Transactions of the Japan Society of Mechanical Engineers
Online ISSN : 2185-9485
Print ISSN : 0029-0270
ISSN-L : 0029-0270
Volume 27, Issue 177
Displaying 1-27 of 27 articles from this issue
  • Komei WATABE
    1961 Volume 27 Issue 177 Pages 543-552
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    Experiments on fluid friction of rotating disc which has protuberances on its surface are performed. The author arranged blades radially or concentrically on the disc and on the vessel, and carried out experiments on some combinations of disc and vessel which have blades or smooth surfaces. The main results obtained are as follows : 1) The relation between Cf and Re can be expressed by Cf=α·Re<-0.1575>. 2) The value of α increases first rapidly with the number of blades and then gradually approaches to a maximum value. 3) The shapes of blades (logarithmic and straight) have little influences on Cf for small clearance. 4) The increment value of Cf induced by the blades on disc is nealy twice as much as the value by the blades on vessel. 5) The boss ratio smaller than 0.5 have little influences on the value of Cf.
    Download PDF (1114K)
  • Tatsuro HOSHINO
    1961 Volume 27 Issue 177 Pages 553-557
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In order to see the effect of width of weir b, experiments are performed for b=0.2, 0.1, 0.05, 0.025, and 0.0125 m when D=0.2 m, and for b=0.025 m when D=0.1 m. According as the magnitude b decreases, the drop of hydraulic gradient (water surface) increases. Consequently, the coefficient of discharge deviates from that of c when b is large. Then, estimating the head h0 on the cross section at the weir plate, and recalculating c0 using h0, it has been found that the differences between c and c0 remarkably decrease. Further, the author obtains an empirical formula [numerical formula] which is based on the Rehbock's formula, where Q is discharge, b is width of weir, g is gravitational acceleration, h : head ; D : the height of weir ; c : coefficient of discharge. The empirical formula can be applied to (i) b >0.60 m, 1.0 m >D >0.1m, 0.6 m >h>0.025 m, D&gnE;h, (ii) 0.60 m&gnE;b&gnE;0.05 m, 0.2 m&gnE;D&gnE;0.1 m, D&gnE;h>0.025 m and the head is measured at the point 1 m above the weir plate and channel is made of wood, whose surface is planed smoothly.
    Download PDF (649K)
  • Tatsuro HOSHINO
    1961 Volume 27 Issue 177 Pages 557-563
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    For the rectangular weir made of thin plate, the effects of height of weir D on the coefficient of discharge c have not been revealed yet, Especially for weirs with a height D≒0, there are verv few experimental results. The author, therefore, for the weir with the range height of D≒0∼0.4 m, has carried out experiments of small rectangular weirs of b=0.10 m, these experiments are performed with respect to certain channel of which B=0.722 and B≒0.20 m, "Full Depth Weir" is the name given to the thin-plate rectangular weir having zero height of weir. Experiments on the full depth weir, of which the values b/B are nearly equal to zero, have been performed for various values of b and for wide range of head h (or m=h/b).
    Download PDF (852K)
  • Ei-ichi HORI
    1961 Volume 27 Issue 177 Pages 564-573
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    Unsteady boundary layer, caused by a fluctuating circulation, of a cylindrical body in a uniform stream, is investigated theoretically. Incompressible, two-dimensional, laminar flow is assumed. Further assumption is made that the magnitude of velocity fluctuation is small compared with the mean velocity and that the frequency of fluctuation is low. Blasius' series method of solution of the boundary layer equations is extended to the unsteady flow. Detailed calculation is shown of the flow around a body symmetrical, with velocity fluctuation outside the boundary layer antisymmetrical, with respect to the undisturbed stream. As in the steady flow problem, the solution of the unsteady boundary layer equation can be expressed as a linear combination of a set of universal velocity distribution functions. Four of the functions are actually computed, by use of which boundary layer Separation from a circular cylinder with circulation fluctuation is discussed. The separation-point movement is found to consist of a quasi-steady and an unsteady components, the latter increasing with frequency.
    Download PDF (995K)
  • Ei-ichi HORI
    1961 Volume 27 Issue 177 Pages 573-582
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    A method of calculating the fluctuating boundary layer on a circular cylinder, performing rotational oscillation in a uniform flow, is presented. Two-dimensional, laminar flow of an incompressible fluid is assumed. Equation of motion is linearized on the assumption that the velocity fluctuation in the boundary layer is small. The frequency of the fluctuation is also assumed rather low, thus retaining terms up to first order in frequency. By extending the Blasius' power series method to unsteady flow, solutions are expressed as a linear combination of a family of universal functions. As an example, the torque fluctuation of the cylinder is calculated. Streamline patterns near the separation point of the boundary layer are calculated, by observation of which a new criterion of the separation of the unsteady boundary layer is proposed. It is based on the asymptotic behavior, at a large distance from the solid wall, of the streamlines.
    Download PDF (1000K)
  • Naomichi HIRAYAMA
    1961 Volume 27 Issue 177 Pages 583-589
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    Boundary layers of a flat plate located parallel to the mean flow in the downstreams of a cylinder row in motion were empirically examined, in order to obtain fundamental know1edges concerning the mutual interference among blade rows working in viscous incompressible fluid. Forced by the fluctuation in the main stream velocity, a unique turbulent boundary layer-forced turbulent boundary layer-is developed at a distance from the leading edge of the flat plate. This boundary layer has several properties different from those of ordinary turbulent boundary layer. The characteristics, such as mean and fluctuation velocity distributions, of the forced turbulent boundary layer as well as of the boundary layer under transition to this type were investigated. With increased fluctuation velocity in the main flow around the leading edge, the thickness of the boundary layer near the leading edge increases. Factors influencing this phenomenon were discussed.
    Download PDF (887K)
  • Naomichi HIRAYAMA
    1961 Volume 27 Issue 177 Pages 590-598
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    A unique turbulent boundary layer-forced turbulent boundary layer-which is found on a flat plate located, with zero incidence angle, in the downstream of a cylinder row in motion in the 1st Report is investigated theoretically in this report. As a result, (1) An unknown function in the experimental formula of velocity distribution in the boundary layer under transition to forced turbulent flow is calculated by means of the equation of mean motion. (2) The mean friction coefficient in the transition region is calculated by means of the equation of momentum. (3) The position where the transition is concluded is estimated theoretically. (4) In the region where the forced turbulent boundary layer is fully developed, the mean friction coefficient is estimated theoretically.
    Download PDF (1043K)
  • Eitaro KISIOKA
    1961 Volume 27 Issue 177 Pages 598-609
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    The present paper is primarily aimed at numerical investigation which follows the author's general theory unfolded in the latest issue of the same publication, especially in the field of the blade lattice of the angle profiles. Also it comprises a section assigned to the multi-valued function for conformal representation which appears in the case of two circles intersecting each other. Then the author deduces another method based on a unit circle instead of two circles for conformal representation of the blade lattice of the angle profiles. Further me analyses the blade lattice of the polygonal profiles by means of a different method, giving some numerical examples. From the results thus obtained, he summarizes some useful considerations that apply to the thick and deep-cambered profiles in cascade, tagether with comparison of the different methods taken above for calculations.
    Download PDF (1197K)
  • Yutaka YAMADA
    1961 Volume 27 Issue 177 Pages 610-618
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    Using spindle oils, the writer studied experimentally the torque resistance and the pressure drop of the flow in co-axial cylinders of which inner cylinder rotates. Comparing the torque coefficient Cf with the pressure drop coefficient λ, the writer clarified the behaviors of the shearing stresses in the region of the small values of axial Reynolds number Re and of peripheral Reynolds number Rω. When Re=0, the critical peripheral Reynolds number Rωe at which instability of laminer flow occurs agrees well with the theory given by G. I. Taylor. When the value of Rω is larger than Rωe for Re=0, the value of Cf approaches to the theoretical value of the laminar flow as the value of Re increases. When the value of Re increases still larger, however, transition to the turbulent flow occurs and the value of Cf increases again. When Rω is larger than the value above mentioned, the value of Cf does not approaches to that of laminar flow and variation of Cf becomes smaller and smaller as the value of Rω increases.
    Download PDF (1039K)
  • Sanae KAWABATA
    1961 Volume 27 Issue 177 Pages 619-631
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    The authors performed a series of experiments for the purpose of investigating the state of the air current at the mouth of the tunnel when the train ran into the tunnel. Next inserting a column instead of the train at the constant velocity into the pipe instead of the tunnel placed horizontally, the velocity of the air current inside the pipe was measured with the hot wire anemometer. And when the experimental result obtained from the experiment above was compared with the theoretical values, the former was found to have been nearly agreeing with the latter.
    Download PDF (1741K)
  • Masanobu YAMAMASU
    1961 Volume 27 Issue 177 Pages 631-637
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    It has often been noticed that a water flow in which air bubbles are mixed shows a similarity to cavitation. Not only these photographic similarities, but the dynamic similarities as found in the following experiment were also noticed. When the flow velocities of cavitation bubbles in a pump impeller were measured by photographic method, the writer assumed the cavitating water flow to be a kind of compressible fluid mixed with vapour bubbles. The higher velocity of bubbles, the discontinuity in pressure distribution, and the generation of shock were explained, but the pressure could not be measured in the pump impeller. In this experiment the velocities of bubbles of cavitation which appeared in a divergent flow were measured by a camera which had been designed to photograph the locus of bubbles on a running film. The velocities of bubbles were not lowered in the divergent flow. Also, in measuring the pressure, a discontinuity in pressure distribution was noticed. By examining the dynamic similarity between cavitation and air mixed water flow, the characteristics of the vapor bubble flow were shown.
    Download PDF (1751K)
  • Sugao SUGAWARA, Shinzo KIKKAWA, Jiro OHIZUMI
    1961 Volume 27 Issue 177 Pages 638-644
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    This paper is divided into two parts. In the first part, we derived a theoretical formula, with which one can calculate the presure drop in the horizontal, vertical and bent pipes not only for the steady case, but also for the accelerating state. In the second part, we observed the behavior of particles in the bent pipe with various curvatures and investigated the re1ations among the mixture ratio of solid particles and air, the accelerating coefficient of particles and the curvature of bent pipe experimentally. From the results of these investigations, it has been concluded that (1) with the increase of the mixture ratio, the accelerating coefficient of particles decreases very rapidly, (2) the effect of the mixture ratio on the accelerating coefficient of particles is remarkable in a bent pipe with intensely shorp curve.
    Download PDF (1261K)
  • Yoshihiro MATSUOKA
    1961 Volume 27 Issue 177 Pages 645-658
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    This paper deals with the flow in the impeller with splitter vanes of centrifugal type hydraulic machinery. Splitter vanes are of partial blades which do not extend to the inlet of blade row. In this investigation, 4 kinds of impellers with splitter vanes and 4 kinds of impellers without splitter vanes were used, and then the head, flow quantity and the pressure distribution along the surface of the vane were measured. Moreover, for the purpose of analyzing the blade-to-blade flow, their stream lines were photographed.
    Download PDF (3144K)
  • Yoshihiro MATSUOKA
    1961 Volume 27 Issue 177 Pages 658-665
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    This paper deals with the theory and the experiment or the three-dimensional impeller, that is the impeller with blades or inconstant height, of centrifugal type hydraulic machinery. The theory is developed by using conformal mapping under the assumption that the flow does not differ much from a two-dimensional potential one, and assuming that the displacement flow is not aflected by the inconstant blade height, but the through flow is affected by it. The calculated results of this theory and two-dimensional one were compared with their respective, experimental results, proving good agreement among them.
    Download PDF (1237K)
  • Susumu MURATA
    1961 Volume 27 Issue 177 Pages 666-673
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In this report, a method is presented for the calculation of the potential flow through straight cascades of thin wings. The calculation of the flow has been based on the method of singularities : normal velocities ν are cubic functions of abscissae, and tangential velocities μ, corresponding to normal velocities ν (=0, 1, x, x2, x3), are determinable by conformal mapping of the cascade into a circle. As the author has determined the integral expressions for velocities μ, an improved accuracy has been obtained by using the velocities μ of higher degree. Numerical examples are included.
    Download PDF (803K)
  • Susumu MURATA
    1961 Volume 27 Issue 177 Pages 674-680
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In this report, the thin wing lattice theory, which was mentioned in the 1st report, has been applied to the calculation of the flow in a centrifugal pump impeller, having wanes in the shape of logarithmic spiral. Compared with other methods, the calculation by this method is comparatively easy and accurate, even when the vane angle is small and the number of vanes is large. Numerical examples of the performance curve and velocity distribution on the vane surface agree with the exact solution obtained by Busemann and Shirakura respectively. Equations and numerical examples are also presented for the vanes in the shape of approximately logarithmic spiral.
    Download PDF (700K)
  • Susumu MURATA
    1961 Volume 27 Issue 177 Pages 681-690
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    The breadth of centrifugal pump impeller is usually a function of the radius. A method for the calculation of the flow in such cases is presented in this report. The author has treated such case by applying the method of series expansion, the method of generalised displacement flow mentioned in the 2nd report, and the thin wing lattice theory presented in the 1st report. Compeared with other methods, this method has no limitations for the number of vanes and the vane angles. Numerical examples are included.
    Download PDF (856K)
  • Susumu MURATA
    1961 Volume 27 Issue 177 Pages 691-698
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In the semi-axial pump impeller, the fluid flows on the conical surface and the breadth of the flow path is not constant usually. In this report the author has treated the cases in which the inclination or the flow path and the variation of impeller breadth are small, and the shape of blades is nearly flat plate. In such cases the flow can be analysed by taking the flow of axial pump impeller composed of the flat plate blades as a basis, and adding the effect of the inclination of the flow path, the effect of the breadth variation, and the effect of the blade camber. The author has analysed these effects by applying the method of complex velocity function in 1st report, the method of generalised distlacement flow introduced in 2nd report, and the method of series expansion mentioned in 3rd report.
    Download PDF (656K)
  • Taiziroo KASAI, Sigenori MATUNAGA, Haruo ISIBASI
    1961 Volume 27 Issue 177 Pages 699-706
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    The flow patterns in an axial-flow pump are observed by the lines of the mixture of pigment and water (Streichlinie) in the impeller flashed by strobo-light. The pressure and velocity distributions on the airfoil sections are culculated from the experimental results. And the impeller-heads computed from these distributions are compared with the input-heads. As the flow rate decreases the stalled states are observed first on the outer radial sections and then, when the performance droops, on successive inner sections. With further decrease of the flow rate the characteristics rise sharply accompanying the radial flow.
    Download PDF (1339K)
  • Tosio KURABAYASI
    1961 Volume 27 Issue 177 Pages 707-712
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In order to develop the fundamental results obtained in the previous reports, experiments were carried out by using the vessel with a small tube projected on its side wall, and the following facts were obtained : (1) The length of the projected tube was changed from 0 to 5 mm, but no marked variation was observed in the diameter of droplets d (S.M.D.). (2) Any inclination of the tube from the radial direction to the direction of an axis of rotation has also no significant effect on the value of d. (3) Though the formation of a thin membrane of a liquid was expected by changing the shape of the nozzle to the tape of a slit or an injector needle, it was slightly observed in the slit type nozzle as shown in Fig. 9 (b). The filmwise disruption like this reduced the value of d. (4) The size distributions of the droplets in all the cases mentioned above have nearly the same tendency as that of the vessel type atomizer described in the previous report. But there is an exception of the case accompanied with the filmwise disruption. (5) In the spray, the maximum diameter dmax, in general, has such a simple relation with the mean diameter d as dmax≒2d, but in the case of the slit type nozzle, the difference between these two diameters becomes larger.
    Download PDF (1003K)
  • Takashi SATO, Ryuichi MATSUMOTO, Kazushige UEDA, Keizo OHIRA
    1961 Volume 27 Issue 177 Pages 713-721
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In this paper, the radiant heat transfer from luminous flame is treated theoretically and experimentally, from the view point that the radiation from the cloud of carbon particles is most effective. But, the combustion reaction that may determine the distribution of carbon particles is not considered. Then, in the case where only the cloud of luminous carbon particles under uniform temperature and uniform particle contents is concerned with radiation, the relation between the total emissivity and x has been found, where x is the function of particle contents, thickness and temperature of flame. And the calculated value agrees with Schack's result. Next, in the luminous flame, the relation between the total emissivity and x, εq has been found, where εq is the total emissivity of non-luminous gases. Moreover, the relation between the emissivity of luminous flame and the quantity of carbon particles in the luminous flame which was formed artificially by inserting the small carbon particles in a non-luminous flame, has been made clear by means of optical pyrometer and radiation pyrometer. As a result of studies, the experimental value has consisted with above mentioned theoretical analysis.
    Download PDF (1043K)
  • Takashi SATO, Itaru MICHIYOSHI, Kiichiro TAKEUCHI, Keigo KONDO
    1961 Volume 27 Issue 177 Pages 722-730
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    This paper deals with the burnout and melt down caused by step-input to a fine nichrome wire (0.2 mm diameter) which is vertically supported at the center of glass pipe (annulus) and cooled by natural circulation of water of 80°C subcooling at the atmospheric pressure. The results obtained are as follows : (1) For the transient nucleate boiling which exists between the biginning of boiling and the burnout point, the correlation between the heat flux and the superheating of surface temperature shows the same tendency, even higher than the burnout critical heat flux, as the results by Nukiyama or McAdams. (2) The time interval τB until the burnout occurs is given by the following equation : [numerical formula] where, C is constant, which is dependent on the annulus diameter, and q is heat flux (kcal/m2h). (3) The burnout critical heat flux which does not cause the burnout is 5.3×105, 8.2 ×105, 1.33×106, 2.4×106, and 2.4×106 kcal/m2h, respectively, in the caee of 2, 3, 4, 6, and 8 mm of annulus diameter. (4) There are four different types of change in the temperature of nichrome wire until it is melted down after the burnout point.
    Download PDF (1104K)
  • Wen-Jei YANG
    1961 Volume 27 Issue 177 Pages 730-735
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    A method for obtaining the fluid and metal temperatures in a plate-fin type heat exchanger is presented. The method is general and can be applied to any configuration operating at any specified fluid conditions. The method is illustrated by its application to the recuperator of the cross-flow type and cross-counter flow (Z) type for small gas turbines. It is concluded that Z type is more compact and will suffer the least from the thermal stresses.
    Download PDF (666K)
  • Ituro KIMURA, Haruo UKAWA
    1961 Volume 27 Issue 177 Pages 736-746
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    The flame length of laminar fuel jets issuing into the atmosphere is analyzed and it is shown that it is proportional to the flow rate of fuel gas and inversely proportional to the diffusion coefficient as that of laminar fuel jets issuing into the air flowing at the same velocity, and that the former is about 1.5 times the latter. The experiments on city gas flames in the air stream are conducted using the normal and the inverted arrangements. Experimental results show that the flame in the atmosphere is influenced by buoyancy and disturbance in the air which is caused by combustion, and that if there are no influences of buoyancy and disturbance, the experimental results on the lengths of flames in the atmosphere seem to agree with the result of the analysis. Furthermore, in this report some discussions for the effects of buoyancy and disturbance on the flame length are presented.
    Download PDF (1557K)
  • Takeshi SATO
    1961 Volume 27 Issue 177 Pages 747-756
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In dieser Abhandlung wird die Moglichkeit des Kreisprozesses thermodynamisch durchgerechnet, indem statt einer Brennkammer der Gasturbine ein den Koks vergasender Gaserzeuger als Energiequelle benutzt wird. In diesem Falle verbrennt man den Brennstoff nicht mit Luft, sondern mit reinem Sauerstoff oder damit angereicherter Luft, so daB sich der Energieverbrauch fur die Verdichtung vermindert und die Zusammensetzung des erzeugten Gases verbessert wird. Zur Verminderung der Eintritts-temperatur zur Turbine und zur Verbesserung des Gases wird Wasser in der Brennkammer zugesetzt. Das erzeugte Gas besteht hauptsachlich CO und H2 und stromt aus der Turbine mit hoherem Druck als dem Atmospharendruck und wird einer weiteren Verwendung zugefuhrt. Wenn man eine auBere Sauerstoffquelle hat, kann man zu einer erheblich gunstigeren Nutzleistung gelangen. Besitzt man dagegen keine Sauerstoffquelle, kann auch die Gasturbine unter bestimmten Bedingungen den Sauerstofferzeuger selbstandig antreiben.
    Download PDF (1005K)
  • Tatsuo EGUSA
    1961 Volume 27 Issue 177 Pages 757-761
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    In the conventional design of stag-tap boiler furnaces, certain complicated procedures based on experience have been required because of their variety of construction. A new comprehensive design is explained here, being deduced from the thermal theory, so that a simplified design method may be introduced from the relation between the furnace heat-release and the form factor, and determined the statistical data. Further, from the latest experimental information, a general viscosity equation of molten coal-ash slag is derived and the thickness of a fluid slag film is determined by the auther. In particular, a simplified equation for calculation is induced from the inherent operating conditions of the slag-tap boilers.
    Download PDF (692K)
  • Fujio NAGAO, Harutoshi KAKIMOTO, Syuji OTA, Masami KONISHI
    1961 Volume 27 Issue 177 Pages 762-767
    Published: May 25, 1961
    Released on J-STAGE: March 28, 2008
    JOURNAL FREE ACCESS
    A well-known phenomenon of the temperature rise due to throttling was applied to a starting aid in a small Diesel engine. Theoretical calculation of the charge temperature at the suction end was carried out for the various throttle ratios. The temperature rise above the ambient increases with the decrease in the mean pressure at the suction stroke and the increase in volumetric efficiency, reaching a maximum at a proper throttle ratio which depends upon the cranking speed. Throttling was performed by two means ; i.e., a throttle plate before the inlet valve and a special cam limiting the opening of the inlet valve to the later part of the stroke. The maximum temperature rise reaches theoretically 21°C in the former while 85°C in the latter. In a practical engine the minimum starting temperature was lowered about 15°C by applying the latter method, which shows an effective aid to cold-starting, but has a drawback of complicating the valve mechanism for the practical use.
    Download PDF (953K)
feedback
Top