat;others might be emaciated or even perish in the alien environs(Dorney et al.,1986;Swenson and Franklin, 2000).Their occasional presence in the otherwise artificial city environment denotes the remnant bequest of a former natural lineage.The relatively higher degree of naturalness and collateral ecological and environmental benefits,in comparison with other urban greenery,signifies their pertinent if not unique functions in the city.As it is usually the large and exceptionally robust and beautiful trees that were chosen for preservation,their vivid contrast with the surrounding developments286C.Y.JIMtends to accentuate their landscape impacts.The inheritance of the cream of nature’s remnants is indeed a windfall for the landscape enhancement of a development area. Where and when development pressure was not too acute,more so in the past than the present,more chances were afforded for the expression of the innate human propensity to protect outstanding and aesthetic natural objects.Trees could also be planted after the advent of urban growth,to co-exist with more pervasive if not overwhelming buildings and artificial surfaces.The unfavor-able growth conditions are often inimical to the nurturing offine specimens trees (Bradshaw et al.,1995;Grey,1996).A small cohort of such cultivated members, by virtue of the chance concurrence of genetic superiority and suitable site char-acteristics,could subsequently achieve meritorious performance by virtue of age, dimension and vigor,and they are often accorded historical and commemorative status.Some local residents literally grow up together with these trees which usually began their urban existence as small and feeble saplings.Unlike pre-urbanization inheritance,these planted trees have witnessed the vicissitudes of the city fabric and the constituent inhabitants,and experienced the harsh treatments imposed on them by people and the stressful environs(Gilbertson and Bradshaw,1985;Loeb,1992). Such outstanding denizens often instill emotional attachment of the community and trigger the protective instinct of citizens and governments.It is not uncommon in different cultures to elevate such respect to the level of veneration or even worship (Hughes,1984),extending to assiduous guarding of the city’s natural gems.Enlightened official policies,combined with sympathetic developers and a green-conscious citizenry are required to usher their preservation and subsequent protection.The success of this endeavor necessitates the confluence of the minds all three partners in the tripartite.In old cities that have often experienced some episodes of development and redevelopment,existing trees of high quality could be lost by felling,or so badly damaged by construction activities that they are no longer worthy of conservation(Goldsmith,1988;Watson and Neely,1995).Intensifica-tion of development is a particularly destructive force that tended to eliminate both cultural and natural heritage in a clean-slate mode,and the problem is often more acute in private lands(Coughlin et al.,1988).The infilling of brownfields and the incursion into greenfields are reducing the green cover of many fast-growing cities (Arnold and Gibbons,1996).In many cities in the developing world,the lack of an effective heritage preservation policy and an adjunct enforcement institution has allowed important historical relics to be eradicated or gravely degraded(Olembo and Rham,1987).In the face of development pressure,cities and citizens are increasingly alarmed by the loss of precious and familiar amenity vegetation,furnishing the impetus to enhance the protection of such living and venerable companions(Bowers,1999). Many municipal authorities regarded them as prized possessions and bestowed statutory protection and special care(Duerksen and Richman,1993;Department of the Environment,1994).It is common for cities to establish official registers to fos-ter their management and to publicize their heritage and environmental–ecologicalHISTORICAL TREES IN URBAN GUANGZHOU287 values as the oldest natural-cum-cultural heritage.They have been accorded spe-cial epithets to echo their importance to the community,namely ancient,beautiful, big,champion,elite,famous,heritage,historic,old,outstanding,remarkable,spec-imen,and veteran trees(Randall and Clepper,1977;May,1990;Mitchell et al., 1990;Alderman and Stevenson,1993;Jim,1994a,b;Pelt,1996;Parkenham,1997; Lewington and Parker,1999;Read,2000;Browne,2001;Meyer,2001).In some cities,such emblem urban trees hitherto have been catalogued in official registers but have seldom been subjected to detailed scientific analysis.A compre-hensive evaluation of the exemplary arboreal population will serve the following objectives:(1)throw light on the fundamental characteristics offloristic composi-tion and biomass structure;(2)permit an understanding of the underlying factors that contribute to their above-and below-par performance in a regime of intense tree–environment interactions;(3)contribute to the effective protection and man-agement of the present cohort,and a planting and management strategy to nurture future candidates in new development areas.This study focuses on the in-depth assessment of the historical trees of a rapidly growing city in developing China, using Guangzhou in south China as a case study.2.Study Area and Methods2.1.S TUDY AREAThe project is based in Guangzhou which is the largest metropolis in south China with a total population of about7million people.The study area,covering eight main urban districts of Guangzhou with116km2of developed area,has about4million residents(Guangzhou Statistical Bureau,2001).The city is situated strategically at the head of the Pearl River Estuary,serving as the gateway and the principal administrative,commercial and industrial hub of south China.With a history of over2500years,the city is a mix of old neighborhoods with a typical tight town plan,and new developments with a more open cityscape.Similar to other Chinese cities,population,building and road densities are high,being typical of the compact-city morphology.Since the inception of the official open and reform policy in1978, the city has undergone a rapid phase of redevelopment of brownfields,infilling of occluded greenfields,and sprawling into the surrounding countryside.Located in the north subtropical zone,the city has a well-defined seasonal pattern of warm-humid summers and cool-dry winters,with annual precipitation reaching 1690mm and mean air temperature of21.8◦C(Huang et al.,1994).The typ-ical monsoon weather regime is occasionally punctuated by thunderstorms and typhoons accompanied by high winds exceeding100km/h and torrential down-pours(Jim and Liu,1997).The topography is mainly alluvial plains laid down by the river,accompanied by scattered coastal wetlands,low terraces and residual hills. The low-lyingflats are covered by immature alluvial soils,whereas the slopes and288C.Y.JIMhills carry a mantle of tropical red earth.The pristine natural vegetation is believed to be the remnant of tropical rainforests with primarily evergreen broadleaved trees plus some deciduous elements,dominated by three botanical families(Lauraceae, Moraceae and Caesalpiniaceae;Jim,2002b).The original vegetation mantle has been almost completely eradicated by centuries of agricultural and recently urban-industrial activities.Some native species,however,have managed to survive in various semi-natural and ruderal habitats in the city.Guangzhou has established a sound foundation of urban forest,with a good tree cover mainly found in three habitat types,namely greenspaces,institutional grounds and roadsides(Jim and Liu,2001a).Centuries of massive human dis-turbances and modifications have left their imprints on the urban forest,which, however,still decidedly carries a strong native character in species composition. The entire city,including occluded enclaves of undeveloped areas,has an exception-ally rich biodiversity of1400vascular plant species,reflecting its inherent tropical biodiversity treasure augmented over the millennia by plant introductions from dif-ferent parts of China and other lands.As one of the earliest Chinese cities opened to other countries,many exotic species,notably from Southeast Asia and Australia, have been introduced into Guangzhou(Wang et al.,1994;Hou,1956).Gauged by the common yardsticks of tree coverage,tree count and species composition, Guangzhou joins the rank of the meritorious green cities of China(Guangzhou Municipal Government,1928;Yang,1991).Recent intensive developments have damaged some urban trees and degraded the quantity and quality of the growth conditions of others.An extensivefield survey of the study area in1995established that a population of115,140trees occurred within the study area,representing254 species in62botanical families;40.8%by roadsides,38.2%in parks,and21.0% in institutional grounds(Jim and Liu,2001b).Many native species,especially by roadsides,dominate the cityscape in terms of species count and tree number.2.2.M ETHODSTwo approaches were adopted to study historical trees in the study area,namely official records andfield assessment.An official register of old trees established by the municipal government in1985and updated in1995recorded348specimens. Targets were selected principally by age>100years regardless of species and provenance,and secondarily on trees with special official commemorative value. The species,locations and environs of individual historical trees were gleaned throughfield surveys.Individual trees were evaluated for tree dimensions(Philip, 1994),habitat,and constraints to growth.Three tree dimensions were measured: crown diameter measured by taping the projected drip line;tree height measured with the help of an Abney level;and trunk diameter at1.4m(DBH)calculated from girth obtained by a measuring tape.Ficus trees required special attention due to the common occurrence of lignified aerial roots entwining around the trunk; if contiguous with the trunk they were measured as part of the trunk diameter.HISTORICAL TREES IN URBAN GUANGZHOU289 For trees with multiple stems,the DBH were the aggregated values of constituent trunks.Habitat characteristics were determined by reference to main land use types (roadside,park,government,education,religious,and other institutional grounds), and to development history.Major artificial growth constraints were recorded as the presence or absence of items such as adjacent infrastructural development or urban renewal,vandalism,poor care and major natural phenomena,e.g.typhoon, thunderstorm,disease and insects.Tree age was estimated using tree-ring analysis, correlation with the history of road and neighborhood development,experience of tree size–age relationships,and evidence from historical photographic and his-torical records.The botanical nomenclature of South China Institute of Botany (1987,1991,1995)was adopted.Statistical analysis was performed using SPSS/PC Version10and Microsoft Excel2000.3.Results and Discussion3.1.S PECIES COMPOSITION AND HABITAT VARIATIONSA total of348historical trees were identified in Guangzhou,of which338were over100years old;the remaining10were officially labeled as“famous trees”due to association with distinguished persons or events.The entire urban-tree population(UTP)in Guangzhou includes254species,dominated by29common species that constitute75%of the population.The historical-tree population(HTP) is represented very unevenly by25species(Table I),with the topfive contribut-ing87.4%of the trees,and with overwhelming presence of Ficus microcarpa at 41.4%.The next four species together took up merely6.0%.A significant num-ber of UTP species,ranging from common to rare,failed to attain the top-notch performance.The three major habitats had dissimilar HTP species distributions(Table II). More HTP trees are found in roadside habitat(44.3%),followed by institutional (31.6%)and park(24.1%).In the UTP,the sequence was roadside,park and in-stitutional.Institutional and park had similar number of historical species(at16 and17,respectively),whereas roadside had a much lower diversity(only seven species).For the HTP,only the top four species are common to all three habitats; six species are common to two and as many as14species are found in only one habitat.In contrast,the UTP had15,7and3species in the same groups.The low level of commonality in the HTP indicated the selective effect of habitats in foster-ing the development of elite trees.Only a handful of species had generalist(versus specialist)ecological requirement to excel in all three habitats and to sustain the excellence over tenure of decades to centuries.Comparing the municipality’s UTP and HTP provides insights on the species composition of the latter.The25HTP species represent16botanical families290C.Y.JIMTABLE ISpecies composition,botanical affiliation and abundance of historical treesSpecies Common name Family Count%fFicus microcarpa a,e Chinese Banyan Moraceae14441.38 Ficus virens a,e Big-leaf Banyan Moraceae4914.08 Cinnamomum camphora a,e Camphor Tree Lauraceae4713.51 Bombax malabaricum b,c,e Red Kapok Bombacaceae3810.92 Cinnamomum burmanii a,e Cinnamon Tree Lauraceae267.47 Ficus religiosa c,d Peepul Moraceae9 2.59 Bischofia javanica a,b,c Chinese Honey Locust Caesalpiniaceae4 1.15 Gleditsia sinensis a,b Flat Mango Caesalpiniaceae4 1.15 Mangifera persiciformis b,d Autumn Maple Anacardiaceae4 1.15 Adenanthera pavonina b,c Red Sandalwood Mimosaceae20.57 Dimocarpus longan b Longan Sapindaceae20.57 Ficus altissima a Mountain Fig Moraceae20.57 Litsea monopetala a Persimmon-leaf Litsea Lauraceae20.57 Michelia alba a,b,d,e White Champak Magnoliaceae20.57 Prunus mume b,d Flowering Plum Rosaceae20.57 Vitex quinata c Orange-bark Vitex Verbenaceae20.57 Artabotrys hexapetalus d Climbing Ilang-Ilang Annonaceae10.29 Cassia siamea a,b,d Kassod Tree Caesalpiniaceae10.29 Celtis sinensis a,e Chinese Hackberry Ulmaceae10.29 Dracontomelon duperreanum a,b,c Yanmin Anacardiaceae10.29 Ficus gibbosa a Humped Fig Moraceae10.29 Pinus massoniana Chinese Red Pine Pinaceae10.29 Quercus robur c,d English Oak Fagaceae10.29 Sterculia nobilis a,b China Chestnut Sterculiaceae10.29 Terminalia chebula c,d Chebulic-myrobalan Combretaceae10.29 Total348100.00a Shading function(with dense and large fully-grown crown>10m diameter).b Vividflowers(with conspicuous and ornamental blooms).c Attractive tree form.d Exotic species.e Species found in the top ranking29speices in the urban-tree population of Guangzhou city, collectively contributing75%of the trees.f The following frequency classes are used in the text:>or=100tree per species is abundant,40–99 is frequeny,10–39is occasional,2–9is rare,and1is solitary(Table I),and accounted for28.1%of the64families found in the UTP.The Moraceae is the most dominant plant family in both HTP and UTP withfive species and205trees.The genus Ficus accounted for20.0%of HTP species and59.0%of HTP trees,but less presence at7.9and16.3%,respectively ofHISTORICAL TREES IN URBAN GUANGZHOU291T A B L E I I A b u n d a n c e o f h i s t o r i c a l -t r e e s p e c i e s i n c o m p a r i s o n w i t h t h e u r b a n -t r e e p o p u l a t i o nH i s t o r i c a l t r e eU r b a n t r e eC o u n t %C o u n t %S p e c i e sI n s t i t u t i o n a l P a r k R o a d s i d e T o t a l R a n k I n s t i t u t i o n a l P a r k R o a d s i d e T o t a l I n s t i t u t i o n a l P a r k R o a d s i d e T o t a l I n s t i t u t i o n a l P a r k R o a d s i d e T o t a lF i c u s m i c r o c a r p a 31437014418.912.420.141.4271996329645630.93.210.614.7F i c u s v i r e n s 334124929.51.13.414.1392682807291461.32.226.129.5C i n n a m o m u m c a m p h o r a 51414731.40.311.813.537647*********.21.51.03.7B o m b a x m a l a b a r i c u m 201713845.74.90.310.9297650225332001.02.17.310.3C i n n a m o m u m b u r m a n i i 1252650.30.07.27.5175140914117250.64.60.55.6F i c u s r e l i g i o s a 630961.70.90.02.637498175<0.10.20.30.6B i s c h o fia j a v a n i c a 211470.60.30.31.158972984530.20.31.01.5G l e d i t s i a s i n e n s i s 130470.30.90.01.114050.0<0.10.00.0M a n g i f e r a p e r s i c i f o r m i s 013470.00.30.91.1013533540.00.01.11.1A d e n a n t h e r a p a v o n i n a 20280.60.00.00.6283310.1<0.10.00.1D i m o c a r p u s l o n g a n 20280.60.00.00.6166297455080.51.00.11.6F i c u s a l t i s s i m a 20280.60.00.00.6592245958780.20.71.92.8L i t s e a m o n o p e t a l a 11280.30.00.30.62680328310.12.6<0.12.7M i c h e l i a a l b a 020280.00.60.00.610911184108733623.53.83.510.9P r u n u s m u m e 020280.00.60.00.6220.0<0.10.0<0.1V i t e x q u i n a t a 110280.30.30.00.61120.00.00.0<0.1(C o n t i n u e d o n n e x t p a g e )292C .Y .JIMT A B L E I I C o n t i n u e dH i s t o r i c a l t r e eU r b a n t r e eC o u n t%C o u n t %S p e c i e sI n s t i t u t i o n a l P a r k R o a d s i d e T o t a l R a n k I n s t i t u t i o n a l P a r k R o a d s i d e T o t a l I n s t i t u t i o n a lP a r kR o a d s i d e T o t a l I n s t i t u t i o n a l P a r k R o a d s i d e T o t a lA r t a b o t r y s h e x a p e t a l u s 010190.00.30.00.301010.00.00.00.0C a s s i a s i a m e a 10190.30.00.00.31411251400.00.00.40.5C e l t i s s i n e n s i s 10190.30.00.00.32021511017130.74.90.05.5D r a c o n t o m e l o n 010190.00.30.00.372132452490.20.40.10.8d u pe r r e a n u mF i c u s g i b b o s a 010190.00.30.00.3325<0.1<0.10.00.0P i n u s m a s s o n i a n a 010190.00.30.00.31352148622890.46.90.07.4Q u e r c u s r o b u r 010190.00.30.00.30110.00.00.00.0S t e r c u l i a n o b i l i s 010190.00.30.00.3767051510.20.20.00.5T e r m i n a l i a c h e b u l a 10190.30.00.00.311112<0.10.00.0<0.1T o t a l n u m b e r o f s p e c i e s 161772521251625T o t a l n u m b e r o f t r e e s 1108415434831.624.144.3100.0345710771167273095511.234.854.0100.00P e r c e n t o f u r b a n -t r e e 0.10.10.10.33.09.414.526.9p o p u l a t i o n aaT o t a l t r e e c o u n t i n t h e u r b a n -t r e e p o p u l a t i o n i s 115,070t r e e s ,o f w h i c h 30,955t r e e s b e l o n g t o t h e 25h i s t o r i c a l t r e e s p e c i e s .HISTORICAL TREES IN URBAN GUANGZHOU293 the UTP.Other important families include Lauraceae and Bombacaceae,to-gether contributing21.6%of HTP species and10.9%of HTP trees.Collectively, these three families represent91.5%of HTP trees.Some common families and species in the UTP did not occur as champions.For example,Palmae and Myr-taceae accounted respectively for13.0and12.0%of UTP,and had no HTP representation.Only7of the29common UTP species were represented among the his-torical trees,namely thefive top-ranking HTP species(Ficus microcarpa,F. virens,Cinnamomum camphora,Bombax malabaricum,and C.burmanii),plus two lower ranking ones(Michelia alba and Celtis sinensis).Most HTP species are also uncommon in the UTP;16HTP species have<1%frequency,10of which also had<1%frequency in mon UTP species,Aleurites moluc-cana,Bauhinia purpurea,B.variegata,Caryota mitis,Casuarina equisetifolia and Melaleuca leucadendra,were absent from the HTP.However,some rare UTP species,Artabotrys hexapetalus,Prunus mume,Quercus robur,Sterculia nobilis and Terminalia chebula,were recognized as champions.Native species dominated the HTP;68.0%of species and94.0%of trees,and the major historical-tree species, F.microcarpa,F.virens,C.camphora,C.burmanii,and B.malabaricum,were na-tive(Table I).These species have been domesticated for urban planting for over 1000years in southern Chinese cities(Qu,1670;Hou,1956).Unlike the HPT,the UTP has comparable number of exotic and native species.Exotic species were un-common in the HTP,but some members have general religious significance,such as Ficus religiosa and Terminalia chebula.3.2.H ABITAT DIFFERENTIATION OF TREE PERFORMANCEDominated by native species,the performance of most historical trees indicated adaptation to the climate and soil conditions of the city.However,recent rapid urban development and rising building density had significantly degraded habitat quality and affected their growth and performance.Since thefirst inventory conducted in 1986(Landscape Bureau,1995),45trees out of the209identified had perished by 1995,and another92showed decline.Major threats are urban development,natural hazards and vandalism.Comparison between UTP and HTP by the three major habitat types indicated differential site conditions in nurturing high-caliber trees(Table II).Institutional contributed31%of the HTP trees,notably higher than the11.2%in UTP,suggesting conditions conducive to the development offine specimens.On the contrary,park and roadside had lower percentage of HTP trees(24.1and44.25%,respectively) than UTP(34.8and54.0%,respectively),suggesting that these two habitats are relatively less amenable to excellent tree growth.That park overall had a smaller proportion of high-quality trees than institutional is somewhat unexpected,and pointed to the need to improve tree management in public greenspaces to maximize294C.Y.JIMtheir potential to raise meritorious trees.The physical and institutional limitationsto superior tree performance in the relatively genial environment should be thor-oughly explored and ameliorated.Despite the stressful roadside habitat(Bassukand Whitlow,1988;Jim,2003),the largest share(44.3%)of the HTP trees dweltthere.Roadside trees endured more physical and physiological constraints,espe-cially in old districts with narrow streets(Jim and Liu,2001b).The tight growingspace sandwiched between the carriageway and adjacent buildings would curtailfuture expansion of roadside trees,many of which are yet to attain theirfinal di-mensions.Champion trees at roadside could provide more prominent landscapeand environmental benefits to compact cities.Finer comparison showed the involvement of individual species in habitat differ-entiation(Table II).In institutional,three common HTP species,Ficus microcarpa,F.virens and Bombax malabaricum,outperformed their UTP counterparts by a wide margin.In park,F.microcarpa and to a lesser extent B.malabaricum in HTPalso surpassed the UTP.Different species were found in roadside;besides the ubiq-uitous F.microcarpa,Cinnamomum camphora and C.burmanii in HTP performedbetter than UTP.Some species achieved notably below par performance.In park,the common F.virens and C.burmanii did poorly in HTP.In roadside,the commonF.virens and B.malabaricum were the under-performers in HTP.Three common UTP species were also poor achievers in HTP.Michelia alba was substandard inall three habitats,whereas Celtis sinensis and Pinus massoniana,rarely planted atroadside,were weak in park.The choice of species in future urban planting en-deavors could take into account the suitability of species in different habitats afterdecades offield testing in real-world situations.Comparing the ranking of species in different habitats could provide a quan-titative indicator of species performance(Table III).Except park,the Spearmancorrelations between the HTP and UTP species ranks were significant(institutional ρ=0.46,p<0.015,roadsideρ=0.58,p<0.01,totalρ=0.51,p<0.01).In institutional,F.virens rankedfirst in HTP,and B.malabaricum ascended to thethird position.An extreme contrast was found in M.alba which rankedfirst in UTPbut last in HTP,showing that most trees failed to achieve well.Of the3362treesof M.alba in UTP,only two were enlisted as historical trees in parks,with nonein institutional and roadside.Other species with a similar polarized performanceincluded C.burmanii and Celtis sinensis.Ficus religiosa did exceptionally wellin institutional,with an eight-step climb on the relative-rank league,mainly assid-uously protected in temple grounds as a Buddhism sacred paring withother habitats,park had the largest number of species,11out of25,experiencing arise in rank.Two common species had an elevated relative rank,namely F.micro-carpa,which preserved its premier position in the HTP,and B.malabaricum,which ranked second.Eight other uncommon species had different degrees of ascendancy, suggesting that the park habitat was favorable to the growth of some species.Pinus massoniana rankedfirst in UTP cohort but had a significant drop to17th rank in HTP;C.burmanii also had a big drop from third to20th.In roadside,all seven HTPHISTORICAL TREES IN URBAN GUANGZHOU295T A B L E I I I C o m p a r i s o n o f u r b a n a n d h i s t o r i c a l t r e e s p e c i e s b y t r e e -c o u n t r a n k i n g i n t h e i r r e s p e c t i v e p o p u l a t i o n sI n s t i t u t i o n a lP a r k R o a d s i d e T o t a lU r b a n -S c a l e d U r b a n -S c a l e d U r b a n -S c a l e d U r b a n -S c a l e d U r b a n t r e e H i s t o r i c a l -h i s t o r i c a l -R e l a t i v e U r b a n -t r e e H i s t o r i c a l -h i s t o r i c a l -R e l a t i v e U r b a n -t r e e H i s t o r i c a l -h i s t o r i c a l -R e l a t i v e U r b a n -t r e e H i s t o r i c a l -h i s t o r i c a l -R e l a t i v e S p e c i e st r e e a s u b s e t b t r e e ct r e e d r a n k e t r e e a s u b s e t b t r e e c t r e e d r a n k e t r e e a s u b s e t b t r e e c t r e e d r a n k e t r e e a s u b s e t b t r e e c t r e e d r a n k e F i c u s m i c r o c a r p a 23525.00.016512.92.14212.4−0.46212.6−0.6F i c u s v i r e n s 13212.5−0.522738.6−1.61124.9−3.91125.1−4.1C i n n a m o m u m143512.5−9.5319617.1−8.121737.3−0.329837.70.3c a m p h o r aB o m b a x 18437.5−3.525825.7−2.38349.7−6.7114410.2−6.2m a l a b a r i c u mC i n n a m o m u m 297717.5−10.573720.0−17.0379512.1−3.1196512.8−6.8b u r m a n i iF i c u s r e l i g i o s a 16218410.08.07414411.42.64611717.0−6.07515615.3−0.3B i s c h o fia j a v a n i c a 5913615.0−2.06413617.1−4.1228614.6−6.65012717.9−5.9G l e d i t s i a s i n e n s i s 18419717.51.514716411.44.611817717.00.019721717.93.1M a n g i f e r a 21420820.00.023020617.12.9196717.0−11.05413717.9−4.9p e r s i c i f o r m i sA d e n a n t h e r a 8614615.0−1.015217720.0−3.011817717.00.013218820.4−2.4p a v o n i n aD i m o c a r p u s l o n g a n 308615.0−7.03710720.0−10.06512717.0−5.04611820.4−9.4F i c u s a l t i s s i m a 5712615.0−3.04711720.0−9.0165717.0−12.0399820.4−11.4L i t s e a m o n o p e t a l a 8715717.5−2.5206720.0−14.09816614.61.44010820.4−10.4M i c h e l i a a l b a 51820.0−19.0134514.3−10.3124717.0−13.093820.4−17.4P r u n u s m u m e 21420820.00.015718514.33.711817717.00.021722820.41.6(C o n t i n u e d o n n e x t p a g e )。