Smithsonian Contributions to Astrophysics
VOLUME 7
PROCEEDINGS OF THE SYMPOSIUM ON THE ASTRONOMY AND PHYSICS OF METEORS
Held at
SMITHSONIAN ASTROPHYSICAL OBSERVATORY Cambridge, Massachusetts August 28-September 1, 1961
Sponsored by Smithsonian Astrophysical Observatory, Cambridge, Mass. and Geophysical Research Directorate, Air Force Cambridge Research Laboratories, Office of Aerospace Research, U.S. Air Force. Coordinated by Wentworth Institute, Boston, Mass.
SMITHSONIAN INSTITUTION Washington, D.C. 1963 Publications of the Astrophysical Observatory This series, Smithsonian Contributions to Astrophysics, was inaugurated in 1956 to provide a proper communication for the results of research con- ducted at the Astrophysical Observatory of the Smithsonian Institution. Its purpose is the "increase and diffusion of knowledge" in the field of astro- physics, with particular emphasis on problems of the sun, the earth, and the solar system. Its pages are open to a limited number of papers by other investigators with whom we have common interests. Another series is Annals of the Astrophysical Observatory. It was started in 1900 by the Observatory's first director, Samuel P. Langley, and has been published about every 10 years since that date. These quarto volumes, some of which are still available, record the history of the Observatory's researches and activities. Many technical papers and volumes emanating from the Astrophysical Observatory have appeared in the Smithsonian Miscellaneous Collections Among these are Smithsonian Physical Tables, Smithsonian Meteorological Tables, and World Weather Records. Additional information concerning these publications may be secured from the Editorial and Publications Division, Smithsonian Institution, Washington, D.C. FBED L. WHIPPLE, Director, Astrophysical Observatory, Smithsonian Institution. Cambridge, Mass.
For sale by the Superintendent of Documents, U.S. Government Printing Office Washington 25, D.C. - Price $2.75 Contents Session 1 Welcoming address: Col. Ernest A. Pinson, USAF 3 Limitations of radar techniques for the study of meteors: J. S. Greenhow 5 The initial radius of ionized meteor trails: B. L. Kashcheyev and V. N. Lebedinets 19 The initial diameter of meteor trails: G. S. Hawkins 23 The relation between visual magnitudes of meteors and the durations of radar echoes: B. A. Lindblad 27 Radio-echo measurements of meteor mass distributions: A.A.Weiss . 41 A preliminary report on radar meteor counts: P. M. Millman and B. A. Mclntosh 45 The Harvard radio meteor project: G. S. Hawkins 53 Meteor rates observed by radio-echo techniques during the IQY-IGC period: B. L. Kashcheyev and K. V. Kostylyov 63 The orbits of meteor streams determined by radio-echo techniques: B. L. Kashcheyev, V. N. Lebedinets, and M. F. Lagutin . . 67 Session 2 The distribution of small interplanetary dust particles in the vicinity of Earth: C. W. McOacken and W. M. Alexander 71 Micrometeorite measurements from Midas II (Satellite 1960 f 1): R. K. Soberman and L. Delia Lucca 85 Micrometeorite collection from a recoverable sounding rocket. I: R. K. Soberman, C. L. Hemenway, T. G. Ryan, S. A. Chrest, J. Frissora, and E. F. Fullam 89 Micrometeorite collection from a recoverable sounding rocket. II: C. L. Hemenway, R. K. Soberman, E. F. Fullam, J. J. Balsamo, J. Cole, D. Hallgren, P. Yedinak, A. Goodman, and G. Hoff 93 Micrometeorite collection from a recoverable sounding rocket. Ill: R. K. Soberman and C. L. Hemenway 99 Rocket and satellite studies of meteoric dust: T. N. Nazarova . . 105 Cosmic dust showers by direct measurements: M. Dubin, W. M. Alexander, and O. E. Berg 109 Discussion: 115 Session 3 A general survey of meteor spectra: P. M. Millman 119 Meteor spectra with high dispersion: Zd. Ceplecha and J. Rajchl . 129 A short note on meteor spectra with low dispersion: J. Rajchl ... 155 Spectrographic observations of meteors in U.S.S.R. in 1967-1960: E. N. Kramer, K. A. Liubarsky, and V. I. Ivanikov .... 157
in Page Diffusion effects observed in the wake spectrum of a Geminid meteor: I. Halliday 161 On the frequency of occurrence of the auroral green line (5577A) in Perseid spectra: J. A. Russell 171 Session 4 Negative ions and luminosity in meteor trains: T. R. Kaiser .... 175 Luminosity from large meteoric bodies: H. J. Allen and K. K. Yoshikawa 181 Preliminary notes on some results of photographic multiple meteorite fall of PHbram: Zd. Ceplecha 195 Results from an artificial iron meteoroid at 10 km/sec: R. E. McCrosky and R. K. Soberman 199 Luminous efficiency of iron and stone asteroidal meteors: A. F. Cook, L. G. Jacchia, and R. E. McCrosky 209 Meteor ionization and the mass of meteoroids: D. M. Lazarus and G.S.Hawkins 221 Statistical verification of the physical theory of meteors: B. J. Levin and S. V. Majeva 229 On the color index of meteors: J. Davis 233 Image-orthicon photographs of meteors: J. Spalding, G. Colter, C. L. Hemenway, J. A. Cole, and J. F. Dugan 237 Meteoritic erosion in space: F. L. Whipple 239 The absence of magnetic micropulsations of meteor origin: C. Ellyett and G. B. Gillion 249 Session 5 The density distribution of telescopic meteors around the earth's orbit: L. Kresak and M. Kresakova 253 Atmospheric trajectories of telescopic meteors: L. Kohoutek and J. Grygar 259 Statistics of meteor streams: R. B. Southworth and G. S. Hawkins . 261 Orbital elements of photographic meteors: P. Babadjanov 287 On the structure of the 5-Aquarid meteor stream: A. K. Terentjeva . 293 Masses of Comet Giacobini^Zinner and the Draconid meteor stream: Y. V. Yevdokimov 297 Dynamical evolution of the Perseids and Orionids: R. B. South- worth 299 A note on the cometary nature of the Tungus meteorites: V. Fessenkov 305 A short note on the origin and age of the Quadrantids: S. E. Hamid and Mary N. Youssef 309
Abstracts of other papers 313
IV PROCEEDINGS OF THE SYMPOSIUM ON THE ASTRONOMY AND PHYSICS OF METEORS
Welcoming Address By Col. Ernest A. Pinson, USAF *
It is my great pleasure to extend a welcome dous potentialities for research in this field. to this distinguished assembly on behalf of the A good example of new opportunities for con- Geophysics Kesearch Directorate of the Air trolled experimentation is the recent artificial Force Cambridge Research Laboratories. We meteor experiment, which will be discussed in are pleased to be associated with the Smithso- detail during the meeting. nian Astrophysical Observatory in cosponsoring Meteor physics experiments have now become this symposium, and with the Wentworth global in nature, requiring ever greater coopera- Institute as coordinating agency. Our long- tion among world scientists. It is encouraging standing interests in astrophysics are empha- to note that international cooperation in this sized this week in this symposium and in field is flourishing; information is being exchang- another international astrophysical symposium ed on a personal basis and appears in the open on the Solar Corona, which we are sponsoring literature as soon as it has been reduced and at Cloudcroft, N. Mex., during this same week. analyzed. Our fervent hope is that such The latter is Symposium No. 16 of the Inter- cooperation will continue and even increase. national Astronomical Union. In this regard, I am happy to offer the use of The U.S. Air Force is interested in studies any space available on our research vehicles to that can be applied to the solution of specific worthy meteor experiments that may be problems of the space environment. However, forthcoming. we fully recognize that theoretical and experi- The extent and degree of international mental studies that appear, on the surface, cooperation that may soon be required is only to increase knowledge of our universe illustrated by our consideration of the possi- invariably lead to useful applications and solu- bility of using a ground-based network of tions of specific problems. photographic and radio observations of natural As most of you know, the last international meteoric phenomena for determining high- meteor physics symposium was held at Jodrell altitude winds and densities. Bank in 1954. A brief examination of the Briefly, the 1954 symposium spoke in terms proceedings of that symposium reveals the of local, ground-based meteor studies, and of tremendous growth in meteor measurement and international cooperation. Soon, the meteor observation techniques that has taken place physicist will be unrestricted—within our during the past few years. This meeting is solar system—as to mode and place of observa- the first opportunity for people specializing in tion. International cooperation is therefore meteor physics to gather at an international even more essential. symposium to discuss measurements made May I again extend the greetings of the Air above the earth's atmosphere. The advent of Force Cambridge Research Laboratories and rocket and satellite vehicles has opened tremen- our best wishes for a fruitful and provocative 1 Vice commander, Air Force Cambridge Research Laboratories. exchange in this important area of research. 3
Limitations of Radar Techniques for the Study of Meteors By J. S. Greenhowl
Radar techniques are widely used for the The scattering theories study of the physics and astronomy of meteors, Linear electron densities less than 1012 cm'1.— and observations of echoes from the ionized When the linear electron density a is less than trails have also been used as a tool for the approximately 1012 cm"1, the scattering equa- study of the upper atmosphere. Various scat- tion derived by Lovell and Clegg (1948) is tering theories have been developed that give appropriate, where the received power to at the amplitudes and durations of the radar time t=Q after the formation of the trail is echoes in terms of the ionization densities in given by the trails, the ambipolar diffusion coefficient, W e2 V .. „. and the parameters of the equipment. watts (1) Although it was realized that departures WJ ' from the simple theories would occur at very where P is the peak transmitted power; 6 the high frequencies (Eshelman, 1956), it has been aerial gain; X the wavelength; R the range; e the assumed hitherto that there are no serious electronic charge; m the electronic mass; and limitations at frequencies of a few tens of c the velocity of light. Mc/sec. In view of recent work on the initial The echo amplitude decays exponentially in radii of the ionized trails (Greenhow and Hall, the manner 1960a,b) and the loss of electrons by attach- A=Ao exp (- (2) ment processes (Davis, Greenhow, and Hall, 1959), it is now apparent that even at fre- where AQ is the initial amplitude, and D the quencies as low as 20 Mc/sec serious deviations ambipolar diffusion coefficient (Herlofson, from theory occur. The deviations are of 1947). 13 sufficient magnitude to introduce very con- Linear electron densities greater than 10 l siderable uncertainties into much of the pre- cm~ .—In the case of the "over-critically dense" 12 1 vious meteor work concerned with measure- trails with a>10 cm" , giving rise to long ments of linear electron density and of mass duration echoes, the scattering equations given and velocity distributions, and with investiga- by Greenhow (1952a) can be applied, where tions of upper atmospheric pressure and scale height. \mc2) (3) The probable errors arising in each of the above measurements made with typical radars with em the maximum received power. operating at frequencies of about 5 to 600 The echo duration TD, determined by the Mc/sec are the subject of this paper. Only time for the effects of diffusion to reduce the the case of backscattering will be considered. electron density n, on the axis of the trail below the critical density ne, is »Dr. Greenhow died in November 1962. At the time of this symposium he was associated with Royal Radar Establishment, (4) Malvcrn, England.
638805—63 2 6 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 13 1 Departures from the scattering theories.—A with X/2TT. When a<10 cm" , this condi- number of simplifications are made in the tion implies that all the electrons in any cross derivations of equations (1) to (4), and more section of the trail scatter in phase. When rigorous solutions, taking account of polariza- rt is finite, the echo intensity given by equation tion and plasma resonance effects, have been (1) is reduced by a factor given by Kaiser and Closs (1952). The modifi- 7,=exp [2(27rr1012 cm"1, the condition is that the meteor should reach the end of the The effect of increasing r above this value will first Fresnel zone before the electron density t ultimately cause the echo to appear as if it on the axis of the trail at the beginning of the were from an underdense trail, with a much zone falls below n , or e reduced amplitude and value of a. Third, while the echo decay times from under- dense trails given by equation (2) are less than about 1 second for most practical values of X The effects of finite meteor velocity on the and D, equation (4) readily permits echo echo characteristics have been computed by a durations of several hundred seconds for bright number of workers (Eshelman, 1956; Loewen- meteors. With times greater than a few thal, 1956; Hawkins, 1956). As the appropri- seconds, the loss of electrons by attachment ate values for V and D for a given meteor are to neutral oxygen molecules is found to be usually known approximately, the magnitude more important than the effects of diffusion of the effect can be determined. Although the (Davis, Greenhow, and Hall, 1959), and it has reduction in echo amplitude or duration may been shown that the echo duration given by be large, there is no particular wavelength or equation (4) may be in error by several orders height at which these reductions become of magnitude. Thus values of a determined catastrophic. from the echo duration will be liable to these The second important assumption made in large errors. The choice of suitable radar the derivation of the scattering equations is parameters for observing these bright meteors that the initial trail radius r, is small compared will be considered in this paper. METEOR SYMPOSIUM—GEEENHOW The height variation of trail radius, attach- ment function, and meteor velocity Previously the tacit assumption had been made that most meteors ionize in a fairly narrow band of heights centered about 95 km, an assumption apparently supported by the radio observations. However, it was over- looked that this concentration in heights may be due primarily to the failure of radars to detect the ionized trails at heights much above 100 km. Recent work (Greenhow and Hall, 1960b) has now shown that even with equip- ment of moderate sensitivity, the majority of echoes above 100 km should be observable if the effect of initial trail radius is negligible. However, echoes above 100 km are not detected in large numbers because of the effects of the large trail radii. As radar observations are widely used for studies of the astronomy and physics of me- teors, it is of considerable interest to investigate the effects of the finite initial radius on the observations. In order to carry out this in- vestigation, we must construct models for the variation of the height of maximum ionization with meteor velocity and linear electron den- sity, and for the variation of trail radius and attachment time constant with height. This 1.3 1.4 1.5 1.6 construction involves a considerable degree of LOG V (km/MC) extrapolation from the limited experimental FIGURE 1.—The model adopted for the variation of the mean evidence available, but the extrapolations must height of maximum ionization for meteors of various am»» be carried out before the results of previous and F. experiments can be assessed, and further ex- periments planned. the Super-Schmidt camera observations of The model jor the variation of meteor heights faint visual meteors with amM^lO1* cm"1 with Vand amax.—Measurements of the variation (Hawkins and Southworth, 1958), and some of height of occurrence with meteor velocity low frequency radar observations of meteors u 1 and the maximum linear electron density amax with amax~5X10 cm" (Greenhow and Hall, have been reviewed by Greenhow and Hall 1960b). This curve has been taken as a stand- (1960a). Between values of amax from about 18 1 ard, and curves for other values of a,n«x have 10" to 10 cm" , and velocities of 15 to 70 been obtained by extrapolation from equation km/sec, the height variation of the point of (9). The curves in figure 1 give only the mean maximum ionization agrees approximately with height of amax for a given velocity, and in any that derived from the theoretical relation velocity group there is a spread of observed heights about this mean value of about ±.H, (9) where H is the atmospheric scale height. Rocket Panel (1952) density gradients have where pm*, is the atmospheric density at the been used in the extrapolations. point of ernnx. The height variation oj r<.—The only available The curve for a^^lO1* cm"1 in figure 1 measurements of the initial trail radius for < lt 1 has been determined from a balance between meteors with am»x O0 cm" , with which we 8 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS are primarily concerned, are due to Greenhow readily be introduced. Curves of the attenua- and Hall (1960b). These measurements were tion in the theoretical echo due to finite r{, for made on meteor trails with values of amax be- meteors of velocity 15, 30, and 60 km/sec oc- tween 10" and 1012 cm"1 at heights between curring at their most probable heights are there- 90 and 115 km; it has therefore been necessary fore given in figure 2. The curves are drawn to extrapolate to greater heights over a further for reductions in intensity up to 80 db, for a 25 km in order to include all values of V and selection of frequencies greater than 5 Mc/sec. «m»x given in figure 1. The extrapolation has That such high attenuations may occur we can been carried out with a reference value of deduce from some very high frequency obser- T{= 140 cm at a height of 98 km, adopting the vations (see p. 12); beyond this point we would experimentally determined relationship expect that departures from an assumed radial gaussian distribution of electron density would r,ocp~ (10) eventually become significant in determinations of echo intensity. This value of r< at a height of 98 km is several The horizontal broken lines in figure 2 show times greater than Manning's (1958) theoretical where r1=X/2?r, and the selection of a suitable estimate of 14 ionic mean free paths, although operating point should be restricted to the re- it agrees approximately with Opik's (1955) gion above this line. For a velocity of 60 value. Both theories, however, predict a more km/sec, it can be seen that 20 Mc/sec is the > 1 rapid height variation of the form r 1ocp~ . highest frequency that can reasonably be used 12 1 A variation of this type would have an even to observe meteors with amax down to 10 cm" . more catastrophic effect on the radar observa- At 40 Mc/sec the echo amplitudes for trails with tions of meteors than that discussed in the this value of amax are already 100 times below following paragraphs. the predicted values. The position becomes The attachment time constant.—The parameter much worse for very faint meteors with amax in of importance for the effects of attachment is the region from 108 to 1010 cm"1, where even the attachment time constant 7^=l//3ri0, where at 20 Mc/sec the echo amplitude is 1000 times j8 is the two-body attachment coefficient, and below the expected value. In order to observe n0 the density of particles capable of forming these meteors, we would need to use a frequency negative ions. There is a possibility that three- between 5 and 10 Mc/sec. body attachment is the relevant mechanism at Decay-type echoes due to fast meteors with meteor heights (Greenhow and Hall, 1961) when linear electron densities apparently of the order 11 12 1 Ty=l/ynona; y is the three-body attachment of 10 to 10 cm" are observed at frequencies coefficient; and na is the density of particles at least as high as 70 Mc/sec. However, these effective in the collision processes. echoes are very probably due to overdense trails Values of To or TV=50 sec at a height of with a of the order 1013 cm"1, but with such 95 km, based on the measurements of Davis large initial radii that the electron density on et al. (1959) have been assumed, again using the axis of the trail is already below the critical Rocket Panel density variations to extrapolate density (Greenhow and Hall, 1960b, para. 5b). to other heights. The situation improves for meteors of smaller velocity that ionize lower in the atmosphere; The effects of finite trail radius upon the for V=15 km/sec, a frequency of 50 Mc/sec radar observations of faint meteors would be suitable for observing meteors with The measurement of linear electron density.— «m»x~1012 cm"1, and 20 Mc/sec for trails with Equation (1) is generally used to determine the
5Mc/sec 10 Mc/sec
20
40 db \40 Mc/sec
60 \ \
\80 Mc/sec 80 \ 80 b \
10" 10 IOe 10" IOI 10°
(a) ___--_^r__35 Mc/sec
20 10 Mc/sec
FIGURE 2.—The reduction in echo intensity over that pre- dicted for an infinitely narrow meteor trail, due to the effects of the finite initial trail radius, a, F=\5 km/sec; h, T=30 km/sec; c, F=60 km/sec. Ordinates: 10 log10ij (db). Abscissae: linear electron density On.^cm"1). 60
80
10" 10" 10 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS
i.o 5 Mc/sec
5 Mc/sec
FIGURE 3.—The fraction of the true number of meteor trails with a^>amia which give detectable echoes, as a function of meteor velocity and frequency, a, For a theoretical limiting sensitivity of ami,,"" 10" cm"1. Experimental points: x,36 Mc/sec; o,17 1 Mc/sec. b, For a theoretical limiting sensitivity oi amiui*»\0P cm" .
/=20 Mc/sec, but only 15 percent of those very rapidly as the frequency increases. Thus with velocities of 70 fan/sec. These percent- at 80 Mc/sec the echo rate for velocities of ages improve at lower frequencies, but decrease 15 km/sec would be 10 percent of the true rate, falling effectively to zero «1 percent) for velocities greater than 30 km/sec. The form of these curves has been verified at frequencies of 17 and 36 Mc/sec (Greenhow and Hall, 1960b); the experimental points are shown in figure 3a. There is reasonable agree- ment with the theoretical curves. Figure 36 is an extrapolation to a lower limiting linear electron density of approximately 108 cm"1, when it can be seen that a frequency of only 5 Mc/sec would be required to give a reasonably accurate velocity distribution. These results are summarized in figure 4, where the maximum usable frequency is shown as a function of a and V. The criterion ri tof1 Measurement of velocity distributions: The amm {ELECTRONS/em) effect of finite meteor velocity.—An effect not FIGURE 4.—Curves giving the maximum usable frequency / previously taken into account in deriving the for the reliable observation of meteors of velocities IS, 30, echo attenuation curves is that of finite meteor and 60 km/sec, as a function of ammx. velocity, since this factor, although it may be METEOR SYMPOSIUM—GBEENHOW 11 large, will generally be negligible in comparison meteor mass and magnitude distributions; it with the effects of trail radius. However, would appear that sensitive radars could be finite velocity limitations will mean that even used to extend the measurements of the mass if large reductions in echo amplitude due to distribution from the faintest photographic the finite initial radii could be tolerated, the meteors (about +3 absolute magnitude, echoes might not show the diffraction effects aw~6X10ia cm"1) to about +15 magnitude 8 1 that enable us to determine meteor velocities. (amax~10 cm" ). This method has been used Curves given by Loewenthal (1956) show by a number of workers, and values of s lying that the zone formation is effectively smoothed within the range s—2±0.2 are usually found, out when the expression c, given by irrespective of radar wavelength or equipment sensitivity. In view of the effects of finite (12) initial trail radius, we must therefore consider the significance of these measurements. is less than about 0.5, at which point the reduc- As an example of the results obtained if this tion in echo amplitude is about 2.7 greater method is used, the differential number- than that given by equation (1). With this amplitude distributions for two entirely dif- criterion, the areas under the broken curves ferent radars working at the very low frequency in figure 2 define the regions in which meteor of 32 Mc/sec (Greenhow and Watkins, un- velocities would become indeterminate. This published) and the very high frequency of condition is also indicated by the broken lines 500 Mc/sec (Barber, Sutcliffe, and Watkins, in figure 4. 1962) are illustrated in figure 5. The limiting Meteor mass distributions.—Determinations sensitivities of the two equipments, obtained of the s in equation (11) have been made by by substituting the equipment parameters in 10 8 1 measuring the distribution of meteor echo equation 1, were amta~2X10 and 10 cm" amplitudes, on the assumption that the scatter- respectively. The slopes of the best straight ing equation (1) is valid. This exponent is of lines in figures 5a and 5b correspond to values considerable interest as it can be related to the of «=1.85 at 32 Mc/sec, and s=2.0 at 500 200 100 90 20 » NUMBCR OF ECHOES NUMBER OF ECHOES FIGURE 5.—Number/amplitude distributions for echoes observed at frequencies of 32 Mc/sec and 500 Mc/sec Ordinates: Echo amplitude A (arbitrary units~receiver noise). Abscissae: Number of echoes with amplitudes >A. a, f— 32 Mc/sec; limiting 10 1 l 1 theoretical sensitivity am|n~2X 10 cm" , b,/=500 Mc/sec; limiting theoretical sensitivity omM««'10 cm" . 12 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS giving a measurement of the meteor mass or electron density distributions. Instead, equa- tion (11) with s~2 must be regarded only as an empirical relationship that shows that for a given radar, more small echoes will be detected than large ones. Echo rate as a function of wavelength.—For equipments with the same theoretical limiting sensitivity, the echo rate will decrease rapidly because of the differential effects of a finite trail radius. The observed echo rate for meteor trails with a apparently greater than 10" cm"1, where a is determined from the echo amplitude and equation (1), is shown as a function of radio frequency in figure 7. The points for 17, 36, and 69 Mc/sec are from Greenhow and Hall (1960b), while the points ,oV for 300 and 500 Mc/sec are derived from the ECHO RATE (No/IO observations of Barber, Sutcliffe, and Watkins FIGURE 6.—Number/apparent linear electron density dis- (1962). It can be seen that the observed echo tributions. Ordinates: Apparent linear electron density. rate falls from 0.4 of the true rate of f=17 Abscissae: Echo rate from trails with apparent linear Mc/sec, to only 0.01 at 100 Mc/sec, and to electron densities ><*. An approximate absolute scale is 0.00001, at/=500 Mc/sec. This is an extremely given, where "No" is the number of echoes observed per z 10* km * in the plane 100 km above the earth per hour. rapid variation, giving an approximate f~ dependence of echo rate between frequencies of Mc/sec in reasonable agreement with previous 30 and 500 Mc/sec. measurements for sporadic meteors. At frequencies significantly greater than However, when the absolute echo rates on the 100 Mc/sec, it is unlikely that the echoes two equipments are compared, an anomalous detected are from underdense meteor trails situation arises. A value s==2 implies that if with a<1012 cm"1. The few echoes that are similar areas of sky are illuminated by the aerial observed are more likely to be due to over- beams, then the echo rate with an equipment of dense trails whose amplitudes have been dras- amta^lO8 cm"1 should be 200 times the rate tically reduced. In the case of the 500 Mc/sec 10 1 8 1 when ami,, is only 2X10 cm" . In fact, when observations with amin~10 cm" , for example, corrections are made for the different polar the echo rate is only about 10~8 of the true value. diagrams, the observed echo rate on the 500 Thus it is probable that the echoes actually Mc/sec equipment is found to be only about 0.2 times the 32 Mc/sec rate (fig. 6), a deviation by a factor 103 from the expected rate at the apparent value of s. Thus it is clear that the curve AB in figure 6 cannot represent the true distribution of values of a in the region 108 FIGURE 9.—Echo durations from the points of maximum ionization as a function of frequency and amBX allowing for the effects of attachment, a—c are for a two-body attachment process with F=15, 30, and 60 km/sec respec- 10" to- tively; d is for a three-body attachment , (ELECTRONS cm"1) process with F«=30 km/sec. AB, CD, EF, GH, JK show where the echo durations are reduced to erl, 10"1,10~*, 10"3, and 10~4 of the values expected considering diffusion alone; xy shows where the effects of the finite initial trail radius become important. (ELECTRONS cm'1) METEOR SYMPOSIUM—GREENHOW 15 (ELECTRONS em"1) 16 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS are for a two-body attachment process, and trail radius is such as to cause the initial axial figure 9d is for a three-body reaction. electron density to be less than the critical 12 1 Let us consider figure 9c for V=30 km/sec value for trails with amax>10 cm" are shown as typical of meteors of average velocity. by the curves X! yi, x2 y2, etc., in figure 9. The series of lines A3 B3, C3 D3, etc., shows Thus, for V=30 km/sec it can be seen that at the points where attachment reduces the echo frequencies above 40 Mc/sec the echoes from duration to e~l, 0.1, 0.01, etc., of the value overcritically dense trails will begin to assume expected considering the effects of diffusion the characteristics of decay-type echoes. How- alone. Below A3 B3, equation (4) can therefore ever, as the initial radii of the ionized columns be regarded as applicable for all points along a will necessarily be >X/2?r, the echo intensities given trail when echo duration is proportional will be very much reduced. At frequencies of 2 x to aX D~ . Above A3 B3, the echo duration 300 to 600 Mc/sec for example, trails with increases very slowly with increasing a, and linear electron densities as high as 10" cm"1 indeed in some cases the echo duration for may be undercritically dense initially. As the a very bright meteor can be less than that trail radii are of the order of 1 meter (~10X/2ir), from a fainter trail ionizing at a greater height. the reduction in echo intensity given by equa- Thus at a frequency of 20 Mc/sec, for example, tion (7) is enormous (~800 db). In this case, a trail with a^i^lO14 cm"1 gives an echo dura- it is probable that the echo intensity will be tion of about 25 seconds from the point of determined by some second-order reflection 16 maximum ionization. Increasing amax to 10 mechanism, such as partial reflection from the cm"1 increases the echo duration to only 30 sharp boundary in ionization near the meteor seconds, instead of to the 3X10* seconds expected itself. Such echoes would have the same from diffusion theory. For such trails, of course, general amplitude-time characteristics as those the longest enduring echoes would normally be from low electron density trails under conditions obtained from some height above the region of of extremely rapid diffusion (Eshelman, 1956). maximum ionization, and curves showing the The two classes of echoes will therefore become magnitude of this effect have been given by indistinguishable. Davis et al. (1959). Thus, in order to make satisfactory observa- For comparison, the effects of a three-body tions of overcritically dense meteor trails, we attachment process for a meteor velocity of would need to select an observing frequency 30 km/sec are shown in figure 9d. The echo such that the operating point lies within the durations are still further reduced, causing the region defined by areas similar to ABxy in line A4 B4 to move to lower values of duration figure 9. Simultaneous measurements of two for a given amM. wavelengths would also be of great value in It is clear that the influence of attachment on determining whether factors other than diffus- echo duration can be so large as to render the sion were affecting the observations. duration of little value for the determination of a. In order to measure the electron den- References sities in the trails of bright meteors, we would need to observe in the region below lines such BARBER, D.; SUTCLIFFE, H. K.; AND WATKINS, CD. 1962. Some observations of meteors and aurorae as A B, where the normal diffusion equation at 300 and 500 Mc/s using a large radio can be applied. Thus no single wavelength telescope. Journ. Atmosph. Terr. Phys., can be used to observe the full range of possible vol. 24, pp. 585-597. values of a. Precise measurements of height DAVIS, J.; GHBENHOW, J. S.; AND HALL, J. E. might be of assistance, but even so the correc- 1959. The effect of attachment on radio echo tion factors can be so large, and the precise observations of meteors. Proc. Roy. nature of the attachment effects so uncertain, Soc. London, ser. A, vol. 253, pp. 130-139 that it is doubtful if any significant improve- ESHELMAN, V. R. ment would be gained. 1956. Short wave length radio reflections from meteoric ionization. I. Theory for low Effect of finite initial trail radius.—The value density trails. Stanford Univ. Contract of linear electron density at which the initial AF 19(604)-1031, Rep. No. 5. METEOR SYMPOSIUM—GREENHOW 17 GREENHOW, J. S. HERLOFSON, N. 1952a. Characteristics of radio echoes from meteor 1948. The theory of meteor ionization. Phys. trails. III. The behaviour of the elec- Soc, Rep. Progr. Phys., vol. 11, pp. tron trails after formation. Proc. Phys. 444-454. Soc. London, vol. 65B, pp. 169-181. KAISER, T. R., AND CLOSS, R. L. 1952b. A radio echo method for the investigation of 1952. Theory of radio reflections from meteor atmospheric winds at altitudes of 80 to trails. I. Phil. Mag., ser. 7, vol. 43, 100 km. Journ. Atmosph. Terr. Phys., pp. 1-32. vol. 2, pp. 282-291. LOEWENTHAL, M. GBEBNHOW, J. S., AND HALL, J. E. 1956. Meteor echoes from underdense trails at 1960a. The variation of meteor heights with very high frequencies. TR Technical velocity and magnitude. Monthly No- Report 132, Lincoln Laboratory, Massa- tices Roy. Astron. Soc, vol. 121, pp. chusetts Inst. Techn. 174-182. LOVELL, A. C. B., AND CLEGG, J. A. 1960b. The importance of initial trail radius on 1948. Characteristics of radio echoes from meteor the apparent height and number dis- trails. I. The intensity of the radio tributions of meteor echoes. Monthly reflections and electron density in the Notices Roy. Astron. Soc, vol. 121, pp. trails. Proc. Phys. Soc London, vol. 183-196. 60, pp. 491-498. 1961. Attachment processes in meteor trails. Journ. Atmosph. Terr. Phys., vol. 21, MANNING, L. A. pp. 261-271. 1958. The initial radius of meteoric ionization GREENHOW, J. S., AND NEUFELD, E. L. trails. Journ. Geophys. Res., vol. 63, 1959. Turbulence at altitudes of 80-100 km and pp.181-196. its effects on long-duration meteor echoes. 1959. Air motions and the fading, diversity, and Journ. Atmosph. Terr. Phys., vol. 16, aspect sensitivity of meteoric echoes. pp.384-392. Journ. Geophys. Res., vol. 64, pp. HAWKINS, G. S. 1415-1425. 1956. Radar echoes from meteor trails under OPIK, E. J. conditions of severe diffusion. Proc. 1955. Meteors and the upper atmosphere. Irish Inst. Radio Eng., vol. 44, p. 1192. Astron. Journ., vol. 3, pp. 165-181. HAWKINS, G. S., AND SOUTHWORTH, R. B. 1958. The statistics of meteors in the earth's ROCKET PANEL atmosphere. Smithsonian Contr. Astro- 1952. Pressures, densities, and temperatures in phys., vol. 2, no. 11, pp. 349-364. the upper atmosphere. Phys. Rev., vol. 88, pp. 1027-1032. Abstract A brief review of the basic theories concerning the scattering of radio waves from the ionized trails produced by meteors is given. Some of the fundamental assumptions used in the derivation of the various scattering equations are discussed, and are shown to be incorrect. Although it has been realized that the simple picture of an initially narrow trail would break down at very high frequencies, it is now apparent that the initial radius can be an appreciable fraction of a wavelength even at fre- quencies as low as 20 Mc/sec. This effect has a profound influence on the observation of meteors fainter than about + 4 absolute visual magnitude. The second major cause of the breakdown of the scattering theories is the loss of electrons by attachment pro- cesses, in addition to ambipolar diffusion. This effect can be serious in the case of long duration echoes, associated with meteors brighter than about +2 magnitude, and the predicted echo durations may be reduced by several orders of magnitude. It is shown that much of the previous work on the determination of linear electron densities in meteor trails, and the measurements of meteor mass distributions, velocity distributions, and upper atmospheric pressures and scale heights, is in error. The interpretation of future radar observations of meteors must be treated with consider- able caution. The Initial Radius of Ionized Meteor Trails By B. L. Kashcheyev * and V. N. Lebedinets * Atoms evaporating from the surface of a ions and atoms of neon and argon, respec- meteoroid possess energies as high as hundreds tively, in the range of meteoric velocities the of electron-volts. The resulting meteor trail following relationship holds true: rapidly diffuses outward until the atoms lose their cosmic energy. From the diffusion equa- <&«4.5&. (3) tion, the trail radius r for a time t after the meteor has traversed a given point is: According to Loeb (1934), for normal tem- peratures the relationship is about the same. 1/2 Hence in Qd(E) diffusion cross section depends upon the value of energy in about the same manner as it does in the case of neutral particles. where D is the coefficient of diffusion, N is From equation (2) the mean free path of number of collisions, and X is the mean free ions can be expressed in the form: path. As X depends upon the velocity of the N meteor, in the above formula 23 M should be *-o where Xo is the mean free path of atmospheric substituted for iVX2, where X* is the mean free molecules, and V is the velocity of the particles path after the kth. collision. Then colliding. For the case when the masses of l/2 ions and of atmospheric molecules are equal, (1) F=1.4X10~6 sec/cm, and <7=1.8X10s cm/sec. Under the above condition, after one collision From their quantum-mechanical computa- an ion retains on the average two-thirds of tions, Massey and Sida (1955) found four values its initial velocity. The velocity of an ion for the cross section of diffusion Qd for col- after N collisions is given by the formula lisions of atoms of Ca and Ne possessing various values of energy E. They proposed an ap- Vff=Voe-«»> (5) proximation of Qd{E), which is, however, too where a=0.4. complex for mathematical computations. A From equations (1), (4), and (5), we obtain better form of approximation is: (2) (6) which does not differ greatly from that pro- The number of collisions before a thermal posed by Massey and Sida for high values of equilibrium is reached is energies, and gives the right value of Qd under normal temperatures. Experimental evidence (Massey and Burhop, 1952) indicates that for collisions between where Va is the mean velocity of atmospheric "Kharkov Polytechnics! Institute, U.S.S.R. molecules. 19 20 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS The initial radius of a meteoric ionized trail is A= FX3/2 Values of r0/Xo for meteors of various velocities are given in table 1. On the average, the All other notations are conventional. initial radius is found to be about 1.5 times From equation (10) we obtain the equation: greater than the first mean free path of the G G ions. 1, P<2 1 h '\ 3 , Xi_, p(AQ In Ax—In m m The time elapsed between two successive Ot2GT2 2 X2 p(A,) N and N-\-1 collisions is given by the expression: During the measurements made at the wharkov Polytechnical Institute in 1960, the The time taken for N collisions is Kavelengths used were Xi=8 meters and X2=4 meters. Identical 5-element director antennas (8) were employed. In order to equalize the effective sensitivities of the equipments, the The time of initial expansion is given by the output power of the transmitter at X2=4 meters equation: was twice as great as its counterpart at Xj=8 meters, and the sensitivity of the receiver was t =t(N )=F(- In &+-K^) X.«3. also somewhat higher. During the observa- o Q tions the powers and the frequencies of the (9) respective transmitters were continually re- corded. Calibration markers enabled the am- where Xa is in centimeters and t0 is in seconds. plitudes to be expressed in terms of microvolts. The initial radius is practically reached after 6 The identity of the patterns of the two antenna fewer than No collisions and o=~~£- when r0>0.6 m, (10) where r0 is in meters. The errors of measurement increase with TABLE 1.—Values of ro/X» for meteors of various greater values of r0, since amplitudes of echoes velocities at X2=4 meters are in such cases lower. From equations (11) it should be plain that for the Velocity Vo (km/sec) 10 20 30 40 50 60 70 case considered, measurements could be made only where O.5m The initial diameter of the column of ion- the initial diameter is given by McKinley's ization produced by a meteor is of great impor- (1961) equation (8-18): tance, since a small increase in its size produces log d=0.075H— 7.6 (meters), (2) a large increase in the attenuation of a radio lo echo. Using equation (8-25) of McKinley where H is the height of the meteor column in (1961), we may write the decibel power loss in kilometers. At a height of 105 km the expected the echo due to the initial diameter d as diameter of the ion column is 1.9 meters, which /d\z is considerably smaller than the value found by loss=85(-j decibels, (1) Greenhow and Hall (1960). For comparison I have plotted in figure 1 the values of Greenhow where X is the wavelength of the radar. At a and Hall and the values predicted by theory. wavelength of 8 meters an initial diameter of The experimental values are considerably 4.0 meters produces an attenuation of 21.25 greater than the theoretical predictions for db, whereas the attenuation is only 5.3 db if heights below 105 km. Greenhow and Hall the diameter is 2.0 meters. Greenhow and attribute this fact to fragmentation of the Hall (1960), and Greenhow (see this sympo- meteoroid and attach significance to the sium, p. 5) measured the amplitude of echoes discrepancy. On the other hand, their measure- at two wavelengths and deduced an average ments may overestimate the diameter of the diameter of 4.0 meters for the column at a column because of an inbalance in the two height of 105 km. It is important to estab- wavelength channels, a disparity in the antenna lish whether this value is correct, for if it is, patterns, or other causes. the attenuation at normal radar wavelengths We can obtain an independent estimate of is so severe as to make worthwhile observations the initial diameter by direct photography. virtually impossible. Our understanding of the meteor process is not Opik (1958) and Manning (1958) have con- sufficient at the moment to distinguish clearly sidered theoretically the initial conditions at between the column of ionization and the coma the meteoroid. Atoms and ions evaporate from that is emitting light. We would not, however, the heated surface of the meteoroid and are expect the region of optical excitation to be any injected into the atmosphere at the velocity smaller in extent than the region of ionization. of the meteor. After approximately 14 colli- In a discussion of meteor spectra, Cook (1955) sions with air particles, each meteor atom or summarizes the levels of optical excitation that ion is reduced to thermal velocities, and there- have so far been observed. Meteoric radiation after normal diffusion takes place. The mean is a mixture of light from both ionized and free path of an atom at these velocities is ap- neutral atoms. We would expect that as a proximately five times greater than that of an minimum the light would be confined to the ion. Thus two concentric columns will be pro- ionized column, where excitation and recombina- duced: a narrow column of ionized particles tion are taking place. At the maximum extent surrounded by a wider shell of neutral atoms. the light could be generated from the core out From the known density of the atmosphere, to the fringes of the neutral atomic column. Although on one occasion Cook and Millman < Harvard College Observatory, Cambridge, Mass., and Boston (1955) observed radiation from the heated University, Boston, Mass. 23 24 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS IU - 9 - Leg end: I X GREENHOW - HALL 1 • - ——— THEORY MANNING I Photographic: 7 - 1 X — O HAWKINS-WHIPPLE / ^' 6 5 " • COOK- STIENON- HAWKINS / y E X^ I - / 3 - / _ 2 - - x—-**"""" T S* - 1 - - 1 1 1 1 1 1 Height , Km FIGURE 1.—Diameter meteor column as function of height. surface of the meteoroid, this source of radiation The photographic data were obtained during may be neglected for meteors that are fainter two series of observations. The first set of than magnitude —5. measurements (Hawkins and Whipple, 1958) was taken from the Palomar Sky Atlas, for which meteors had been photographed with the a 48-inch Schmidt telescope during the sky survey IB METEORS MAG 0 by the Mount Wilson and Palomar Observa- JKHIPR.E- HAWKINS tories and the National Geographic Society. Photographic plates were exposed in red and blue light; the average magnitude of the meteors was 0 for red and +3 for blue. The results are summarized in figure 2a,b. The mean width of the 18 meteors photographed on the red plates was 1.32 meters; and of the 33 meteors on the blue plates, 1.25 meters. Although these measurements suggest that the brighter meteors 33 METEORS MAG *3 produce wider trails, the result is not conclusive since the estimated error of a single measure- ment was ±0.5 meters. We may safely con- clude, however, that the width of the optical coma for meteors with magnitude between 0 and +3 is less than the initial diameter found 9 GEMIMDS by Greenhow and Hall. MAS *6 COOK-STItNON-HM The second set of measurements was obtained by Cook, Stienon, and Hawkins (1960) during the Geminid meteor shower of December 1957. The 48-inch Palomar Schmidt was directed towards the Geminid radiant so that the meteor FIGURE 2.—Photographic measures of meteor trail widths. trails were considerably foreshortened. In this METEOR SYMPOSIUM—HAWKINS 25 manner the limiting visual magnitude was ex- unobservable by the radio-echo technique. We tended to +9, and the average magnitude of must regard such meteors as "abnormal." meteors measured in this program was +6. In Their excessive width is probably due to ex- their original measurements, Hawkins and plosive fragmentation. Whipple (1958) focused the camera at infinity To explore further the disagreement between to record the stellar images, and employed a the radar measurements of Greenhow and Hall sidereal drive. These factors introduced some (1960) and the optical data, we can consider uncertainty in the work since a correction had to the dependence of trail width on height. In be applied to allow for the fact that the meteors, table 1, I have used the published optical data not being at infinity, were slightly out of focus. to group the meteors according to velocity. Also, the width of the meteor trails had to be The mean height of these meteors can be taken compared with the diameter of the circular star from previous photographic data (Hawkins and images. In the measurements of 1957 several Southworth, 1958). The Geminid data of photographs were obtained with the plate dis- Cook, Stienon, and Hawkins (1960) are given, placed 50 microns from the focal plane so that where the mean value for the height of a the meteor trails were correctly "in focus" and Geminid of magnitude +6 was taken to be 100 the star images were "out of focus." Since the km. The two meteors of excessive width were sidereal drive was not employed, we were able excluded from this sample. In addition, the to compare meteor trails directly with star trails were reanalyzed to obtain values for the trails. The widths of the nine Geminids that mean diameter of the first hah* of the trail, and were measured are given in figure 2c. Seven that of the second half of the trail. From table have widths that are distributed symmetrically 1 we obtained the five values of diameter that around a mean value of 0.97 meters. The other are shown in figure 1. The optical coma cor- two meteors were considerably wider, with responds in size to the initial diameter of the values of 5.8 and 5.1 meters. The estimated ionization column as predicted by Manning error for the Geminid meteors was ±0.4 meters, (1958). It is difficult to see how the initial slightly less than the value obtained by Hawkins ionization column can be any larger in diameter and Whipple (1958). than the measured optical values, i.e., of the If we define the 58 meteors in the quasi- order of 1.0 meter over the normal range of symmetrical distributions to be "average meteor heights from 90 to 105 km. Above a meteors," then the width of the optical coma height of 110 km, the trails may indeed be for an average meteor is about 1.2 meters for extremely wide, but photographic meteors magnitudes between 0 and +6. If the ionized seldom appear in this region of the atmosphere. column is comparable in size to the optical The majority of meteors with visual magni- coma, then the attenuation for an 8-meter radar tudes between 0 and +6 are narrow enough to signal for these "average meteors" is less than avoid the extreme attenuation effects suggested 2 db. On the other hand, the two Geminids of by Greenhow and Hall (1960). There is no excessive width would produce an attenuation optical evidence to support the idea that faint of more than 40 db, and would be virtually meteors produce an optical coma that is wider than the coma of the brighter photographic TABLE 1.—Photographic width as a function of height meteors. Mean width Mean height (meters) (kilometers) References (Hawkins, Whipple) Perseids, Orionids 0.79 104 COOK, A. F. 1955. The nature of meteoric radiation. In (Hawkins, Whipple) Meteors, ed. T. R. Kaiser, Spec. Suppl. 8 Aquarids, Taurids, 1 Journ. Atmosph. Terr. Physics, vol. 2, Quadrantids, Lyrids, [ 1.20 94 pp. 8-15. Pergamon Press, New York. Geminids J COOK, A. F.; STIENON, F. M.; AND HAWKINS, G. S. (Cook, Stienon, Hawkins) 1960. Meteor trail widths. Interim Report 47, Mean value, Geminids 0.97 100 Harvard Radio Meteor Project, Harvard First half 1.53 102 College Observatory, Contract AF 19 Second half 0.44 98 (122)-458, Subcontract no. 57. 26 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS COOK, A. F.f AND MILLMAN, P. M. HAWKINS, G. S., AND WHIPPLB, F. L. 1955. Photometric analysis of a spectrogram of a 1958. The width of meteor trails. Astron. Perseid meteor. Astrophys. Journ., vol. Journ., vol. 63, pp. 283-291. 121, pp. 250-270. MANNING, L. A. GREENHOW, J. S., AND HALL, J. E. 1960. The importance of initial trail radius on 1958. The initial radius of meteoric ionization the apparent height and number distri- trails. Journ. Geophys. Res., vol. 63, butions of meteor echoes. Monthly pp. 181-196. Notices Roy. Astron. Soc, vol. 121, pp. MCKINLET, D. W. R. 183-196. 1961. Meteor science and engineering. McGraw- HAWKINS, G. S., AND SOTJTHWOBTH, R. S. Hill, New York. 1958. The statistics of meteors in the earth's OPIK, E. J. atmosphere. Smithsonian Contr. Astro- 1958. Physics of meteor flight in the atmosphere. phys., vol. 2, no. 11, pp. 349-364. Interscience, New York. The Relation Between Visual Magnitudes of Meteors and the Durations of Radar Echoes By B. A. Lindblad1 Observations of the luminous and ionized gram are designed to cover as completely as trails of meteors have shown that meteors in possible a limited area of the sky. The radio the visual brightness range give rise to radar investigations are conducted with a beamed echoes with durations ranging from a fraction radar equipment, and the field of the visual of a second to several minutes. Observational observers is restricted to the beamed sky area. evidence for a relation between the apparent During the observations the beam is directed luminosity of a meteor and the duration of the to a point in the sky whose elevation and associated radar echo was presented by Millman azimuth differ by 90° and 180°, respectively, (1950). This relation has been discussed by from those of the shower radiant. various authors (Greenhow and Hawkins, 1952; Visual recording is made by a group of three Kaiser, 1953). or four trained meteor observers. The The most extensive observations relating the observers estimate meteor magnitude and train magnitudes of visual meteors to the lifetimes of duration, and also plot the meteors on star the corresponding radar echoes have been made charts. Meteor brightness is estimated to the by Millman and McKinley (1956). From a nearest half magnitude by direct comparison total of 1,405 coincidences obtained during with standard magnitudes within the field of observations of the Perseid shower, these observation. workers found a linear relationship between Radar observations are made with a pulsed the logarithm of the echo duration and the equipment. The transmitter generates 20-50 absolute magnitude of the meteor. A similar kw peak power with a pulse width of 20 n sec investigation by Lindblad (1956) was based on at a wavelength of 9.2 meters. Repetition a rather small sample. The purpose of the frequency is 50 c/s. The receiver has an present communication is to repeat and extend overall bandwidth of 75 kc/s. The equipment the previous work, using the larger amount of used was essentially the same during the entire data which is now available. program, but the recording technique has been The observational data have been collected improved from time to time in order to increase in the course of Perseid observations carried out the detectability of faint meteors. The present at the Onsala Wave Propagation Observatory 2 arrangement uses two separate displays, one in Sweden between 1953 and 1960. The experi- intensity-modulated range-time display, and mental technique, which has been described in one amplitude-time display. Both displays are more detail elsewhere (Lindblad, 1956, 1961), simultaneously photographed at high film speed is summarized below. (25-50 cm/min). The arrangement permits the individual echo pulses to be recognized and Observational techniques gives good discrimination against the noise background. The shortest echo duration which The radar and visual observations of the pro- can be resolved is between 0.04 and 0.02 i Lund Observatory, Sweden. second. Some typical registration results are * A division of the Research Laboratory of Electronics of the Chalmers shown in plate 1. Technical University, Gothenburg. 27 28 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS IN OO CO m 35 CO oo 35 35 oc oo 68 8 Tota l •N - - r--. to +3. 9 +3. 5 - CO lO IN i-l i-H to +3. 4 +3. 0 CO IN o (N m CO (N t o +2. 9 +2. 5 CO CO co 35 * O CO CO IN CO oo o CO 10 4 to +1. 4 +1. 0 35 •e IN co to oo +0. 9 +0. 5 CO IN CO o o 35 to ^ oo CO i—i CO CO CO (N t o co -0. 6 -1. 0 i-H (N to IN -1. 1 -1. 5 35 to -1. 6 -2. 0 1*1 I—1 co IN IN t o -3. 0 -2. 6 t o -3. 5 -3. 1 .—1 to -4. 0 -3. 6 6T0 - N O M 35 35 35 35 35 35 35 (N 35 CO CO O o o O 1-H ^ CM <—I f—I T-l .-i O O i- 1 1 1 1 I o o o 3 o 3 O 3 o 5 o o o o o © O •*-> +J -u o t- i o i o o o o o (N i-i 1—1 i—( O1- 1 1-COt CT-DH o o o o o o Dta l 1 1 1 1 1 1 H LINDBLAD PLATE 1 Registrations obtained with fast amplitude-time and range-time radar displays. Time difference between consecutive pulses 0.02 second. Lower display, amplitude-time (~0-10jiF); upper display, range-time with superimposed amplitude-time (•-'0--100 (iF). Top photos: Short duration echoes with no associated visual meteor. Bottom photo: Long duration echo produced by apparent magnitude 2"P0 Perseid at 22h3Sm18» M.E.T. on August 12, 1959 (Meteor No. 59-279). Range 180 km. A marker between 40-70 km indicated simultaneous visual recording. METEOR SYMPOSIUM—LINDBLAD 29 Reduction of data the scatter diagram of log echo duration versus zenithal magnitude. This diagram for 688 The criterion for accepting an association be- Perseid meteors is shown in figure 1. The data tween a visual meteor and a radar echo is of this sample, arbitrarily grouped in intervals mainly one of time coincidence. As a rule, of 0.3 in log echo duration and 0™5 in zenithal no echo is accepted as a coincidence if the time magnitude, are collected in table 1. Inspection difference between the signal of the visual of the table shows that in any one array there observer and the echo is more than 0.5 second. is a considerable scatter about the mean. During the period of maximum Perseid activity, Since only a moderate correlation exists between the echo rate on the present equipment rarely the variates, statistical methods must be used exceeds 120 echoes per hour. The number of in the presentation and discussion of the data. spurious associations in our data is thus less than 2 percent. The possibility of a spurious Means of rows and columns for the data of association is further reduced by stipulating table 1 are plotted in figure 2. Inspection of the that the accepted radar range should produce a diagram shows that the points approximate well reasonable mean height for the luminous trail to two straight lines, although for very bright of the meteor. meteors the means log r are somewhat low Within the limited magnitude range covered by For accepted coincidences the observed this investigation, one may, however, assume meteor magnitudes are corrected to a zenithal that the regression is linear. or absolute value by adding a correction Regression lines, computed according to Am=10—5. log R, where R is the radio range standard procedure, are included in figure 2. in kilometers of the associated echo. No cor- The correlation coefficient for the magnitude- rection for absorption was included, since the duration data is —0.725, and the slopes, comparison method of magnitude estimates b and b , of the regression lines are —0.44 was used. x v and —0.83, respectively. These results are The recorded echo lifetimes are durations based on computations from ungrouped data. above a fixed threshold level. Duration is Error analysis.—For a comparison with nominally measured to within 0.1 second, with physical theory it is desirable to express the stricter tolerances for echoes lasting less than relation between the variates as a single 0.4 second. Logarithms of echo lifetimes were straight-line relation. The problem is best computed to within 0.01. No corrections were treated by error analysis. An unknown struc- applied to the echo durations. tural relation is postulated and it is assumed The magnitude-duration data for each ac- that the scatter about this relation is caused by cepted association were collected on a punched errors of various kinds. The structural re- card together with other relevant information lation is then found if we correct one (or both) about the meteor. The data were then grouped of the observed regression lines for the effect in various ways and correlations between of these errors. magnitude, echo duration, and height were investigated. Estimates of meteor magnitudes are avail- able from visual observations only, and the observational error in these estimates is not Experimental results I negligible. There is, however, fairly general Observational data.—A total of 808 meteors with agreement that the probable error of a magni- associated radar echoes was available for analy- tude estimate for a trained group of meteor sis. Of these, 688 belonged to the Perseid observers is not more than 0^3 or 0^4. In the shower and 120 were either members of various present investigation special attention was other showers or sporadic objects. Since the given to correct magnitude estimates, and all Perseid sample was the largest and most homo- observers had previous experience of astro- geneous group, only that group will be dis- nomical observations. It therefore appears cussed. reasonable to adopt as the probable error in The first investigation was concerned with our investigation for the combined observa- 638806—63 3 00 o +3.00 ZENITHAL MAGNITUDE FIGURE 1.—Scatter diagram of log echo duration in seconds versus zenithal magnitude of meteor. METEOR SYMPOSIUM—LINDBLAD 31 Correction for error.—Two different correction procedures were investigated. A numerical method for correcting an observed frequency curve, which is described by Eddington (1913), was considered to be the best method. An alternative method given by Seares (1944) was used as a check on the computations. Eddington's method was applied to the data in the following way: Each separate horizontal array in table 1 was corrected for error dispersion in the magnitudes. A magni- ZENITHAL MAGNITUDE tude-error curve based on a slow increase in FIGURE 2.—Array means and regression lines for table 1. the magnitude error with meteor brightness was assumed. Eddington's second-order dif- tional and reductional error a value of 0^35 ferences were computed from smoothed error (standard deviation +0^50). On the basis of curves, deduced from 3-array running means this assumption it is possible to correct the in log T. Corrected values of the correlation log r-line for error in the meteor magnitudes. coefficient and the gradient bx were then com- Measurements of echo durations are made puted. The results of these computations for directly from the radar film. With the display two models, denoted A and B, are given in used in our work, sizable errors in these meas- table 2. (Both these models were based on a urements occur only for a few exceptionally mean probable error of 0^35 and differed only long persistent echoes. The duration scatter in the method of applying the second-order which is present in meteor data is instead corrections.) In the computations A and B, related to the atmospheric and geometrical Sheppard's corrections for grouping have been conditions at the echoing point. It thus included. appears difficult at this stage to give a numerical The correcting of a bivariate distribution by estimate of the "errors" in the echo durations, the numerical method and the computing of and even more difficult to attempt to correct the corrected regression lines is a rather tedious the M*-line for these errors. procedure. The necessity to include correc- In view of the above considerations the tions for grouping introduces a slight uncer- structural relation behind the experimental tainty in the final result. The error-correction data of table 1 is sought by correcting the formula given by Seares was therefore also log T-line for magnitude error. The correction employed. In this method one directly com- procedure adopted is summarized in the next putes the corrected gradient. However, it is section. necessary to make the assumption that the TABLE 2.—Correlation data: bx, magnitude-duration gradient; x, mean zenithal magnitude; y, mean log echo duration p, correlation coefficient b, X V P Observed values -0. 435 1*00 0.52 -0. 725 Corrected values: A -0. 519 1.14 0.52 -0. 761 B -0. 480 1.05 0.54 -0. 754 C -0. 521 1.00 0.52 Mean of A and B —0.500 1>P10 0.53 32 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS magnitude error is independent of magnitude. Inspection of table 2 shows that the results A constant mean probable error in the magni- of the different methods are in surprisingly tudes of 0^35 was therefore assumed. The good agreement, all indicating similar mass corrected gradient bx is listed in table 2 (Model centers and a magnitude-duration gradient C). after error correction of very nearly —0.50. 1 The correction for magnitude error has thus 1— 1 r by -to - only slightly increased the correlation and the slope. It follows that most of the scatter in * „ the diagram of figure 1 must be related to some other factor than magnitude error. ° Observational selection.—In a following inves- -0.5 tigation the influence on the computed gradients of different observational selections was studied. bx A priori, one might expect that meteors observed at large distances and low elevations 100 150 200 250 nn would exhibit peculiar effects and that they RANGE might show unduly large scatter when plotted in a magnitude-duration diagram. The Per- seid data were therefore grouped according to T i elevation and range. Correlation coefficients -to and gradients were then computed for each of * o these groups. The results of these computa- by tions are shown in figure 3. The dependence of the correlation quantities on angular distance • from beam is also shown. In view of the -05 N paucity of the data, no attempt was made to correct the results of figure 3 for magnitude error. Although minor irregularities exist, the gen- 80o 70° 60° 50° 40° 30° 20° 10° cxo eral trend of the curves of figure 3 is clearly ELEVATION very similar. The trend is such that the correlation coefficient approaches values be- tween —0.8 and —0.9 as the variates approach 90 km, 90°, and 0°, respectively. However, 1.0 - - this increase in correlation is associated with 4M « by a marked decrease in the gradient bv, while the gradient bx is not affected to the same extent. 0 If Gaussian error functions are assumed, it can be shown that the observed lower value of 05 - by indicates a reduced scatter in the echo b, durations. It thus appears that part of the scatter in the diagram of figure 1 is related to various factors, which influence the recorded 10* *>• 30* 40' SO* 60* echo durations of distant or low-elevation 00 • I ANGULAR DISTANCE FROM BEAM AXIS meteors. FIGURE 3.—Variations of correlation quantities with range, Some investigations were made in which elevation, and elongation from beam. Symbols:p= correla- other groupings of the meteor data were tion coefficient; bz and &,= gradients of regression lines attempted. One point of interest revealed by log T and A/,, respectively. these investigations was that the derived METEOR SYMPOSIUM—LINDBLAD 33 Experimental results II Echo duration and meteor height.—According to the diffusion theory of meteor echoes, the duration T of a persistent meteor echo is proportional to the electron line density a and inversely proportional to the diffusion coeffi- cient D. For an isothermal atmosphere, D is inversely proportional to density and pressure, and both log D and log T will vary linearly with height for a group of meteors with similar electron line densities. ZENITHAL MAGNITUDE It thus appears of interest to investigate the FIGURE 4.—Adopted linear relationship between zenithal relation between log T and height in a sample magnitude and log echo duration (Perseid meteors). of meteors with similar visual magnitudes. magnitude-duration relation is independent of In the present investigation approximate meteor trail orientation; i.e., independent of whether heights were obtained from the radar range the meteors are recorded in a specular reflecting measurement and from the plot of the trail by condition or not. the visual observer. Since these latter observa- Adopted magnitude-duration relation.—The tions are subject to large errors, the individual mean elevation of the recorded meteors in our height determinations are uncertain, and only investigation is 40°. If meteors below 40° mean heights will be used. The magnitude- were rejected, it follows from figure 3 that a duration data were arbitrarily grouped in slight improvement in the correlation would intervals of 0.3 in log echo duration and l'Hi in result. A somewhat higher gradient bx would zenithal magnitude. The mean height com- also result. However, the correction for mag- puted for each of these cells is listed in table 3. nitude error is very difficult to apply to a small Means computed from less than five observa- sample. The overall accuracy of the corrected tions are given in parentheses. linear relation between log echo duration and Inspection of the vertical arrays of table 3 magnitude would therefore not be improved shows that there is a statistical relation be- by the rejection procedure. tween the mean height and the logarithm of In view of these considerations, the finally the echo duration. It is seen that meteors that adopted magnitude-duration relation was com- lie high in the magnitude-duration diagram are puted from the total Perseid sample. A mean observed at mean heights below 100 km; was taken of the two computations A and B. meteors that lie low in the diagram, are observed The resulting coefficients are included in table 2. at mean heights above 100 km. Inspection of The adopted linear relationship between magni- the marginal array indicates, as one would tude and log echo duration for a radio wave- expect, that the relation between mean height length of 9 meters is thus and log echo duration is less pronounced when an average is taken over all magnitudes. log T=—0.50M.+1.08. Meteor height versus magnitude.—Meteor The adopted line is plotted in figure 4. From theory predicts that the heights of meteors the diagram it is seen that a Perseid meteor of should decrease with increasing meteor bright- zenithal magnitude 0^0 produces an echo of ness. An experimental verification is provided duration very nearly 10 seconds and a 4^0 by the marginal means of table 3. The meteor an echo of duration 0.1 second. The variation of mean height with magnitude is relation of these experimental results to the shown in figure 5 (broken line). theory of meteor echo durations will be dis- In a subsequent investigation those meteors cussed in a following section. were selected which showed small scatter about 34 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABUB 3.—Mean height of Perseid meteor trails as a function of log echo duration and visual magnitude. Meteor heights in kilometers log* Mt -10to -3.0to -2.0 to -1.0 to -0.0 to +1.0 to +2.0 to +3.0 to Mean height -3.1 -2.1 -1.1 -0.1 +0.9 +1.0 +2.9 +3.9 2.00 to 2.29 (102) (99) 100.5 1.70 to 1.99 (117) 96 99 95 (98) 98.5 1.40 to 1.69 (91) 97 97 102 (99) 98.4 1.10 to 1.39 93 97 96 98 96.7 0.80 to 1.09 (102) 100 101 100 96 99.8 0.50 to 0.79 99 104 101 91 101.1 0.20 to 0.49 109 97 100 98 (107) 99.6 -0.10 to +0.19 (93) 104 101 98 (105) 101.0 -0.40 to -0.11 108 105 100 99 102.6 -0.70 to -0.41 (94) (55) 105 100 103 101.0 -1.00 to -0.71 (107) 102 (96) 101.4 -1.30 to —1.01 (100) -1.60 to -1.31 (91) Mean height 101.7 96.5 98.6 100.4 101.4 98.8 101.2 the magnitude-duration relation. A maximum dence of the meteor). This point requires variation in log echo duration of ± 0.20 from the further study. adopted straight-line relation in figure 4 was Duration of persistent visual train.—Meteors allowed. This group, consisting of 199 objects, are often accompanied by a persistent train, showed a somewhat more pronounced variation having a visually observed duration of the order of the mean height with magnitude (fig. 5, of a second or less. In the present radar-visual solid line). program, regular observations of these trains Individual meteors exhibit large scatter about were made, and time durations were estimated the mean height-magnitude relation of figure 5. by the observers. It is not possible to explain this scatter as the The mean train duration in seconds, com- result of observational errors alone. The puted for each of the cells of table 3, is shown question arises whether the height scatter for in table 4. A meteor with no recorded visual given visual magnitude is the result of a random train has been given zero duration in the breakup of the meteor or if some additional averaging. Inspection of the marginal mean variate is involved (such as the angle of inci- durations of the table reveals a dependence on visual magnitude and on echo duration. Inspec- tion of the vertical arrays of the table indicates a correlation with echo duration, which is independent of visual magnitude. From table 4 it may be concluded that meteors that lie high in the magnitude-duration diagram exhibit visual trains of unusual persist- ence. The durations of meteor trains and meteor ionization are thus intimately connected. The data of table 4 are published to illustrate qualitatively a general relation. They are rep- resentative of observations made at sea level. ZENITHAL MAGNITUDE The measured duration of a persistent train FIGURE 5.—Mean height of Perseid meteor trails as a function depends critically on the atmospheric trans- of visual magnitude. parency—and thus on both locality and meteor METEOR SYMPOSIUM—LTNTDBLAD 35 TABLE 4.—Mean duration of persistent train as a function of log echo duration and visual magnitude. Train durations in seconds logr Mz -4.0 to -3.0 to -2.0 to -1.0 to -0.0 to +1.0 to +2.0 to +3.0 to Mean duration -3.1 -2.1 -1.1 -0.1 +0.9 +1.9 +2.9 +3.9 2.00 to 2.29 3.5 7.0 5.3 1.70 to 1.99 3.3 3.6 2.3 1.2 2.4 1.40 to 1.69 1.5 1.0 1.5 0.4 1.3 1.04 1.10 to 1.39 0.6 0.4 0.4 0.2 0.38 0.80 to 1.09 0.3 0.6 0.3 0.4 0.02 0.32 0.50 to 0.79 0.4 0.4 0.14 0. 17 0.26 0.20 to 0.49 0.6 0.4 0.30 0.20 0.32 —0.10 to +0.19 3.5 0.3 0.03 0.04 0.12 -0.40 to -0.11 0.2 0.04 0.02 0.04 -0.70 to -0.41 0.2 0.08 0.09 -1.00 to -0.71 0.02 0.02 -1.30 to —1.01 -1.60 to -1.31 Mean duration 2.42 1.50 0.98 0.37 0. 16 0.05 elevation angle. Because of the lack of pre- trail by attachment to neutral oxygen molecules' cision of the experimental data, and since the to form negative ions. Although this reduction number of trains available for analysis is rather in duration is not of great importance for fast small, no attempt is made in this paper to treat Perseid meteors, which ionize at a mean height the train data quantitatively. of 100 km, a slight indication of some such ceil- ing effect is evident in the diagrams of figures 1 Comparison with other investigations and 2, where the recorded echo durations for m m From radar-visual data obtained on a wave- magnitudes — l and —2 are slightly less than length of 9.2 meters by Millman and McKinley predicted from diffusion theory. (1956), linear relations between visual magni- A comparison of Millman and McKinley's tude and log echo duration were reduced for (1956) table III with our table 1 shows that several meteor showers. For Perseid meteors, bright meteors are more common in the Cana- which represent the largest group, these workers dian observations than in our survey. The dis- crepancy in magnitude-duration gradient be- found a magnitude-duration relation (after tween the two investigations could thus possibly correction for error) of logr=—0.406 M.+0.87. be explained as due to the attachment process, The gradient value —0.406 is considerably which would systematically lower the durations lower than that found in the present investiga- of the bright meteors in the Canadian investiga- tion. The reason for this discrepancy is not tion and thus also cause a lowering of the clear at present, but one factor which may computed slope. be of importance in this connection will be A detailed comparison of the two investiga- mentioned. tions is complicated by the fact that various On the basis of the diffusion theory of meteor elevation-dependent corrections are introduced echoes, one expects a linear relation between log in the reduction of the Ottawa data. It is not echo duration and visual magnitude. However, known to what extent these corrections influ- Davis, Greenhow, and Hall (1959) have shown ence their observed gradient. that the echo durations of very bright meteors Since a selection according to either elevation are far less than those to be expected on simple or beam elongation is mainly one of distance, diffusion theory. This has been explained by the results of figure 3 lend some support to a the removal of electrons from the ionization suggestion of McKinley (1953) that corrections 36 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS for attenuation should be applied to echo ration of the meteor is related to the atmos- durations. However, an attempt to apply pheric density at the same point by the rela- this type of correction to our data was not very tion 1/3 successful, the net result being that the Pmax OC a mal. (4) magnitude-duration scatter diagram was shifted upwards, with no pronounced change in the Meteor magnitudes are defined by the relation correlation quantities. log a= const—0.4 Mz. (5) Theoretical relations between magnitude, duration, and height Let TmAX be the predicted echo duration at the point of maximum evaporation. Combining The ionized trails of meteors can be divided equations (3), (4), and (5) gives into two main groups, generally referred to as underdense and as overdense trails, in which log Tmax=const—0.533 Mz (6) the linear electron density a is less or greater than 2X1012 electrons/cm. for the predicted variation of echo duration For low density trails with a<2X1012 with magnitude. It is thus assumed that the electrons/cm, Herlofson (1948) has shown that final echo signal is returned from a segment of the radar echo duration to 1/e of the initial trail near the point of maximum ionization. amplitude is independent of a, and is given by Height-magnitude relation.—A relation be- tween meteor height and visual magnitude is derived as follows. An isothermal atmosphere (1) is assumed with atmospheric density p given by where D is the diffusion coefficient and X is the wavelength. The echo duration is thus inde- log p= const —^ log e, (7) pendent of a, being determined by the wave- length and the diffusion coefficient. where H and $? are height and scale height, When a]>2X1012 electrons/cm, the volume respectively. Introducing Hmwx, the predicted electron density in a trail exceeds the critical height at maximum evaporation and substi- density Ne. Under these conditions the dura- tuting equations (5) and (7) in equation (4), tion of a meteor echo is proportional to a, and we find that with the assumption that the ionization is dis- sipated mainly by diffusion, Greenhow (1952) H 04 and Kaiser and Closs (1952) have shown that —^r log e=-^r-Mz, (8) the echo duration is given by or (2) ^-x=0.306M,+const, (9) 4irDNe Magnitude-duration relation.—Meteors in the for the relation between meteor height and magnitude range — 2^0 to +4^0 will mainly zenithal magnitude. produce overdense ionization trails, and we Echo duration versus meteor heights.—It is of will assume that equation (2) applies to the interest to deduce the predicted variation of majority of the meteors in our survey. We echo duration with meteor height for a con- will further assume that the diffusion coefficient stant magnitude group. Substituting equa- is inversely proportional to the atmospheric tions (5) and (7) in equation (3), we find that density p. The radar duration of visual meteors is then given by the relation log T=const-0.4 Mz——^ log e, (10) log T=log a+log p-f const. (3) where T, Mz, and H are echo duration, meteor Meteor theory predicts that the electron line luminosity, and meteor height at the echoing density at the point of maximum rate of evapo- point. METEOR SYMPOSIUM—LINDBLAD 37 Discussion of results equation (2) is valid for all visual meteors. Comparison oj experimental and predicted gra- The validity of this assumption will now be dients.—The theoretical relationships derived discussed. For this purpose it is necessary to in the preceding section will now be compared estimate the approximate echo duration and with the available experimental observations. visual magnitude at which the transition from Physical interest is focused mainly on the underdense to overdense ionization trails occurs. gradients in the magnitude-duration and height- For an underdense trail, the echo duration magnitude relationships. to 1/e of the initial amplitude may be computed Comparison of the predicted and observed from equation (1), provided the diffusion slopes of the magnitude-duration relation shows coefficient is known. For conversion of this that the experimental data (table 2) are in very duration to threshold duration, an empirical good agreement with the predicted relation (6). graph of signal-to-noise ratio against zenithal The experimental height-magnitude relation- magnitude was constructed and used together ship for Perseid meteors is given in figure 5, with the magnitude-duration relation (fig. 4). which shows that bright meteors penetrate Assuming an exponential decay of echo ampli- lower into the atmosphere than do faint tude, appropriate values of threshold duration, meteors, as predicted by theory. The observed relative duration, and visual magnitude may value of the slope is, however, not consistent be computed for a number of different values with the value inferred from equation (9). For of the diffusion coefficient. a scale height of 6.5 km the theoretical relation 2 (9) predicts a slope of approximately 1.9 km/ With Z>=4.0 m /sec we find from equation magnitude. From the experimental curve (B) (1) a relative duration of 0.14 second. This a gradient of about 1.2 km/magnitude is duration corresponds to a meteor of approxi- deduced. However, taking means of the meteor mate magnitude 3*5, a mean signal-to-noise heights artificially lowers the experimental ratio of 4, and a threshold duration of 0.201 gradient somewhat, owing to the large scatter second. The transition between underdense which is present in the data. An approximate and overdense trails thus occurs near zenithal correction for magnitude error raises the magnitude 3*5. experimental slope to about 1.4 km/magnitude. For the faint visual meteors in the transition It should be observed that the experimental region it may, however, be more appropriate heights refer to the midpoint of the visual track, to use a higher value of the diffusion coeffi- while the height in equation (9) is the predicted cient, since these meteors may be assumed to height at maximum light. A correction to the ionize rather high in the atmosphere. If observed heights may be derived from photo- Z?=13m2/sec, as determined by Greenhow and graphic meteor data (Millman, 1958). This Neufeld (1955) for a mean height of 100 km, correction, which is —0.5 km for a 3*5 Perseid the computations give relative and threshold and —4.0 km for a —2*5 Perseid, is sufficient durations of 0.04 and 0.05 second, respectively, to remove the remaining discrepancy between and a visual magnitude between 4*5 and S"H). the experimental and predicted slopes. In this case, all meteors in a radar-visual For the variation of log echo duration with survey are of the long persistent type. It height in a constant magnitude group, the thus appears correct to apply equation (2) to previous theory predicts a slope of approxi- all meteors in the present investigation. mately —0.07 log r/km. An experimental The electron-line density in the trail may be verification of this prediction is, however, not computed from the magnitude-duration relation possible since the scatter diagram of log r and equation (2), provided the visual magnitude against H indicates very large scatter in the exceeds the transitional value. For Z>=13m2/ data even when a narrow magnitude interval 7 sec and iVe=1.33Xl0 , the relation between is considered. echo duration and line density is given by Transitional duration and line density.—In the present investigation it is assumed that T=4.5X1O"" a, ©38805—63 4 38 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS where a is the line density per centimeter path. Acknowledgments For a fourth-magnitude meteor with an echo This investigation was carried out at the Onsala duration of 0.1 second, the corresponding 12 Wave Propagation Observatory and at the Lund density is about 2.2 XlO electrons per centi- Observatory. The author is indebted to the meter. The main uncertainty in these compu- directors of these stations, Professors O. Ryd- tations rests on the assumed value of the beck and C. Schalln, respectively, for their diffusion coefficient. constant interest in the work. The investiga- Limiting meteor magnitude.—An experimental tion has been financially supported by the curve of the signal-to-noise ratio against Swedish Natural Science Research Council. meteor magnitude indicated a slow decrease with magnitude. Extrapolation of this curve References to faint meteors of zenithal magnitude 6^0 DAVIS, J.; GBEENHOW, J. S.; AND HALL. J. E. indicated that these meteors should still be 1959. The effect of attachment on radio echo detectable. A mean signal-to-noise ratio of observations of meteors. Proc. Roy. two was assumed as the detection limit. Soc. London, ser. A, vol. 253, pp. 130-139. However, the shortest duration which is EDDINGTON, A. S. recorded cannot be appreciably less than the 1913. On a formula for correcting statistics for the effects of a known probable error of time interval between two consecutive trans- observation. Monthly Notices Roy. mitter pulses. For a pulse repetition frequency Astron. Soc, vol. 73, pp. 359-360. of 50 c/s, this limiting echo duration is 0.02 GBEENHOW, J. S. second. The corresponding visual magnitude, 1952. Characteristics of radio echoes from meteor deduced from figure 4, is very nearly 5^5. It trails: III. The behaviour of the elec- thus appears that the limiting meteor magnitude tron trails after formation. Proc. Phys. Soc. London, vol. 65B, pp. 169-181. of the present equipment is determined by the GBBBNHOW, J. S., AND HAWKINS, G. S. pulse repetition frequency and that fifth-magni- 1952. Ionizing and luminous efficiencies of me- tude meteors represent the limit of detectability. teors. Nature, vol. 170, pp. 355-357. Summary of results.—An analysis has been GBEBNHOW, J. S., AND NBUFBLD, E. L. given of the magnitudes, echo durations, and 1955. The diffusion of ionized meteor trails in the upper atmosphere. Journ. Atmosph. heights of Perseid meteors. A relation between Terr. Phys., vol. 6, pp. 133-140. visual magnitude and log echo duration has HBBLOFSON, N. been deduced, and some observational factors 1948. The theory of meteor ionization. Phys. influencing the determination of this relation Soc, Rep. Progr. Phys., vol. 11, pp. have been discussed. The experimental rela- 444-454. tion is found to be linear over the range of KAISEB, T. R. ip 1953. Radio echo studies of meteor ionization. zenithal visual magnitudes from —2 0 to +4^0. Advances in Physics (Phil. Mag. The observed slope is very nearly —0.50, in SuppL), vol. 2, pp. 495-544. agreement with theoretical predictions based KAISER, T. R., AND CLOSS, R. L. on equation (6). A variation of meteor height 1952. Theory of radio reflections from meteor trails: I. Phil. Mag., ser. 7, vol. 43, with magnitude has been found in accordance pp. 1-32. with equation (9). Meteors that brighten up, LlNDBLAD, B. A. at heights above the mean height predicted 1956. Combined visual and radar observations of from their magnitude, have shorter echo dura- Perseid meteors. I. Observations in tions than those given by the magnitude- 1953. Medd. Lund Astron. Obs., ser. 1, no. 189. duration relation; meteors which penetrate 1961. Combined visual and radar observations of lower into the atmosphere have echoes of un- Perseid meteors. II. Observations in usual duration. These meteors also exhibit 1955. Medd. Lund Astron. Obs., ser. 1, persistent visual trains. The experimental re- no. 198. sults as to echo durations are in general accord MCKINLBT, D. W. R. 1953. Effect of radar sensitivity on meteor echo with the results one might expect on the basis durations. Canadian Journ. Phye., vol. of diffusion theory. 31, pp. 758-767. METEOR SYMPOSIUM—LINDBLAD 39 MILLMAN, P. M. tudes. Canadian Journ. Phys., vol. 34, 1950. Meteoric ionization. Journ. Roy. Astron. pp. 50-61. Soc. Canada, vol. 44, pp. 209-220. 1958. The heights of meteors. Journ. Roy. SEABBB, V. H. Astron. Soc. Canada, vol. 52, pp. 230-232. 1944- Regression lines and the functional rela- MILLMAN, P. M., AND MCKINLBY, D. W. R. tion. Astrophys. Journ., vol. 100, pp. 1956. Meteor echo durations and visual magni- 255-263 Abstract This investigation is concerned with the statistical relation between visual magnitudes of meteors and radar echo durations. For 688 Perseid meteors observed at a radio wavelength of 9.2 meters, a linear relation is deduced between the logarithm of the duration of the echo and the absolute magnitude of the meteor. The adopted magnitude- duration relation indicates that a Perseid meteor of zenithal magnitude 0*0 produces an echo duration of 10 sees, and a meteor of zenithal magnitude +4>PO a duration of 0.1 sec. The scatter of points in the magnitude-duration diagram is discussed. It is shown that this scatter, for constant visual magnitude, is related to the mean atmospheric height of the meteor. It is further shown that meteors which lie high in the magnitude-duration diagram are associ- ated with persistent visual trains. Radio-Echo Measurements of Meteor Mass Distributions By A. A. Weiss * Method of measurement appropriate to the fainter group of sporadic Mass distributions have been determined for meteors can be determined. Our observations sporadic meteors and for meteors belonging to show, however, that for 6.0 TABLE 1.—Values of mass distribution exponents s (o) Di«trlbut!on of acho durations Shower 7.5>Mr>6.0 2.5>itfr>1.0 1.0>Mr>0 • 1 July 28-31 Geminids 1.5 1.7 2.7 5 Aquarids 1.5 1.9 2.7 n Aquarids 1. 6 1.4 2.0 Sporadics 2.0 2.5 2.5 Two values of s are given for the present dura- Echo duration T (Me) (b) Proportion of tions measurements, since it was found that with Mr < 2 5 even after allowance for attachment the results ( persistent KhOM) could not be fitted by a single value of s over the whole range of echo durations. Kashcheyev and Lebedinets (1959) have also observed an abrupt increase in the value of s at about the second radio magnitude, which is most pro- nounced over the time of maximum activity of the shower and is then too large to be attrib- uted solely to the effects of attachment, which they neglected. A similar trend is also apparent 6 Aquoridi in the Jodrell Bank results, but only the weight- ed mean for the whole shower has been given FIGURE 2.—Increases in the proportion of bright and very in the diagram. The new results suggest that, bright meteors near the time of maximum activity of the 5-Aquarid shower, indicating a central condensation of at least near the time of maximum activity, massive meteors within the stream. the maximum concentration of mass per unit magnitude range occurs around Mr= + 1, a strong concentration of bright meteors to which is rather brighter than suggested by the the center of the stream. The evidence for Jodrell Bank authors. such a central condensation is presented in The 8 Aquarids.—The mass distribution figure 2. The proportion of persistent echoes appears to be identical with that for the Gem- increases, and the value of s for the brightest inids (see table 1). Like the Geminids meteors decreases, for the 4 or 5 days centered (Browne et al., 1956), the 5 Aquarids exhibit on the time of peak activity. At the same time, there is no change in the value of * for meteors fainter than 5th magnitude. The i} Aquarids.—This shower, which is prominent visually, gives a comparatively low echo rate on the radar when the limiting mag- tt MrlkvtU nitude is reduced to MT=7.5. In fact, during the 1957 apparition, some 30 percent of all echoes detected had MT<2.5. These results alone suggest that s<2.0, and this is con- firmed by the more accurate determinations given in table 1. With these values of s, the predicted proportion of persistent echoes (Mr <2.5) to all echoes is about 10 percent. This FIGURE 1.—Comparison of mass distribution exponents s for is only a third of the observed number; the the Geminid stream, measured at Jodrell Bank and at discrepancy is almost wholly due to prolonga- Adelaide. tion of echo durations by atmospheric motions. METEOR SYMPOSIUM—WEISS 43 Discussion time of maximum shower activity (Browne et These southern hemisphere results confirm the al., 1956). general picture which had emerged from the In conclusion, it is well to emphasize that, so northern hemisphere radio measurements, that long as the evaporation theory is used to for the major showers the value of s increases interpret radio-echo data, any appreciable from about 1.5 for meteors of 7th magnitude fragmentation, especially amongst the fainter to values >2.0 for meteors of zero magnitude, meteors, will result in inconsistencies in the mass and that the mass per unit magnitude range distributions derived by different methods (for reaches a maximum well above the visual example, from relative echo rates and from the threshold. The more detailed analysis of the widths of measured height distributions). This durations of persistent echoes has shown that provides us with a means for comparing the physical structure of meteoroids belonging to in some cases, notably the Geminids and 8 different streams. Aquarids, there is a very steep rise in the value of s for Mr Abstract Mass distributions for sporadic and shower meteors are derived from radio-echo observations at 67 Mc/sec. The data used are total echo rates for the brightness range 6.0 A Preliminary Report on Radar Meteor Counts By Peter M. Millman 1 and Bruce A. Mclntosh* As part of the Canadian program of the Inter- visual observers recorded visible meteors. On national Geophysical Year, a meteor patrol certain specified nights a team of eight observers radar was designed and constructed for use at manned the visual station and recorded all the Springhill Meteor Observatory a few miles meteors seen, for correlation with the radar south of Ottawa, latitude 45°ll/.8 N, longitude records. 75°28'.3 W, elevation 310 feet above sea level. Both the radio circuitry and the photographic This radar operates at a frequency of 32.7 me/ recording unit of this equipment were designed sec with a peak power approximately 20 kw, to give maximum reliability of performance and pulse recurrence frequency 117 cps, and a pulse uniform operating characteristics over long length 12 n sec. Independent transmitting and periods of time. The radar has been in opera- receiving aerials consist of crossed horizontal tion on a 24-hour-per-day schedule since the dipoles, fed 90° out of phase and mounted 0.4 X beginning of October 1957, and the total time above a wire mesh 105 feet in diameter and some lost due to breakdowns of all kinds has been less 3 to 4 feet above ground level. The antenna than 3 percent. pattern given by this system is essentially uni- The radar film is read by means of a specially form at all altitudes and azimuths down to an designed viewer coupled with a standard IBM- elevation of 20°. Neale gives a more complete 2 024 alphabetic key punch. The complete unit description of this meteor radar system. is illustrated in plate 1. The projection of the An air view of the Springhill Meteor Observa- film moves horizontally across the viewing tory appears in plate 1. screen past a transparent strip which carries Continuous recording of meteor echoes, using the alphabetic range coding. The operator can a range-time presentation, is made on 35-mm vary the speed of motion with a knee pedal. film. A sample record is reproduced in plate 2. The range of each meteor echo is recorded to Range markers appear as horizontal lines at the nearest 10 km from 70 km to 370 km. The 20-km intervals, the 100-km range markers durations of all echoes are also recorded on a being bright, the others dark. Individual 5-step scale—<1 sec, 1 to 2 sec, 2 to 4 sec, seconds-markers appear at top and bottom of 4 to 8 sec, and >8 sec. This information is the record, with a more prominent marker every punched on primary data cards, which are then tenth second. The serial day number, 713, and processed in the IBM-650 computer and give a the hour and minute of Universal Time, 06.18- set of 10-minute summary cards, on each of 06.19, are placed on the film above the trace which the totals are recorded for a 10-minute each minute, while various calibration dials to interval. These cards then are put through the check the performance of the equipment are IBM-402 accounting machine, which gives a set photographed at the bottom of the trace every of cards with hourly totals and a printed record hour. The small triangular markers just below similar to the one shown in plate 2. minimum range indicate the times when the A single sheet of this record tabulates the 1 National Research Council, Ottawa, Ontario, Canada. data for one-quarter of a day. Both individual 1 M. J. Neale, "Radar Equipment for Continuous Meteor Observa- tions," In press. 10-minute totals and hourly totals are repro- 45 46 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS duced, the column headings being self-explana- Information for finding correction (a) has tory in most cases. The numbers and letters at been collected, but the correction itself has not the top of the range column correspond to the yet been determined. In any case it will have punching code used. The "CLASS LONGS" little effect on the type of results reported here. columns record the numbers of echoes in various The whole instrumental system was designed duration classes. The nature of the interference to keep correction (6) as small as possible, and evident on the film is coded in the next-to- we believe that, in general, this aim was last column. Since the sample reproduced here achieved. The major exception is that, for is from a period near the peak of the Geminid some reason not entirely explained, the meteor shower, both the echo rates and the proportion radar seems to have operated somewhat more of long-duration echoes are considerably higher efficiently during the 1959-60 period than in than average. 1957-58. No attempt has yet been made to The data for the period from October 1957 allow for this. through December 1960 have now been proc- A preliminary determination of correction essed as described above, and approximately (c) has been made by studying discontinuities 4,400,000 meteor echoes were recorded during in the mean diurnal rate curves, the variation these 39 months. We have begun a detailed in rates for specific local times in the day, and statistical study of this material and give here the variation in the ratios of long duration to only a brief outline of some of the preliminary total echoes. These data were correlated with results. the index figures for interference as estimated The tabulated rates should be corrected for: by the filmreaders , and a mean correction factor (a) The personal equation of the film readers; was found. It varies from 1.0 for no inter- (b) variation in the performance characteristics ference to 2.5 for maximum interference. In of the radar equipment; and (c) interference in plotting diurnal rate curves the observed rates the radar reception. have been corrected by multiplying by this 1 ' 1 r- 1 ' : '——1 '——1 1 i 1 1 1 1 1i 1 1 1 1 1 1 o 300 ;.' - PE R if) LJJ •• • - ' : '•. ' ':.'• • • : r X 200 ••"••.'. • .;.'•- V'.•»'";;"." --.vV - " - -% LJJ ' • •' • • ..".*••-' ••!•' . < . • •.« . • •." •. :TEO R '•'^ • "" -E IOO - . v. . •• •• •••" ;.- 1957 | 1958 [ 1959 | I960 1 300 400 500 600 700 800 900 1000 1100 1200 1300 I40O I.G.Y. DAY FIGURE 1.—The mean hourly echo rate for each day. The IGY serial day is numbered from 1 on January 1,1957, METEOR SYMPOSIUM—MILLMAN AND McINTOSH 47 300 JAN I FEB MAR APR MAY JUN JUL I AUGI SEP I OCT NOV I DEC I FIGURE 2.—The annual variation of mean hourly echo rate for each day: upper curve, all echoes; middle curve, echoes with durations 1 second or greater; lower curve, echoes with duration 8 seconds or greater. factor. Periods where the correction factor serially from then on. Note that the mean was 2.0 or more have been omitted from the hourly rates for a day vary from a minimum plotted results. If we consider only the cases near 50 during the first few months of each where the rates were altered by more than 10 year to peak rates between 250 and 300 in percent, for the first 18 months a correction June and December. was made in the case of 40 percent of the ob- To give a general picture of the average served rates, while for the last 21 months, variation in meteor rates throughout the year, which were further removed from the sunspot the values from figure 1 were averaged for each maximum, a correction was made for less than 7 calendar day and are plotted as the upper curve percent of the observed rates. in figure 2. Below this are plotted the mean Uncorrected mean hourly rates of meteor hourly rates for each day for all echoes with echoes for each day are plotted in figure 1. durations of 1 second or greater, and all echoes The IGY serial day number is based on the with durations 8 seconds or greater. Most of system used at all the Canadian meteor and the minor fluctuations in these curves are prob- auroral stations where January 1, 1957, is ably accidental and should not be considered as taken as Day No. 1 and the days are counted particularly significant. Some of the major 48 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 1958 1959 I960 JAN FEB MAR en f-1-2 1 r- Q 200 APR LU 100 300 200 MAY 100 400 300 JUN 200 !00 19 I- 7 13 19 I 7 13 19 I 7 13 EASTERN STANDARD TIME FIGURE 3.—The monthly mean diurnal echo rate curves for January to June: solid line, all echoes; dots, echoes with durations 1 second or greater. showers, however, stand out clearly, particularly to estimate the background rate, excluding for the lower curves of the long-duration echoes. these major showers, and the values read are Good quantitative values for the percentage of given in the first half of the table. The second meteors in each of these showers must await a half of the table shows the peak rates that cor- much more detailed study of the tabulated respond to the shower meteors only. records, but qualitative values can be found It is necessary to remember here that all from figure 2, and are listed in table 1. Eleven these rates are based on averages for a complete well-marked shower peaks were chosen, includ- day. In cases where a shower radiant is below ing the strong daylight shower complex in June, the horizon or at low elevation for part of the and, in addition, two other peaks near Septem- day, the peak strength of the shower will appear ber 8 and September 30. An attempt was made smaller in the table than it actually is. MILLMAN AND MclNTOSH PLATE 1 Springhill Meteor Observatory near Ottawa, Ontario. The meteor radar antennas appear in fore- ground and left background. The visual observing station is at upper right. Meteor radar film reader coupled with IBM-024 alphabetic key punch. MILLMAN AND MclNTOSH PLATE 2 Typical meteor radar range-time record. Over 30 meteor echoes and 2 visual marker appear on this portion of the film. 1. 0. T. MITIOI 1 ADIO CANADA Nl M E« IN 1 AhIG C A! SE S CL/ ONOI UNI VEIS Al NUMMH LONG lO.r DAT I Tl 25 0 * 150 3 3 « OF Out 350 bi• 300 0 100 OAT METE CMS T a 12 14 06 77 47 32 4 4 5 1 3 3 4 3 8 3 1 2 1 4 2 2 7 4 2 6 15 87 52 2 2 1 2 2 4 4 6 7 3 4 9 6 5 3 1 1 1 10 17 3 74 54 36 13 4 81 47 1 1 1 4 I 66 38 4 35 1 1 2 1 5 1 1 5 4 3 3 3 4 1 3 8 2 •? 2 10 1O 3 66 32 1 1 1 2 1 4 2 5 3 4 4 5 1 4 6 4 "5 2 a 5 10 5 1 6 451 27O 69 41O 1O 10 10 5 2O 14 24 13 27 16 22 27 22 21 29 16 24 27 18 15 19 6 66 72 94 28 713 7 3 3 13 •3 4 a 14 07 0 72 36 14 3 | 1 1 1 2 3 4 6 3 5 1 4 3 7 11 71 32 3 1 4 3 3 1 2 1 6 ; 2 2 4 2 3 6 7 3 2 4 6 9 14 3 69 25 4 1 1 1 1 1 2 2 4 2 5 6 2 3 4 3 6 6 4 3 3 9 7 •S 3 3 73 22 14 3 1 7 3 4 65 25 6 14 5 5 84 33 313 2 4 10 7 !"» 3 434 173 58 8212 9 13 7 6 11 15 16 1O 22 8 19 26 24 22 33 17 33 32 30 20 17 49 59 44 21 713 1 76 36 12 1 1 86 37 11 5 19 4 1 1 16 2 3 3 101 33 12 31 - 1 2 5 6 6 3 6 1 6 8 5 12 14 2 4 97 36 1 11 H 3 81 26 3 1 1 5 5 8 1 4 6 5 5 2 •5 8 * 3 2 518 21O B t 3 6 5 I 6 8 17 16 ZO 15 2B 16 23 IS 43 18 22 37 36 33 41 18 23 9 1 1 92 37 69 12 713 inmm r 8 12 O9 O 83 26 1 6 4 11 L 72 25 1 1 1 3 5 3 1 6 1 4 7 5 6 •3 5 1 10 5 7 3 90 25 12 4 9 11 3 3 85 27 4 80 22 12 22 5 81 11 1 1 1 6 5 4 4 3 5 i 5 7 6 8 8 6 5 1 1 10 1 i 491 147 53 34 3 4 12 8 IS 11 24 19 L3 23 29 30 23 3O 33 39 3O 46 31 35 7 4 59 22 49 17 713 8 It? 14. 1O 0 83 2 4 1 1U y 3 1 57 14 1 1 1 6 5 3 1 6 1 p 78 11 1 1 a e 6 8 7 2 4 3 fl7 1 3 I : 1 6 6 10 6 6 1 4 1 4 68 16 1 1 t 6 E 3 7 4 5 82 1 3 l 3 1 6 i 10 7 7 lO 6 i 6 4 3 6 455 91 31 2 1 17 14 IS 19 16 12 30 3O 33 25 28 43 31 47 2 9 23 11 2 39 13 27 12 713 a 14 11 U 70 1 7 1 1 6 8 6 •> 1 1 72 12 1 1 9 a 1 ] 4 2 2 73 1 1 1 1 i 1 6 1 c 1 i i 4 3 61 i 7 1 ] 6 6 lO 1 4 54 6 1 ] 1 4 6 1 i 1 77 12 6 4 f 11 2 6 4O7 65 21 2 6 11 6 5 17 22 17 14 23 21 12 30 35 33 24 36 27 36 11 3 1 21 11 1 1 713 Typical sheet of data from the meteor radar, covering the period 6-12 hours U.T., December 14, 1958 METEOR SYMPOSIUM—MTTiLMAN AND McINTOSH 49 1958 1959 I960 JUL NOV DEC EASTERN STANDARD TIME FIGURE 4.—The monthly mean diurnal echo rate curves for July to December: solid line, all echoes; dots, echoes with durations 1 second or greater. On the basis of the absolute numbers of all When we study the differences between the echoes, the Geminid shower is by far the meteors of the major showers and those of the strongest, followed by the Quadrantid shower, general background, the percentage of long- the June showers, and the peak at September duration echoes is of particular interest. 30. But if we look at the long-duration echoes These have been listed in parentheses in table 1. only, we see that Perseids and Geminids are For the background, roughly 10 percent of all about equal in strength, while the Quadrantids echoes have durations 1 second or greater, and and June showers are only half as strong, and 2 percent have durations 8 seconds or greater. the rest of the showers much weaker still. We see immediately that most of the showers 638805—63 5 50 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Mean hourly echo rate for 1 day Background echoes Shower echoes Date Shower all >lsec >8sec all >lsec >8sec Jan. 3 Quadrantid 135 14 2 79 29(37)* 9 (11) Jan. 20 110 11 1.5 — — — Feb. 20 90 10 1 — — — Mar. 20 105 11 1 — — — Apr. 22 Lyrid 146 16 2 17 6(35) 3. 5(20) May 5 V Aquarid 160 17 2.5 53 15(28) 2.5 (5) Jun. 11 Daytime 169 17 3.5 81 35(43) 9.5(12) Jul. 29 S Aquarid 170 17 3.5 40 11(28) 2.5 (6) Aug. 12 Perseid 169 17 3.5 31 29(94) 19. 5(63) Sep. 8 165 17 3.5 50 2 (4) 0.5 (1) Sep. 30 161 16 3 79 3 (4) 0.5 (1) Oct. 22 Orionid 156 16 3 49 6(12) 3 (6) Nov. 20 150 16 3 — — Dec. 13 Geminid 142 15 2.5 128 52(41) 18.5(14) Dec. 22 Ursid 137 14 2 13 5(38) 4 (31) Mean 145 14(10) 2. 5(2) 'Figures in parentheses are long-duration echo percentages in each category. have a very much higher percentage of long- on this occasion. The problem of correcting duration echoes. The Perseid shower is quite for interference became much less acute in 1959 unique in this respect, with over 90 percent of and 1960. its members having echoes of duration 1 second We do not here propose to discuss the or greater, and over 60 percent, 8 seconds or monthly diurnal rate curves in detail. Despite greater. Most of the rest of the showers have various fluctuations from year to year, it is from 30 to 40 percent of echoes with durations clear that the diurnal curve for any given of 1 second or greater, and 10 to 20 percent, 8 month shows certain general characteristics seconds or greater. The Orionids and the two throughout. This is true both for all echoes September peaks seem to have about the same recorded and for the much smaller quantity of percentage of long-duration echoes as the data represented by the long-duration echoes. background. The two diurnal maxima which appear near A mean diurnal hourly rate curve was now the equinoxes are particularly prominent in drawn for each month, and corrected for March. The first of these maxima, in the interference using the correction factor de- early morning hours, tends to be more promi- scribed earlier in this paper. These curves, for the entire 39 months, are reproduced in figures nent before March; the second, during the 3 and 4. The solid lines represent values for daylight hours, becomes stronger after March all echoes, and the dots represent echoes with and has a high absolute value as the complex durations of 1 second or longer. Where heavy of summer daylight showers becomes evident. interference prevented the use of data for more The diurnal peak for the long-duration echoes than half the days in a month, the solid line is does not always coincide with that for all echoes. replaced by a broken line. Examination of the This is particularly true in the last quarter of diurnal curve for December 1957 makes it the year, and is no doubt due in part at least quite obvious that the mean interference to the effect of showers. The general shape of correction factor used was not adequate to the curves for all echoes reproduced in figures correct for the actual interference experienced 3 and 4 bears a remarkable similarity to the METEOR SYMPOSIUM—MILLMAN AND McINTOSH 51 plotted in figure 5. Here the high peak rates in June are a prominent feature, together with the December peaks and less regular ones in September and October. In the minimum rates the low values during February and March stand out. To illustrate the changes that occur through- out the year in meteor activity for various hours in the day, the monthly mean rates for four specific hours of E.S.T. are plotted in figure 6. The highest rates definitely tend to occur near 7 hours and the lowest near 19 hours. FIGURE 5.—Monthly mean rates for all echoes: top curve, Here again the high rates for the daylight hours diurnal peak rate; middle curve, overall average rate; lower in June are outstanding. The mean rate for curve, diurnal minimum rate. 1 hour is definitely higher for the last half of the year, but does not show as much variation as might have been expected. We have begun a much more detailed study of all these observational data. In particular, the analysis of the range distribution will make it possible to estimate with much greater accuracy both the contributions and the indi- vidual characteristics of the various showers. Correlation with the visual observations will aid in calibrating the radar records in relation to the visual data. Acknowledgments Many individuals have contributed to the FIGURE 6.—Monthly mean rates for all echoes taken at 4 specific hours of E.S.T. success of this Canadian IGY-IGC program. M. J. Neale was chiefly responsible for the predicted diurnal curves computed by Hawkins,8 design of the meteor radar equipment, and although the latter were plotted for assumed members of the staff of the Upper Atmosphere observational conditions somewhat different Research Section at NRC constructed and from those of the Canadian program reported maintained it. Those assisting with the film here. reading and the IBM processing include Miss The mean hourly rates for each month, to- A. M. Baker, Mrs. S. Baron, Mrs. H. Bone, gether with the peak rates and minimum rates Mrs. E. Bradley, Miss J. Fagen, Miss D. taken from the curves of figures 3 and 4, are Gibson, Mrs. H. MacLean, Mrs. M. McCall, Mrs. D. M. McKiernan, Miss M. Panet, Miss * O.S. Hawkins, "A Radio Echo Survey of Sporadic Meteor Radi- J. Proudman, and Miss M. Watson. ants." Monthly Notices Roy. Astron. Soc, vol. 116, pp. 92-104,1960. Abstract A meteor patrof radar equipment was put in operation at Springhill Meteor Observatory near Ottawa at the beginning of October 1957, and a continuous recording of meteor echoes has been carried out since this date. The antennas are omnidirectional, with operating frequency 32.7 mc/sec and peak power 20 kw. Echoes are recorded on 35-mm film in the form of a range-time record. Echo counts for each 10-minute interval, echo ranges and durations, and the intensity of background noise are punched manually on IBM cards from the film record. All film has been read for the 39 months of operation from October 1, 1957, to December 31, 1960, and over 4,400,000 meteor echoes have been recorded. Mean hourly echo rates range from a diurnal minimum of less than 100 to a diurnal maximum which in June at 8h local time may be over 400. The maxima of the major meteor showers show up more distinctly in the rates for the long-duration echoes. There is a tendency for the shape of the monthly mean diurnal rate curve to repeat itself in the same months for successive years. The Harvard Radio Meteor Project* By Gerald S.Hawkins2 The Radio Meteor Project has been described The antenna pattern for the single trough is by Whipple and Hawkins (1956). It is a shown in figures 3 and 4: the measurements multistation radar system in which a pulse were obtained from a scale model. The main train is reflected from the ion column of a beam, directed at azimuth 113° E of N, has meteor. The strength of the echo varies with an elevation of 43°, and the side lobes are more time and the pulses form a Fresnel pattern as than 20 db below the gain of the beam center. described by Davies and Ellyett (1949). The The antenna pattern of the double trough in velocity of the meteor can be found from the the vertical plane is also represented by figure 3, period of amplitude variations and the measured while the horizontal pattern is shown in figure 5. range, while the angle of approach of the meteor These patterns were obtained from measure- is given by the time delay between the stations. ments on the model antenna. The beam width Thus the velocity, radiant, and orbit of indi- between half-power points is ± 14° for the vidual meteors can be found. Gill and Davies double trough, and ± 25° for a single trough. (1956) developed a 3-station system for measur- The double trough antenna is used both for ing orbits at Jodrell Bank Experimental Station transmission and reception and a TR/ATR in England. The Harvard system has been switch is used to effect the changeover. designed to give the same data as the Jodrell Receivers and radio links.—Superheterodyne Bank equipment and in addition to provide a receivers with an intermediate frequency of 3 measurement of height, deceleration, and dis- me were used. Each receiver was fed from a tribution of ionization along the trail. A high voltage stabilizer. sensitivity was chosen to make possible the The microwave units were standard Raytheon detection of meteors at magnitudes -j-12, some links (MCR-1000) with a transmitter power of three magnitudes fainter than the limit of Gill 1 watt. The pulses were carried on the links and Davies. at IF frequency of 3 me to the Long Branch field site of the Bureau of Standards at Havana, Construction of the system 111. The work carried out during the construction Frequency controls, gate circuits, and display period has been described (Whipple and Haw- units.—The master oscillator unit, containing kins, 1960). The completed system is shown the time base generator, the 150 kc oscillator, in figure 1. The various components of the and the associated divider units are all housed system are described below. in the console, shown in plate 1. Plate 2 con- Antenna.—A double trough antenna was used tains a close-up view of the coaxial lines feed- at the main site. Two 13-element Yagis were ing the double trough antenna, and also a view used at the remote sites. Figure 2 shows a of the gating, triggering, and bright-up equip- sketch of the double trough, and a photograph ment. The Fresnel and range pulses were dis- of this antenna is shown in plate 1. played on single-beam, three-inch tubes. > This research was supported by Lincoln Laboratory, MIT, AFl9(122)-468/57; National Bureau of Standards, CST-7282, CST-7067; Computing program and National Science Foundation Grant Q-14699. It was published In detail as Harvard Radio Meteor Project Final Beport of CST-7282; A detailed computational program had been Dec. 13, 1957, to June 30,1901. prepared for the IBM-704 computing machine 'Harvard College Observatory, Cambridge, Mass., and Boston University, Boston, Mass. (Whipple and Hawkins, 1956). The reduction 68 54 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS HARVARD RADIO METEOR PROJECT STATION , BLOCK DIAGRAM Trough-Guide MASTER MASTER Antenna N.B.S. 40.92 Me P.R.F. OSCILLATOR MEGA-WATT CONSOLE GENERATOR 40.92 Me TRANSMITTER STANDBY MONITORING 30 Kw EQUIPMENT TRANSMITTER TRANSMITTER Trough -Guide 13- Element 13- Element Antenna Yag'i Antenna Yagi Antenna No.3 BASE No. 4 REMOTE No. 5 REMOTE STATION STATION STATION 1 r-- 40.92 Me 40.92 Me 40.92 Me RECEIVER RECEIVER RECEIVER 6 Kmc 6 tone VIDEO MICROWAVE MICROWAVE LINK LINK I L I I J_ I VIDEO Link Antenna Lpnk Antenna GATE AND TRIGGER UNITS 6 Kmc 6 Kmc VIDEO MICROWAVE MICROWAVE LINK LINK RECORDING CONSOLE FILM (Na4) (NO5) 6ange) RECORDING UNIT (SCOPE INDICATORS) RECEIVERS FIGURE 1.—Diagram of Harvard radio meteor project system. METEOR SYMPOSIUM—HAWKINS 55 FIGURE 2.—Artist's sketch of double trough antenna. of the Fresnel patterns is based upon a com- The majority of the records were obtained parison of the measured pattern with the set with the National Bureau of Standards Collins of theoretical Fresnel patterns derived by transmitter with a power output of 25 kilo- Loewenthal (1956). By comparing the obser- watts. A small amount of filming was carried vations with the theoretical pattern, the 90 out at a power level of 1 megawatt. We esti- percent confidence limits are computed for each mate the limiting visual magnitude was +6 for Fresnel pattern. In this way we can determine the low-power transmitter and +9 for the high- whether or not significant deceleration has power transmitter. taken place. The program was converted from IBM-704 Results to IBM-7090. A Benson-Lehner film viewer A total of 407 meteors has been completely complete with digitizer was used and this was reduced to give radiants, velocities, and orbital coupled to an IBM card punch. parameters, and the results were published in 56 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS HALF-POWER POINTS 0° 20° 40° 60° 80° 100° 120° H0° 160° ELEVATION ANGLE (degrees) FIGURE 3.—Single trough antenna gain pattern (voltage), 2 4 6 8 10 12 14 16 vertical plane. Hours central standard time FIGURE 6.—Histogram of hours of observation. periods will introduce a statistical bias unless HALF- POWER we operate on a continuous 24-hour schedule. POINTS Not all of the films have been analyzed and the histogram therefore covers only those films selected for analysis. Our observations were made most frequently between 4.00 and 11.00 hours Central Standard Time. This non- uniform coverage was due to many factors, mainly electrical interference. Our observers -80° -60° -40° -20° 0° 20° 40° 60° 80° found that interference was more intense during AZIMUTH ANGLE (d«er««») the daylight hours from 11.00 to midnight, and FIGURE 4.—Single trough antenna gain pattern (voltage), we therefore chose to operate during the early azimuthal plane. morning hours when our yield of echoes was higher. We decided that this coverage was satisfactory for preliminary work since we were observing both the antisolar radiants and the - HALF- POWtn POINTS apex area of the sky. 1 \ 1/ to- The distribution of velocity is given in figure f \ 7. The most frequently observed velocity was I \ between 30 and 35 km/sec, which coincides with \ the peak found by other radio investigators. / s \ / 20- The histogram from the Harvard Radio Meteor Project differs markedly from other radio data at the high-velocity end. The Harvard data contain few meteors with velocities greater than FIGURE 5.—Double trough antenna gain pattern (voltage), 50 km/sec, whereas other observers have found azimuthal plane. (Pattern measured from model by a subsidiary peak between 50 and 70 km/sec. Radiation Engineering Laboratory.) We attribute this difference to a statistical bias in the Harvard data due to the fact that our condensed form (tables 1 and 2, Whipple and observations were made during the months of Hawkins, 1961). Before proceeding with the December, January, February, March, and discussion of the results we give a histogram of April. At this time of the year the apex of the observing hours in figure 6, since our observing earth's path is in a low declination region of the HAWKINS PLATE 1 a, Console housing the master oscillator unit, b, Double trough antenna. HAWKINS PLATE 2 a, Close-up view of the coaxial lines feeding the double trough antenna, b, View of the gating, triggering, and bright-up equipment. METEOR SYMPOSIUM—HAWKINS 57 VELOCITY DISTRIBUTION 1 1 1 1 1 90 — — 80 1 70 - — - - o >,r o f meteor s JO J40 - - -30 -20 -10 0 10 20 30 40 50° 6 30 - FIGURE 8.—Distribution of observed declinations. 20 - In" observations have shown that the radiants are 10 — fairly uniformly scattered over the sky and that 1 there is a broad concentration of radiants n 1 1 1 1 1 20 30 40 50 60 70 between declinations —30° and +30°. We Velocity, km/sec presume that the discrimination of our system FIGURE 7.—The distribution of velocity. against low altitude radiants has caused this asymmetry in the distribution of declination. ecliptic. Our sensitivity decreases as the During the period of observation covered in radiant altitude decreases, and thus our obser- this report the apex was between declination vations of the apex will not be representative 0° and —23°, and the number of meteors until the latter part of the year. The poor observed from this region of the sky was there- observability of low altitude radiants is shown fore considerably reduced. clearly in figure 8, which gives the distribution Figure 9 shows the distribution of the semi- of observed declinations. At the latitude of the major axes of the orbits. We have chosen I/a Long Branch Site a radiant with a declination as our parameter rather than a since the ob- of —50° never rises above the horizon. A servational errors in I/a are almost constant. radiant with 0° declination has a maximum The most frequent value of a is 1.0 in our data, altitude of 50°, and one with +30° declination whereas the corresponding value obtained at is in the zenith. Photographic and visual Jodrell Bank was 3.0. In the Harvard data 40 - JODRELL (2400) HARVARD (406 meteors) n 35 30 L-u~U « 25 o I 20 o l .5 z 10 5 0 -1.0 -.5 0 .5 1.0 1.5 2.0 '/a FIGURE 9.—Distribution of the semimajor axes. 638805—63 6 58 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 30 27 24 12. D • o (407) RADIO (360) PHOTOGRAPHIC XI I 12 9 Q 6 Toroidal Geminld. group 3 til. m m r~-i m 0 8 10 12 14 16 25 27 36 Number of meteors in each stream FIGURE 10.—Meteor groups separated by small differences in their orbital elements. no correction has as yet been made for the Geminids obtained from the Super-Schmidt deceleration of the meteor in the atmosphere. camera data (Hawkins and Southworth, 1958). This deceleration is thought to be insufficient to The agreement in the radiant distribution is account for the difference in the two distribu- satisfactory, though the difference in declina- tions. A small fraction (3 percent) of the or- tion is a little larger than we would expect. bits in the Harvard survey was hyperbolic. The difference in velocity shows the effects of We do not consider this percentage significant; deceleration; the photographic data have been it is the value to be expected from the error corrected for deceleration whereas the radio in velocity determination. data have not been corrected. Thus we may South worth (Hawkins and Southworth, 1960) assume that the average deceleration of the has applied a criterion for identifying minor radio meteors in this sample of Geminids is streams amongst the orbits in this survey. 2.3 km/sec. The orbital data were put on punched cards and sorted so as to find meteor groups which TABLE 1.—Geminid data were separated by small differences in their orbital elements. The results are shown in Radiant Position Velocity figure 10. From the 407 orbits tested, 53 km/sec groupings were found This compares with a R.A. Dec. total of 35 groups found from a similar search of 360 Super-Schmidt orbits. The percentage Radio 110?9 + 35?0 33.7 of streams in the two sets of data is comparable Photo 111?4 + 32?5 36.0 because allowance must be made for the larger number of radio orbits and the fact that these orbits were concentrated in a 4-month period The stream with the largest membership (36 of the year. meteors) represents a very complex system of One of the streams found by this automatic orbits of high inclination and small eccentricity. selection was, of course, the Geminids of De- These orbits were similar to those found by cember. This well-known major stream of Davies (1957) in the survey conducted at the nighttime sky serves as a useful check on the Jodrell Bank. This grouping of meteors is radio system. The mean radiant and velocity shown more clearly in figure 11, where we have for twelve Geminids is given in table 1. For plotted the distribution in inclination for three comparison we include the mean value of the values of e. As the eccentricity of the orbit is METEOR SYMPOSIUM—HAWKINS 59 30 1 • • 1 1 • • 1 eccentricity. These orbits are difficult to ac- HARVARD count for in terms of our present ideas of - -—JODRELL interplanetary material. Previously the photo- graphic observations had shown a marked 20 - TOROIDAL e<0.3 concentration of meteor orbits in the plane of GROUP —-I 67 METEORS the ecliptic, and the high inclination orbits were not detected in any great numbers. For I—1_ r- J '— convenience we call this set of orbits the - i—i — "toroidal group" because the orbits form a r — i—i i i cylindrical toroid in space. i i The time of appearance of the toroidal group I is concentrated at 9.00 hours Central Standard 1 I —• 60 120 180° Time, as can be seen from figure 6. Thus the radiant points of this group are approximately 1 • i 1 1 at the same longitude as the apex of the earth's path, since the apex passes through the sensitive sector between 9.00 hours and 10.00 hours 20 - Central Standard Time. TOROIDAL 0.3<»<0.7 GROUP 1 '— As far as we can tell there is no reason why 178 METEORS the photographic survey should not have located this type of orbit. The radiant occurs at the same longitude as the apex, some 30° 10 -- r*"» - i • north and south of the ecliptic. Figure 12 i shows the distribution of radiant points ob- tained in the photographic survey (Hawkins 1 H~^ , and South worth, 1958) plotted in ecliptic coordi- 60 120 180* nates with the sun chosen as a fixed zero point. The photographic data contain a concentration of radiants north of the apex, but the orbits of these meteors were not circular in nature as the ones found by the radio survey. We do not think that the high inclination orbits found by the radio survey are a result of selection effects due to the orientation of the antenna. To investigate this possibility we have plotted in figure 13 the position of radiants on an altitude-azimuth diagram together with the position of the sensitive sector of the antenna. From the diagram we see that the radiants of meteors are uniformly scattered 0 £ • 60 120 180* along the sector (the lack of meteors in the FIGURE 11.—Inclination distribution of meteors. northern half of the sector is caused by our early-model triggering system which could only record meteors coming from the southern decreased, there is a tendency for the meteors quarter of the sky). The radiant points of the to occur with inclinations of approximately high inclination near-circular orbits are scat- 60° to the plane of the ecliptic. For comparison tered fairly uniformly around the sector in a we show the Jodrell Bank results, and the manner similar to that of the other types of agreement is good, though the Harvard data orbits. Thus we cannot attribute this grouping contain a higher percentage of the circular, in- of orbits to a side lobe or to an area of high clined orbits in the intermediate values of sensitivity in our antenna sector. OS o • Sporadic meteor ° Minor stream o Major stream FIGURE 12.—Distribution of radiants of major streams, minor streams, and sporadic streams. bJ .EVA!"ION 8 80° 90° 60° 50° 70° 30° 40° I0 0' NORTH 0 • NormalRadiants ToroidalRadiants NORTHWEST 0 <.~/ FIGURE 13.—Positionofradiants. WEST SOUTHWES o . Antenna Sector SOUTH C_eJ_e^tial Equator I SOUTHEAST - 0° - 10° -20° -30° -40° -50° -60° -70° -80" -90° CD 62 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS References Project, Harvard College Observatory, Contract AF 19(122)-458, Subcontract DAVIES, J. G. no. 57. 1957. Radio observations of meteors. In Ad- LOEWENTHAL, M. vances in Electronics and Electron 1956. Meteor echoes from underdense trails at Physics, vol. 9, pp. 95-128. Academic very high frequencies. Technical Report Press, New York. no. 132, Massachusetts Institute of DAVIES, J. G., AND ELLYETT, C. D. Technology, Lincoln Laboratory. 1949. The diffraction of radio waves from meteor WHIPPLE, F. L., AND HAWKINS, G. S. trails and the measurement of meteor 1956. The Harvard radio meteor program. Final velocities. Phil. Mag., ser. 7, vol. 40, Summary Report, Harvard Radio Me- pp. 614-626. teor Project, Harvard College Observa- GILL, J. C, AND DAVIES, J. G. tory, Contract AF 19(122)-458, Subcon- 1956. A radio echo method of meteor orbit de- tract no. 57. termination. Monthly Notices Roy. 1960. Radio meteor project. Final report, Har- Astron. Soc, vol. 116, pp. 105-113. vard Radio Meteor Project, Harvard HAWKINS, G. S., AND SOUTHWORTH, R. B. College Observatory, Contract CST- 1958. The statistics of meteors in the earth's 7067. atmosphere. Smithsonian Contr. Astro- 1961. Harvard radio meteor project. Final Re- phys., vol. 2, no. 11, pp. 349-364. port, Harvard Radio Meteor Project, 1960. Statistics of meteor streams. Interim Re- Harvard College Observatory. Contract port No. 44, Harvard Radio Meteor CST-7282. Meteor Rates Observed by Radio-Echo Techniques During the IGY-IGC Period By B. L. Kashcheyev1 and K. V. Kostylyov2 Meteor rates were measured at the Kharkov the echoes can be attributed to major meteor Polytechnical Institute according to the IGY- streams—primarily the Arietids, Geminids, IGC program. The measurements were started 8 Aquarids, rj Aquarids, Perseids, Lyrids, and in December 1957. The wavelength used was others. The rest can be attributed to sporadic 8 meters. Twice since the beginning of the meteors and the numerous minor meteor measurements new types of antenna systems streams that cannot be distinguished from the have been introduced, and the sensitivity of the sporadic background unless highly directive an- receiving equipment has also been increased. tennas are used for the meteor-rate measure- During the first period, from December 1957 ments. to August 1958, a 5-element director antenna and It is worth noting that the Perseid stream, a half-wave dipole were employed; in the although one of the most active for photo- second period, from September 1958 to March graphic and visual observations, produces a 1960, an 8-element director antenna and a scarcely noticeable increase in the general num- half-wave dipole were employed for trans- ber of radio echoes. On the other hand, the mission and reception, respectively. During Perseid stream is clearly discernible because the third period, since April 1960, 5-element the radio echoes produced by Perseid meteors director antennas have been used for both have a duration of more than 1 second. This transmission and reception. During the ob- can be explained by the fact that high-velocity servation period the antenna systems were meteors with a brightness less than +5 magni- usually oriented eastward. tude produce weak radio echoes, and that there The sensitivity of the equipment and the are few faint meteors in the Perseid stream analysis of the records show that during the (Kresakova, 1958). Radio-echo techniques first period mentioned the overwhelming ma- lead to discoveries of new meteor streams that jority of echoes recorded had an absolute visual cannot be observed by optical methods. The magnitude brighter than +6*5. For the highest meteor rates are produced by the second period this critical value lay between Arietid stream. Next highest are those of the +7*5 and +8*0, while for the third period it Geminid, Quadrantid, and other streams that was +9m. The successive improvements in are well known from optical observations. the equipment resulted in an increase in the At the Engelhardt Astronomical Observatory diurnal rates of meteors recorded, which were in Kazan, systematic radio-echo measurements about 1,800, 4,000, and 8,400 for the three of meteor rates were made with fixed and periods, respectively. rotating antennas (Kostylyov, 1958). The fixed From December 1957 to December 1960 antenna was directed to the west and the wave- observations were made during 330 24-hour length was 4.2 meters. With the rotating periods and about 1,300,000 echoes were re- antennas two wavelengths were used: Xi=4.2 corded, some 85,000 of which had a duration of meters and X2=8.7 meters. The antenna sys- more than 1 second. About 10 to 15 percent of tem was rotated in steps of 30°, and stayed in each position for 5 minutes. Thus during 1 i Kharkov Polytechnical Institute, U.S.S.R. > Kazan State University, U.S.S.R. hour the antenna system turned 360° in a 63 64 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS h h h h h h h h the antennas were 5-element directors with o 3h 6 9 I2 .5 I8 2. 24 i horizontal jibs, and the height above the ground 400 - — was X2/2. For the 17 months of observation in the period 200 / - 200 1958 to 1960, the equipment with fixed antenna v and Xi=4.2 meters was employed during 205 n 1 i i 24-hour periods; 35,000 echoes were recorded h h I2r I5h I8h 2l 24 and the mean daily rate was 160. The equip- FIGURE 1.—Yearly means of diurnal variations of meteor ment with the rotating antenna system (Xi=4.2 rates as measured in Kharkov in 1959 (X=8 meters); ordi- meters) was employed during 229 24-hour nate, hourly rates; abscissa, hours local time. periods in 1958 to 1960; about 30,000 echoes were recorded and the mean daily rate was 130. 10 The equipment with rotating antenna system JAN and X2=8.7 meters was employed during 290 10 24-hour periods in the period 1958 to 1960; 0 about 369,000 echoes were recorded and the 10 mean daily rate was 1,270. Figures 1 and 2 show the mean diurnal 0 10 variations of meteor rates, and figures 3 and 4 show the seasonal variations. The diurnal and 0 seasonal rates as obtained in Kazan and in 10 0 10 MAY 10 JUN 0 10 JUL IO00 - - 1000 0 I 31 SI EC 3E 21 KI 10 AUG n. x 31 HE FIGURE 3.—Monthly variations of meteor rates as measured in Kharkov in 1959 (X=8 meters); ordinate, monthly means of daily meteor rates; abscissa, numbers of the months. 600 DEC 400 16 18 20 22 0 2 4 6 8 10 12 14 FIGURE 2.—Diurnal variations of meteor rates as measured in Kazan in 1959 (X=4 meters); ordinate, hourly rates; 200 abscissa, hours U.T. horizontal plane. With X!=4.2 meters, both the transmitting and the receiving antennas FIGURE 4.—Monthly variations of meteor rates as measured were 7-element directors, raised to the height of in Kazan in 1959 (X=4 meters); ordinate, monthly means 2XX above the ground. With X2=8.7 meters, of daily meteor rates; abscissa, numbers of the months. METEOR SYMPOSIUM—KASHCHEYEV AND KOSTYLTOV 65 Kharkov are in good agreement, so that the brighter than +6m move directly in orbits data can be discussed together. The discus- whose inclination to the ecliptic is small. sion here, however, will be based mostly upon In May and June, the former of the two the data obtained in Kharkov, since most of maxima predominates (10b to llh), while in the echoes were recorded there. The survey August and September, the latter (lh to 2h). of seasonal variations is based upon the meas- As has been shown by Kashcheyev, Lebedinets, urements made in Kharkov in 1959 because the and Suvorov (1960), twice within the year (in sensitivity of the equipment was kept constant May-June and August-October) the earth throughout that year. crosses in opposite directions a wide belt of The minimum monthly means of diurnal meteor orbits which adds to the usual back- meteor rates in 1959 occurred in February and ground of sporadic meteors observed in Decem- March and were 2,834 and 2,695, respectively, ber through April. while the maximum monthly mean, 7,832, occurred in June. Thus there was a threefold References change of monthly means within a year. The HAWKINS, G. S. difference between certain daily rates is still 1956. A radio echo survey of sporadic meteor greater: 2,000 echoes for some dates in March radiants. Monthly Notices Roy. Astron. and 10,000 echoes in June during the Arietid Soc, vol. 116, pp. 92-104. shower. In December 1959, during the Gem- KASUCHETEV, B. L.; LEBEDINETS, V. N.; AND SUVOROV, inid shower, daily rates reached 8,500. Y. I. Hourly meteor rates vary widely from 30 1960. Meteor rates from observations in Kharkov h m h m during 1957-1960. In Meteors, collec- echoes for the period 17 00 to 19 00 in January tion of papers, V. V. Tolstov, ed., vol. 1, through March, and up to 1,000 for the Arietid, pp. 11-20. Publ. House of Kharkov Geminid, and Quadrantid showers. In broad Univ., Kharkov. outline, meteor rates show the same daily KOSTTLTOV, K. V. variations from year to year. During the 1958. The equipment of the Engelhardt Astro- year, meteor rates reach their maximum in nomical Observatory for radio-echo the neighborhood of the moment of the culmi- meteor measurements with automatic nation of the apex (6h local time). Two more recording. Astron. Journ. U.S.S.R., vol. maxima can be observed in the periods about 35, pp. 643-651. KBESAKOVX, M. 10h to 11* and lh to 2h. Hawkins (1956) has 1958. On the presence of faint meteors within shown that these maxima are explained by the the Perseid stream. Bull. Astron. Inst. fact that most meteoroids producing meteors Czechoslovakia, vol. 9, pp. 82-88. 638805—63 The Orbits of Meteor Streams Determined by Radio-Echo Techniques By B. L. Kashcheyev,1 V. N. Lebedinets,1 and M. F. Lagutin1 During the IGY-IGC period, special equip- ing the 5 nights of the period from December ment was developed at the Kharkov Poly- 9 to December 14, 298 Geminid meteors were technical Institute for radio-echo determination recorded. Table 1 shows coordinates of the of orbits of individual meteors. radiants and orbital elements given as mean The observation technique (Kashcheyev, values for each night. Lebedinets, and Lagutin, 1960) was similar to The root-mean-square error of velocity meas- that used at Jodrell Bank by Gill and Davies urements for one meteor is ±1.8 km/sec. (1956). The equipment consisted of a pulse From an analysis it is seen that the following transmitter (X=8 meters) and three receivers. factors contribute to the errors in velocity One of the receivers was situated at the trans- measurements: mitter site; the other two receivers were at 1. Errors made in the evaluation of the distances of 4.5 and 7.1 km, and the directions atmospheric deceleration involved to them from the central station made an (± 0.6 km/sec). approximately right angle. The signals as received by the two receivers were retrans- 2. Metering errors (±0.75 km/sec). mitted to the central station. Three ampli- 3. Errors due to turbulent winds in the tude-time patterns were recorded for every atmosphere (±0.5 km/sec). meteor. From these patterns, meteor velocity 4. Errors due to diffusion expansion of a and time shifts between the corresponding meteor trail during the period when maxima and minima of the diffraction patterns principal Fresnel zones are being were determined. The orbits were computed formed. according to the scheme proposed by Kleiber The root-mean-square error in the deter- (1891). mination of the coordinates of meteor radiants The electronic computer "Ural" was used for is ±2°5. There are two error terms, the first this purpose and it took about 5 minutes to being due to errors in determination of velocity compute each orbit. Meteors brighter than and the second being due to errors in measure- about +8 magnitude were recorded. ments of time delays. The latter, in their Systematic measurements were started in turn, are due to turbulent winds in the atmos- November 1959. In this paper the data given phere. On the average, the two errors terms and discussed are those obtained for meteors are equal. of the Geminid, Arietid, and S-Aquarid showers Throughout the Geminid shower, systematic as well as for 360 sporadic meteors recorded in variations of the elements of the orbit occur November and December 1959. as follows: When observations of the Geminid shower a=110°36'+52' (Xo-2600) of 1959 were planned, their principal purpose g=+32°18'—14' (Xo-260°) was considered to be the checking of the ac- Fa, =36.15+0.30 (Xo-260°) curacy obtainable during observations. Dur- a=1.320+0.036 (Xo-260°) > Kharkov Polytechnical Institute, TJ.8.S.R. e=0.892+0.0045 (Xo-260°) 67 68 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Radiants, orbital elements, and meteor rates for Geminid shower, December 9-14, 1959 Dates Dec. 9-10 Dec. 10-11 Dec. 11-12 Dec. 12-13 Dec 13-14 No. 31 40 90 74 62 meteors Orbital elements a 108T90 ±0?44 110T28 ±0?52 lll?05 ±0?28 111?59 ±0?33 112?50 ±0!3 S +32!57 ±0?41 +33?30 ±0?45 +32T43 ±0?23 + 32 TOO ±0?27 +32?07 ±0?31 v 35.48 ±0.43 35.50 ±0.31 35.83 ±0. 19 36.06 ±0.21 36.66 ±0.20 a m 1.26 ±0.04 1.26 ±0.36 1.29 ±0.02 1.34 ±0.03 1.41 ±0.04 e 0.881 ±0. 005 0.881 ±0.005 0.889 ±0.003 0.891 ±0. 003 0.898 ±0.003 q 0.138 ±0.004 0. 143 ±0.005 0. 137 ±0.003 0. 139 ±0. 003 0. 136 ±0.003 i 22 °A ±1?1 24!2 ±0!9 23 ?3 ±0!5 22 ?3 ±0?6 23!5 ±0?7 a 257?2 258? 1 259 ?2 260?2 261 ?2 326 °.O ±l?0 325?7 ±0?8 326 !8 ±0!4 325T6 ±0?5 325 "A ±0?5 These variations agree with similar data From July 14 to August 14, 1960, orbits ot on 15 orbits determined by Whipple (1954) 153 meteors of the 5-Aquarid stream were from photographic observations of Geminid determined. Table 3 shows coordinates of meteors: the radiants and orbital elements, given as cr=lll°35'+46' (Xo—260°) mean values for the 153 meteors, and compared 8=+32°35'-9' (Xo-260°) with those obtained from five photographic F.=35.95+0.27 (Xo-260°) meteors (Wright, Jacchia, and Whipple, 1954). 0=1.301+0.050 (Xo-260°) The daily motion of the radiant is Aa= + 51'0; e=0.892+0.0033 (XQ—260°) A5=+22'5. The coordinates of the radiant As was shown by Kashcheyev and Lebedinets are given with respect to the longitude of the (1959), towards the end of the Geminid shower sun, XO=126?9. A satisfactory agreement the proportion of large particles increases. of the data obtained by the two methods is An increase in a and e as well as a simultaneous seen, but the scatter of the individual data in increase in the proportion of large particles both cases is rather high. towards the outer periphery of the stream seems In November and December 1959 measure- to be due to the evolution of the meteor stream ments were made to obtain orbits of 360 spo- resulting from the Poynting-Robertson effect radic meteors; of these, 19 orbits were found to or from the effect of deceleration by inter- be hyperbolic, but this may have been due to planetary matter. measurement errors. In the period from June 6 to June 18, 1960, Comparison of our data with those for 347 meteors of the Arietid stream were recorded. photographic meteors (Whipple, 1954) shows Table 2 shows coordinates of the radiants and a rather good agreement in the distribution orbital elements given as mean values for the of orbits according to their eccentricities. period of observations, and compared by For most of the orbits 0.7<«<1.0. For 25 statistical methods with those obtained at percent of the orbits «<0.7; with photographic Jodrell Bank from radio-echo observations in meteors the percentage is lower, namely 10 the period 1950 to 1952 (Lovell, 1954). The percent. The difference is greater when the daily motion of the radiant is Aa=+48'4; distribution of the orbits according: to their A3= + ll!l. The data obtained from the two inclinations is analyzed; that is, there is a large stations show good agreement. group of orbits where 40 ° TABLE 2.—Radiants and orbital elements of Arietid meteors observed from two stations Orbital elements Station a S a e Q i n u Kharkov 44?2 23?3 38.9 1. 98 0.941 0.098 18!5 77 t o Jodrell Bank 44 22?7 38.3 1. 57 0.943 0.093 17?3 77?5 C O O 0 TABLE 3.—Orbital data of 5-Aquarid meteors, July 14-August 14, I960 Orbital elements Source a 8 vm a e q i n Kharkov 341 "A -16?4 41. 1 2.22 0.96 0.08 28?2 306 !9 151?3 Wright, et. al. (1954) 339 ?2 -16?7 43.0 2.60 0.976 0.061 29 ?3 302 °.9 154".7 shows the distribution of the meteor orbits References according to their inclination and eccentricity. DAVIES, J. G., AND GILL, J. C. 1960. Radio echo measurements of the orbits of e faint sporadic meteors. Monthly Noti- ces Roy. Astron. Soc, vol. 121, pp. i.o 437-462. GILL, J. C, AND DAVIES, J. G. 1956. A radio echo method of meteor orbit determination. Monthly Notices Roy. Astron. Soc, vol. 116, pp. 105-113. 0.5 KASHCHETEV, B. L., AND LEBEDINETS, V. N. 1959. On the structure of the Geminid meteor stream. Astron. Journ., U.S.S.R., vol. 36, pp. 629-640. KASHCHETEV, B. L.; LEBEDINETS, V. N.; AND LAGUTIN, M. F. o.o 1960. The orbit of the Geminids of 1959. In 50 100 150 V. V Tolstov, ed., Meteors, Collection of FIGURE 1.—Distribution of meteor orbits according to their papers, vol. 1, pp. 25-37. Publ. House inclinations to the plane of the ecliptic and to their of Kharkov Univ., Kharkov. eccentricities. KLEIBEH, I. A. 1891. Determination of orbits of meteor streams. St. Petersburg. A particular group of orbits (about 13 percent LOVELL, A. C. B. of the total) appears that are not found from 1954. Meteor astronomy. Clarendon Press, Ox- photographic observations. These orbits are ford. characterized both by 40° The Distribution of Small Interplanetary Dust Particles in the Vicinity of Earth By C. W. McCracken1 and W. M. Alexander1 One of the components of the solar system is mation about the influx rate or spatial density a cloud of dust particles surrounding the sun. of dust particles over a limited but astronom- The importance of the cloud arises not only from ically important range of particle mass. More cosmogonal aspects such as the sources and sophisticated experiments capable of defining composition of the dust particles, but also from selected parameters of individual dust particles cosmological aspects such as the rapidity with are possible and technically feasible in the very which changes occur in their dynamical charac- near future. teristics. Properties of the dust particles and Radar, visual, and photographic techniques the general shape and extent of the dust cloud are used in observing meteoric phenomena have been inferred from observations of mete- produced by meteoroidal dust particles. The ors and from photometric studies of the zodiacal dust particles studied by direct measurement light and the solar corona. Uncertainties in techniques are too small to produce meteoric conclusions reached on the basis of such in- phenomena and cannot be observed by the ferences arise because of the differences that pos- ground-based techniques common to studies of sibly exist between the classes of dust particles meteors. The magnetic spherules found in responsible for the two vastly different phen- deep-sea deposits and in high-altitude collec- omena. New and independent methods of tions of atmospheric aerosols may include determining selected parameters of the dust extraterrestrial material, but positive identifi- particles are needed in order to remove the cation of this component is extremely difficult. uncertainties that presently exist in the knowl- Volatile cometary debris does not appear in edge of their physical characteristics and either of the collections, and there has been distributions. little or no success in the identification of A new technique that holds great promise for siliceous particles as being of extraterrestrial supplying some of the most-needed parameters origin. has been developing very rapidly during the One is left with the conclusion that there are past few years. This technique is that of basically two effective techniques of experi- measuring (in space) various parameters of mentally acquiring reliable information about statistical samples of dust particles. The the distributions and physical characteristics of direct measurements that are presently avail- small interplanetary dust particles. The first able have been obtained as dust particles technique is photometry of the zodiacal light impacted onto special sensors mounted on and the solar corona (from outside the atmos- rockets, earth satellites, and other spacecraft. phere of Earth as well as from the surface of The sensors that have been flown have Earth). The second technique is that of making necessarily been of very simple design and have statistical samples within the dust cloud by the provided only limited information about the direct-measurements technique. The direct- dust particles. The experiments have been measurements technique can include the re- designed primarily to give preliminary infor- turning to Earth of collections of dust particles from outside the atmosphere of Earth as well as 1 Ooddard Space Flight Center, National Aeronautics and Space Administration, Oreenbelt, Md. the direct measurements of selected parameters 71 72 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS of dust particles in space. The two techniques of the two interpretations of the presently must remain complementary. Each should available direct measurements is the correct supply information unique to the method as one. The second interpretation is not com- well as provide information to be used in ad- patible with an analysis in which the direct vancing the other technique. measurements from Explorer VIII are used as a Spacecraft carrying instruments for obtaining basis for analyzing other direct measurements. direct measurements can be used for probing The first interpretation should, therefore, be comets and meteor streams simultaneously with regarded as correct until direct measurements studies of the dust cloud. Probes sent through of the vectorial velocities of impacting dust comets and meteor streams would provide particles have been obtained. The direct valuable information about the physical con- measurements from Explorer VIII serve as the ditions and processes within such collections first firm basis for analyzing existing direct of dust particles. The direct-measurements measurements of interplanetary dust particles. technique represents an interesting, useful, and A review of the available direct measurements almost unique means of studying small partic- (including those from Explorer VIII) obtained ulate aggregates of matter in selected regions with microphone systems is sufficient to of the solar system as well as in the vicinity of demonstrate the validity of the foregoing state- Earth. ments about the interpretation of the direct measurements. Interpretations of direct measurements The available direct measurements must be Direct measurements from Explorer VIII periodically reviewed in order to determine how A definitive set of direct measurements recently consistently they fit into one or another pattern was obtained with the satellite Explorer VIII. that has physical meaning. The direct measure- Two independent systems for studying inter- ments have, until just recently, been of such a planetary dust particles were mounted on the nature that they could be given either of two satellite. The systems provided data from interpretations. The choice of interpretation launch on 3 November 1960 until the power depended critically upon the assumptions that supplies were exhausted on 13 December 1960. one was willing to accept during an analysis of One of the systems included a photomultiplier the data. tube as the sensor for measuring the luminous The two interpretations of direct measure- energy generated in hypervelocity impacts of ments that could be given prior to obtaining the dust particles. A microphone was used as the new direct measurements from Explorer VIII sensor in the other system for the purposes of (Satellite 1960 Xi) were: (1) The mass distribu- counting impacts and determining the magni- tion and spatial density of small interplanetary tudes of the impulses delivered by impacting dust particles indicated by the direct measure- dust particles. ments differ significantly from those predicted on Two sounding boards of the microphone the basis of linear extrapolations (over about system were mounted on the lower cone of the six orders of magnitude) of results from obser- spin-stabilized satellite. The spin vector of the vations of meteors; or (2) the direct measure- satellite lay within approximately 40° of the ments indicate the existence of an appreciable apex of Earth's motion during the active life- geocentric concentration of interplanetary dust time of the satellite. The "sounding boards" particles. The conjectural geocentric concen- were arranged to detect dust particles for which tration mentioned here is of the type proposed the radiants generally lay on the hemisphere by Beard (1959), and should not be confused centered on the antapex of Earth's motion. An with concentrations arising from the temporary average particle velocity of 25 km/sec has been suspension of dust particles above the temper- assigned for Explorer VIII until the velocity ature inversion layers in the atmosphere of distribution of interplanetary dust particles Earth. can be determined. The direct measurements obtained with The best calibrations that are at present Explorer VIII strongly indicate that the first possible for the microphone system show that it METEOR SYMPOSIUM—McCRACKEN AND ALEXANDER 73 is sensitive to a quantity which is closely related paper, to avoid the unnecessary propagation of to the momentum of an impacting dust particle. preliminary numerical results that may change Calibrations have been accomplished with low- slightly. The three datum points obtained velocity particles impacting elastically onto with the microphone system on Explorer VIII sounding boards to which the microphones are corrected for the effect of shielding by Earth, were attached. Hypervelocity particle accel- and are plotted as a cumulative mass-distribution erators are now being used in order to obtain curve in figure 1. The three datum points more realistic calibrations. The mass sensi- represent the impacts of more than 3,100 dust tivity of a microphone system can be computed particles registered during an interval of 40 from the momentum sensitivity if an average days. The large number of impacts provides particle velocity is assumed. statistical weight for the datum points from the The microphone system of Explorer VIII two ranges of highest sensitivity. The spacing was designed to count impacting dust particles, of the threshold sensitivities allows one to and to determine in which of three ranges the establish definitively the shape of a segment of momentum of each particle lay. A preliminary an average distribution curve that extends readout of the data has established approxi- over approximately three orders of magnitude mately the number of dust particles encountered in particle mass. for each of the three ranges of sensitivity. In addition to determining the average influx rate Direct measurements from other satellites within each range, the data have defined the and spacecraft shape of a section of an average mass-distribu- Direct measurements have been obtained with tion curve. This section extends over a limited rockets, earth satellites, and space probes flown but significant range of particle mass, and by both the United States and the Soviet Union. provides a basis for analyzing all the other Dust-particle sensors carried by these payloads available direct measurements of interplanetary have included microphones, photomultiplier dust particles. tubes, thin films, erosion grids, wire grids, and A tabulation of the preliminary data from pressurized chambers. The microphone system Explorer VIII has not been included in this is probably the best calibrated of all the sensors, and certainly has been flown more times and on more payloads than any other sensor. It is sufficient, for the analysis summarized in this paper, to consider only those direct meas- urements obtained with microphone systems. The data obtained with microphone systems are entirely adequate to establish an average dis- tribution curve for interplanetary dust particles in the vicinity of Earth. It may be qualita- tively stated that the direct measurements ob- tained with other systems do not indicate that a distribution curve different from the one estab- lished by the direct measurements obtained with microphone systems should be considered. The U.S. satellites and spacecraft for which direct measurements with microphone systems have been obtained are listed, together with the relevant data, in table 1. The data for Ex- plorer I (Satellite 1958 Alpha) and Pioneer I are those given by Dubin (1960). The total PARTICLE MASS IN GRAMS number of impacts (145) for Explorer I was used FIGURE 1.—The average distribution curve established by in computing an average influx rate even though preliminary results from Explorer VIII. approximately one-half the impacts probably 74 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS represented a "shower" component (Dubin, was 40 km/sec. McCracken (1959) used 40 1961). The high influx rate during the ' 'shower" km/sec in an early analysis of direct measure- is approximately counteracted by an interval ments, but this value is almost certainly too with a low influx rate. An influx rate computed high. from the total number of impacts serves very Whipple (1957) assigned different values of well as the average influx rate from Explorer I average velocity to different sizes of dust to be used in establishing an average distribu- particles. The velocities varied from 28 km/sec tion curve. A factor of 2 was used in correcting for meteoroids to 15 km/sec for small dust the influx rate from Explorer I for the shielding particles of the direct-measurements range of effect of Earth. particle size. In order to avoid prematurely The microphone system on Pioneer I regis- constraining the model distribution of inter- tered 25 impacts, of which 17 probably repre- planetary dust particles, we have assigned a sented dust particles. The average influx rate value of 30 km/sec as the average particle was computed on the basis of 17 impacts. No velocity when the spacecraft is not oriented correction was made for the shielding effect of or when the dust particle sensors are open to Earth because Pioneer I spent most of its time an omnidirectional flux. Average particle ve- at large geocentric distances (2 to 19 Earth's locities for oriented sensors are chosen accord- radii). ing to the orientation of the solid viewing Preliminary results for the microphone sys- angle of the sensor. tem on Vanguard III (Satellite 1959 Eta) were The mass sensitivities for the systems on the reported by LaGow and Alexander (1960). Soviet spacecraft are reduced by the square Reading of the telemetered data is still in prog- of the ratio of 40 to 30 in order to compensate ress, so the preliminary results shown in table 1 for the difference in the assigned average may change slightly. Approximately 6,000 particle velocities. The spacecraft—Lunik I impacts were recorded during 80 days of (Mechta, 2 January 1959), Lunik II (Lunar operation. Approximately 2,800 of the impacts Probe, 12 September 1959), and Lunik III occurred during a 70-hour interval on 16-18 (Satellite 1959 Theta)—operated at large geo- November (Alexander, McCracken, and La- centric distances, so corrections for the shielding Gow, 1961). An average influx rate was com- effect of Earth are not necessary. puted for Vanguard III on the basis of ~ 3,500 The influx rates measured by Sputnik III impacts. This number allows for the sporadic (Satellite 1958 52) underwent tremendous background during the 70-hour interval. A changes during the first 3 days of operation factor of 1.5 was used to correct for the shielding of the equipment. The influx rates, as re- effect of Earth. An average particle velocity of ported by Nazarova (1960), were 4 to 11 parti- 30 km/sec was used in computing the threshold cles m~2 sec"1 on 15 May (launch day), 5X10"4 mass sensitivities for the microphone systems particles m~2 sec"1 on 16-17 May, and less on all the spacecraft listed in table 1. than 10~4 particles m~2 sec"1 during the in- Direct measurements reported by Nazarova terval 18-26 May. Nazarova attributes the (1960) from the microphone systems mounted high influx rates during the first few days to a on Soviet satellites and spacecraft are included meteor shower, but the manner in which the in table 2. Portions of the information shown influx rate changed could also be suggestive have been computed from information given by of payload interference. Only the influx rate Nazarova in order that data from United States for the last 9 days of operation of Sputnik III and Soviet spacecraft may be included in the is of any value in establishing an average dis- same analysis. tribution curve. It is not clear whether The sensitivities for the microphone systems Nazarova corrected the influx rate from Sput- on the Soviet spacecraft were expressed by nik III for the shielding effect of Earth, so the Nazarova in terms of particle mass. Nazarova influx rate is left as it was given. computed the mass sensitivities by assuming The method of encoding information into that the microphone systems were energy- the telemetered signal for Lunik I was such sensitive and that the average particle velocity that only very crude upper limits for the TABLE 1.—Direct measurements obtained with microphone systems on the U.S. satellites and space probes Cumulative Influx rate Momentum (particles/m»/sec) Spacecraft sensitivity Mass sensitivity Effective area Exposure time Erposuro Number of (dyne seconds) (grams) (meter') (seconds) (m> sec) particles Observed Corrected for earth shielding Explorer I > 2. 5X10-8 8. 3X10-" 2.3X10-» 7.9X10* 1. 8X104 145 8.4X10-* 1. 7X10-J Pioneer I >1. 5X10-* 5. OX 10-'1 3. 9X10-* 1. 1X105 4.2X10* 25 (17) 4.0X10-8 4.0X10-* Vanguard III >i. oxio-» 3.3X10-' 4. OX 10-' 6. 9X10« 2.8X10" ~3500 1.3X10-8 2.0X10-* TABLE 2.—Direct measurements obtained with microphone systems on the Soviet satellites and space probes Mass sensitivity (grams) Effec- Exposure Num- Influx rate (partlcles/m'/sec) Spacecraft tive time Exposure ber of area (seconds) (m> sec) particles ro-40 km/sec 0O-3O km/sec (meter') (Nazarova) Cumulative 8 Sputnik III 8. 0X10-'—2. 7X10-8 1.4X10-8—4.8X10-8 0.34 ~8X10» ~3X10 ? (see text) TABLE 3.—List of rockets with microphone-type dust-particle sensors Zenith Rocket Launch time and date altitude Type of data (kilo- meters) V-2 Blossom IV-D 8 Dec. 1949 ~135 First direct measurement V-2 Blossom IV-G 31 Aug. 1950 ~135 Qualitatively fair data Aerobee No. 58 14 Sept. 1955 100 Qualitatively fair data Aerobee No. 77 0819 MST 9 Apr. 1957 27 No data, vehicle failure Aerobee No. 80 0630 MST 16 July 1957 122 Quantitatively good data Aerobee No. 81 0730 MST 18 July 1957 No data, vehicle failure Aerobee No. 87 0808 MST 14 Oct. 1957 146 No data, telemetry difficulty Aerobee No. 88 2212 MST 16 Oct. 1957 114 Quantitatively good data Nike-Cajun AF-1 0944 MST 24 Apr. 1958 167 No data, thermal interference Nike-Cajun AF-2 0533 MST 1 May 1958 137 Quantitatively fair data Nike-Cajun AA6.203 2200 CST 15 Oct. 1958 151 Quantitatively good data Nike-Cajun AA6.204 0600 CST 14 Oct. 1958 137 Quantitatively good data Nike-Cajun AA6.205 2125 CST 17 Oct. 1958 143 No data, mechanical failure Nike-Cajun AA6.206 2145 CST 21 Oct. 1958 157 Quantitatively fair data Spaerobee 10.01 1047 CST 22 Oct. 1958 177 Quantitatively fair data Spaerobee 10.02 1327 CST 25 Oct. 1958 420 No data, telemetry failure influx rates could be set. Only that influx rate but of no use in quantitative discussions of measured by the scale of highest sensitivity is interplanetary dust particles. of any value in the present analysis. The spare equipment from the V-2 flights was mounted on Aerobee No. 58 and flown on Direct measurements from rockets 14 September 1955, as has been reported by Signals that could be ascribed to no other Dubin (1956). The rocket did not reach a source than the impacts of dust particles onto sufficiently high altitude for observation of dust the skins of rockets were first found on the particles that had not been appreciably deceler- telemetry records for two V-2 rockets. The ated by atmospheric drag forces. Signals that rockets were V-2 Blossom IV-D flown on 8 probably were caused by the impacts of dust December 1949, and V-2 Blossom IV-G flown particles were recorded, but the data must be on 31 August 1950. Microphones and the as- classified as qualitatively fair. sociated electronics had been designed and in- Bohn and Nadig (1950) also proposed that stalled by personnel of Temple University under rockets carrying special acoustical systems for the direction of Prof. J. L. Bohn. The equip- counting the impacts of dust particles be flown. ment was flown for the purpose of measuring Preferably, the rockets were to have nose tips acoustical intensities in the warheads and on the that could be ejected in order to expose a skins of the rockets. Descriptions of the circular diaphragm. Shielding the diaphragm instrumentation and the data have been given during the passage of the rocket through the in detail by Bohn and Nadig (1950). lower atmosphere served to eliminate the Bohn and Nadig tentatively suggested that possibility that spurious counts were being the unexplained pulses observed on the tele- introduced by the aerodynamic heating of the metry records were caused by the impacts of sensor. The type of rocket program that was meteoric particles on the skins of the rockets. suggested was earned further under the direc- This explanation may now be regarded as the tion of Mr. M. Dubin, then of the Air Force correct one. The data from these flights were Cambridge Research Center. the first direct measurements of interplanetary A list of the rockets that have carried micro- dust particles ever obtained. Both sets of data phone systems is given in table 3. Aerobee should be regarded as being qualitatively fair No. 77 was the fourth rocket to carry a micro- METEOR SYMPOSIUM—McCRACKEN AND ALEXANDER 77 TABLE 4.—Basic corrected data for the successful OSU rockets Range of momentum sensitivity Effective area Time Exposure Number of Rocket (dyne seconds) (meter») A>110km. A > 110 km. impacts (seconds) (m« sec) Aero bee No. 80 6.0X10-«-3.0X10~8 0.05 100 5.0 39 3.0X10-»-8.0X10-» 0.05 5.0 10 1.0X10-»-3.0X10-» 0.5 50 2 3.0X10-*-1.8Xi0-J 0.5 50 1 Aerobee No. 88 1.3X10-*-2.0X10"» 0.05 60 3.0 5 2.0X10-*-2.6X10-* 0.05 3.0 1 4.7X10-*-1.0X10"» 0.5 30 10 1.0X10-»-2.8X10-» 0.5 30 7 Nike-Cajun AF-2 6.0X10-*-1.2X10-' 0.2 154 31 30 I.2X10-*-4.0X10-3 12 4.0X10-»-3.0X10-» 3 Nike-Cajun AA6.203 s.oxio-^-s.oxio-8 0.2 186 37 52 3.0X10-»-1.8X10-a 3 Nike-Cajun AA6.204 7.0X10-*-3.0X10-» 0.2 167 33 31 3.oxio-«-3.oxio-» 1 Nike-Cajun AA6.206 1.5X10"*-1.0X10-» 0.2 118 24 11 1.0X10-»-6.0X10-» 1 7.0X10-*-5.0X10"» 6 Spaerobee 10.01 5.0X10-*-7.5X10-» 0.04 202 8. 1 20 phone system and was the first of a series of 13 measurements by McCracken (1959). A final rockets instrumented with microphone systems readout of the unconnected data from the seven and flown by Oklahoma State University under rockets was made available as a final report on contract with the Air Force. The series of the contract, being dated 14 April 1960 (Buck, rockets was flown especially for the purpose of 1960). directly measuring influx rates and momenta of The data for the seven rockets are given interplanetary dust particles. Aerobee No. 77, in table 4. Corrections have been applied for like the other Aerobees in the series, had two the counts that possibly could have causes crystal microphones mounted on the skin of the other than the impacts of dust particles. The instrument compartment, and two microphones data from Nike-Cajun AF-2 are slightly mounted on a circular diaphragm that could be rearranged in order to compensate for a change exposed by ejecting the nose tip of the rocket at in sensitivity that occurred as a result of a an altitude of about 60 km. The diaphragm low battery potential during the flight. A had been highly polished by Bohn in the hope corrected tabulation of the data from the OSU of observing hypervelocity craters on the rockets has not previously been made available recovered diaphragm. The rocket reached in the open literature because of the series of zenith at an altitude of 27 km, which was too corrections that were in progress and have low for proper operation of the microphone only recently been completed. system. The recovered diaphragm was ex- Only those impacts that occurred when the amined photochemically by Yagoda (Dubin, rockets were at altitudes greater than 110 km 1957), who found several tiny craters resembling are counted. The deceleration of dust particles those formed by hypervelocity impacts of by atmospheric-drag forces becomes quite small dust particles. appreciable at altitudes below approximately Seven of the thirteen rockets provided data 100 km. The velocities of the dust particles that are of use in quantitative discussions of become functions of altitude and zenith angle, interplanetary dust particles. A preliminary making the determination of the velocities of readout of the data from six of these rockets the particles relative to a rocket an almost was included in an early analysis of direct impossible task. Setting the lower limit on 78 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 5.—Direct measurements obtained with the OSU rockets Momentum Particle Mass Number Exposure Cumulative Rocket sensitivity velocity sensitivity of Impacts A>110km. Influx rate (dyne seconds) (km/sec) (grams) (m> sec) (partlcles/mtysec) Aerobee No. 80 6. 0X10-* 70 8. 6X10-" 49 5.0 9.8 3. 0X10-* 70 4. 3X10-10 10 5.0 2.0 1. 0X10-* 40 2. 5X10-10 3 50 6.0X10-2 3. 0X10-* 40 7. 5X10-10 1 50 2. OXlO-i Aerobee No. 88 1. 3X10-4 20 6. 5X10-" 6 3.0 2.0 2.0X10-5 20 1. 0X10"» 1 3.0 3. 3X10"1 4.7X10-4 35 1. 3X10-10 17 30 5.7X10-1 1. OXIO"3 35 2. 9X10-10 7 30 2. 3X10"1 Nike-Cajun AF-2 6. 0X10-* 40 1. 5X10-10 45 31 1.5 1.2X10-' 3. 0X10-10 15 4.8X10-' 4. 0X10-* 1. 0X10-9 3 9. 7X10~» Nike-Cajun AA6.203 3. 0X10-* 35 8. 6X10-" 55 37 1.5 3. OX10-* 8. 6X10-10 3 8. lXlO-2 Nike-Cajun AA6.204 7. OX 10-* 40 1. 8X10-10 32 33 9.7X10-» 3. 0X10-» 7. 5X10-10 1 3.0X10-2 Nike-Cajun AA6.206 1. 5X10"4 35 4 3X10-" 12 24 5. 0X10-1 1. 0X10-" 2.9X10-1° 1 4.2X10-2 7. OX 10-* 2. 0X10-10 6 2. 5X10"1 Spaerobee 10.01 5. OX 10-* 60 8. 3X10"11 20 8. 1 2.5 altitude at 110 km allows one to assume that ticles. The data were of considerable impor- the majority of the dust particles impacting tance in the design of the dust-particle onto the sensors had not been appreciably experiments for the early satellites and space decelerated by atmospheric-drag forces. Geo- probes. The degree of importance may be centric velocities of dust particles above the partially illustrated by the fact that the rocket atmosphere of Earth lie between 11 and 72 data showed influx rates higher by a factor of km/sec, so velocities of the dust particles 104 than was expected on the basis of an (relative to the rocket) lie approximately extrapolation of the tabulation by Watson between the same limits. (1941) of the results from observations of It seems reasonable to assign average particle meteors, and higher by a factor of 102 than was velocities for each sensor of each rocket until expected on the basis of a similar extrapolation the velocity distribution becomes known. The by Whipple (1957). The weight and power average particle velocities have been assigned allotments on the early satellites did not permit on the basis of the region on the celestial one to build in enough dynamic range to allow sphere (relative to the apex of Earth's motion) for such uncertainties in the influx rate. It to which each sensor was exposed. The limit- is interesting and perhaps of moderate quanti- ing mass sensitivity of each sensor is computed tative value if the rocket data agree with and from the momentum sensitivity and the support the more definitive direct measure- assigned average particle velocity. Table 5 ments now available from satellites. The contains the limiting momentum sensitivities, rocket data probably should not, however, be the assigned average particle velocities, the used as the decisive element in a critical test limiting mass sensitivities, and the numbers of of the validity of a particular model of the impacts from which datum points for a cumu- distribution of dust particles in the vicinity lative mass distribution curve were computed. of Earth. The direct measurements made with the rockets served as an early and preliminary Analysis of the direct measurements determination of the influx rate and mass The available direct measurements that have distribution of small interplanetary dust par- been obtained with microphone systems on METEOR SYMPOSIUM—McCRACKEN AND ALEXANDER 79 IV tion curve that can be used until more sophis- ticated experiments have established the curve 1 EXPLORE* Sm (PRELIMINARY) • • ® VANOUJAROID over a considerably larger range of particle 1 <•> EXPLORE* I mass. o V • OSU ROCKETS 0 \ • SPUTNIK m It is quite advantageous to the present M -1 SPACE ROCKET I analysis to point out the uncertainties that ]>* • SPACE ROCKET It K m 0 INTERPLANETARY STATION are encountered in displaying the direct meas- • PIONEER I V urements as a mass distribution. Qualitative MET I discussions of the uncertainties are followed by * Nk CLES / attempts to assign them numerical values. w1 - The values of average particle velocity z assigned in converting from momentum sensi- AT I tivity to mass sensitivity for the systems may x'at °-4 - be slightly too high. The correct values 3 \ depend on the velocity distribution that exists Z \ for dust particles in the direct-measurements Itf* - range of particle size. Lower average veloci- ties would apply if most of the dust particles irf* were in geocentric orbits or direct circular PARTICLE MASS IN GRAMS heliocentric orbits. The use of lower particle velocities would tend to shift the direct meas- FIGURE 2.—The average distribution curve established by direct measurements from microphone systems for inter- urements toward slightly higher values of planetary dust particles in the vicinity of Earth. (Space particle mass without markedly changing the Rocket I, Space Rocket II, and Interplanetary Station are shape of the distribution curve. Even a referred to in the text as Luniks I, II, and III, respec- radical reduction from about 30 km/sec to the tively.) value 15 km/sec used by Whipple would lower the threshold mass sensitivities by only a the OSU rockets, U.S. spacecraft, and Soviet factor of 2. spacecraft are shown plotted as a cumulative The impulse delivered to a sensor as a dust mass-distribution curve in figure 2. The vari- particle impacts at hypervelocity is almost ous direct measurements, with the exception of certainly greater than the impact-momentum the one from Pioneer I, fit remarkably well of the dust particle, but the value remains onto the distribution curve, regardless of the unknown. A value greater than unity (which geocentric distance or sensitivity at which the is used in tables 1, 2, 4, and 5) for the ratio measurement was made. The departure of of the impulse to the impact-momentum in- the direct measurement for Pioneer I from the creases the momemtum sensitivities for micro- average distribution curve can easily be ex- phone systems. The proper correction factors plained on the basis of the actual fluctuations probably lie between 2 and 3, and are now in the influx rates measured with Explorer I, being established by hypervelocity accelerators. Vanguard III, and Explorer VIII. Pioneer I Allowing for the impacts of dust particles was active for about \% days and easily could at angles of incidence appreciably different have flown during an interval of low dust- from 0° effectively decreases the momentum particle activity. sensitivity, but the decrease is probably less The direct measurements Us ted in tables 1, than a factor of 2. This uncertainty can be 2, and 5 encompass a variety of altitudes, largely removed through the use of more sensitivities, lengths of sampling interval, and directional sensors. times of the year. The fact that all except The magnitude of the correction for the the direct measurement from Pioneer I fit the shielding effect of Earth depends on the velocity distribution curve established by Explorer VIII distribution of the dust particles. The correc- lends substantial support to the validity of the tions that have been applied involve the curve shown in figure 2 as an average distribu- assumption that the dust particles are pre- 80 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS dominantly in heliocentric orbits and are not distances (Lunik II, 3.6X10* km; Lunik III, necessarily confined to direct circular orbits. 4.7 X105 km) should fall appreciably below those If this assumption is not valid, then slightly from the earth satellite Explorer VIII. The smaller corrections would be applied. The low altitude (~150 km) rockets should have factors applied to correct for Earth's shielding shown influx rates considerably higher than of the spacecraft were all less than or equal to 2. those obtained with Explorer VIII. With the Smaller correction factors would shift the direct exception of the direct measurement from measurements toward slightly lower values of Pioneer I, the departures of direct measure- influx rate without changing the shape of the ments from the distribution curve established distribution curve. by Explorer VIII are negligible. Knowing which of the minor corrections to The distribution curve of figure 2 is shown, apply requires a better knowledge than exists together with the model distributions of at present of the mass distribution, velocity meteors from Watson (1941) and Whipple distribution, and spatial density of the inter- (1957, 1960), in figure 3. None of the linear planetary dust particles. All of the direct extrapolations of the results from observations measurements have been handled in as con- of meteors fits the direct measurements. The sistent a manner as possible, so that introducing curves labeled "Watson (1941)" and "Whipple minor corrections will change only the position (1957)" approximately represent limits for the of the direct measurements on a plot of mass uncertainty encountered in placing average distribution without changing the shape of the influx rates for meteors on a mass-distribution distribution curve. It seems reasonable to curve. The uncertainty of approximately 200 assume that likely combinations of the minor arises because the mass-to-magnitude relation- uncertainties will leave the distribution curve ships are not well known for meteors, and as it is shown in figure2 . Direct measurements exceeds by one to two orders of magnitude the obtained at large distances from Earth and/or uncertainties that exist for the direct measure- a determination of the velocity distribution ments. The curve labeled "Whipple (I960)" are needed before much further progress can represents a revision |by Whipple of the curve be made in defining a model distribution of the interplanetary dust particles. The distribution curve shown in figure 2 can be approximated reasonably well by a straight- line segment. The equation of the straight line is log /= —17.0 —1.70 log m, where / is the influx rate in particles m~2 sec"1, and m is the particle mass in grams (Mc- Cracken, Alexander, and Dubin, 1961). The equation should not be used outside the range 10"10 gm Abstract Existing direct measurements of small interplanetary dust particles in the vicinity of Earth are analyzed on the basis of a new set of data obtained with the satellite Explorer VIII. All but one of the direct measurements made with microphone systems fit remarkably well onto the distribution curve derived from the Explorer VIII data. This good fit allows one to establish an average distribution curve for small interplanetary dust particles in the vicinity of Earth. The equation of a straight-line segment that approximately fits the new distribution curve for particle masses in the range lO"10 gm Micrometeorite Measurements from Midas II (Satellite 1960 r 1) By R. K. Soberman * and L. Delia Lucca 2 To detect micrometeorites in the vicinity failed. Approximately 4,000 seconds of data of the earth, instruments similar to those used were acquired from the acoustic detectors, on earlier satellites (Cohen, 1960) were carried while the grids provided data for 25,000 seconds on the Midas II (1960 fl) earth satellite. of exposure. Midas II was launched, on May 24, 1960, into Although the amount of data is small com- an approximately 500-kilometer circular orbit pared to that accumulated from other satellite which was inclined 33° with respect to the vehicles, these are the first reported measure- equatorial plane of the earth. ments at this altitude. In the light of recent Five detectors were mounted on the vehicle hypotheses (Whipple, 1961; Hibbs, 1961) re- as shown in figure 1. Three of the detectors garding the altitude variation of micrometeorite were acoustic or microphone type and two were flux, it is felt that these results, although scanty, wire-grid type. The dotted lines in the sketch are worth mentioning. indicate a detector facing outward on the other Sixty-seven impacts were recorded on the side of the vehicle. three acoustic units, and were distributed over The vehicle was designed to maintain a sta- the four partial orbits for which data were bilized attitude, with the end shown on the right acquired, as shown in figure 2. The solid lines in figure 1 pointing radially outward from the indicate the portions of the flight path where a earth. Attitude stabilization was not achieved. "clean" or noise-free signal was received. Since Data obtained from other experiments on the the total exposed area of the three acoustic vehicle indicate that it was tumbling, but the plates was 686 square centimeters, the data data are inadequate to determine the attitude represent a mean flux of 0.25 particles per history of the vehicle. square meter per second. The mean momentum Data from the acoustic detectors were ac- sensitivity of the three acoustic unite as deter- quired only for impacts that occurred while the mined by dropping spherical glass beads on vehicle was being interrogated by a receiving the plates was 3±lXlO~* gm cm/sec. No station. Provision had been made to acquire shielding corrections have been applied. On data over other portions of the orbit, but the the assumption of a mean velocity of 15 km/sec, vehicle recorders did not operate. The wire the sensitivity represents a flux measure- grids are inherently data-storage devices. ment of particles of mass greater than 2 X 10~10 Breaks in the wires that occur during any part gm. Further, if spherical particles of mean of the orbit can be detected by periodically density 3 gm/cms are assumed, the impacting monitoring the continuity of the wires, and the particles would be larger than 5 microns in number of breaks is accumulated. diameter. Data (from stations located in Hawaii, No wire grids were broken. An area of 790 California, and Florida) were obtained from square centimeters of grids was exposed. The parts of four orbits before the telemetry system wire diameter was approximately 20 microns. > Geophysics Research Directorate, Air Force Cambridge Research Previous work (Cohen, Corman, and Dubin, Laboratories, Bedford, Mass., and Northeastern University, Boston, Mass. 1959) has indicated that hypervelocity particles * Geophysics Research Directorate, Air Force Cambridge Research 10 microns in diameter or larger should break Laboratories, Bedford, Mass. 85 86 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS MICROMETEORITE They are presented here for comparison pur- DETECTOR poses only. Satellite 1961 al was in a circular ACOU STIC polar orbit at approximately the same altitude (about 500 kilometers) as Satellite 1960 fl. Similar experiments of approximately the same sensitivities were carried out. Data were obtained from parts of 23 orbits, again only over readout stations. The data from the acoustic units yield a mean influx of 0.34 parti- cles per square meter per second. It should MICROMETEORITE also be pointed out that while the measured OE TECTOR- GRID acoustic flux shows statistical variations over FIGURE 1.—Earth Satellite 1960 f 1 the 23 orbits, it does not appear to decrease with time. Also on Satellite 1961 al eight this size wire. Therefore, the flux of particles breaks occurred in the wire grids. These all larger than 10 microns in diameter is less than occurred after the first interrogation somewhat 5XlO~4 particles per square meter per second. as follows: No break during the first orbit; If one assumes the type of distribution func- three breaks in the next 7 orbits; 1 break in the tion that is commonly arrived at from various second 7 orbits; and finally 4 breaks in the last data considerations such as Whipple's meteoric 8 orbits This leads to a flux of particles detectable by the wire grids (i.e., larger than risk curve (Whipple, 1958), then the data from 2 l the two types of detectors appear to be incon- 10 microns in size) of 8X10~* m~ sec~ . The sistent. That is, the ratio of fluxes from the ratio for the Satellite 1961 al acoustic-grid two types of measurements is greater than 500 measurements is about 400 as compared with where the particle sizes differ by a factor of greater than 500 for Midas II. A complete 2 or so, depending upon the assumptions made report on the Satellite 1961 al measurements regarding velocity and density. will be published shortly. One might thus disregard the acoustic It is felt that the similarity of the measure- measurements with the assumption that the ments from Satellites 1960 fl and 1961 al units were not working properly, or that either cannot be ignored. Despite the apparent in- vehicle noise was being introduced or thermal consistency the values appear to be reliable. noise was occurring within the units themselves. Varying the assumed values for velocity and This is the usual manner of accounting for high density within reasonable limits does not change flux rates observed from the first day's data of the particle diameters appreciably since the practically all satellite vehicles. cube root of these values is involved in the For several reasons it is felt that the acoustic calculation. It may be necessary to re- data should not be discarded. First, the micro- examine the calibration of these devices by phones and plates were acoustically isolated means of hypervelocity techniques now avail- from the vehicle. There was only one instance able to resolve the apparent discrepancy. The where all three acoustic units were triggered possibility of a large variation in the distribu- during the one-second interval between read- tion function can be eliminated. The acoustic outs. This triple coincidence, shown in figure units on board Satellite 1961 al had two 2, was not included in the total count or in the momentum sensitivities. The lower sensitivity flux calculations. Second, although the micro- range (with the previously assumed values for phone and grid measurements appear to be velocity and density) should have a threshold inconsistent with each other, more recent infor- of 10 microns. The recorded fluxes are not mation is now available that shows that the consistent with the wire-grid data. values obtained are reproducible. A prelimi- The remaining possibility that can be sug- nary reduction of the data from the Satellite gested is that fluffyparticle s of very low density, 1961 al has been completed. These results that might not be able to break the wire grids, are not yet in final form and may vary slightly. are the major constituent of the acoustic flux. METEOR SYMPOSIUM—SOBERMAN AND DELLA LUCCA 87 35 ORBITS FOR SATELLITE I960 2O'N IOM 1 5 Jo'e 'so* TTC** iTo * MTOTW iiWWw no * /©5*w LONGITUDE FIGURE 2.—Occurrence of micrometeorite impacts. This possibility does not agree with empirical o.o formulas of hypervelocity penetration, but cannot be entirely discounted. It is not presently possible to check this hypothesis in To - 1.0 the laboratory since techniques for accelerating fluffy particles do not yet exist. — - 2.0 Finally, the relationship between the acoustic I960 data reported here and the altitude variation of micrometeorite flux should be pointed out. - 3.0 Figure 3 shows a curve similar to that given by Whipple (1961). The acoustic flux as measured from Satellite 1960 fl (corrected to a value of - 4.0 10~9 gm by assuming a mass distribution func- L2 ZOD tion which is proportional to m~ and a velocity -5.0 CLD of 15 km/sec) has been added. The corrections are the same as those applied to the other satellite points in Whipple's paper. The early 2.0 Z.0 4.0 5.0 acoustic rocket data, about which there is some LOG ALTITUDE (km) dispute, have been deleted and a new curve FIGURE 3.—Meteoritic impact (mass = 10~* gm) versus alti- with a slope of —1.1 has been fitted to the tude from the earth. (Revised from Whipple, 1961.) 88 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS points. It should be pointed out that with the References exception of the zodiacal cloud, all the points COHEN, H. A. were obtained with satellite acoustic detectors. 1960. Measurements of flux of small extrater- In conclusion, the data reported here estab- restrial particles. Geophys. Res. Dir., lish that a discrepancy exists between the acous- Research Notes, no. 28. tic and the wire-grid measurements which must COHEN, H. A.; CORMAN, A.; AND DTJBIN, M. 1959. Calibration of micrometeoritic detectors be resolved before the dust content of the used in satellites and rockets. In interplanetary medium is understood. How- Proceedings 3d Symposium on Hyper- ever, if one considers the acoustic measurements velocity Impact, vol. 1, pp. 405-421. in a purely relative manner, it appears that Armour Research Foundation, Chicago. there is a geocentric variation in the micro- HIBBS, A. R. meteorite flux. 1961. The distribution of micro meteorites near the earth. Journ. Geophys. Res., vol. 66, pp. 371-377. Acknowledgment WHIPPLB, F. L. The authors are indebted to Dr. L. Bohn of 1958. The meteoritic risk to space vehicles. In Temple University for his cooperation and Proceedings 8th International Astronau- tical Congress, Barcelona, 1957, pp. 418- assistance in the preparation of this experiment, 428. Springer-Verlag, Vienna. and to Mr. H. A. Cohen, who began this experi- 1961. The dust cloud about the earth. Nature, ment. vol. 189, pp. 127-128. Abstract Five micrometeorite sensing devices were carried on board the Midas II (Satellite 1960 f 1) which was launched into an approximately 500-kilometer circular equatorial orbit on May 24, 1960. The five devices consisted of three acoustic detectors and two wire-grid detectors. Data were obtained for four partial orbits. These data were recorded in real time. There were no breaks in the wire grids which could record impacts of particles larger than 10 microns. Sixty-seven impacts were recorded on the acoustic detectors which had a momentum threshold of 3 X 10~4 gm cm/sec and a total area of 0.0686 m3. These results indicate a flux of 0.25 m~2 sec"1 of particles 5 microns in diameter or larger if one assumes a mean velocity of 15 km/sec and a density of 3 gm/cmJ for the particles. The discrepancy between acoustic and wire-grid data will be discussed. These results will be compared with those obtained from other satellite measurements. SOBERMAN, HEMENWAY, et al. PLATE 1 Venus Flytrap nose cone. Left to right: fully closed, fully open, and partly open. SOBERMAN, HEMENWAY, et al. Typical micrometeorite collecting box. Micrometeorite Collection from a Recoverable Sounding Rocket. I By R. K. Soberman,1 C. L. Hemenway,2 T. G. Ryan,3 S. A. Chrest,4 J. Frissora,4 and E. F. Fullam6 On June 6, 1961, at 5:31 a.m., M.S.T., an two layers were Mylar films 6 microns thick. Aerobee 150 rocket, descriptively named the These films were examined prior to the flight "Venus Flytrap," was successfully fired from by means of a helium leak detector to insure the White Sands Proving Ground in New that there were no holes. In addition to flux, Mexico. Photographs of the nose cone or penetration, and cratering data, it was hoped "payload" section of that rocket are shown that this experiment might yield some zenith in plate 1. The photograph at the left angle information if more than one layer were shows the nose cone in the fully closed position penetrated by the impinging particles. in which it was fired and traveled up through A second type of experiment carried on the the lower atmosphere. An electric motor then vehicle were surfaces which were intended to raised the outer skin or ogive assembly (photo- trap particles of lesser momentum (that is, graph at right), and finally another motor, aided submicron particles if such existed, or some- by the spin of two revolutions per second, what larger particles which had been suffi- which was applied to the rocket, extended the ciently slowed down in the atmosphere). eight "leaves" outward to complete the opening These surfaces were mounted in aluminum cycle (middle photograph). The nose cone boxes for cleanliness. Eight such boxes, each was designed to fly in this last position through having an open area of 160 square centimeters, the upper atmosphere, exposing collectors were mounted on four of the leaves of the nose mounted on the tops of the leaves and on the cone. The lids of the boxes, on which more internal structure. Prior to reentry the rocket collection surfaces were mounted, were fastened was "de-spun," thus allowing the nose cone to to the internal structure of the nose cone. be resealed, and the booster section was severed The boxes and their covers were prepared with explosively from the payload. At an altitude high-purity materials for this experiment. All of 6 kilometers a parachute was deployed and boxes were treated and loaded in an identical the nose cone descended to the ground where fashion and were completely interchangeable. recovery was effected. Since the fabrication left the inside of the Four of the leaves carried a penetration boxes extremely rough, it was found to be experiment (sketched in fig. 1) which covered impractical to try to clean the internal surfaces. a total area of 0.24 square meter. The first In order to prevent the collection surfaces 1 Geophysics Research Directorate, Air Force Cambridge Research from being contaminated with debris from the Laboratories, Bedford, Mass., and Northeastern University, Boston, box, the debris was sealed to the surface with Mass. 1 Dudley Observatory, Albany, N.Y., and Union College, Schenec- a plastic coating, a high-purity nitrocellulose. tady, N.Y. All the covers, clamps, screws, and washers * Geophysics Research Directorate, Air Force Cambridge Research Laboratories, Bedford, Mass. were cleaned in benzene to remove grease and * Northeastern University, Boston, Mass. other contaminants. The open area inside the * Ernest F. Fullam, Inc., Schenectady, N.Y., and Dudley Observa- tory, Albany, N.Y. gasket on the cover was coated with a nitro- 88 638805—63 8 90 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS serve as controls for contamination. In ad- 60.0 will i m I • rt dition, as can be seen from figure 2 and plate 2, there were controls mounted in each box. 2 .4 •! Hint I .3 Mi Mix These controls consisted of an area approxi- mately one-fourth of the surface area which was 3 mi Ili mat ir protected by an aluminum shield placed a half pltiigl att millimeter above the surface so that air could flow freely under the shield. These shielded FIGURE 1.—Penetration experiment. surfaces were exposed in the same environments and subjected to the same handling procedures, cellulose film 0.003 inch thick. This film was but could not be impinged upon from above. held in place with 10-mesh stainless-steel screening. Eight different surfaces were loaded in the boxes. There were seven lucite slides, each % by 2 by #6 inch, covered with various sub- strates. The eighth surface was a large piece of bare lucite, 1% by 5 by KB inch. A schematic drawing of the positioning for the various substrates used is shown in figure 2. All the thin films, each approximately 200 A thick, were supported with 200-mesh copper screening. The coated screening and the self- supporting thick nitrocellulose films (0.003 inch thick) were glued to the lucite slides which were held in place by clamps. Since there was FIGURE 2.—Location of substrates for collection experiment: no past experience to go by, a number of sub- (A) lucite slide shadow-cast with aluminum, (B) silicon monoxide on thin formvar, (C) thick nitrocellulose film, strate materials and handling techniques were (D) thin formvar, (E) thick nitrocellulose coated with tried in an attempt to find the most effective and carbon, (F) thin formvar shadow-cast with aluminum, (G) unambiguous collection system. A photograph thin nitrocellulose shadow-cast with aluminum, (H) bare of a typical box and cover is shown in plate 2. lucite. As can be seen in this photograph, the boxes and their lids were protected by sealed plastic covers. The boxes were opened and closed during the Extreme care was necessary in handling these flight by the action of the leaves. Upon descent boxes to insure that contamination would be of the nose cone the boxes were sealed by at- kept to a minimum. Therefore the nose cone mospheric pressure. Filtered ports at the base was thoroughly cleaned prior to the installation of the nose cone allowed pressure equalization of all the experiments. After installation of the inside the skin of the nose cone. The boxes how- experiments, the nose-cone operation was ever, were designed to maintain a vacuum inside. checked by cycling it twice in a precleaned These extreme precautions were not taken room. During these operations the protective with the penetration experiments, but care was covers were still on the boxes and lids. Finally, exercised that the film surfaces were protected a large sealed plastic tent was erected over the from any direct contact. nose cone. The air in the tent was allowed to The "Venus Flytrap" system performed as stagnate for some 2 hours, the nose cone was planned. The telemetry record which was cycled for the final ground test, and during this used only to monitor the rocket functions in- test the protective covers were removed from dicated that everything operated perfectly and the boxes and lids. The boxes and lids were at approximately the preset times. From this exposed for less than 30 seconds in the tent telemetry record the angles at which the leaves before being sealed in the nose cone. An iden- were exposed were determined. The four tical box was exposed in the handling room to leaves containing the penetration experiments METEOR SYMPOSIUM—SOBERMAN AND HEMENWAY 91 VENUS FLYTRAP 200 JUNE 6, I 9 6 I 15 0 100 50 50 100 150 200 250 300 350 400 450 500 TIME (ttcoods) FIGURE 3.—Vehicle altitude plotted against time. were exposed at an angle of 73.5° with respect fuel and oxidizer tanks were valved shut in to the vertical axis of the rocket. The four accordance with standard rocket procedure. leaves on which the boxes were mounted were The opening of the nose cone was accomplished exposed at an angle of 66.5° with respect to this in about 1 second at an altitude of 88 kilo- same axis. meters. The rocket then ascended for 131 An aspect camera was installed to record the seconds to an apogee of 168 kilometers, and attitude of the rocket. These data have not then descended for 105 seconds to an altitude yet been fully reduced. However, it appears of 116 kilometers where closing was accom- that the rocket axis precessed from about 10° plished again in about 1 second. During this on one side of the vertical to about 45° on the entire exposure time the rocket remained in a other side of the vertical during the portion near vertical attitude (i.e., it precessed, but did of the flight where the nose cone was open. not turn over). The spin of the rocket was reduced from 2 In figure 4 the velocity of the rocket is r.p.s. to 1.4 r.p.s. by the opening of the leaves. plotted against time and the curve shows that A device known as a "yo-yo" threw off two the rocket approximated a free-fall trajectory weights and reduced the spin to essentially during the time that the nose cone was open. zero just before the nose cone was closed. The nose cone was found sealed and intact. This was necessary, since the motor did not It lay in the desert for about 1 hour after landing have sufficient power to close the leaves while before being placed in a protective plastic bag. the rocket was spinning. Samples of the desert sand were taken for con- In figure 3 the altitude of the rocket is plotted trol purposes. On return to the handling room against time, and the curve shows that rocket the outer skin was washed down with acetone burnout occurred at an altitude of 35 kilometers. and the nose cone was wrapped in a clean plastic It should be pointed out that at burnout the bag for shipment. One week later in a "dust 92 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 2 .0 VENUS FLYTRAP JUNE 6, 196 I I .5 1.0 .5 50 100 150 200 250 300 350 400 450 500 Tl ME (seconds) FIGURE 4.—Vehicle velocity plotted against time. free" environment the nose cone was opened Gustofson of the Geophysics Research Direc- for examination. Five of the eight boxes were torate at the Air Force Cambridge Research found to be still under sufficient vacuum to Laboratories. We also express our thanks to require opening by injection of clean air with the staff at Aerojet-General Corp., particularly a hypodermic syringe. to Messrs. N. Migdal, J. Smith, and J. McCabe This experiment has been described in great who fabricated the nose cone and assisted in detail because we believe that these details have the preparations for the experiment. Finally an important bearing upon the validity of the we thank Capt. T. Smith (USAF), Lt. J. results which are presented in the next paper. Turpel (USN), and the officers and men of the U.S. Navy Missile Testing Center at White Acknowledgments Sands, N. Mex., for their wonderful cooperation We acknowledge the debt that this project and assistance. But for the help of these and owes to Mr. H. A. Cohen, who originally began many other support groups, the experiment the program. Further, we acknowledge the described above would still be "on the drawing assistance of Messrs. R. Skrivanek and P. boards." Abstract The "Venus Flytrap" micrometeorite collector rocket was fired from White Sands, N. Mex., on June 6,1961, at 05:31 a.m., local time. Specially prepared particle impactors were exposed between the altitudes of 88 and 168 kilometers and successfully recovered. The rocket design and performance are discussed. The experimental surfaces consisted of 0.24 square meter of 6-micron-thick Mylar foil for impact and cratering studies, and 0.13 square meter of sealed boxes which were loaded with high-purity materials and surfaces suitable for electron microscopy. In all, eight different materials were used to permit an evaluation of the effectiveness of different materials and techniques for the collection and study of micrometeoritic particles. Some of these materials were shadowed with aluminum before and after the flight to aid in the discrimination of micrometeorite particles from contaminants. Micrometeorite Collection from a Recoverable Sounding Rocket. II By C. L. Hemenway,1 R. K. Soberman,2 E. F. Fullam,3 J. J. Balsamo,4 J. Cole,1 D. Hallgren,5 P. Yedinak,6 A. Goodman,6 and G. Hoff1 As described in the preceding paper, two studies are presently being carried out to verify types of experiments were carried out on the these inferences and to determine the strength "Venus Flytrap" rocket (i.e., the penetration of these films under low-velocity impact and the trapping experiments). In this paper conditions. Two holes found in an equal the results obtained to date from these experi- number of control films appear to have been ments are presented. caused by handling, and careful examination The mylar films from the penetration ex- shows no similarity with the holes that were periment were tested by a helium leak detector found on the flight samples. after return to the laboratory. Three holes, A careful microscopic examination of the shown in plate 1, were found in top-layer films. top-layer mylar films was made in the hope of There was no evidence of damage on the second- finding imbedded particles or craters. It was layer films beneath these holes. The holes immediately noticed that the upper layer of shown in plate la,b were found on the same films had more particles on them than the mylar film. The first hole (pi. la) measures second layer or the controls. But as was 90 microns across the short axis and 330 mentioned in the preceding paper, extreme microns across the long axis. The second hole cleanliness had not been maintained with these (pi. 16) measures 450 microns across the short films and therefore this information could not axis and 1160 microns across the long axis. be used. This scanning did not result in any The third hole (pi. lc) found on another film further unambiguous information, since the which was mounted on the same leaf measures experiment was designed to look for punctures 330 microns across the short axis and 550 and no records were kept of film defects. microns across the long axis. Note the flap There have been many attempts in the past structure and curling of the mylar in all three to collect micrometeorite particles at ground photographs. This flap structure and curling level, on mountain tops, by balloons, and by are believed to be indicative of low-velocity high-altitude aircraft. However, all these tech- impact. From the appearance of the holes niques have been open to criticism because of it is inferred that they were caused by particles contamination by terrestrial particles. Since comparable in dimension to the hole size this is a very serious problem, we next discuss which impacted at relative velocities of less the various safeguards which were built into than about 2 kilometers/second. Laboratory the "Venus Flytrap" experiment to ensure re- liable identification of micrometeorites. There * Dudley Observatory, Albany, N.Y., and Union College, Schenec- tady, N.Y. were five controls for this portion of the experi- 1 Geophysics Kesearch Directorate, Air Force Cambridge Research ment. Some have been described in the pre- Laboratories, Bedford, Mass., and Northeastern University, Boston, Mass. ceding paper. All five controls are listed »Ernest F. Fullam, Inc, Schenectady, N.Y, and Dudley Observatory, below with brief comments: Albany, N.Y. * Control Equipment Corp., Needham, Mass. 1. Laboratory exposures.—These were spare ' Ernest F. Fullam, Inc, Schenectady, N.Y. * Dudley Observatory, Albany, N.Y. boxes prepared at the same time and in the 93 94 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS same way as the flight boxes. One of these preceding paper shows the metal shield in the spares was exposed in the laboratory for an middle of an open box. extended period to study contamination parti- This study utilized three independent teams cles from the laboratory. Micrometeorite par- of electron microscopists using three electron ticles as described below were not found in the microscopes. It may be relevant here to point laboratory controls. out some of the characteristics of electron 2. Rocket preparation exposure.—A spare box microscopes. The magnification of an electron was exposed at White Sands during final microscope is generally very high in comparison rocket testing. As described in the preceding with optical techniques and one can scan only a paper, it was necessary to remove the mylar few square millimeters per day. With an elec- covers from the boxes at the time the nose cone tron microscope, the limit of resolution is of the was sealed. The boxes had to be fabricated order of 10 to 200 A, which is a factor of 100 or in such a way that the nose cone could be more better than the resolution possible with opened and closed and yet not contaminate the optical-microscope techniques. With an elec- samples while the final checkouts were being tron microscope, one can distinguish surface made on the nose cone. In the room where the details and textures of particles, can measure nose cone was being cycled for the final time, particle sizes accurately, and can see how trans- a spare box was opened for the full time during parent the particles are to electrons. Finally, which the mylar covers were off prior to the while optical microscopes can scan large areas, final sealing. This operation was carried out it is generally extremely difficult to distinguish in a plastic tent as described in the preceding one small particle from another. When study- paper. ing micrometeorites one is always working with 3. Desert dust examination.—A sample of a "noise" or "background" level, and con- taminant particles will always be present. It desert dust was examined to see what kind of will be shown that if the textures of the particle particles could be expected from the air at surfaces, the shapes and sizes of the particles, White Sands. To minimize desert dust con- and so on, can be studied, contaminants can be tamination, the outside of the nose cone was recognized and readily ruled out of consideration. thoroughly cleaned in the laboratory and kept Perhaps the most effective technique utilized in a plastic bag until one hour before flight and in this experiment for micrometeorite identifi- removed only when the rocket was pressurized. cation is the shadowing technique wherein, just No evidence of desert dust contamination was before the boxes were loaded, certain of the detected. slides had a coating of aluminum evaporated on 4. Rocket particle examination.—In a clean the collection surface at an angle of approxi- environment wipings were made of the inside mately 80° from the normal to the surface. of the nose cone and were mounted on electron- Thus any particles that were on the collection microscope screens for examination. Some surface before the boxes were loaded would cast metal chips and some oil were found. None of a shadow in the direction in which the aluminum the types of particles considered to be micro- was deposited. After the flight, another shad- meteorites were found in these wipings. owing at 75° was put on at roughly right 5. Shielded flight samples.—The best controls angles to the previous shadowing. If a particle were the shielded areas inside each box. Ap- was deposited between the two shadowings it proximately one quarter of the area of each box should have only the second shadow, and if it was blocked off in such a way that air flow was was deposited during the scanning process it allowed beneath the shield, by separating the should have no shadows. Contaminant parti- shield approximately a half millimeter from the cles within the films will also be shadowless. control surfaces. Thus, if small particles were This shadowing or flagging technique is a useful drifting around in the air they could wander way of reducing the ever present contamination beneath the shield, yet any particles that would level. The flagging techniques had disadvan- hit at high altitude would be prevented from tages. For example, if one attempts electron- bitting these control surfaces. Plate 2 of the diffraction analyses then one may be bothered METEOR SYMPOSIUM—HEMENWAY, SOBERMAN, AND OTHERS 95 by the presence of the aluminum, although it niques to eliminate this type of contaminant may be convenient for some purposes to have from electron microscope specimens. These a calibration standard on the same sample. particles can be readily identified; they are After a preliminary examination of the flat, of low density, and have sharp edges. various collecting surfaces it was decided, for Plate 3a shows a rounded, medium-density, the initial study of the particles, to concentrate submicron particle of the type initially diffi- on the surfaces on which the flagging technique cult to identify as a contaminant. Notice the was employed. The observer scans first with fuzzy edge of this particle. These types of a magnification of the order of 2000 to 5000 submicron particles were found on both the times. He looks for single-flagged particles. controls and the flight samples. One reason However, a single flag is considered to be a for using high magnification to aid in the iden- necessary but not a sufficient requirement, tification of micrometeorites was to discover since it was possible for contaminant particles whether such particles had sharp or fuzzy edges. to be deposited between the two flaggings. Plate 36 shows a small spherical micromete- The single-flagged particle is examined to deter- orite with a sharp edge and high density. mine whether it can be recognized as a con- Notice that it can be readily distinguished from taminant at low magnification. If not, then the fuzzy-edged spherical particle in plate 4a. the magnification is increased to between 15 Plate 4a shows a large irregular type of par- and 20 thousand times, and the particle ticle which looks almost like a museum mete- examined very carefully. In order to be count- orite. Notice that the surface is characterized ed as a micrometeorite the particle has to be by rounded irregularities and is partially trans- of a type not found on any of the controls. parent to electrons (medium density). This There are three types of particles that did particle was counted as a large irregular micro- not appear on the controls. These were dense meteorite; in fact, it is one of the largest spheres with sharp edges ranging in size down single-flagged particles found to date. to a few tenths of a micron or even less, irreg- Plate 46 shows an extremely irregular par- ular submicron particles, and extremely irreg- ticle which is termed a "fluffy" particle. All ular particles which were termed "fluffy." In particles not having a compact structure were addition, some smashed fluffy particles were labelled "fluffy." Like the others obtained in found. Also a number of holes were found in the Venus Flytrap experiment, this type of some of the thin collecting surfaces. The re- particle is found in U-2 aircraft collections. sults described here were obtained with surfaces Plate 4c shows a fluffy particle obtained from of nitrocellulose and formvar (approximately a U-2 collection. 200 A thick). Plate 5a shows a "sphere" with a hole. No- Plate Id shows an optical micrograph of a tice that the "sphere" is not quite spherical. portion of a 200-mesh per inch copper screen It is approximately 2.25 microns in diameter which has an open area roughly 100 microns and is the largest of the spheres found thus along an edge. Plates 2-6 are electron micro- far. The hole is about 2.4 microns in diame- graphs of a small portion of these open areas ter and apparently the sphere bumped into in the mesh screen. this 200 A-thick nitrocellulose film, did not Plate 2a shows the flagging technique de- get through, and made a flapped hole about scribed above. The particle on the left has the same size and shape as its own body. A two shadows and is therefore clearly a contam- number of flapped holes of this type have inant; the particle on the right has one shadow been found. and was identified as an irregular type sub- Plate 56 shows a flapped hole where a smaller micron micrometeorite. spherical particle broke through the thin nitro- Plate 26 shows mica-like contaminants. This cellulose completely. Apparently the flapped- type is found on all electron microscope slide hole phenomenon appears over a wide range materials and apparently is in the air we of dimensions when one deals with low-velocity breathe. There have been no successful tech- collisions. 96 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS dN d8 0 .25 .5 75 .25 .5 .75 1.0 DIAMETER (MICRONS) DIAMETER (MICRONS) FIGURE 1.—Size distribution of irregular micrometeorites, FIGURE 2.—Size distribution of "fluffy" micrometeorites, 0.1 0.1 to 1 micron. to 1 micron. Plate 5c shows an elliptical hole where size range is still increasing at that point. the particle apparently penetrated the thin There are also some larger spherical particles film at an angle to the surface. Notice the which are not shown on this distribution. flaps as well as the cracks radiating from the Figure 4 shows the distribution of spherical ends of the hole. holes. Notice that there appears to be a Plate 6a shows a portion of a smashed par- cutoff at about 0.3 micron. The fact that the ticle, probably a fluffy type. The debris distribution extends down to sizes as small extended over a length of more then 100 as this, and that spherical particles much microns. Several of these smashed particles larger than the minimum size hole are observed, have been observed. Such particles are also indicates that there must have been a diversity found in impact-type collections made with in the velocities of particles hitting the surfaces. U-2 aircraft. An attempt will be made to To date, 133 particles which are believed measure the breaking strength of fluffyparticle s to be of extraterrestrial origin have been found to determine whether the values obtained in the range of 0.1 to 1 micron in diameter. correspond to those required to crumble They were found in a total area of 17.84 cometary meteoroids. square millimeters. Eleven particles larger Plate 66 shows a group of particles. Since than 1 micron have been found in an area it was uncertain whether such particle groups of 18.52 square millimeters. This latter area represented debris from the breakup of a is a little larger because the less serious lab- larger particle, this type of particle group oratory contamination problem associated with was not considered in the determination of larger particles permitted faster scanning. size distributions. The total number of particles larger than Figure 1 shows a distribution function for 0.1 micron per unit area was 7.3 per square the irregular type of particles. In the figure millimeter. If we separate these particles only particles in the region of 0.1 to 1 micron according to type, 72 percent are irregular have been plotted. A few larger and smaller with sharp edges and rounded texture, 16 individual particles were found. percent are high-density spheres, and 12 Figure 2 shows a distribution of the "dia- percent are of the fluffy type. Of the single- meters" of the fluffy-type particles. It is flagged particles examined, two out of every rather difficult to determine mean diameters. three were judged to be micrometeorites, For this determination the diameters in two although this number varied somewhat from perpendicular directions were measured and sample to sample. In a few cases there was the particle was assumed to be ellipsoidal. some difficulty in deciding whether or not In figure 3 the size distribution of the a particle was a micrometeorite. Each oper- spherical-type particles has been plotted. ator was asked how frequently he experienced Notice that the size extends down to one-tenth this difficulty, and the largest number of of a micron and apparently the number-per- doubtful cases was found to be 10 percent. HEMENWAY, SOBERMAN, et al. PLATE 1 a-c, Mylar films showing holes caused by meteorites, d, 200-mesh-per-inch copper screen used to support collecting surfaces. VAV HEMENWAY, SOBERMAN, et al. PLATE 2 a, Illustration of flagging technique used to identify micrometeorites. b, Mica-like contaminant particles. Scales, 1 micron. SOBERMAN, HEMENWAY, etal. PLATE 3 a •.-. • »- .».'• -,•• a, Submicron contaminant particle, b, Submicron spherical micrometeorite. Scales, 1 micron. HEMENWAY, SOBERMAN. et al. PLATE 4 <3, Irregular micrometeorite. b, "Fluffy" micro- meteorite. Scales of a and b, 1 micron. c,"Fluffy" particle obtained from U-2 collection. HEMENWAY, SOBERMAN, et al. PLATE 5 a, Spherical micrometeorite and hole, b, Circular \ hole created in 200 A-thick film by micro- meteorite, c, Elliptical hole created in 200 A-thick film by micrometeorite. Scales, 1 micron. HEMENWAY, SOBERMAN, et al. PLATE 6 a V . » V a, Remnants of a smashed micrometeorite. b, Cluster of irregular micrometeorites. Scales, 1 micron. METEOR SYMPOSIUM—HEMENWAY, SOBERMAN, AND OTHERS 97 dN d8 .5 1.0 15 2 0 DIAMETER (MICRONS) FIGURE 4.—Spherical hole size distribution from 200 A-thick films. O .25 .50 .75 chosen and lend confidence to the results. DIAMETER (MICRONS) Second, the particles were collected at altitudes FIGURE 3.—Size distribution of spherical micrometeorites, where terrestrial particles are very unlikely to 0.1 to 1 micron. reside for any length of time. The third fact is the distribution of holes found. By dropping All doubtful cases were automatically discarded. particles through a vacuum tube 2 meters long, Actually, a few particles which really came it was determined that particles of the order of from beyond the earth's atmosphere were 7 to 10 microns diameter will just break through probably discarded because of this conserva- 200 A-thick nitrocellulose collection film with tive policy. a velocity of 6 meters/second. The analysis of these particles has now Figure 4 shows holes 0.3 micron in diameter begun. There are three techniques which offer in the flight-collection films. If the strength of hope for determining the composition of these the films is energy sensitive, this would suggest particles: electron diffraction, neutron acti- relative velocities of the order of 1 or 2 kilo- vation, and electron-beam probe techniques. meters/second for a particle comparable in size Thus far no positive results have been to the smallest hole found. Finally, it should obtained with electron diffraction. We are be pointed out that the observers have had a beginning to suspect that micrometeorite considerable amount of experience looking at particles do not have a crystal structure and the high-altitude particles and at the contami- that any original crystal structure has been nants which appear in electron microscopy. In scrambled by cosmic-ray bombardment. We fact, four of the observers among them have may find that the lack of crystal structure is an had a total of 40 years of experience in electron appropriate criterion for distinguishing mi- microscopy. This experience increases our crometeorites. More work is needed to estab- confidence that the particles described and lish this point. pictured in this paper are indeed nonterrestrial A second type of analysis involves neutron in origin. activation. Gamma-ray spectra studies of ac- Although there is much more to be done, tivated samples are now in progress. The particularly on particle analysis, and these third technique which will be employed is the results are only preliminary, we feel that further electron-beam probe wherein the particles are study is not likely to change the present numer- used as X-ray targets. The wavelengths of ical results significantly. the characteristic X-radiation are used to identify the composition of the particle. Acknowledgments In summary, we believe that extraterrestrial The authors thank Miss L. Delia Lucca, particles have been collected. This belief is Mr. P. Manning, and Mr. R. Oppenheim for based on several facts. First, the experiment their assistance and cooperation in obtaining had a number of controls which were carefully some of the results described in this paper. 638805—63 9 98 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Discussion ant particles, on the other hand, we have always E. L. Fireman.—Are these particles different been able to see some diffraction pattern. from the ones you reported as giving electron A. F. Cook.—The low velocities with which diffraction patterns for nickel and nickel oxide? you feel most of the particles impacted could C. L. Hemenway.—Yes. In that paper, we be comparable to the rocket velocity at the were reporting on extremely small particles altitudes where the collector opened. found in large profusion on a U-2 flight on R. K. Soberman.—That would be the right November 15, 1960. There was a sufficient order of magnitude. concentration of these particles in a cluster to give identifying diffraction patterns. I am Unidentified speaker.—The large sphere now saying that in this experiment and also in which you found would have had to strike the the other U-2 flights we have been unable to film with a very low velocity. obtain a diffraction pattern from any of the C L. Hemenway.—That would have to be larger individual particles of suspected extra- the case. There were a number of similar terrestrial origin. With terrestrial contamin- events. Abstract Three independent electron microscopes and observer teams have been used to evaluate, count, electron micro- graph, and measure the micrometeorite particles. Typical particles and size distributions are shown. Particles have been found in structure and sizes similar to those obtained at lower altitudes with U-2 aircraft and balloon collection techniques. Approximately seven particles per square millimeter were collected during the flight. Most of the particles were submicron in size. The micrometeorite particles collected generally fell into three types: high- density spheres; medium-density irregular particles; and extremely irregular medium-density particles (fluffy parti- cles). Some of the larger particles had sufficient momentum to rupture exposed films. Laboratory and nose-cone control surfaces were carefully studied to permit identification of micrometeorite particles. Micrometeorite Collection from a Recoverable Sounding Rocket. Ill By R. K. Soberman,1 and C. L. Hemenway 2 The preceding papers have presented ex- would be observed at the corresponding alti- perimental details and preliminary results of tudes. However, if one assumes that the the "Venus Flytrap" experiment, and the particles had been in orbit about the earth, reasons for confidence in these results. The then tangential deceleration at altitudes of the present paper introduces interpretations, ten- order of 500 km or less would result in radial tative conclusions, and speculations. velocities equal to, or even less than, the ter- The results indicate first, the apparent minal velocities at the altitudes shown. From existence of submicron particles in the extra- the lack of high-velocity impacts observed in terrestrial flux. To account for this, one must the experiment, one must accept some form suppose a breakup mechanism for larger par- of orbital trapping, or seek a mechanism of ticles or speculate on the apparent inefficiency radial deceleration or "slow down" which is of radiation pressure in removing submicron more effective than atmospheric drag for particles from the solar system. micrometeoritic particles. Electrostatic decel- A second surprising result is the low velocity eration by the ionosphere could conceivably with which the particles impinged on the col- be such a mechanism and should be investi- lectors. Table 1 lists the terminal velocities gated. Whatever the mechanism, it will be of particles of various diameters over the assumed for the present analysis that the altitude range covered by the Venus Flytrap particles observed were moving downward rocket. If one argues that the particles should with approximately terminal velocities. be arriving at the earth with large radial The third result is the size distribution and velocities, then the velocities presented in numbers of particles collected per unit area. table 1 (particularly for the larger diameter Figure 1 shows the preliminary results from particles) are much lower than those that the collectors which were mounted in the boxes. TABLE 1.—Particle terminal velocities (km/sec) The number of particles found (larger than a given diameter) are plotted against the diam- Diameter (microns) eter. The points were normalized to a total Altitude area of 1 square meter, and a line with slope (km) 0.1 1.0 10.0 100 of —1.3 was drawn through these points. It can be seen from the ordinate values that the assumption of low velocities is practically 168 1.5 4.7 a necessity if these results are to be under- 148 1.0 3.2 128 0.52 1.6 5.2 stood in the light of other meteoric and micro- 108 0. 15 0.47 1.5 4.7 meteoric work. 88 0.021 0.066 0.21 0.66 With certain assumptions, a value for the extraterrestrial flux can be calculated. If a collector is moving through a column of unit * Geophysics Research Directorate, Air Force Cambridge Research laboratories, Bedford, Mass., and Northeastern University, Boston area in which particles are falling (fig. 2), Mass. then the number of particles of given size * Dudley Observatory, Albany, N. Y., and Union College, Schenec- tady, N.Y. collected can be written as: 09 100 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 10' (1) where (2) 3-1.3 10' where v9 is the velocity of the particles, and N na is the ambient particle density. The minus sign is used when the collector is moving downward. If one assumes a constant ambient flux The first integral is the exposure time of the collector. The second term can be evaluated if one assumes that vP is the terminal velocity of the particles. That is (4) 4 10 .2 2 .5 I 2 3 10 DIAMETER (MICRONS) where pa is the ambient air density, and (5) where vP=ve when the collector was moving downward. Implicit in these calculations was the assumption of 100 percent collection where 5 and p are the diameter and the density v efficiency. Theoretically this should be true for of the particles. For an exponential atmosphere particles larger than 0.01 micron at the altitudes this equation, on integration, yields: of interest here. The number of particles of each size collected per square meter divided (6) by the effective collection time for that size is then the flux at all altitudes. The results where H is the scale height, and the subscripts are plotted in figure 4. The ordinate is the t and / refer to initial and final values. The integrated flux (i.e., the flux of particles larger quantity in the square brackets (the effective than a given diameter is plotted as a function exposure time T of the collectors) is plotted in of the diameter). The correction factors for figure 3. the exposure angle and any shielding effect of A mean particle density of 3 gm/cm3 has the ogive assembly have been neglected. In been assumed. The calculation was performed the lower right-hand corner of figure 4 the by summing equation (6) over regions where three points from the mylar films (i.e., the the scale height remained reasonably constant. particle sizes assumed to have caused the holes) Densities, scale heights, and gravitational ac- were plotted. These points were normalized celerations were obtained from the standard in the same manner as the others. Notice atmosphere tables (Minzner, Champion, and that the upper slope has changed to —1.2. For Pond, 1959). The summation was cut off particle diameters larger than 2 or 3 microns METEOR SYMPOSIUM—SOBERMAN, HEMENWAY, AND OTHERS 101 1 1 1 ft » 1 1 1 1 » • TRUE OPEN TIVE 10 10 30 .90 1.0 CO JO 5.0 20 30 SO BO FIGURE 3.—Effective collection time for Venus Flytrap experiment. .-1.2 LOG FIGURE 2.—Moving collector in unit area column. -36 (the solar radiation cutoflF) a shaded area with slope —3.6—which is the best estimate from satellite data for these sizes (Whipple, 1961b)— -2 was somewhat arbitrarily drawn. Thus from 0 I 2 the foregoing a flux estimate of approximately LOG DIAMETER (MICRONS) 300 particles per square meter per second larger FIGURE 4.—Mircometeorite flux as calculated from Venus than 3 microns in diameter is obtained. This Flytrap results. estimate is based, of course, on preliminary data and may change slightly. It must be remembered, however, that the calculation is were included, the flux would increase by based on the assumption of particles moving approximately 20 percent. at terminal velocity. Were the velocities sub- At this point the present estimated fluxes stantially higher, the flux estimate would be can be compared with measurements made raised by the ratio of the assumed effective from satellites. At first glance the derived flux exposure time to the actual open time of the value appears to be orders of magnitude higher collector, i.e., the flux estimate for particles 3 than any recent measurements or extrapola- microns and larger would be multiplied by a tions. One might, however, compare the factor of approximately 4/3. On the other present values with a curve for a geocentric hand, if it were shown that the actual particle distribution of particles. Such a curve is velocities were lower than terminal velocities shown in figure 5. This curve, presented by in this altitude range, then the estimated flux Sobennan and Delia Lucca (see this sym- would be too high. It should also be noted posium, p. 87), is similar to that given by that the present flux estimate does not include Whipple (1961a). For an assumed spherical the particles which penetrated the 200 A films particle with uniform density of 3 gm/cm3, and were not recovered. If these particles this curve represents particles 9 microns in 102 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS will decide this question), then one is forced to 0.0 look for either a radial "slow-down" or an orbital trapping mechanism. Whatever the choice, one appears to be led to the conclusion To - 1.0 that a dust layer (either moving orbitally or settling slowly) exists about the earth. This - 2.0 conclusion follows from the number of particles I960 observed and the low relative velocities. The possible origin or origins of such a dust layer - 3.0 have been discussed by Whipple (1961a) and will not be treated here. In summary, the three results of this experi- -4.0 ment are: (1) A preponderance of submicron particles was observed; (2) the particles were ZOD -5.0 CLD apparently falling at low velocity; and (3) an unexpectedly large number of particles were collected. These results are consistent with 2.0 3.0 4.0 5.0 one or both of the following assumptions: (1) LOG ALTITUDE (km) The breakup of larger low-density "fluffy" particles; and (2) the existence of a dust layer FIGURE 5.—Meteoritic impact plotted against altitude. or a geocentric distribution of micrometeoritic particles about the earth. diameter and larger. The curve in figure 5 Acknowledgments would predict a flux of such particles of 0.25 m~2 sec"1 at 100 km. From figure 4, the present The authors thank Miss P. Bumham and estimate would be about 6 particles per square Mr. R. Sarkesian for then* assistance with the meter per second. If one assumes, however, computations. that the particles measured from satellites were References low-density loose agglomerates, and if one uses a density of 0.1 gm/cm3, then a particle with MINZNER, R. A.; CHAMPION, K. S. W.;and POND, H. L. 9 1959. The ARDC model atmosphere. AF Sur- a mass of 10~ gm would have a diameter of veys in Geophysics No. 115, Geophys. 27 microns. The present estimate (from fig. 4) Res. Dir., Air Force Cambridge Res. Lab. 2 1 of this particle size would be 0.1 m~ sec" , WHIPPLE, F. L. which is in much better agreement with the 1961a. The dust cloud about the earth. Nature, satellite observations. vol. 189, pp. 127-128. 1961b. The particulate contents of space. In P. To speculate briefly on the source of these Campbell, ed., Medical and biological particles, it is felt that terrestrial origin can aspects of the energies of space, pp. be ruled out. A cataclysmic event imme- 49-70. Columbia University Press, diately prior to the firing would be required to New York. raise particles of this size to a 90-km altitude. Discussion No such event was observed or reported. A C. W. McOracken.—If we can believe the mi- possibility which cannot be ruled out at the crophone measurements that you made on the moment is that of a sharp increase in flux Midas II, any of these results you have sub- during the time of the experiment. On June mitted give added support to the early rocket- 6th, two daylight radio-meteor showers were in probe data. progress, the Arietids which peak on June 8 R. K. Soberman.—We have little, if any, con- and the f Perseids which peak on June 9. No fidence in those early rocket results. Although unusually high activity has been reported for it might appear, at first glance, that our present this period. If the present experiment was measurements lend support, one must recall performed during a period of relatively normal that we had, in this experiment, 0.25 square activity (future experiments of the same type meter of puncture area exposed over a trajec- METEOR SYMPOSIUM—SOBERMAN, HEMENWAY, AND OTHERS 103 tory similar to that of the early acoustic rockets. craters one might expect. The mylar was alu- We do not see the large number of high velocity minum coated and we found small petals of this punctures, which one would expect from the aluminum around the hole. Therefore, I would interpretation of the acoustic-probe data. question the statement that flapped holes in- C. W. McCracken.—Do you have any hyper- dicate low-velocity impact. velocity measurements showing that particles Unidentified speaker.—What was the rela- moving with 6 to 8 km/sec velocities do not form tion between hole size and particle size in those holes with flaps of the type you see in this Space Technology Lab tests? experiment? C. W. McCracken.—The holes were not more R. K. Soberman.—We do not have such meas- than a factor of two larger than the particles. urements, but Richards and Gehring at the This was another surprise in that we had Ballistic Research Laboratory measured holes expected them to be larger. made in mylar by particles down to 5 microns M. Dubin.—You suggested that many of the travelling at 12 km/sec and have shown that at particles you observed might have been the this velocity one obtains typical hypervelocity result of an earlier breakup. Where might punctures, that is, small circular holes with no this have occurred and how? flap structure. R. K. Soberman.—This could have occurred C. W. McCracken.—We have recent data at altitudes below 300 km, where particles un- from tests performed at the Space Technology dergo deceleration of 1 g and more due to Laboratories with 1 to 3 micron particles im- atmospheric drag. Loosely agglomerated parti- pacting 0.25-mil mylar at 6 to 8 km/sec veloc- cles could break up under such forces. Such a ities. It was a shock to us to find small mechanism would explain why solar light pres- cylindrical cavities instead of the more spherical sure does not remove the submicron particles. Abstract The interpretation is given of the preliminary results obtained from the "Venus Flytrap" rocket. These results are compared with those from satellite observations. The possible existence of a dust layer is discussed. Rocket and Satellite Studies of Meteoric Dust By T. N. Nazarova1 During the last 3 years, investigators in fore, although earlier publications used the value both the Soviet Union and the United States of 40 km/sec for the particle impact velocity, have studied meteoric dust by means of in this paper we recalculated the recorded instrumentation carried aloft by rockets and masses using a particle impact velocity of 15 artificial earth satellites. Despite the limited km/sec. In some cases, a satellite encounter amount of experimental data obtained, and with meteor streams is possible; actual velocities the considerable difficulties associated with can then differ considerably from the assumed both the experiments and the interpretation of velocity. the data, the results are of interest to science The observational results are listed in table as a whole and in their application to some 1. In the experiments which utilized geophysi- engineering problems. cal rockets, the calibration of the instrumenta- We shall discuss the results of direct investi- tion was redetermined. For these experiments gations of meteoric dust carried out by means there is an uncertainty of half of an order of of Soviet geophysical rockets, artificial earth magnitude in the recorded meteoric particle satellites (Nazarova, 1961a), and space probes impact rate, since the detectors reacted to (Nazarova, 1961b). Data provided by non- impacts not only against the operating surface Soviet publications are also taken into account. but also against the rocket body. In experi- In the U.S.S.R., measurements were made ments on Sputnik III and Space Probe I by means of ballistic piezoelectric detectors (Lunik I), results were carefully corrected for which measure the pulse caused by ejection of this effect. The detectors used on Space detector material when a meteoric particle Probes II and III (Luniks II and III) reacted explodes on the surface of the detector (Isako- only to impacts against the operating surface. vich and Roi, 1960; Komissarov et al, 1960). Table 1 lists, for all experiments, the operating From the number of such pulses, the particle surface area S (for the geophysical rockets the impact rate is determined. effective area is given), the time t of the instru- In analyzing the experimental results we use ment operation, the impact rate, and the the theoretical relationship, derived by Stanyu- altitude range over which the measurements kovich (1961), that the pulse recorded by the were made. dectector is a function of the particle energy and To compare data obtained from different can be expressed as I—AE. This relation experiments, it is necessary to equate them to (IccE) is not generally accepted. Lavrentyev the same sensitivity. It is known that the (1961) believes the relation should be Iccmv16, integral mass distribution of meteoric particles while American investigators usually assume it has the form N(m)=km~0, where N(m) is the to be Iccmv. number of particles with masses greater than m. The estimate of the masses of recorded Observations of F-corona, gegenschein, and meteoric particles naturally depends on the meteors have yielded for /? the values 0.5, 0.6, velocity assumed. We regard Whipple's (1961a) and 1.3, respectively. Comparison of direct estimate for the impact velocity of meteoric measurements from Explorer VI, Vanguard III, particles with a satellite as reasonable. There- and Explorer I has led to the value 0^1 The U.S.S.R. Academy of Sciences, Moscow, U.S.S.R. (Moroz, 1962). 638805—63 10 105 1U6 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Rocket and satellite meteoric dust data Mass of recorded Vehicle Dates Area S Time s.t. Alt. range particle (with »=• Impact rate (mi) t (sec) (m« sec) (km) 15 km/sec) (gm) (m-» sec-') 1. Geophysical May 24, 1957 4 134 536 100-200 io-' 0.06 rocket 2. Geophysical Aug. 8, 1957 4 148 592 100-210 10-* 0.05 rocket 3. Geophysical Feb. 22, 1958 4 85 340 126-297 10-' 0.75 rocket 4. Sputnik III May 15, 1958 0. 34 1.8X10* 6. 0X103 400-1, 880 6X10-*—2X10~7 4-11 May 16, 17, 1958 400-700 5X10-* May 18-25, 1958 400-700 <10~4 5. First Cosmic Jan. 2, 1959 0.2 3. 6X104 7. 2X10* 2, 000-360, 000 2X10-8-10-' <2X10"» rocket 10-7_l0-« <5X10"< 8 6. Second Cosmic Sept. 12, 1959 0. 2 1. 1X 10s 2. 2X10* 2, 000-360, 000 10-8-4X10" <5X10-» rocket 4X10-8-4X10-7 <5X10-» >10~7 9X10-* 7. Third Cosmic Oct. 4-18, 1959 0. 1 2.3X10* 2. 3X104 102, 000-470, 000 7X10-»-2X10-8 4X10-4 rocket 2X10-8-6X10-8 2X10"» >6X10-8 4X10-* Figure 1 presents Soviet and American ex- these values the computed impact rates are 10 5 perimental data (LaGow and Alexander, 1960; and 10~* m"2 sec"1, respectively (Moroz, 1962). Dubin, 1960; Whipple, 1961a) equated to the Despite the nonhomogeneity of the experi- i same sensitivity (tnmia=10~ gm) through the mental data (due in large part to differences in relation Nitr^ccm'1. experimental technique, instrument calibration, The recorded impact rate is referred to the and assumptions regarding the mass-velocity mean height for each experiment, a procedure dependence of the detectors) and the limited which in the case of space probes may lead to amount of such data available, consideration of additional errors. (The graph was plotted on a all the available data allows the supposition logarithmic scale, hence the points are not that a region of increased concentration of centered between the altitude limits.) As is meteoric dust exists at a height of 100 to 300 km evident from figure 1, the observations made by above the earth's surface. As already noted, at altitudes between 400 and 2,000 km the impact means of geophysical rockets definitely reveal a 3 2 1 zone of increased density of dust particles near rate is of the order of 10~ m~ sec" (i.e., ap- the earth at an altitude of 100 to 300 km. The proximately one order of magnitude larger than the density of the zodiacal cloud would indi- existence of this increased concentration is in- cate). Because of the great dispersion of points dependently confirmed by observations of the at distances of the order of 10* to 105 km from twilight sky (Moroz, 1962). At altitudes be- the earth (fig. 1), it is difficult to say at present tween 400 and 2,000 km the impact rate is 3 2 1 whether the decreased concentration of meteoric approximately 10~ m~ sec~ . Measurements matter (compared with the lower altitudes) is made at greater distances from the earth are real or whether the dispersion in the experi- characterized by a large dispersion in the nu- mental data is indicative of fluctuations in the merical data obtained. Explorer VI and spatial density. Pioneer I yielded impact rates of 5X10"8 and 8 2 1 On the basis of direct experiments it may be 4 X10" m~ sec" , while the values for Luniks II concluded that the density of meteoric matter and III were 9X10"4 and 1X 10"3 m"2 sec"1 (all 8 in the vicinity of the earth is not constant. At values are corrected to mmln=10~ gm). altitudes of 100 to 300 km, variations in the It should be noted that estimates of the impact rate N are not greater than one order of density of matter in the zodiacal cloud, ob- magnitude. At the same time, at altitudes of tained from photoelectric measurements, range 23 8 400 to 2,000 km, sporadic increases in N are between p=3 X 10" and 3 X lO^gm/cm . From observed. METEOR SYMPOSIUM—NAZAROVA 107 10 - -• io 10"' o o i or _3 12 UJ 10 1 13 • 14 LJJ 15 I 16 io-5- ,7 _6 10 10' 10 10 10 10' h (KILOMETERS) FIGURE 1.—Integral impact rate of meteoric dust N as a function of altitude h for masses greater than 10~8 gm. Key: 1, 2, 3, Soviet geophysical rockets; 4-9, U.S. Air Force rockets; 10, Sputnik III, May 15, 19S8; 11, Sputnik III, May 16, 17, 1958; 12, Explorer I; 13, Vanguard III; 14, Soviet cosmic rocket II; 15, Soviet cosmic rocket III; 16, Pioneer I; 17, Explorer VI. On May 15, 1958, instruments carried by may be almost totally excluded. Explorer I Sputnik III recorded 4 to 11 impacts per instruments also recorded sporadic increases square meter per second.2 On May 16-17, the of the stream in this region. These increases impact rate decreased to 5X10"4m"2 sec"1 were about an order of magnitude higher than 8 (wmm=6X 10~ gm according to present velocity the average impact rate. Of great importance assumptions), which is apparently normal for are the diurnal variations in the impact rate these altitudes. Since the impact rate changed which were also observed on Explorer I (Dubin, with a period equal to the precessional rate of 1960). the satellite, instrumental origin for this effect According to Whipple (1961a; 1961b) there is a dust envelope about the earth in which li 1 It is likely that the phenomenon observed on May 15 was due to a Noch~ in a broad region from 100 to 100,000 stream with an unknown radiant. If we assume that the velocity of the km. As seen from the above considerations, stream was 70 kra/sec, then the mass of the particles detected in the stream was apparently ~5X10"w gm. it is difficult at the present time to encompass 108 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS this entire range with a single relationship LAVBENTYEV, M. A. between N and h. In the altitude range from 1961. The problem of piercing at cosmic veloci- 100 to ~2,000 km, we can write Ncchr1. From ties. In Artificial earth satellites, L. V. Kurnosova, ed., vol. 3, pp. 85-91 (transl. 2,000 km to higher altitudes, N decreases more from Russian). Plenum Press, New slowly. Future experiments will determine the York. dependence of impact rate on height and lati- MOBOZ, V. I. tude. 1962. On the dust envelope of the earth. In Several hypotheses have been suggested to Artificial earth satellites, L. V. Kurno- explain the existence of the dust cloud about sova, ed., vol. 12 (transl. from Russian). the earth; at the present time, however, no one Plenum Press, New York. of these hypotheses can be regarded as final NAZABOVA, T. N. (Moroz, 1962). 1961a. Results of a study of impacting of meteoric matter by means of instruments mounted References on space rockets. In Artificial earth satellites, L. V. Kurnosova, ed., vol. 5, DUBIN, M. pp. 524-527 (transl. from Russian). 1960. IGY micrometeorite measurements. In Plenum Press, New York. Space research, H. K. Bijl, ed., pp. 1042- 1058. North-Holland Publishing Co., 1961b. Study of meteoric particles through instru- Amsterdam. ments on the third Soviet artificial ISAKOVICH, M. A., AND Roi, N. A. satellite. In Artificial earth satellites, 1960. An acoustical method for the measurement L. V. Kurnosova, ed., vol. 4, pp. 402-410 of the mechanical parameters of meteor- (transl. from Russian). Plenum Press, ites. In Artificial earth satellites, L. V. New York. Kurnosova, ed., vol. 2, pp. 105-107 STANYUKOVICH, K. P. (transl. from Russian). Plenum Press, 1961. Elements of the theory of the impact of New York. solid bodies with high (cosmic) velocities. KOMISSAROV, O. D.; NAZABOVA, T. N.; NEUGODOV, In Artificial earth satellites, L. V. L. N.; POLOSKOV, S. M.; AND RUSAKOV, L. Z. Kurnosova, ed., vol. 4, pp. 292-333 1960. Micrometeorite studies using rockets and (transl. from Russian). Plenum Press. satellites. In Artificial earth satellites, New York. L. V. Kurnosova, ed., vol. 2, pp. 68-74 (transl. from Russian). Plenum Press, WHIPPLE, F. L. New York. 1961a. The participate contents of space. In LAGOW, H. E., AND ALEXANDER, W. M. Medical and biological aspects of the 1960. Recent direct measurements of cosmic dust energies of space, P. Campbell, ed., pp. in the vicinity of the earth using satellites. 49-70. Columbia University Press, New In Space research, H. K. Bijl, ed., pp. York. 1033-1041. North-Holland Publishing 1961b. The dust cloud about the earth. Nature, Co., Amsterdam. vol. 189, pp. 127-128. Cosmic Dust Showers by Direct Measurements By M. Dubin,1 W. M. Alexander,2 and O. E. Berg The cometary origin of meteor streams is an "s" value (the exponent in the distribution well established. During the process of disin- function indicating rate of increase in number tegration of comets, particles are dispersed with decreasing mass) equal to 2.7 (compared along the heliocentric orbit, and it appears to "s"=2.0 for sporadic meteors). This is the reasonable to assume that the age of the meteor only shower with "s" greater than 2.0. stream is related to the uniformity in the Direct measurements of micrometeorites or counting rate of shower meteors for annually interplanetary dust have been made in the recurring meteor streams. The large number United States on a number of satellites, rockets, of meteors observed in the Draconid display in and probes. Significant data on four satellites 1946 is evidence of the recent formation of the have been obtained: 1958 Alpha, Explorer I, stream from the familiar comet Giacobini- launched February 1, 1958, perigee 355 km, Zinner. Dispersion of dust particles in the apogee 2550 km (Dubin, 1960a, b, 1961; and meteor stream from interplanetary forces re- Hibbs, 1961a, b); 1959 Eta, Vanguard III, sulting from the Poynting-Robertson effect, launched September 18, 1959, perigee 509 km, corpuscular bombardment, and other radiative apogee 3751 km (La Gow and Alexander, 1960); forces, is expected to be increasingly pronounced 1960 Delta, Explorer VI, launched August 7, with decreasing dimensions of the dust particles. 1959, perigee 186 km, apogee 39,000 km; and The extension of the range of measurements of 1960 Xi, Explorer VIII, launched November 3, meteoritic material to smaller masses by direct 1960, perigee 425 km, apogee 2300 km measurements with satellites may, therefore, (McCracken, Alexander, and Dubin, 1961, and improve our knowledge of the history and age McCracken and Alexander, this Symposium, of meteor streams. p. 71). The number distribution of meteors as a The direct measurements of cosmic dust function of visual magnitude has been measured refer to particle of mass less than 10~8 grams by visual and optical methods. Similar dis- or greater than +20 on the visual magnitude tributions as a function of electron line density scale. Daily variations in the flux of cosmic along the trail have been determined by using dust particles greater than an order of magni- radar. Browne, Bullough, Evans, and Kaiser tude are often observed. We have excellent (1956) determined such distributions for the evidence that two cosmic dust showers or Perseids, Quadrantids, and the Arietids. For streams have been detected by these direct the Perseids, for example, they found that measurements. One of these dust streams, compared to the distribution of sporadic meteors detected on 1958 Alpha, permitted a deter- there was a depletion of the small meteors for mination of an approximate radiant; the other visual magnitudes greater than 7. Weiss has been detected on more than one occasion (1961) measured distribution functions for the and apparently is related to an annually Geminids, the S Aquarids, and the 17 Aquarids, recurring meteor stream. as well as for the sporadic meteors. Although the distribution function varied for different Results of Measurements meteor streams, only the daytime Arietids has Cosmic dust shower, February 1958.—Evidence of a cosmic dust shower from the data of 1958 1 National Aeronautics and Space Administration, Washington, D.C. :O Time of February Greenwich No. hits pass Station Hits/sec date time (sees) X 103 2 1512 2 087 Santiago, Chile 2. 9 2 1833 5 159 Woomera, Australia 31. 0 2 1930 2 233 Quito, Equador 8. 5 2 1933 1 09 Fort Stewart, Ga. 14. 5 2 2033 4 787 Woomera, Australia 5. 1 2 2139 2 221 Fort Stewart, Ga. i). 1 2 2238 2 691 Woomera, Australia 2. •) 2 2338 5 192 Ehrlick, Kansas City, Mo 26. 3 2 2339 11 366 Fort Stewart, Ga. 30. 0 2 2340 7 174 Havana, Cuba 40. 0 :i 0043 7 832 Woomera, Australia 8.4 3 0132 2 213 Temple City, Calif. 9. 3 3 0138 4 259 Blossom Point, Md. 15. 5 3 0139 5 187 Ehrliek, Kansas City, Mo 26. 5 3 0141 12 261 Fort Stewart, Ga. 46. 0 3 0142 2 7 Havana, Cuba 286. 0 3 0258 3 180 Woomera, Australia 16.7 3 0334 4 213 Temple City, Calif. 18. 7 3 0338 ] 120 Ehrlick, Kansas City, Mo 8. 1 3 0344 1 71 Fort Stewart, Ga. 14. 0 3 0344 2 192 Havana, Cuba 10.4 3 0349 1 — Antigua, B.W.I. 3 0539 7 338 JPL, Calif. 20.8 :i 0750 2 142 Quito, Ecuador 7.0 3 0958 2 527 Santiago, Chile 3.8 3 1200 0 952 Santiago, Chile 0 3 1405 5 663 Santiago, Chile 7.5 Average impact rate, Feb. 1-Feb. 12, 12 days 1.5 Average impact rate, Feb. 5-Feb. 12, 8 days 0.43 ning on February 2, 1958. The detector was of time. The inner filled-in circle in the polar a piezoelectric crystal with an average thresh- plot is the zero level, or zero hits per second. old sensitivity to micrometeoroid impacts of The next ring is the average impact rate of the 2.5X10"3 dyne-seconds. This is equivalent to last 8 days of the 12-day experimental period, detecting particles of mass 8XlO~10 grams and i.e., 0.43 XlO"3 hits/sec. The average impact greater with an average impact velocity of rate over this period is shown by the next 30 km/sec. The data from Explorer I were in circle, and was 1.5X10"3 hits/sec. The plotted real-time, recorded only while the satellite points are the real-time readout rates during was over a telemetry receiving station. Table the shower period. It is quite evident that 1 shows the data for each pass over a telemetry the impact rates during the shower period station in the interval from February 2, 1512 were nearly two orders of magnitude greater hours, to February 3, 1405 hours, Greenwich than the average of the latter two-thirds of time. Impacts were recorded for every pass the measurement period. but one during the interval of the cosmic dust Except for Woomera, Australia, which is shower. The table includes the station loca- represented by the squares, the circular points tion, the number of hits during each pass, the in figure 1 represent the impact rate for all time of each pass, and the number of hits per stations in essentially the same time zone second. covering about a three-hour period. The local The impact rate from table 1 has been time for these stations in the United States and replotted as a polar plot in figure 1 as a function South America during the maximum of the METEOR SYMPOSIUM—DUBIN, ALEXANDER, AND BERG 111 GREENWICH TIME-HOURS 0 FIGURE 1.—Impact rate for Satellite 1958 Alpha, February 2-3, Greenwich Time. shower intensity was early evening. This rocket. Additional data from 1960 Xi are still indicates a meteor stream in a direct helio- of a preliminary nature, but support the fact centric orbit. It is evident that the stream is that this shower recurs annually. in the ecliptic since it was detected by receiving A number of rocket flights were carried out stations in both the Northern and Southern by Berg in 1955 (Berg and Meredith, 1956) and Hemispheres. The square-plot points represent in 1960. He used a micrometeorite detector the impact rate from five passes over the which detected the impact-flash of light result- station at Woomera, Australia. They provide ing from the hypervelocity impact of the further evidence of a direct heliocentric orbit; particle on an aluminum-coated lucite or quartz the impact rate at Woomera as a function of surface. time is first high, then decreases, and then Figure 2 describes the measurements on three increases again. This impact rate was near separate rocket firings. The first firing of a minimum when at the other stations it Berg's experiment was carried out on Aerobee was near a maximum. Since the longitude at NRL^25 on November 17, 1955, at 0215 Woomera differs from other receiving stations M.S.T. A 1P21 photomultiplier was used to by nearly 180°, the earth should have shielded detect light flashes on impacts of an aluminized the satellite from this stream. The apogee of lucite cone with a light sensitivity of 10~4 1958 Alpha at 2550 km was located near lumen seconds/m2. On the assumption that Woomera on this date. Hence, the earth's 0.002 percent of the impact energy is trans- shield there would be incomplete. Figure 1 formed into visible light, a detector would be has been shadowed to suggest the extent of the able to detect particles of mass greater than 13 meteor stream, which appeared to have lasted 10" grams. Berg found that 101 impacts about 10 hours. There is no known meteor were observed on the telemetry record in the stream which may be related to this cosmic 84-second period while the rocket was above an dust shower detected on 1958 Alpha. altitude of 85 km. Almost none were observed Annually Recurring Cosmic Dust Shower in below this altitude, and the distribution of November.—A cosmic dust shower has been de- impacts on the upward and downward tra- tected with both satellites and sounding rockets jectory of the rocket was fairly symmetrical. The area of the detector was 75 cm2, and the in November. It was first recognized from the 2 2 micrometeoroid data from 1959 Eta, although impact rate was 1.6 X 10 impacts/m /sec. it was first detected in 1955 from an Aerobee This rather high impact rate observed on 112 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS AEROBEE AEROBEE JUPITER r AM-2E NRL-25 t ~ <:>* \ \ \ LUCITt ^ ^ OUAKTi ^ \^OUARTZ OUADTZ •m & 0 Oft F1.WS215 fin* 13<0 Fl»4 If 47 K... 17. 1»55 M-. M. IH» .. 25. 19W A». • 1«5C»2 FIGURE 2.—Description of measurements on three separate rocket firings. NRLr-25 remained an anomaly for a number 1X 10~2 dyne-seconds. If we assume an average of years. The experiment was repeated by impact velocity of 30 km/sec, particles of mass Lovering (1959) in Australia, and no impacts greater than 3.3X10"9 grams could trigger the were observed on a single rocket flight. In counter on the satellite. Except in rare in- 1960 Berg repeated the measurement on stances, real-time data were not obtained on Aerobee NASA 4.12 and on Jupiter AM-28. 1959 Eta, although the sample was much greater He used the 1P21 with lucite and with quartz than on 1958 Alpha. Data were recorded for cones to check for Cerenkov radiation. Aero- 78 days, with the total number of events ap- bee NASA 4.12 was fired on March 25, 1960, at proximating 5000 impacts. 1340 E.S.T., with a 75 cm2 detecting surface Figure 3 is a plot of the impact rate (log of and with 1P21 and 6199 photomultipliers. the number of impacts per hour) as a function One impact was observed on the 1P21 and two of the date from November 10 through Novem- events were observed on the 6199. This is ber 20: "C" is the average rate from September equivalent to rate of 1.2 impacts/m2/sec and an 18 through October 9; "B" the average from exposure area-time of 2.4 m2 sec. On NRL- November 10 to November 20; and "A" the 25 the exposure area-time was 0.64 m2 sec. average from November 16 to November 18. Similarly, on Jupiter AM-28, fired on Jan- It is evident that from November 16 to 18 uary 25, 1960, a space-oriented detector with there are periods when the impact rate was two 185 cm2 was exposed at 1947 E.S.T. The 6199 orders of magnitude greater than the average photomultiplier with a quartz cone was used. Included also was a calibrated light source, giving a 50 microsecond light pulse every 26 seconds. The experiment was known to be working correctly, and the area-time of the exposure was 2.4 m2 sec. Four events were detected, giving an impact rate of 1.6 impacts/ m*/sec. Thus, of three separate rocket flights similar to the earlier experiment not one could confirm the 1955 results for the impact rate within a factor of 100. A AVERAGED 16 NOV. - 18 NOV. Although it was surmised that a dust shower B AVERAGED (0 NOV. - 20 NOV. C AVERAGED 18 SEPT. - 9 OCT. might explain the rocket data, Alexander, McCracken, and LaGow (1961) provide rather clear confirmation of this hypothesis with their micrometeorite data from 1959 Eta. This n II is is M s experiment consisted of four piezoelectric crys- FIGURE 3.—Impact rates during the November 1959 tals with a threshold impact sensitivity of interplanetary dust particle event. METEOR SYMPOSIUM—DUBIN, ALEXANDER, AND BERG 113 shown in "0", and also than the average of the of approximately 5000 impacts, about 2800 78-day period. It is also interesting that there occurred during the cosmic dust shower between were rapid fluctuations in the rate during the November 16 and November 18. Even if the period of the shower, and that several orbits number of impacts so far recorded on 1960 Xi showed no impacts—or very few—even though and all other space vehicles were included, the extremely high rates were observed for small number of events occurring in these streams segments of the orbit. These rapid fluctuations represent a considerable fraction of this total. are probably real fluctuations in the spatial By comparison, the number of meteors in density of interplanetary dust particles. meteor streams observed optically, visually, The Aerobee NRLr-25 was fired on Novem- and by radar is only about 10 percent of the ber 17, 1955, at 0215, which was during the number of sporadic meteors similarly observed. same annual period as that of the intense impact On the other hand, Gallagher and Eshelman rate observed on 1959 Eta. Hence we arrive (1960) have reported that a large fraction of the at the hypothesis that both measurements meteors observed with their highly sensitive detected a shower, and that this cosmic dust radar equipment, capable of detecting meteors shower recurs annually and may possibly be of approximately visual magnitude +14, ap- related to the Leonid meteor stream. Pre- pear to be in streams. Does this perhaps sup- liminary analysis of the data from the light- port the hypothesis of a continuous generation flash micrometeorite experiment using a 6199 of dust by disintegration of conglomerates of photomultiplier with 1000 A evaporated layer interplanetary material and a short lifetime of of aluminum on 1960 Xi indicates that an im- the dust in interplanetary space? If so, it pact rate similar to that observed on NRL—25 would explain the relative unimportance of the was also observed in November 1960 during space or background density of sporadic dust approximately the same time period. The particles. On the other hand, if the rate of analysis, however, has not been completed, dispersion of dust particles after disintegration and it is still not possible to prove or disprove from a large conglomerate were rather rapid the annual recurrence of this cosmic dust and if the rate of removal of dust were slow, shower. The piezoelectric experiment on Sat- the large majority of the dust particles would be ellite 1960 Xi was so oriented as to be shielded isotropically dispersed; then the variations in from the Leonid radiant; in fact, the micro- the flux of cosmic dust as detected with satel- phone data of this spin-stabilized satellite did lites would be small, and the number of sporadic dust particles would far outweigh the com- not detect a shower at the time. ponent of dust in showers. Additional measure- Discussion and conclusions ments are certainly required to determine the geueration and destructive characteristics of From the available data on direct measurements interplanetary dust. of cosmic dust made with satellites and rockets we have presented evidence of large fluctuations References in the flux of cosmic dust. In addition to daily ALEXANDER, W. M.; MCCBACKEN, C. W.; AND LAGOW, variations, which are often greater than an H. E. order of magnitude, we have observed sets of 1961. Interplanetary dust particles of micron- conditions which may be attributed to cosmic size probably associated with the Leonid dust streams. Large fluctuations are also meteor stream. Journ. Geophys. Res., apparent in the streams themselves. vol. 66, pp. 3970-3973. BKHG, O. E., AND MEREDITH, L. H. Although the total amount of data available 1956. Meteorite impacts to altitude of 103 kilo- from direct measurements is still fairly small, meters. Journ. Geophys. Res., vol. 61, these two cosmic dust streams represent a large pp. 751-754. fraction of the total number of impacts recorded BROWNE, I. C; BULLOUGH, K.; EVANS, S.; AND KAISER, on all the satellites. The total number re- T. R. 1956. Characteristics of radio echoes from meteor corded on Satellite 1958 Alpha was 145, of trails. II. The distribution of meteor which nearly half may be associated with the magnitudes and masses. Proc. Phys. cosmic dust shower. Similarly, on 1959 Eta, Soc. London, vol. 69B. pp. 83-97. 114 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS DUBIN, M. LAGOW, H. E., AND ALEXANDER, W. M. 1960a. Meteoritic dust measured from Explorer I. 1960. Recent direct measurements of cosmic dust Planetary Space Sci., vol. 2, pp. 121-129. in the vicinity of the earth using satel- 1960b. IGY micrometeorite measurements. In lites. In H. K. Bijl, ed., Space research, H. K. Bijl, ed., Space research, pp. 1042- pp. 1033-1041. North-Holland Publish- 1058. North-Holland Publishing Co., ing Co., Amsterdam. Amsterdam. 1961. Remarks on the article by A. R. Hibbs, LOVERING, J. F. "The distribution of micrometeorites 1959. Micrometeorite impacts to an altitude of near the earth." Journ. Geophys. Res., 135 km. Planetary Space Sci., vol. 2, vol. 66, pp. 2592-2594. pp. 75-77. GALLAGHER, P. B., AND ESHELMAN, V. R. 1960. "Sporadic shower" properties of very MCCRACKEN, C. W.; ALEXANDER, W. M.; AND small meteors. Journ. Geophys. Res., DUBIN, M. vol. 65, pp. 1846-1847. 1961. Direct measurement of interplanetary dust HIBBS, A. R. particles in the vicinity of earth. 1961a. The distribution of micrometeorites near Nature, vol. 192, pp. 441-442. the earth. Journ. Geophys. Res., vol. 66, pp. 371-377. WEISS, A. A. 1961b. Author's reply to the preceding discussion 1961. The distribution of meteor masses for on the article, "The distribution of sporadic meteors and three showers. micrometeorites near the earth." Journ. Australian Journ. Phys., vol. 14, pp. Geophys. Res., vol. 66, pp. 2595-2596. 102-119. Abstract By means of rockets and satellites, direct measurements of cosmic dust over the past several years have yielded two separate sets of data which indicate interplanetary dust showers consisting of particles of mass less than 10~* grams. These cosmic dust showers were observed with micrometeorite detectors using piezoelectric crystals and photomultipliers for detecting the light flash resulting from particle impacts. The Explorer I interplanetary dust stream was detected for about a 10-hour period beginning on February 2, 1958. The peak rate of particles was nearly 50 times the average impact rate over the remainder of the 12-day period of the experiment. From the limited amount of data it may be shown that this stream was nearly in the ecliptic in a direct heliocentric orbit. The shower does not correspond to a known annually recurring meteor shower. The second stream of interplanetary dust, detected with satellites and from a sounding rocket, appears to recur annually with a peak intensity on November 17. It was detected first in 1955 from an Aerobee rocket and recognized in the November 1959 data from the Vanguard III satellite. Preliminary analysis of the Explorer VIII data indi- cates that it was probably also detected in 1960. The dust stream corresponds to the Leonids with a heliocentric velocity of 72 km/sec. The data supporting the detection of these two interplanetary dust streams are presented and discussed. The physical significance of these measurements is also reviewed. Although direct measurements from satellites have been made over a period of about four months and during the time of other known meteor streams, so far only these two interplanetary dust streams have been detected by direct measurements. Discussion F. L. Whipple.—I think that we should throw would predict an increase in concentration of the discussion open to all of the papers in this something like 107 at a 15-km altitude. There- session since they seem to be so intimately fore, I am surprised that people haven't seen related. many more of these particles in the aircraft I have plotted the number of counts versus collections. the time since launch in a number of these C. L. Hemenway.—We do find much higher acoustic experiments and have found that this concentrations of these same types of particles drops off rapidly, so that it is possible that on aircraft collectors. some of these effects are due to noise. F. L. Whipple.—Some calculations made M. Dubin.—This is not true for many of the from the aircraft collections are not completely satellites. out of line with these other estimates of the I would like to comment on the paper by influx rate. R. Soberman and C. Hemenway. I believe C. L. Hemenway.—I might interject that that there is a fundamental difficulty in relating when aircraft collections are made by the these results to the rocket and satellite experi- filtering technique where the air expands ments. The rocket in question was opened through a hole at subsonic velocities you don't below 90 km, where the particles have been find the fluffy particles breaking up. On the slowed down by the atmosphere. There is a other hand, when you collect by the impactor large uncertainty in the settling rate at these technique you find a large number of smashed altitudes, and I believe that where the exact particles that look similar to the ones we have altitudes of collection are unknown in this shown today. experiment, we should not place too much J. Manson.—We have been working on air- faith in the flux calculations. craft collections for some time. We think that L. O. Jacchia.—I have seen some figures possibly the secondary precipitation mechanism quoted for dust at altitudes as high as 90 km of forming sulfate particles at altitudes of about for the bright sunsets after the explosion of 20 km may hide some of these particles and Krakatoa in 1883; those bright sunsets are would certainly speed up their fall rate. said to have persisted for something like 3 or 4 C. L. Hemenway.—I would support that. I years. would say that the extremely small particles F. L. Whipple.—I don't feel that we can that we have seen in aircraft collections and accept the validity of those height estimates at have named nanometeorites for a convenient that time. Dr. Goody's work has shown that label were found at U-2 altitudes only in you cannot make a reliable estimate of the clumps. This would suggest that such a clump amount of material at high altitudes by the had been brought down as a part of a larger sunset method. The fact that something was particle that partially evaporated under elec- seen after sunset at that altitude does not tron-beam bombardment. prove, I think, that it wasn't secondary scatter- C. Ellyett.—There is evidently some question ing. I don't see anything to hold larger par- as to a possible particle increase at the time of a ticles up there. meteor shower, thus associating the small par- E. L. Fireman.—One can go the other way ticles with the larger meteoroids. You propose and say that these results should lead to a to fire into a quiet period and a pronounced tremendous concentration at the airplane col- shower period. I would like to suggest an lection altitudes. The relative settling rates alternative method, that is, to fire in both the 115 116 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS northern and southern hemispheres at the same through the atmosphere they should have been time since we have an entirely separate set of separated by meters at least. I might suggest showers. that as the rocket goes up there is a lot of de- P. M. Millman.—I would like to remind the gassing going on and the air density in the audience of the evidence of the noctilucent vicinity of your collectors may be as much as clouds and the trains left by large meteors, such 1000 times greater than ambient. It is quite as the Tunguska fall, which tend to collect at a possible that the particles that survived entering preferred height due to the temperature profile. the atmosphere may have broken up when they One of the best theories to explain the noctilu- hit this much denser gas envelope. cent clouds suggests a possible increase in size M. Dubin.—Dr. Soberman said in his papers of these particles due to the formation of ice on that he did not trust the earlier rocket acoustic a very small solid nucleus. When one is talk- measurements because he did not find any large ing about the possibility of particles being held number of holes in the penetration experiments. at certain levels I think that this evidence should Evidence was presented that many of the par- be carefully considered. ticles do break up and may not be able to pene- R. K. Soberman.—We deliberately opened trate the mylar films. These particles would above the noctilucent cloud altitude, which I still show up as a pulse on the microphone sys- believe is consistently in the 80 to 85 km altitude tems. There were cases where we observed region. clustered pulses that we did not believe. This M. Dubin.—The estimates of particle con- may be the explanation. Therefore, I do not centration in noctilucent clouds (that is, 1 to 5 feel that one can say that these experiments particles per cubic centimeter) give sufficient rule out the validity of the earlier acoustic numbers at about the altitude of opening to measurements. The velocity of impact, I feel, account for your results, and this would explain can still be questioned since the rocket did open the low velocity. A higher altitude opening close to the so-called noctilucent cloud region. would solve this question. Another point is R. K. Soberman.—With regard to noctilucent that Dr. Weiss mentioned in his paper that Dr. clouds, there is a half hour per day at these Kaiser has measured the distribution of the latitudes in which you could see the clouds if Arietids in the radio region and found an s num- they existed in this area. They have not been ber greater than 2.5, which would give a large observed this far south, let alone at 30° N. number of small particles at the time of that F. L. Whipple.—But the dust that may form shower. the clouds may exist at all latitudes. C. W. McCracken.—Earlier Dr. Soberman M. Dubin.—If these fluffy particles were to made the remark that his results from the two strike a wire grid on a satellite, what would be satellites did not fit the mass-distribution curve the result? that we presented. I find that if I treat his G. L. Hemenway.—I would say that the fluffy four datum points as I did the other data, then particles have no chance of breaking the grid. those four points fall within an order of magni- M. Dubin.—The microphones would still tude of the curve. Also the 300 particles per detect them and you could develop a curve of square meter per second with diameters greater the type that McCracken has shown. But than 3 microns, which he finds from the rocket because of the fluffy particles I don't think measurement, is within one order of magnitude you can compare the microphone data with the of an extrapolation of the mass-distribution wire-grid breakage rate. curve. Therefore, I don't see any tremendous E. L. Fireman.—If the statistics are as Dr. discrepancies that would lead one to conclude Hemenway reported, then fluffy particles are that an appreciable geocentric concentration only about 12 percent of the total number. has been measured. R. K. Soberman.—This is pure speculation, F. L. Whipple.—Referring to the clusters that but these irregular particles which we observe Dr. Hemenway showed, I have been considering and which form the majority of the particles how these particles can be spread out only over may have been part of larger loose agglomerates. a few microns. If they had fallen any distance C. L. Hemenway.—Because of our stringent METEOR SYMPOSIUM—DUBEST, ALEXANDER, AND BERG 117 requirements and the difficulties in classifica- the absolute rate for a line density of 1012 per tion, I wouldn't trust the percentage break- centimeter, we find about 30 per 104km2per down too far. I would trust the size distribution hour. We only detect, of course, those at ap- and the over-all numbers of particles per unit proximately right angles to us; in order to get area. a true rate we have to multiply by a factor of M. Dubin.—One more point. If you believe about 20. This line density corresponds to a the distribution function obtained by micro- mass of about 10~2 gram. For micrometeorite, phone systems is as McCracken presented it, I obtained from the previous graphs a rate of plus the fact that there are irregularities in the about 10~3 per square meter per second for a flux of particles, then you do not have to impose mass of 10"8 gram. If we take this range of 6 7 a dust belt to tie in the rocket and the satellite 10 in mass and something like 5 x 10 in rate data. when one puts in the various factors, we find J. S. Greenhow.—I would like to show how that the s value required to account for this the measurements of radar rates extrapolate difference in rates is 2.4. So it seems reasonable to the micrometeorite measurements. Dr. that we have an s value of about 2.4 all the way from these large meteors that are about Weiss showed yesterday that with electron line 2 8 densities of 1014 to 1015 per centimeter, s is about zero magnitude and 10~ gram down to 10~ 2.5. These are fairly reliable measurements. gram. It seems to me, therefore, that it would I have extended this down to 10" with a value be unreasonable to expect the slope to change of s of about 2.4. So over this range of electron in the micrometeorite range and become some- line densities s is of this order. If we look at thing else. A General Survey of Meteor Spectra By Peter M. Millman l The study of the spectra of meteors dates velopment of new photographic emulsions, from the 1860's when an English astronomer, which are faster and cover a wider range of A. S. Herschel, commenced his visual observa- wave length, and to the use of gratings instead tions with specially mounted direct-vision bin- of prisms for the dispersing element. The orig- ocular prisms. For about 20 years observa- inal spectra were all photographed on blue- tions were reported by a number of astronomers sensitive plates. Fast plates sensitive to the including Herschel himself, A. Secchi, H. A. yellow-green became available in 1932 (Mill- Newton, and N. von Konkoly. For a bibliog- man, 1935) and panchromatic plates sensitive raphy of the subject during this period see to the red in 1933 (Millman, 1934). By 1950 Millman (1932). it was possible to use an infrared emulsion to Through this early work it was realized that photograph meteor spectra (Millman, 1953). in general meteors had bright-line spectra, It was also in 1950, on the Ottawa program, though sometimes the intense luminosity of that replica transmission gratings were first the meteor head caused the lines to blend into successfully used. With a grating relatively each other giving the appearance of a continu- high dispersons are possible in some of the ous spectrum with a few predominant lines orders, and lenses of longer focal length can be superposed. The orange-yellow lines of neu- employed without materially increasing the ab- tral sodium and the green lines of neutral mag- sorption of the dispersing element. Plates la,b nesium were correctly identified by these nine- and 2a, b illustrate spectra photographed with teenth century observers. gratings on the Canadian programs. The first meteor spectrum photograph was The photographic observations have con- made on a Harvard patrol plate at Arequipa, firmed the fact that the luminosity of meteors Peru, on June 18, 1897 (Pickering, 1897). The results primarily from a bright-line atomic first program of meteor spectro-photography spectrum with a small percentage of the light was carried out by S. Blajko (1907) at the coming from simple molecules, and possibly a Moscow Observatory during the spring and continuous background in certain cases. Table summer of 1904. However, at this time there 1 lists the multiplets that have been identified was no general activity in this field of astro- with some reliability, the reference numbers nomical photography, and by 1930, when I being those from the table by Moore (1945). first became interested in the subject, only Multiplets found in some recent spectra of high eight photographic meteor spectra were known. dispersion are included in table 1, while some Five of these had been photographed by chance identified earlier on the basis of poor quality on programs conducted for other purposes. prismatic spectra have not been included. An During the last 30 years the number of photo- attempt has been made to select only those graphic meteor spectra has grown slowly but multiplets for which it is felt there is inde- steadily and the world total now stands at pendent evidence of their presence. Since well over 400. meteor spectra as a whole give a consistent Even more important than the increase in picture, an analysis of a weak spectrum of poor numbers of spectra has been the advance in definition has been judged in view of what is quality. This has been due chiefly to the de- present in high dispersion spectra of good defini- tion. In other words, the multiplets listed > National Research Council, Ottawa, Canada. 119 120 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Multiplets identified in meteor spectra (reference numbers according to Moore, 1945) (Bold face type indicates the most prominent multiplets; parentheses, those Identified with less certainty.) H I 1 N I 1 2 3 8 16 (20) 21 (25) O I 1 4 5 9 10 11 12 14 23 3F Na I 1 5 6 (8) (12) Mg I 1 2 3 9 11 14 15 Al I 1 (5) Si I (2) (3) (9) (10) (ID (16) (17) Ca I 1 2 3 4 5 (9) 18 19 (20) 21 22 (23) 32 (47) Cr I 1 7 (9) 22 (31) (40) Mn I 2 5 (6) 16 Fe I 1 2 3 4 5 6 13 15 16 18 19 20 21 22 23 24 25 26 36 37 38 39 41 42 43 45 62 68 71 72 (74) 152 168 169 207 217 268 278 (280) 318 350 (352) 354 355 357 383 (385) 409 (476) 482 522 523 553 558 559 (597) 686 (687) 693 733 (772) 801 816 965 984 1062 (1066) 1067 1092 1094 1107 1146 1150 1163 1165 1178 Co I (28) (29) (31) (34) (150) Ni I (1) (3) 5 (6) (17) 18 (19) (20) (30) (32) 33 (35) (36) (52) 86 N II 3 5 8 15 28 0 II 1 2 Mg II 4 (9) (10) (20) Si II 1 2 3 4 5 (8) (9) Ca II 1 2 Fe II 27 28 37 38 42 48 49 74 Sr II 1 x2 1st positive group, 2d positive group. represent my selection on the basis of a study shutter. Plate lb illustrates a good example of all the observational data available to date. of this effect in a Perseid meteor. The train There is no doubt that it will soon be possible or wake spectrum exhibits a similar character to add to the list a substantial number of addi- to the spectrum of the meteor head, but the tional multiplets. low excitation multiplets and intercombination Table 1 contains multiplets from 13 neutral lines are much stronger relative to the other atoms and 7 singly ionized atoms, and band multiplets. Except for Ca II the ionized systems from one molecule. The ionization elements are almost entirely lacking. Table potentials of the upper levels involved are in 2 lists the multiplets reliably identified in general low, particularly for the stronger, more train or wake spectra. Their excitation poten- prominent multiplets. Over 80 percent are in tials average 4 eV compared with an average the range 2 to 8 eV and most of the remainder of about 6.5 eV for the multiplets in table 1. are in the range 10 to 14 eV. Only the multi- The durations of the luminosity photographed plets from N II and O II have excitation poten- in meteor trains or wakes probably range from tials above 20 eV. a few hundredths of a second up to a second or Any trailing train luminosity of reasonable two. Unfortunately, the spectrum of a persist- strength will be photographed in the breaks of ent or long enduring train has not yet been the meteor track, produced by a rotating photographed. METEOR SYMPOSIUM—MILLMAN 121 Of particular interest in meteor trains is the given by Urey and Craig (1953). Of the 15 presence of the forbidden line of oxygen at abundant elements listed by these authors, 5577A (multiplet 3F of O I), often called the all except sulphur, potassium, phosphorus, and auroral green line (see pi. 2b). This was first titanium appear in table 1 above. Neither identified by Halliday (1958) and has since sulphur nor phosphorus has good low excitation been found in at least 25 spectra including a lines in the observed region, and the position number photographed by Russell (1959, and of the potassium line in the far red is over- this Symposium, p. 171). This feature tends lapped by first-order iron lines in the spectra to appear at greater heights than most of obtained up till now. Titanium, as a minor the other meteor spectrum lines, and often constituent, is difficult to identify because of disappears before the rest of the spectrum has the multiplicity of lines and the overlapping reached maximum intensity. On the average, with other elements. Three elements not com- this oxygen line is brightest between heights mon in meteorites—hydrogen, nitrogen, and of 110 km and 100 km, while the maximum strontium—are listed in table 1. The occur- light of the remainder of the meteor spectrum rence of hydrogen in meteor spectra may occurs at an average height of 90 km. The possibly indicate the presence of ices of some green line appears chiefly in the spectra of form in the cometary meteoroids. Nitrogen fast meteors, and this height of 90 km is typical must certainly be contributed by the atmo- of the maximum light for the spectra of meteors sphere, and it is not possible to say whether with velocities greater than 50 km/sec. Gemi- any of it comes from the meteoroid itself. nid spectra have the light maximum some 15 Strontium, though not common, is found in km lower on the average, and the spectra of very small amounts in some meteorites. In slow sporadic fireballs with velocities near summary, it appears that tbe cometary 20 km/sec are brightest at a height of around meteorites do not differ too markedly in general 60 km. Giacobinid spectra, brightest near a composition from the stony meteorites. The height of 90 km, are abnormally high for their absence of sulphur and potassium from meteor velocity of 23 km/sec. spectra and the presence of hydrogen are Of the 318 photographic meteor spectra probably the most significant indications of listed up to the end of 1958 (Millman, 1959), possible differences. over 70 percent are from the well-known Meteor spectra vary greatly in quality, from meteor showers. Since meteors of this category faint records showing only a single line to high do not seem to fall as meteorites, meteor dispersion spectra in which 100 or more sepa- spectra provide the only direct evidence of the rate items have been measured. In any sta- composition of these objects. The atomic tistical study of spectra it is important to abundance of the elements in the largest class consider this feature, and the following quality of stony meteorites, the chondrites, has been scale has been found convenient: a - detail of 50 lines or more, TABLE 2.—Multiplets in meteor train and wake spectra b - 20 to 49 lines, (Bold face type Indicates the most prominent multiplets; parentheses, c - 10 to 19 lines, those identified with less certainty.) d - 9 lines or less. Meteor spectra exhibit a fairly consistent OI 3F pattern with a few dominant characteristics, Nal 1 and I proposed a simple system of classification Mgl 1 2 3 (9) some years ago (Millman, 1935). A study of Cal 1 2 3 18 (19) (21) all available data has indicated that this Mn I 2 Fel 1 2 3 4 5 15 (16) 20 original classification scheme is still useful and 21 (23) 41 42 43 (268) it is summarized below, revised slightly to adjust it to the panchromatic emulsions now Call 1 2 generally used. Classification of a spectrum is based on its character at light maximum and N, (1st positive group.) involves identification of the strongest features 122 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 3.—Classification of 259 meteor spectra Meteor Vel Y X Z W km/sec a b c d a b c d a b c d d Leonid (L) 72 7 4 Orionid (0) 66 2 4 3 3 Perseid (P) 60 9 27 11 45 1 4 4 2 Lyrid (Y) ] Quadrantid (Q) > 31-48 1 1 1 1 1 1 6 Aquarid (D) Taurid (T) Geminid (G) J 35 1 3 1 1 2 4 12 3 Giacobinid (J) 23 3 9 9 1 2 2 Miscellaneous (N) 12-70 1 5 9 13 1 6 10 14 2 5 3 4 1 Totals 12 35 35 66 2 13 28 39 2 6 5 10 6 148 82 23 6 in two wavelength regions, the blue-violet within each type. It is suggested that these from 3500 to 4500 A and the orange-green should be used in the same sense as for stellar from 5000 to 6000 A. Four main types are spectra, that is, lower numbers referring to defined below: higher excitations. Type Y: The H and K lines of Ca II are the strongest All the meteor spectra secured prior to 1959, features in the blue-violet. for which information is available, have been Type X: If not as above, then either the D-lines of classified in the above system. Their distri- Na I, or the lines near 5180A or 3835A bution in type, listed according to velocity, is of Mg I, give the strongest feature in either the orange-green or the blue-violet. given in table 3. It will be noted that the fast Type Z: If not as above, then the lines of Fe I, or meteors are predominantly Type Y, meteors of those of Cr I, give the strongest feature medium speed tend to be Type X though some in either the orange-green or the blue- are of the other types, while slow meteors are violet. confined to Types X and Z. Giacobinid Type W: Not as above. spectra (see pi. \c) are good examples of what This classification, though developed on a we find in the light of very slow shower meteors. strictly empirical basis, has some physical In the miscellaneous group the Type Y significance since Type Y spectra are of higher examples probably represent chiefly cometary excitation than Type X. In Type Z spectra meteoroids, since asteroidal meteoroids as a most of the elements common in stony meteor- class are medium to low velocity objects oids are weak or absent, and hence a reason- (Whipple and Hughes, 1955). In table 3 able percentage of Type Z spectra must cor- there are probably not more than 30 miscella- respond to iron meteoroids. Type W will neous spectra produced by asteroidal meteoroids, include peculiar spectra and those too faint or and at least two-thirds of these will be Types Y ill-defined to have other classification features. and X, produced by stony meteoroids. Of Since spectrum quality has a direct bearing on the miscellaneous spectra Type Z, it is likely what can be identified, it is suggested that a that not more than six represent iron meteor- letter for the quality scale be prefixed to the oids. Two examples which are probably in type. For example, a spectrum of 35 lines this category are reproduced in plates Id and 2c. showing very strong H and K of Ca II would be These spectra exhibit very little other than classified as bY, one of 58 lines showing just the strong low level iron multiplets and sometimes iron multiplets would be aZ. If it should be the ultimate lines of chromium. found advisable, numbers could be suffixed to A further insight into the relations among me- the type letters to indicate levels of excitation teor spectra may be obtained by an examination METEOR SYMPOSIUM—MILLMAN 123 100 o b c d a Mgl + Si I + CaE + Fell 100 bed bed Fel CrI + Mnl Na I + Mg I + Ca I FIGURE 1.—The frequency of identification of various elements in 259 meteor spectra photographed prior to 1959, plotted against the spectrum quality from a to d. The plot has been smoothed to show general trends and should not be used for quantita- tive values. Abbreviations: L, Leonid; O, Orionid; P, Perseid; G, Geminid; J, Giacobinid; Sh, shower; N, miscellaneous. of the frequency with which various atoms are laneous group which probably contains a sig- identified in any group. Figure 1 illustrates nificant percentage of nonshower meteors. some of these data. The observational material Another way of looking at meteor spectra used was the same as that included in table 3. is to list the features which are strongest in Percentage identification of the atoms is plotted each example. Naturally, this will be greatly against spectrum quality. These curves are modified by the wavelength sensitivity of the smoothed averages and give qualitative infor- grating or prism-lens-emulsion combination. mation only and should not be used to give If we consider only the good quality spectra quantitative values. They do, however, clearly studied here, ionized calcium is the strongest demonstrate the danger of drawing general feature in all Perseid spectra. The Geminid conclusions from spectra of low quality only. spectra are roughly divided into four equal The correlation of the appearance of the groups with ionized calcium, neutral mag- ionized elements with the velocity of the meteor, nesium, neutral sodium, and neutral iron the and the universal presence of iron in all spectra strongest in each group respectively. Mag- of good quality, are quite evident features of nesium is somewhat ahead of the other three the plot. Another interesting point is that the elements in frequency, suggesting the possibility elements Ca, Na, and Mg are more frequently that this element is a significant constituent in identified in the shower meteors, while Cr and Geminid meteoroids. Giacobinid spectra are Mn are found more frequently in the miscel- about equally divided between sodium and iron 124 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS EXPLANATION OF PLATE 1 Meteor Spectrum No. 340, Type aY, a bright (1961). Note high relative strength of multi- Perseid photographed at the Newbrook Meteor plets Fe I 1, 2, 3, Mg I 1, and Ca I 1 in wake Observatory at 9h 27» 55« UT, August 13, 1959. spectrum filling next to last shutter break. Lens, focal length 20 cm, aperture 7 cm; grating, Meteor Spectrum No. 86, Type bX, a Giacobinid 400 lines/mm; emulsion, Kodak Tri-X Aerecon; photographed at North Bay, Ontario, at 3h dispersion in second order 60 A/mm, no rotating 51m 42» UT, October 10, 1946. Lens, focal shutter, meteor moving down in figure, observed length 15.0 cm, aperture 3.3 cm, prism angle visually magnitude —7 or —8 with a persistent 30°, flint glass; emulsion, Kodak XXX-B train of 10-second duration. Halliday (1961) plate; mean dispersion H5—H-y 410 A/mm, has measured over 100 lines in the spectrum. rotating shutter occulting 10/sec, open to closed Note particularly the late appearance in the ratio Yi, meteor moving down in figure, observed burst of the low level iron multiplets Fe I 2 visually magnitude —2 with persistent train and 3 as compared to the higher level multiplets duration 20 sec, photographic trajectory heights Fe I 41, 42, 43, and the ionized elements. This from 109 km to 90 km, light maximum of sodium is a similar effect to that found by Russell (1960) line 102 km, of magnesium and iron lines 98 in the spectrum of the terminal burst of another km. Note early appearance and disappearance Perseid meteor. of sodium line. Meteor Spectrum No. 272, Type aY, a bright d: Meteor Spectrum No. 24, Type aZ, nonshower Perseid photographed at the Springhill Meteor photographed at Flagstaff, Ariz., at 6h 35™ Observatory at 6h 20° 15', August 12, 1958. 23» UT, February 27, 1933. Lens, focal length Lens, focal length 15.0 cm, aperture 3.3 cm; 15.0 cm, aperture 3.3 cm, prism angle 30°, grating, 200 lines/mm; emulsion, Kodak 103 aF flint glass; emulsion, Cramer Iso Presto; mean spectrographic; dispersion in first order 320 dispersion Ha—Ry 410 A/mm, no rotating shut- A/mm, rotating shutter occulting 10/sec, open ter, meteor moving down in figure violet end of to closed ratio Yt, meteor moving up in figure, spectrum at left, observed visually magnitude observed visually magnitude —7 or brighter — 9 with train 2 sec duration. Visually deter- with persistent train duration 180 sec, photo- mined heights of major bursts at 75, 73, 71, graphic trajectory heights from 112 km to 83 69, 66, 65, 63, 59 km respectively. This spec- km with light maximum at 90 km. Infrared trum has been studied by Miliman (1935). spectrum of this meteor, secured with another Spectrum shows very little besides iron and camera, studied by Miliman and Halliday nickel. MILLMAN PLATE 1 U Fe 43 Fe 42 Fe 41 Fe 3 Fe 2 U I I I Ca+ I Sr+l Ca I Mg+ 4 For explanation see facing page. MILLMAN PLATE 2 For explanation see facing page. METEOR SYMPOSIUM—MILLMAN 125 EXPLANATION OF PLATE 2 Meteor Spectrum No. 273, Type bY,& bright Perseid shutter occulting 11.6/sec, open to closed ratio %, photographed at the Springhill Meteor Observa- meteor moving up in figure, approximate heights h tory at 6 39" 25" UT, August 12, 1958. Lens, of photographic trajectory 120 km to 100 km. focal length 20 cm, aperture 7 cm; grating, 200 This spectrum has been studied by Halliday lines/mm; emulsion, Kodak infrared aerographic; (1960). Note the auroral O I line at 5577A no filter, dispersion in first order 240 A/mm, which carries across all the shutter breaks, rotating shutter occulting 12/sec, open to closed indicating a minimum duration of an appreciably ratio y%, meteor moving down in figure, observed fraction of a second. The sodium D-line appears visually magnitude -5 with a persistent train just to the right of the auroral line, and the duration 30 sec, photographic trajectory heights magnesium green line just to the left. from 104 km to 79 km with light maximum at Meteor Spectrum No. 26, Type bZ, nonshower 84 km. First order infrared overlaps second photographed by the Texas Observers at Fort order violet. This spectrum has been studied by h Millman and Halliday (1961). Note the early Worth, Tex., at 7 03- UT, August 12, 1933. appearance of the O I line, and the increase in Lens, focal length 30 cm, aperture 7.5 cm; prism relative strength of the Ca II lines towards the angle 20°, flint glass; emulsion, Kodak Hyper end of the path. Press; mean dispersion H5—H7 333 A/mm, no rotating shutter, meteor moving down in figure Meteor Spectrum No. 294, Type dX, a Perseid violet end of spectrum at left, observed visually photographed at the Newbrook Meteor Observa- h magnitude —5 with train 1.5 sec duration. tory at 7 30" UT, August 15, 1958. Lens, focal length 6.2 cm, aperture 8.7 cm; grating, 80 Spectrum shows mainly low level multiplets of lines/mm; emulsion, Kodak I-D spectroscopic; neutral iron. Note similarity to Spectrum 24 dispersion in first order 1950 A/mm, rotating (see pi. Id). 638805—«3 11 126 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS as the strongest feature. For the miscellaneous In conclusion I express my thanks to I. S. group of spectra, sodium comes first, followed by Astapovich, E. N. Kramer, and K.A. Liubarsky, ionized calcium and neutral iron. U.S.S.R.; Zd. Ceplecha and J. Rajchl, Czecho- Several investigators, including myself, have slovakia; I. Halliday, Canada; H. B. Ridley, noted that the character of a meteor spectrum England; and J. A. Russell, U.S.A., all of whom frequently changes along the photographed have provided me with unpublished details of path. Usually this change is in the sense of meteor spectra for this survey. progressing to higher excitations, with the associated greater relative strength of the References ionized atoms such as Ca II, Mg II, and Si II. BLAJKO, S. Bursts near the end of the path for bright 1907. On the spectra of two meteors. Astro- meteors of the faster showers demonstrate phys. Journ., vol. 26, pp. 341-348. particularly well this enhancement of the ionized CEPLECHA, Zd. atoms (Russell, 1949). However, in the case of 1959. On the colour index of meteors. Bull. slow nonshower fireballs it is quite possible for Astron. Inst. Czechoslovakia, vol. 10, the neutral iron vapor to show a lower excita- p. 39. tion at the bursts (Millman, 1935). COOK, A. F. Since a meteor generally increases in luminos- 1955. The nature of meteoric radiation. In T. R. Kaiser, ed., Meteors, Spec. Suppl. ity over most of the visible path, the change Journ. Atmosph. Terr. Physics, vol. 2, in excitation will be correlated with luminosity. pp. 8-15. Pergamon Press, New York. The early part of the path of a fast, bright HALLIDAY, I. fireball may show only the nitrogen bands, 1958. Forbidden line of O I observed in meteor sodium D-lines, and some iron. This will give spectra. Astrophys. Journ., vol. 128, pp. 441-442. an intensity distribution strong in the red and 1960. Auroral green line in meteor wakes. Astro- yellow. As the fireball brightens the lines of phys. Journ., vol. 131, pp. 25-33. ionized calcium and ionized magnesium, as well 1961. A study of spectral line identifications in as other features, become relatively stronger Perseid meteor spectra. Publ. Dominion and will tend to increase the intensity in the Obs. Ottawa, vol. 25, pp. 1-16. blue and violet. It is found that spectra of JACCHIA, L. G. the faint meteors of high-velocity showers often 1957. On the "color index" of meteors. Astron. appear similar to the early parts of the paths Journ., vol. 62, pp. 358-362. of much brighter meteors of the same shower. MILLMAN, P. M. 1932. An analysis of meteor spectra. Ann. This agrees with the observational results of Harvard Coll. Obs., vol. 82, pp. 113-146. Jacchia (1957), Ceplecha (1959), and Davis 1934. Meteor spectra—list I. Journ. Roy. (this Symposium p. 233), which indicate a Astron. Soc. Canada, vol. 28, pp. 279-281. higher color index, and therefore a redder color, 1935. An analysis of meteor spectra: second for the faint meteors. There are a few very paper. Ann. Harvard Coll. Obs., vol. 82, pp. 149-177. faint Perseid spectra that lack the strong lines 1953. The typical Perseid meteor spectrum. of Ca II, a characteristic feature of most Nature, vol. 172, pp. 853-854. spectra of this shower. Cook (1955) has 1959. Photographic meteor spectra (appendix suggested a two-parameter dependence of 7). Journ. Roy. Astron. Soc. Canada, ionization and excitation in meteor spectra, the vol. 53, pp. 271-276. major controlling factors being meteor velocity MILLMAN, P. M., AND HALLIDAY, I. and the ratio of characteristic meteoroid 1961. The near-infra-red spectrum of meteors. dimension to the mean free path of an evapo- Planetary Space Sci., vol. 5, pp. 137-140. rated atom. Since the fainter, and hence smaller MOOBE, C. E. meteors, disappear higher in the atmosphere, 1945. A multiplet table of astrophysical interest. Contr. Princeton Univ. Obs., no. 20, the above suggestion could explain some of the parts 1, 2. observed correlations, but the situation is PICKERING, E. C. probably somewhat more complex in many 1897. Spectrum of a meteor. Astrophys. Journ., cases. vol. 6, pp. 461-462. METEOR SYMPOSIUM—MILLMAN 127 RUSSELL, J. A. UBBT, H. C, AND CRAIG, H. 1949a. A composite spectrum of a Perseid 1953. The composition of the stone meteorites meteor of 1948. Pop. Astron., vol. 57, and the origin of the meteorites. Geo- pp. 187-192. chim. et Cosmochim. Acta, vol. 4, pp. 1949b. A note on revised wavelenths for a com- 36-82. posite spectrum of a Perseid meteor of 1948. Pop. Astron., vol. 57, pp. 344-345. WHIPPLE, F. L., AND Hughes, R. F. 1959. Some Perseid meteor spectra of the past 1955. On the velocities and orbits of meteors, decade. Sky and Tel., vol. 18, pp. fireballs, and meteorites. In T. R. 549-551. Kaiser, ed., Meteors, Spec. Suppl. Journ. 1960. A time-resolved spectrum of the terminal Atmosph. Terr. Physics, vol. 2, pp. 149- burst of a Perseid meteor. Astrophys. 156. Pergamon Press, New York. Journ., vol. 131, pp. 34-37. Abstract Over 400 meteor spectra have now been photographed at various observatories, with most of the data having been obtained in three countries—Canada, the U.S.S.R., and the U.S.A. The range of wavelengths covered extends from 3000A to 9000A. The bright-line emissions for roughly 200 multiplets have now been identified with some certainty. These belong to 13 neutral atoms, 7 singly ionized atoms, and one molecule. The composition of come- tary meteoroids does not seem to differ markedly from that of stony meteorites. The excitation in a meteor spectrum depends chiefly upon the velocity of the meteor and its absolute magnitude. A simple system of meteor spectrum classification is suggested, based on an earlier proposal made by the author in 1935. Meteor Spectra with High Dispersion By Zd. Ceplecha* and J. Rajchl* This paper is based on the analysis of two The spectral plate is measured with the y-axis photographs of meteor spectra with high placed roughly parallel to the meteor motion, dispersion obtained by transmission diffraction thus leaving the x-coordinate of any one line gratings in the autumn of 1960 at the Ondfejov only slightly dependent on the position along Observatory (Rajchl, 1961). The cameras used the track where it is measured. The coordinate were provided with Tessar objectives having system for the whole plate is computed from focal ratios of 1:3.5 and 1:4.5 and focal lengths measurements of zero-order images of stars of 30 cm and 36 cm. The first camera was (Ceplecha, 1951; 1954). The spectrum must be equipped with a transmission objective grating photographed with a rotating shutter. The with 600 grooves per millimeter, and the second measurements of one break in three different one with a grating with 400 grooves per milli- lines then give the diffraction hyperbola, and meter. The photographic materials used were thus the connection between the coordinate Agfa Press plates 25° D with dimensions 18X24 system of the plate and that of the grating. cm. A shutter rotating approximately 450 The coordinate system of the grating directly times per minute was used with these cameras. determines the wavelengths of individual lines if we know the precise number of grooves per The method of wavelength determinations millimeter and the direction cosines of the zero- Since a meteor spectrum photographed with order image of the meteor. Identification of an optical transmission grating is usually off one line will yield this precise number of grooves the optical axis of the camera, the ordinary per millimeter, and comparison with the number method of wavelength determination is hardly of grooves specified by the manufacturer will feasible. The diffracted ray falls in the plane check the identification. Thus it is possible to determined by the incident ray and the optical say that the wavelengths of all the lines are axis when the spectrum passes through the computed without previous identification, and optical axis. On the other hand, when the that they are determined absolutely by the spectrum does not pass through the optical measurements of angles. Of course, because of axis, the diffracted rays lie on the surface of a optical distortions, the coordinate system, de- cone, the axis of which is parallel to the grooves. fined by zero-order images of three stars and by The section of this diffraction cone made by their apparent spherical coordinates with cor- the focal plane is the diffraction hyperbola, rections for atmospheric refraction added, is along which the spectral record is distributed. less accurate than necessary for absolute deter- Absolute wavelengths can be found from such minations of wavelengths. It is possible, how- a spectrum if it includes the zero-order image. ever, to correct the coordinate system established A method based on Rowland's (1893) was from three stars at different places of the meteor developed to exploit the full accuracy that spectrum by using additional stars. For could be obtained. Complete details, with example, with an undriven camera the inter- equations, have been published (Ceplecha, section points of star images with the meteoric 1961). We present here a brief description of spectral lines may be used for this purpose. the principal features of the method. Since the directions of spectral lines in various parts of the spectrum differ slightly, it is neces- 1 Astronomical Institute of the Czechoslovak Academy of Sciences, Ondfejov, Czechoslovakia. sary to compute the point of intersection of 129 130 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS diffraction hyperbolae with each line, a some- what laborious procedure, obviated in the case plate of hundreds of lines only by the use of an electronic computer. We used NE-803, com- pleting our work on the two spectra in one hour of computer time. -24 cm It is very interesting to note incidentally FIGURE 1.—Overlapping of orders in meteor spectrum S 6. that the dispersion can be chosen along an ellipse, if the grating is inclined to the optical axis of the camera. For our cameras the angle n of 54° is sufficient to photograph all meteor 30 spectra with dispersion along an ellipse. Such an experimental arrangement might offer some advantages for photography of meteor spectra. 20 Meteor spectrum S 6 This spectrum was photographed September 13, 1960, at 18* 59m UT. The spectrum of a single bright flare was well exposed. The flare was 10 shorter than one break (the number of breaks per second was about 15.2). The spectrum is 24 cm long and is near the longer edge of the plate, with the hyperbolic form visible at first -1.0 _•_. sight. Some of the brightest spectral lines also -0.5 + 0.5 -1.0 A appeared three breaks before the flare. Thus it was possible to determine that the angular FIGURE 2.—Distribution of the deviations of measured velocity was 6.9°/sec. A direct photograph of wavelengths from tabular wavelengths of 116 unblended this meteor is, however, not available since the lines. meteor flare was only 28° above the horizon at azimuth 118°, where none of our direct cameras The disentanglement of the different over- was pointed. The spectrum begins with a part lapping orders was very laborious. The wave- of the second order and continues with the lengths were computed for every possibility, third, fourth, fifth, sixth, and seventh orders. and identifications of the individual lines The main order contained is the fourth, but the were sought in different orders. Some factors, orders overlap, as shown in figure 1. They are for example the different optical quality of the on the opposite side from the "blazed" first image depending on the wavelength, simplified order. Thus the distribution of the intensity the choice of the spectral order, but the most of one line in different orders is relatively important factor considered was the repetition smooth. The approximate dispersion, which of one line with corresponding intensities changes slightly with the place on the plate, is in all spectral orders. The results appear in presented in table 1. The dispersion in the table 2, which contains 166 lines with their fourth order is about 19 A/mm. With such a identifications. It is interesting to note that high dispersion, it is possible to compute the 70 percent (116 lines) are not blended at this wavelengths with relatively great precision. high dispersion. The mean deviation of the TABLE 1.—Approximate dispersions of meteor spectrum measured wavelength from the laboratory S6 wavelength for these unblended lines was only Order 2 3 4 5 6 7 ±0.21 A. The distribution of the deviations is Dispersion A/mm 38 25 19 15 13 11 given in figure 2. METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 131 The basic spectral tables used for the identi- meteor belongs to the Southern Taurid stream. fications are the Russian tables of Zajdel et al. Then the radiant aR=5°, 5R=0°, which lies (1952), from which the intensities 7tab presented on the great circle of the meteor, and an as- in table 2 are taken for the arc spectra of indi- sumed velocity of 29 km/sec compared with the vidual elements. The observed intensities 70bl! observed angular velocity will yield 11° of of lines of the meteor spectrum were estimated inclination to the horizontal plane and 76 km by eye (the scale is given with table 2). The for the height of the flare. This is quite accept- mean wavelengths of table 2 were calculated for able. If the height of the flare is assumed to be each line from wavelengths of different orders in the interval from 90 to 60 km, the velocity by weighting with the number of the order. lies within a short interval between 34 and 23 The mean intensities were computed for the km/sec. In any case this meteor can be classi- intensity level of the fourth order. The inten- fied as slow. sities from other orders were reduced to the level of the fourth order by multiplying by the factors Spectrum of meteor No. 27471 derived from comparison of observed intensities The meteor was observed October 27, 1960, of all lines in different spectral orders (the cor- at 0h13m048 UT. Since double-station photo- rection factors are given with table 2). Since graphs are available, complete data on the the spectral sensitivity of the plate was not atmospheric trajectory and the orbit can be taken into account, the intensities are only 2 presented. The detailed study of the velocity provisional values. changes of this meteor is discussed in a paper The majority of the lines are iron lines be- by Ceplecha and Davies.3 longing to 76 different multiplets. The multi- The height of the beginning of this meteor plet numbers presented in table 2 are from was 86.8 km with the original velocity of Moore's (1945) tables. The strongest lines are »„,=25.95 km/sec, which was possible only if Na doublet 1; Mg triplet 2; ionized Ca doublet the composition of the meteor was iron (Ceple- 1; Ca line 4226.73; Fe lines 4375.93 of multiplet cha, 1958). The observed radiant was aR= 2; 4383.55 of multiplet 41; 5328.37 of multiplet 39?09, 5R=10?87, and the elements of the orbit 15. Other elements observed are Mn multi- were plets 2, 4, 5, 6, 22, 23, 24; Ni multiplets 45, 86, 131, 143; Al multiplet 1; Cr multiplets 1, 7, a=2.08 a.u., co=98? 64, 31, 96; and Co multiplet 34. 6=0.759, Q=33?44, 2=0.501, i= 4?89. The elements usual in meteorites and air, which are not observed in this spectrum, are This orbit is identical with the association Cu, P, S, C, Si, K, N, O. No. 85 observed by Jacchia and Whipple (1961). The quality of this meteor spectrum is well The brightness of the meteor increased from demonstrated by the neighborhood of the Ca II —0*5 at the beginning to the maximum light line 3933.67 A. The following lines are sepa- of — 5*5 at height 65 km, where the velocity rated: Fe line 3927.92; Fe line 3930.30; weak Fe of the meteor was 25.5 km/sec. This part of line 3931.12; weak Fe line 3932.63; and Ca the trajectory was composed of a series of II line 3933.67. This is the first instance in individual pulsations, the last of which was the which these iron lines were separated from the maximum itself. After the maximum there Ca II line. The closest lines separated in this were several pulsations; then a relatively long spectrum are only 0.43 A apart. part of the trajectory exhibited almost constant The Ca II doublet in this spectrum is very brightness between —3m and —4m. The end strong, although the meteor is slow. of the trajectory occurred quite suddenly at As already indicated, direct double-station the height of 41.1 km. The velocity measured photographs of this meteor are not available. at the end was 13.2 km/sec. It is possible, however, to estimate that this » Zd. Ceplecha and J. Q. Davies, "On the Change of the Heat-Transfer 'The boloraetric intensities of lines will be published in Ball. Astron. Coefficient of Bright Meteors," to be published in Bull. Astron. Inst. Inst. Czechoslovakia, vol. 14. no. 2, 1963. Czechoslovakia. 132 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6 [EXPLANATION: Scale of observed Intensities I»bi: 0, not observed; 0.5, weakest lines; 1, very weak; 2, weak; 3, medium intensity; 4, strong; 5, very strong; 6, strongest lines. The correction factors for the level of intensities of the 4th order were derived from comparison of observed intensities in different orders of all observed lines: Order: 2 3 4 5 6 7 Correction factor: 0.7 0.8 1.0 1.5 2.0 3.3 The symbol + means that the line was observed but that it was not measured because it was too weak (or not sharp enough) or because it blended with another line. The column headed "Mult." gives the number of multiplet according to Moore (1945). The figures In the last column (ev) represent the energy of the upper excitation level.] Line observed Blended Identification with line No. Order lohn Xob. No. Order ^tmb /,.b Mult. ev 1. 4 1 3644. 27 5 1 3643. 62 6 0. 5 3644. 87 7 1 3644. 75 Mean: 1. 7 3644- 44 3644. 41 200 Cal 9 5.30 2. 4 0 5 3681. 60 /158 3 2 \ll9 4 6 0 3682. 23 400 Fel 772 6. 91 7 0 3679. 92 500 Fel 5 3. 36 Mean: (0.8) (3681.60) 3683. 06 200 Fel 5 3.41 3. 4 1 + 5 1 3696. 99 160 3 6 0 7 0 3697. 43 100 Fel 389 6.35 Mean: 0.6 3696. 99 3696. 57 100 Mn I 24 6. 23 4. 4 1 3700. 72 5 0. 5 + 6 0 7 0 Mean: 0.4 3700. 72 3701. 09 300 Fel 385 6.35 5. 4 1 3705. 25 5 0.5 + 6 0 7 0 Mean: 0.4 3705. 25 3705. 57 700 Fel 5 3.39 6. 4 o 3719. 95 5 1 3720. 09 6 1 3719. 94 7 0 Mean: 1.4 3719. 99 3719. 94 1000 Fel 5 3.32 7. 4 1 3733. 93 5 1 3733. 79 6 2 3733. 86 7 0 3734. 87 1000 Fel 21 4. 18 Mean: /. 6 3733. 86 3733. 32 400 Fel 5 3.43 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 133 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order 7ob8 Xob« No. Order Xt«b /tab Mult, ev 8. 4 2 3737. 27 5 2 3737. 97 6 1 + 7 0 Mean: 1. 8 3737. 66 3737. 13 1O00 Fe I 5 3. 35 9. 4 1 3743. 00 5 0 6 1 3743. 48 7 1 3743. 37 Mean: 1. 6 3743. 32 3743.36 200 Fel 21 4.30 10. 4 2 3745. 72 5 0. 5 + 6 1 3745. 01 7 1 3745. 17 3745. 56 500 Fe I 5 3. 39 Mean: 2. 0 3745. 24 3745.90 150 Fel 5 3.43 11. 4 2 3749. 20 5 1 -f- 6 1 3749. 64 7 1 3749. 81 Mean: 2. 2 3749. 61 3749. 49 1000 Fe I 21 4.22 12. 4 1 3816.06 5 0 6 0 Mean: 0. 3 3816. 06 3815. 84 700 Fe I 45 4. 73 13. 4 1 3820. 36 5 0 6 0 Mean: 0. 3 3820. 36 3820.43 800 Fel 20 4. 11 14. 3 1 3822. 41 4 1 3823. 14 5 0.5 + 6 0 3823. 51 75 Mn I 6 5. 38 Mean: 0. 6 3822. 83 3823. 89 50 Mn I 6 5. 40 15. 3 2 3824 37 4 2 3824. 68 128 3 5 1 + 6 0 Mean: 1. 3 3824. 55 3824.44 150 Fe I 4 3. 24 16. 3 0 4 1 3826. 27 129 3 5 0 6 0 Mean: 0. 2 3826. 27 3825. 88 500 Fe I 20 4. 16 6SSS05—63- 12 134 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /Ob» \>b« No. Order Xtab 7tl>b Mult, ev 17. 3 0 4 1 3827. 78 5 1 3828. 11 6 0 Mean: 0.6 8827.96 3827. 82 200 Fe I 45 4. 79 18. 3 0 4 1 3829. 36 5 1 3829. 91 6 0 Mean: 0. 6 8829. 67 3829. 35 10 Mg I o 5. 94 19. 3 1 3831. 69 4 1 3832. 40 5 0 6 0 Mean: 0.4 8832. 10 3832.31 250 Mg I 3 5. 94 20. 3 0 4 1 3832. 93 5 0 6 0 Mean: 0. 2 8832.98 3833. 31 100 Fe I 221 5. 79 21. 3 1 3834.04 4 1 3834. 10 5 1 3834. 96 (3834. 36 75 Mn I 6 5.39) 6 0.5 + (3833. 86 75 Mn I 6 5.41) Mean: 1. 1 8834. 44 3834. 22 400 Fe I 20 4. 19 22. 3 0 4 1 3836.49 5 1 3837. 04 6 0 Mean: 0.6 8886.80 3836.33 100 Fe I 664 6.53 23. 3 1 3837. 86 4 1 3838. 19 5 0.5 + 6 0.5 + Mean: 0. 9 8888. 05 3838. 29 300 Mg I 3 5. 94 24. 3 0 4 1 3839. 36 5 0 6 0 Mean: 0.2 8889. 86 3839. 26 100 Fe I 529 6. 27 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 135 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /obi ^obs No. Order ^tab ^tab Mult. ev 25. 3 0 4 1 3839. 77 5 1 3839. 80 6 0 Mean: 0.6 S8S9. 79 3839. 78 100 Mnl 6 5.41 26. 3 0 4 1 3840. 62 5 0 6 0 Mean: 0.2 SS40. 62 3840. 44 400 Fel 20 4.22 27. 3 0 4 1 3841. 21 5 0 6 0 Mean: 0.2 3841. 05 500 Fel 45 4.83 28. 3 1 3849. 83 4 1 3850. 32 5 1 3850. 27 6 0 3849. 97 500 Fel 20 4.23 Mean: 0.8 S850. 18 3850. 82 200 Fel 22 4.21 29. 3 0 4 1 3852. 28 5 0 6 0 Mean: 0.2 3852. fS 3852. 58 150 Fel 73 5.99 30. 3 (3) 32 3 4 3 3856. 52 5 2 3856. 80 6 0 Mean: 2.0 3856. 68 3856. 37 500 Fel 4 3.25 31. 3 (3) 32 3 4 1 3858. 38 5 1 3858. 63 6 0 Mean: 0.8 3858. 52 3958. 30 800 Ni I 32 3.63 32. 3 3 3859. 79 4 3 3859. 97 5 2 3859. 69 6 1 + Mean: 3859. 81 3859. 91 1000 Fel 4 a 2i 136 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6— Continued Line observed Blended Identification with line No. Order /obs Xoba No. Order Xtab /tab Mult. ev 33. 3 2 3872. 15 4 1 3872. 58 5 1 3872. 55 6 1 + Mean: 1.5 3872. 46 3872. 50 300 Fel 20 4. 19 34. 3 3 3878. 21 4 (5) 133 3 5 2 (3877. 47) 165 3 6 0 3878. 02 400 Fel 20 4. 16 Mean: 1. 8 8878. 21 387S. 58 300 Fe I 1 3. 28 35. 3 4 3886. 39 4 3 3886. 47 5 1 3886. 66 6 0. 5 + 3886. 80 125 Cr I 23 4. 19) Mean: 2.2 3886. 53 3886. 28 600 Fe I 4 3.24 36. 3 1 + 4 1 3892. 34 5 1 3891. 47 6 0.5 + Mean: 1. 1 3891. 86 3891. 93 100 Fel 733 6.60 37. 3 1 (3894. 17) 38 3 4 1 3893. 40 135 3 5 0.5 + 6 0 Mean: 0.6 3893. 40 3S93. 39 100 Fel 430 6. 13 38. 3 1 (3894. 17) 37 3 4 1 3894. 06 5 1 3894. 12 6 0.5 + Mean: 1. 1 3894. 09 3894. 07 1000 Col 34 4.23 39. 3 2 3895. 94 4 2 3895. 76 5 (3) (3897. 01) 166 3 6 1 + Mean: 1. 9 3895. 84 3895. 66 400 Fel 4 3. 29 40. 3 1 3898. 07 4 1 3897. 83 5 5 (3) / 39 \166 3 6 0.5 + Mean: 0.9 3897. 93 3897. 90 100 Fel 280 5.87 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 137 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order Job. ^obs No. Order Xt»b A«b Mult. ev 41 3 3 3899. 82 4 2 3899. 73 5 2 3899. 37 6 1 _j_ Mean: 3899. 60 3S99. 71 500 Fel 4 3.26 42 3 2 3902. 91 4 2 3903. 11 137 3 5 1 3903. G3 6 0 Mean: i. S 3903. 28 3902. 95 500 Fel 45 4.73 43 3 2 3905. 47 4 1 3906. 52 139 3 5 0. 5 + 6 0. 5 + Mean: 1. 1 3906. 07 3906. 48 300 Fel 4 3.28 44 3 2 3920. 46 4 3 3920. 26 5 0.5 + 6 0 Mean: 1.3 3920. So 3920. 26 500 Fel 4 3. 28 45 3 3 3922. 97 4 3 3922. 85 5 2 3922. 97 6 1 + Mean: 2. 6 3922. 93 3922. 91 600 Fel 4 3.21 46 3 (6) 156 2 4 3 3927. 90 5 1 3927. 74 6 (6) 156 4 Mean: 9 9 3927. 81 3927. 92 500 Fel 4 3.26 47 3 (5) 50 3 4 3 3930. 34 5 1 3930. 49 6 (6) 157 4 Mean: 2. 2 3930. 42 3930. 30 600 Fel 4 3.24 /50 3 48 3 (5) \47 3 4 0.5 3931. 66 5 0. 5 + 6 0 Mean: 0.4 3941. 66 3931. 12 35 Fel 565 6.42 138 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order lob* Xoba No. Order ^tab 7, >b Mult. ev /50 3 49. 3 (5) 4 0.5 3932. 63 \47 3 5 0.5 + 6 0 /280 5.85 3932. 63 3932. 63 80 Fel Mean: 0.4 \652 6. 40 50. 3 5 3933. 64 4 4 3933. 71 5 3 3933. 54 /(488 6.22) 6 200 Fel 3 3933. 55 (3933. 61 \ 562 6.42) Mean: 4.6 3933. 60 3933. 67 600 Call 1 3. 15 51. 3 1 3944. 25 4 0.5 + 5 0.5 + 6 0 Mean: 0.5 3944. 25 3944. 01 2000 All 1 3. 14 52. 3 0.5 + 4 1 3956. 33 3957. 05 80 Cal 6 5.02 5 0 3957. 03 50 Fel 562 6.40 6 0 3956. 68 150 Fe I 278 5.80 Mean: 0.6 3956. 33 3956. 46 100 Fel 604 6.37 53. 3 1 3961. 20 4 1 3961. 51 5 1 3962. 29 6 0.5 + Mean: /. 1 3961. 76 3961. 52 3000 All 1 3. 14 54. 3 5 3968. 45 4 4 3968. 51 5 3 3968. 38 6 3 3968. 08 3969. 26 600 Fel 43 4.59 Mean: 4.6 3968. 32 3968. 47 500 Call 1 3. 12 55. 3 1 + 4 0.5 + 5 0 6 1 3984. 68 3985. 39 125 Fel 661 6.41 Mean: 0.8 3984. 68 3983. 96 200 Fel 277 5.83 56. 3 2 4005. 38 4 2 4005. 18 145 3 5 1 + 6 0 Mean: 1.3 4005. 27 4005.25 250 Fel 43 4. 65 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 139 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order hbs Xob, No. Order Xtab / tab Mult. ev. 57. 3 3 4030. 78 4 3 4030. 73 5 2 /58 5 + \59 5 6 0.5 (4030. 50 120 Fel 560 6.26) Mean: 2. 1 4030. 75 4030. 76 500 Mnl 2 3.08 58. 3 2 4032. 97 4 2 4033. 13 5 2 /57 5 + \59 6 0.5 5 Mean: /. 5 4033. 06 4033. 07 400 Mnl 2 3.08 59. 3 2 4034. 46 4 2 4034. 41 5 2 /57 5 6 0. 5 + \58 5 Mean: 1.5 4034. 43 4034. 49 250 Mnl 2 3.08 60. 3 (4) 61 3 4 1 4042. 55 5 0 6 0 ? Mean: 0.3 4042. 55 4041. 36 100 Mnl 5 5. 18 61. 3 4 4045. 97 4 3 4045. 85 5 2 + 6 1 -f Mean: 2.8 4045. 90 4045. 82 400 Fel 43 4 55 62. 3 3 4063. 69 4 2 4063. 68 5 1 4064. 05 6 0 4063. 53 100 Mnl 5 5. 21 Mean: 1.6 4063. 84 4063. 60 400 Fel 43 4.61 63. 3 1 + 4 1 4066. 21 5 1 4065. 33 6 0 4066.60 40 Fel 424 6.03 Mean: 0.8 4065. 72 4066. 98 100 Fel 358 5.87 64. 3 3 4071. 71 4 3 4071. 86 5 0.5 -f 6 0.5 + Mean: 1.8 4071. 80 4071. 74 300 Fel 43 4.65 140 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order 7ob. Xob8 No. Order Xtab /tab Mult, ev 65. 3 1 4105. 13 4 1 4104. 90 5 0. 5 + 6 0 Mean: 0. 6 4105. 00 4104.13 100 Fel (j£J J * 66. 3 1 4127. 82 4 0. 5 + 5 0. 5 + 6 1 4126. 33 161 4 4126. 19 80 Fe I 695 6. 34 Mean: 1. 0 4126. 83 4127. 61 100 Fe I 357 5. 86 67. 3 2 4132. 02 4 1 4132. 11 5 (5) 132 4 6 0. 5 + Mean: 1. 2 4132.07 4132.06 300 Fe I 43 4. 61 68. 3 (3) 69 3 4 1 4143. 31 5 (2) 69 5 6 (1) 69 6 Mean: 1. 0 4143. 31 4143.42 200 Fe I 523 6. 03 69. 3 3 4143.80 4 3 4143. 67 5 2 4143. 96 6 1 4144.07 Mean: 2. 6 4143. 91 4143. 87 400 Fe I 43 4. 55 70. 3 1 4154. 46 4 0. 5 + 5 0 4153.91 120 Fe I 695 6. 38 6 0 4154. 81 100 Fe I 694 5. 81 Mean: 0. 3 4154. 46 4154.50 100 Fe I 355 6. 35 71. 3 1 4199. 12 4 1 4198. 99 5 0 (4199.97 - Fe I 3 3. 03) 6 0 4198. 31 250 Fe I 152 5. 35 Mean: 0. 4 4199.05 4199. 10 300 Fe I 522 5. 99 72. 3 3 4202.02 4 2 4201. 82 5 1 + 6 1 + Mean: 2. 0 4201. 91 4202.03 400 Fe I 42 4. 44 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 141 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order Job8 \ohs No. Order *tab /tab Mult. ev 73. 3 1 + 4 1 4206. 87 5 1 + 6 0 (4207. 13 80 Fe I 352 5. 77) Mean: 0.8 4206.87 4206. 70 125 Fe I 3 3.00 74. 3 1 4210. 66 4 1 4210. 21 5 (5) 143 4 6 1 4210. 42 Mean: 1. 3 4210. 41 4210. 35 300 Fe I 152 5. 42 75. 3 1 4213. 42 4 0. 5 4213. 82 5 (5) 143 4 6 0. 5 4213. 73 Mean: 0. 8 4213. 69 4213.65 100 Fel 355 5.78 76. 3 3 4216.02 4 2 4216. 24 5 (5) 143 4 6 1 + Mean: 2. 1 4216. 15 4216. 19 200 Fe I 3 2.94 77. 3 5 4226. 74 4 4 4226. 78 5 3 4227. 00 6 3 4226. 98 4227. 43 300 Fe I 693 6. 25 Mean: 4. 6 4226. 90 4226. 73 500 Ca I 2 2.93 78. 3 1 + 4 1 4242. 69 5 0.5 + 6 1 4242. 91 1 Mean: 1. 1 4242. 82 79. 3 1 + 4 1 4245. 09 5 0. 5 + 6 0.5 + 4246. 09 80 Fe I 906 6. 56 Mean: 0. 9 4245.09 4245. 26 80 Fe I 352 5. 77 80. 3 1 + 4 1 4247. 19 5 0.5 + 6 1 4247. 51 4248. 23 150 Fe I 482 5. 98 Mean: 1. 1 4247. 88 4247.43 200 Fe I 693 6. 29 142 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order 7ob9 Xob3 No. Order Xt.b Jtab Mult, ev 81. 3 (2) + 82 3 4 (2) + 82 4 5 0.5 + 6 1 4249. 73 Mean: 1. 4 4249. 73 4250. 13 250 Fe I 152 5.38 82. 3 (2) + 81 3 4 (2) + 81 4 5 1 4251. 13 6 1 4250. 92 Mean: 1.8 4251.02 4250. 79 400 Fe I 42 4.47 83. 3 3 4254. 48 4 2 4254. 43 5 1 4254. 33 6 1 4254 42 Mean: 2.0 4254. 41 4254. 35 5000 Cr I 12. 91 84. 3 1 + 4 1 4268. 96 5 1 4269. 07 6 1 4268. 76 4268. 74 30 Fe I 649 6. 20 Mean: 1. 3 4268. 92 4267. 83 125 Fe I 482 6.01 85. 3 4 4271. 30 4 3 4271.80 5 3 4272. 20 6 2 4271. 67 4271. 16 400 Fe I 152 5. 35 Mean: 3. 7 4271. 78 4271. 76 1000 Fe I 42 4.39 86. 3 3 4274.33 4 2 4274. 68 5 2 4275. 03 6 1 4274. 70 Mean: 2.4 4M4- 73 4274.80 4000 Cr 1 1 2.90 87. 3 1 4282. 23 4 1 4282. 47 5 0 6 0 (4283. 01 40 Ca I 5 4. 78) Mean: 0. 4 4282. 37 4282.41 600 Fe I 71 5.07 88. 3 1 + 4 1 4285. 92 5 0 6 0 Mean: 0. 4 4285. 92 4285.45 125 Fe I 597 6. 12 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 143 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /ob, Xob» No. Order Xt.b / * Mult. ev 89. 3 3 4289. 64 4 2 4289. 71 5 1 4289. 81 6 0.5 + (4289. 36 35 Cal 5 4.77) Mean: /. 7 4289. 73 4289. 72 3000 Crl 1 2.89 90. 3 3 4293. 94 4 2 4294. 20 5 1 4294. 09 6 1 + Mean: #. 0 4294- 09 4294. 13 700 Fel 41 4.37 91. 3 1 4299. 89 4 1 4299. 93 5 0.5 + 6 0 Mean: 0. 6 4299. 91 4299. 24 500 Fel 152 5.29 92. 3 1 + 4 1 4302. 48 5 1 4301. 64 6 0 4302. 53 50 Cal 5 4.78 Mean: 0.8 4302. 01 4302. 19 50 Fel 520 5.92 93. 3 1 + 4 1 4305. 16 5 0 6 0 Mean: 0.4 4305. 16 4305. 46 100 Fel 476 5.89 94. 3 3 4308. 25 4 3 4307. 90 5 2 4307. 96 6 2 4308. 03 (4307. 74 45 Cal 5 4.76) Mean: 3. 1 4308. 02 4307. 91 1000 Fel 42 4.44 95. 3 0 4 0 5 1 4312. 70 6 0 Mean: 0.4 4312. 70 4312. 55 100 Mnl 23 5.81 /97 3 96. 3 1.5 + \98 3 4 1 4314. 95 5 (4) + 148 4 6 0 Mean: 0.5 4314- 95 4315. 09 500 Fel 71 5.07 144 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /obs No. Order ^tab »b Mult. ev /96 3 97. 3 1.0 + \98 3 4 1 4317. 98 5 (4) + 148 4 6 0 Mean: 0.5 4317.98 4318. 65 60 Cal 5 4.77 /96 3 98. 3 1. 5 + \97 3 4 1 4319. 31 5 (4) -f 148 4 6 0 4320. 59 125 CrI 96 5.76 Mean: 0.5 4319. 31 4319. 64 100 Crl 06 5.75 99. 3 4 4326. 07 4 3 4325. 66 5 2 4325. 58 6 1 4325. 96 Mean: 2.8 4325. 81 4325. 76 1000 Fel 42 4.47 100. 3 0 4 0. 5 + 5 0.5 + 6 1 4374. 78 Mean: 0.8 4374- 78 4374. 95 150 Mn I 5.69 101. 3 5 4376. 56 4 4 4376. 08 5 3 4375. 96 6 3 4376. 00 Mean: 4376. 10 4375. 93 500 Fel 2 2.83 102. 3 5 4383. 74 4 4 4383. 51 5 3 4383. 53 6 3 4383. 60 3/ean: 4. £ 4383. 68 4383. 55 1000 Fel 41 4.31 103. 3 0. 5 4. 4 1 4392. 79 5 1 4393. 72 6 (4) -f. 143 5 1 Mean: 1.0 4393. 31 4393. 45 20 Nal 17 4. 92 104. 3 1 + 4 1 4401. 48 5 1 + Mean: A / 4401. 48 4401. 55 1000 Nil 86 6.00 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 145 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. *i)rder IobS Xob8 No. Order ^tab /tab Mult. ev 105. 3 3 4404. 14 4 3 4404. 64 5 3 4404. 72 Mean: 3. 3 4404. 55 4404. 75 1000 Fel 41 4.37 106. 3 2 4406. 58 4 1 4407. 43 5 0.5 + Mean: 1. 1 4407. 07 4407. 72 100 Fel 68 4.99 107. 3 1 4409. 10 4 0. 5 4409. 47 5 0 Mean: 0.4 4409. 31 4408. 42 125 Fel 68 5.01 108. 3 2 4414. 88 4 (6) + 156 3 5 2 4415. 29 4414. 88 150 Mn I 22 5.69 Mean: 2. 3 4415.14 4415. 12 600 Fel 41 4.42 109. 3 1 4421. 81 4 (6) + 157 3 5 1 4421. 90 Mean: 1. 2 4421. 87 4422. 57 300 Fel 350 5. 64 110. 3 3 Mil. 50 4 3 4427. 31 5 4 4427. 37 Mean: 3.8 4427. 38 4427. 31 500 Fel 2 2.85 111. 3 0.5 + 4 0. 5 + 5 1 4456. 11 Mean: 0.8 4456.11 4455. 89 100 Ca I 4 4.68 112. 3 1 + 4 1 5 0 4458. 31 Mean: 0.6 4459. 12 400 Fel 6S 4.95 4458. 31 113. 3 2 4461. 39 4 3 4461. 72 5 3 4461. 58 Mean: 3.0 4461. 58 4461. 65 300 Fel 2 2. 86 114. 3 0.5 + 4 1 4466. 44 5 1 4466. 23 Mean: 1.0 4466. 32 4466. 55 500 Fel 350 5.60 146 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /ob. Xob» No. Order Xt,b /t»b Mult, ev 115. 3 0. 5 + 4 1 4471. 81 5 0 Mean: 0. 5 4471.81 4472. 79 100 Mn I 22 5. 72 116. 3 0. 5 + 4 1 4475. 49 5 0 Mean: 0. 5 4475. 49 4476.02 500 Fe I 350 5. 61 117. 3 1 4481. 65 4 2 4481. 71 5 2 4482. 21 4481. 33 100 Mgll 4 11. 63 Mean: 1. 9 4481. 90 4482. 17 150 Fe I 2 2. 82 118. 3 0 4 1 4578. 59 5 0 Mean: 0. S 4578. 59 4578. 56 80 Ca I 23 5. 23 119. 3 1 + 4 2 4601. 81 /l58 3 \ 2 5 5 1 + Mean: 1.4 (4601.81) 4602. 94 300 Fe I 39 4. 18 120. 3 1 4829.32 4 1 4828. 87 5 0. 5 4828. 97 4829. 38 200 Cr I 31 5. 11 Mean: 0. 8 4829. OS 4829.03 300 Ni I 131 6. 10 121. 3 1 + 4 1 4891. 23 5 0 4891. 50 70 Fe I 318 5. 38 Mean: 0. 6 4891. 23 4890. 76 100 Fe I 318 5.41 122. 3 0 4 1 4902. 60 5 0 Mean: 0. 3 4902. 60 4903. 32 500 Fe I 318 5.41 123. 3 0. 5 + 4 1 4957. 32 5 0. 5 + 4957. 60 300 Fe I 318 5. 31 Mean: 0. 7 4957. 32 4957.30 100 Fel 318 5.35 124. 3 0. 5 + 4 0. 5 + 5 1 4965. 99 Mean: 0. 8 4966. 99 4966. 10 300 Fe I 687 5. 82 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 147 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order /„bs Xob. No. Order ab Mult. ev 125. 3 1 4984. 41 4 0.5 -f 5 1 4984.24 4984. 13 500 Nil 143 6.29 Afean: 0. 4984. SO 4983. 86 200 Fel 1066 6.59 128. 3 1 5005. 96 4 1 5005. 96 5 1 -|- 5006. 13 300 Fel 318 5.31 Mean: 5005. 96 5005. 72 200 Fel 984 6.36 *• 127. 3 0.5 + 4 2 5012. 14 5 1 5011. 91 Afean: 1.3 6012. 01 5012. 07 300 Fel 16 3.33 128. 3 2 + 15 4 4 1 5 1 5098. 87 Afean: 1.4 5098. 87 5098. 70 200 Fel 66 4. 61 129. 3 1 + 16 4 4 1 5101. 89 5 1 5102. 37 Afean: 1.1 5102. 16 5102. 24 SO Fel 65 4.63 130. 3 1 5109. 93 4 2 5110. 30 5 1 5110. 39 Afean: 1.4 5110. 24 5110. 41 300 Fel 1 2.42 131. 3 (3) + 30 4 /1092 6.64 4 1 5142. 21 5142. 54 100 Fe I \1090 6.68 5 0 5141. 75 100 Fel 114 4.83 Afean: 0.5 5142. 21 5142. 93 125 Fel 16 3.36 132. 3 5 + 133 3 4 5 5167. 33 5167. 49 700 Fel 37 3. 89 5 3 5167. 38 (5166. 29 125 Fel 1 2.40) Afean: 4-5 6167. 36 5167. 32 100 Mgl 2 5. 11 133. 3 5 + 132 3 4 5 5172. 64 5 3 5172. 66 (5171. 60 300 Fel 36 3.88) Afean: 4-5 5172.65 5172. 68 200 Mgl 2 5. 11 134. 3 5 + 4 5 5183. 72 5 3 5183. 60 Afean: -4-5 6183. 65 5183. 60 500 Mgl 2 5. 11 148 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line Mult. ev No. Order /ob* Xobs No. Order / ab 135. 3 1 5191. 25 37 4 4 1 5191. 42 5 1 99 6 Mean: 1. 1 5191. 35 5191. 46 400 Fe I 383 5.42 136. 3 1.5 5195. 20 4 1 5195. 25 5 2 5195. 15 (5195. 47 100 Fe I 1092 6.61) Mean: 1.7 5195. 20 5194. 94 200 Fe I 36 3.95 137. 3 2 5204. 21 42 4 4 2 5205. 20 5 1 5204. 53 5204. 58 125 Fe I 1 2.47 Mean: 1. 7 5204- 64 5204. 52 400 Cr I 7 3.32 138. 3 1 5205. 76 4 2 5206. 34 5 1 5205. 63 Mean: 1.4 5205. 90 5206. 04 500 Cr I 7 3.32 139. 3 1 5208. 71 43 4 4 2 5208. 44 5 1 5208. 28 5208. 60 200 Fe I 553 5.62 il/ecm: ./. A 5208. 44 5208. 44 500 Cr I 7 3.32 140. 3 1 + 4 1 5216. 28 5 0 Mean: 0.(5 5216. 28 5216. 28 300 Fe I 3G 3.99 141. 3 3 5227. 03 4 3 5227. 25 5 2 5227. 13 5226. 87 200 Fe I 383 5.41 Mean: 5227. 14 5227. 19 400 Fe I 37 3.93 142. 3 2 5229. 48 4 1 5228. 92 5 1 5228. 87 1090 6.56 5229. 86 200 Fe Mean: 1.4 5229. 04 < 553 5.63 143. 3 5 5269. 42 4 5 5269. 54 (5270.27 20 Ca I 22 4.88) 5 4 5269. 49 5270. 36 400 Fe I 37 3.96 Mean: 5.0 5269. 49 5269. 54 800 Fe I 15 3. 21 144. 3 5 5328. 75 4 5 5328. 08 5328. 53 150 Fe I 37 3.87 Mean: 5328. 37 5328. 04 400 Fe I 15 3.23 145. 3 2 5339. 98 56 4 4 2 5340. 04 Mean: f. « 5340. 01 5339. 94 200 Fe I 553 5.58 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 149 TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line No. Order Job, Xob. No. Order *t«b /t«b Mult, ev 146. 3 1 5341.08 4 (3) + 85 5 5341.06 200 Mn I 4 4.44 Mean: 0. 8 6341. 08 5341. 03 200 Fe I 37 3.91 147. 3 3 5371.26 4 4 5371.40 Mean: 3. 2 6371. 34 5371.49 700 Fe I 15 3. 26 148. 3 3 5397. 07 4 4 5397. 10 Mean: 3. 2 5397. 09 5397. 13 400 Fe I 15 3. 21 149. 3 3 5405.59 4 4 5405. 87 Mean: 3. 2 5405. 75 5405.78 400 Fe I 15 3.28 150. 3 4 5430. 11 4 4 5430.02 Mean: 3. 6 5430. 06 5429. 70 500 Fe I 15 3. 24 151. 3 2 5434.42 4 3 5434.41 Mean: 2. 3 6434- 41 5434. 53 300 Fe I 15 3. 29 152. 3 4 5447. 19 4 4 5446.97 Mean: 3. 6 5447. 06 5446.92 300 Fe I 15 3. 25 153. 3 3 5455.83 4 3 5455. 77 Mean: 2. 7 5456. 80 5455. 61 300 Fe I 15 3. 28 154. 3 0. 5 + 4 1 5497. 71 Mean: 0. 7 5497. 71 5497. 52 150 Fe I 15 3. 26 155. 3 0. 5 + 4 1 5615. 63 Mean: 0. 7 5615.63 5615.65 400 Fe I 686 5.54 156. 2 6 5890. 15 3 6 5889.95 4 6 5889. 95 Mean: 5. 0 5889.99 5889.95 9000 Na I 1 2.11 157. 2 6 5895. 66 3 6 5895. 92 4 6 5895. 97 Mean: 6. 0 5895.88 5895. 92 5000 Na I 1 2. 10 150 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Observed wavelengths and identifications of meteor spectrum S 6—Continued Line observed Blended Identification with line ev No. Order /obs X* No. Order >lab Mult. 158. 2 1 6137. 79 /119 4 3 (2) 6136. 80 r \ 2 6137. 70 100 Fe I •JO 7 4. 61 4 1 6135. 77 Mean: 0.8 159. 2 1 + 3 0 4 1 6154. 11 Mean: 0. 6 6154. 11 6154. 23 500 Na I 5 4. 12 160. 2 1 6162. 56 3 (2) 6162. 34 :i f> 4 1 4- 6162. 17 40 Ca I 3 3. 91 Mean: 0.8 6162. 43 6160. 76 500 Na I 5 4. 12 161. 2 1 6191. 10 3 0. 5 4- 4 1 6190. 17 66 6 6191. 19 500 Ni I 45 3.67 Mean: 0. 7 6190. 48 6191. 56 100 Fe I 169 4.44 162. 2 (3) + 86 3 3 1 6411.96 4 1 6411.36 Mean: 0.9 6411. 62 6411. 66 100 Fel SI 6 5. 58 163. 2 (3) 89 3 3 1 6430. 24 4 0 Mean: 0.4 6430. 24 6430. 85 100 Fel 62 4. 11 164. 2 /89 3 \90 3 3 1 6437. 87 4 1 6437. 60 ? Mean: 0.9 6437. 72 6439. 07 150 Cal 18 4. 45 165. 2 3 6462. 22 3 2 6462. 39 34 5 4 1 4- Mean: /. 6 6462. 32 6462. 57 125 Ca I 18 4.44 166. 2 2 6495. 37 3 (3) 6494. 98 39 5 4 1 6495. 15 Mean: 1.2 6495. 14 6494. 98 400 Fel 168 4.31 METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 151 The spectrum of this meteor was photographed TABLE 3.—Spectrum of meteor No. 27471 in the zero order, in the first order with dis- [The lines with numbers from 1 to 22 are present in the second order only. persion of 72 A/mm, and in the blue part of The lines with numbers from 23 to 33 are present in the first and second order. If one measured wavelength only is presented, it refers to the the second order with 36 A/mm. [The dis- first order. All other lines from Nos. 34 to 99 are present in the first order persions given by Rajchl (1961) are those of only. The dispersion of the meteor spectrum in the first order is approxi- mately 72 A/mm and In the second order 3d A/mm. The intensity scale the spectograph and not of the meteor spectra.] is the same as is used in table 2.] The spectral lines of the Na doublet are the Line observed Identification most intense lines and are separated one from No. /ob. Xobs ^t»b Mult another in the first order. The beginning of 1 1 3679. 6 3679. 9 Fel 5 these lines is almost the same as the beginning 2 1 3681. 1 3682. 2 Fel 772 of the meteor trail photographed by normal 3 1 3683.5 3683.0 Fel 5 cameras. The whole spectrum is photographed 4 1 3686. 8 3687. 5 Fel 21 5 1 3689. 6 3689. 5 Fel 369 in the height interval from 85 km to 56 km. 3689. 5 Fel 386 The spectral lines were measured at the maxi- 6 1 3705. 4 3705. 6 Fel 5 mum brightness at the height of 65 km, and the 7 1 3707. 3 3707. 8 Fel 5 wavelengths were computed by the method 8 4 3721. 3 3719. 9 Fel 5 briefly described in this paper. The wave- 9 1 3724. 1 3722. 6 Fel 5 lengths of all 99 measured lines are given in 10 1 3734. 7 3733. 3 Fel 5 11 2 3736. 1 3734. 9 Fel 21 table 3. 12 3 373R 4 3737. 1 Fel 5 The majority of the lines are iron lines 13 1 3743. 0 3743. 4 Fel 21 belonging to different multiplets. The strong- 14 3 3746. 8 3745. 6 Fel 5 3745. 9 Fel 5 est lines are: Na doublet 1; Mg triplet 2; Fe 15 2 3749. 8 374a 3 Fel 5 lines 3720 of multiplet 5; 3856 of multiplet 4; 3749. 4 Fel 21 3860 of multiplet 4; 4045 of multiplet 43; 16 1 3816. 6 3815. 8 Fel 45 4383 of multiplet 41; 5269 of multiplets 15 and 17 1 3818. 8 3816. 2 Crl 40 37. Other elements observed are Ca I and 3817. 8 CrI 40 Ca II, Mg I, Ni I, Cr I, Mn I. The presence 3818. 5 Crl 40 18 2 3821. 3 3820.4 Fel 20 of Co I and Si I is questionable. The elements 19 2 3825. 4 3825. 9 Fel 20 usual in meteorites and air but not observed in 3824. 4 Fel 4 this spectrum are: Al, Cu, P, S, C, K, N, O. 20 1 3830. 1 3829. 4 Mgl 3 The iron line 3930.3 is well separated from 21 2 3833. 1 3832. 3 Mgl 3 22 1 3835. 0 3834. 2 Fel 20 the Ca II line 3933.7. This Ca line, in com- 23 3 3838.9 1 3838. 3 Mgl 3 parison with spectrum S 6, is relatively weak, a 3 3839. 2 J fact that may be connected, as previously 24 4 3856. 5 1 3856. 4 Fel 4 indicated, with the iron composition of the body. 2 3857. 0 J A feature of this Ca II line is quite different 25 4 3860. 8 1 3859.9 Fe I 4 3860. 6 J 3858. 3 Nil 32 from the other lines. It is observable in the 4 26 1 3866. 4 1 3865. 5 Fel 20 maxima of brightness only. In comparison 1 3865. 5 J with the others the changes of brightness of 27 1 3868. 2 t this line are very strong. The Ca II doublet is 28 2 3878. 0 1 387a 0 Fel 20 present in this spectrum, although the meteor 2 3879. 0 J 387a 6 Fel 4 can be classified as slow. 29 3 3885. 6 1 3886. 3 Fe I 4 3 3887. 2 j 3886. 8 Crl 23 Artificial spectra of meteoritic material using 30 1 3891. 5 3891. 9 Fel 733 shock tubes (Nicholls, et al., 1959) also show 31 1 3894. 4 3895. 7 Fel 4 the Ca II doublet at velocities less than 10 3894. 1 Co I 34 32 2 3899. 2 1 km/sec. Thus it is possible to conclude from 3899. 7 Fel 4 1 3900. 2 J all these experimental results that the Ca II 33 2 3905.2 1 3906. 5 Fe I 4 doublet is emitted not only by high-velocity 1 3905.9 J 3905. 5 Si I 3 meteors but also by slow meteors. 34 1 3913. 1 3913. 6 Fel 120 152 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 3.—Spectrum of meteor No. 27471—Continued TABLE 3.—Spectrum of meteor No. 27471—Continued Line observed Identification Line observed Identification No. /ob. Xob. Atab Mult No. lob* Xobs Xt»b Mult 35 3 3921. 3 3922. 9 Fel 4 73 2 4462. 2 4461. 6 Fel 2 3920. 3 Fel 4 74 1 4482. 5 4482. 2 Fel 2 36 3 3928. 4 3927. 9 Fel 4 4481. 3 Mgll 4 3930. 3 Fel 4 4481. 1 Mgll 4 37 3 3932. 8 3933. 7 Call 1 75 1 4835. 6 ? 3933. 6 Fel 488 76 1 4878. 4 4878. 2 Fe I 318 3933. 6 Fel 562 4878. 1 Cal 35 38 1 3948.9 3948. 8 Fel 604 77 1 4932. 6 ? 39 3 3967. 7 3968. 5 Call 1 78 1 4942. 0 ? 3969. 3 Fel 43 79 1 4989. 6 4989. 0 Fel 1066 40 1 4004. 5 4005. 2 Fe I 43 4988. 0 Col 41 2 4032. 0 4030. 8 Mn I 2 80 1 5011. 9 5012. 1 Fel 16 4033. 1 Mn I 2 81 1 5132. 5 5133. 7 Fe I 1092 4034. 5 Mn I 2 82 4 5168. 2 5168. 8 Fel 1 42 4 4045. 2 4045. 8 Fe I 43 5167. 5 Fel 37 43 1 4052. 2 ? 5167. 3 Mgl 2 44 1 4056. 6 4055. 0 Fel 218 83 4 5172. 4 5172. 7 Mgl 2 4056. 5 Fel 320 5171. 6 Fe I 36 45 1 4058. 4 4058. 0 Mn I 29 84 4 5183. 8 5183. 6 Mgl 2 4059. 4 Mn I 29 85 1 5206. 7 5208. 4 Crl 7 46 3 4063. 1 4063. 6 Fel 43 5206. 0 Crl 7 47 3 4070. 8 4071. 7 Fel 43 5204. 5 Crl 7 48 1 4078. 4 4076. 6 Fel 558 5204. 6 Fel 1 4078. 4 Fel 217 86 1 5227. 1 5227. 2 Fel 37 4080. 2 Fel 558 87 4 5270. 0 5270. 4 Fe I 37 49 1 4131. 6 4132. 1 Fel 43 5270. 3 Cal 22 50 2 4143. 2 4143. 9 Fel 43 5269. 5 Fel 15 4143. 4 Fel 523 88 1 5319. 6 ? 51 1 4200. 4 4198. 3 Fel 152 89 1 5321. 4 5322. 0 Fel 112 4199. 1 Fe I 522 90 2 5327. 9 5328. 0 Fel 15 4202. 0 Fel 42 5328. 5 Fel 37 52 1 4216. 6 4216. 2 Fel 3 91 1 5396. 3 5397. 1 Fel 15 53 1 4227. 0 4226. 7 Cal 2 92 1 5405. 9 5405. 8 Fel 15 4227. 4 Fe I 693 93 1 5542. 3 ? 54 1 4250. 7 4250. 1 Fel 152 94 1 5572. 0 5572. 8 Fel 686 4250. 8 Fel 42 5569. 6 Fel 686 55 2 4254. 5 4254. 4 Crl 1 95 1 5622. 2 5623. 6 Fel 625 56 1 4260. 8 4260. 5 Fel 152 96 1 5681. 1 5682. 6 Nal 6 4261. 4 Crl 96 97 1 5882. 8 ? 57 3 4271. 6 4271. 2 Fel 152 98 5 5889. 9 5889. 9 Nal 1 4271. 8 Fel 42 99 4 5896. 2 5895. 9 Nal 1 58 1 4290. 0 4289. 7 Crl 1 59 1 4294. 2 4294. 1 Fe I 41 A wave feature in some spectral lines was 60 3 4308. 3 4307. 9 Fel 42 4307. 7 Ca I 5 observed in both spectra. It reaches up to 61 1 4320. 4 4319. 6 Crl 96 about 1 A in the spectrum of meteor 27471, 4320. 6 Cr I 96 and up to about 0.3 A in the spectrum S 6. 62 1 4323. 0 4322. 7 Fe I 215 We have previously observed the wave-feature 63 3 4325. 8 4325. 8 Fel 42 of some spectral lines in the low dispersion 64 1 4332. 9 4331. 6 Ni I 52 65 1 4368.2 4367. 9 Fel 41 spectrograms of meteors where it reached up to 66 1 4371. 3 4369. 8 Fe I 518 several Angstrom. A provisional explanation 4371. 3 Crl 22 of this phenomenon may lie in the blending 67 3 4376. 5 4375. 9 Fel 2 of very narrow lines, when one line and another 68 4 4383.3 4383. 6 Fel 41 change places as a result of local variations in 69 1 4395. 4 ? 70 2 4404. 5 4404. 8 Fel 41 microsensitivity of the photographic emulsion. 71 2 4414. 6 4415. 1 Fel 41 The spectrum of meteor No. 27471 is suitable 72 2 4427. 8 4427. 3 Fel 2 for study of spectral changes during the trajec- METEOR SYMPOSIUM—CEPLECHA AND RAJCHL 153 tory. We know accurately the changes of veloc- JACCHIA, L. G., AND WHIPPLE, F. L. ity during the trajectory of this meteor. As 1961. Precision orbits of 413 photographic me- a result, we have a dynamic determination of teors. Smithsonian Contr. Astrophys., vol. 4, no. 4, pp. 97-129. the heat-transfer coefficient and of the mass ablation. Thus it will be possible to determine MOOHE, C. E. 1945. A multiplet table of astrophysical interest. the monochromatic luminosity coefficients. Contr. Princeton Univ. Obs., no. 20, This is a preliminary report on the subject. parts 1, 2. Detailed data will be published in Bulletin of NICHOIXS, R. W.; WATSON, M. D.; and PARKINSON, the Astronomical Institute of Czechoslovakia. W. H. 1959. The laboratory excitation of meteoritic References spectra in shock tubes. Journ. Roy. CEPLECHA, Zd. Astron. Soc. Canada, vol. 53, pp. 223- 1951. The centre of a plate. Bull. Astron. Inst. 231. Czechoslovakia, vol. 2, pp. 105-106. RAJCHL, J. 1954. Meteor photographs. III. A reduction 1961. Two meteor spectra with high dispersion. method of meteor negatives. Bull. As- Bull. Astron. Inst. Czechoslovakia, vol. tron. Inst. Czechoslovakia, vol. 5, pp. 12, p. 167. 9-13. 1958. On the composition of meteors. Bull. ROWLAND, H. A. Astron. Inst. Czechoslovakia, vol. 9, pp. 1893. Gratings in theory and practice. Astron. 154-159. and Astrophys., vol. 12, pp. 129-149. 1961. Determination of wavelengths in meteor ZAJDEL, Z. N.; PBOKOFEV, V. K.; and RAJSKIJ, S. M. spectra by using a diffraction grating. 1952. Tables of spectrum lines. Moscow. Bull. Astron. Inst. Czechoslovakia, vol. (English translation published by VEB 12, pp. 246-250. Verlag Technik, Berlin, 1961.) Abstract Analysis of photographs of two meteor spectra with dispersions from 72 A/mm to 11 A/mm is presented in this paper. The method of reduction is briefly described. Absolute wavelengths were computed without previous identification of any line, and the zero order of stars for determination of the coordinate system was used. The mean deviation of computed wavelengths from the wavelengths of identified lines is only ± 0.2 A in the case of spec- trum S 6. The lines of neutral iron predominate in both the meteor spectra. A Short Note on Meteor Spectra With Low Dispersionl By J. Rajchl2 Up to the present, 16 spectra of meteors and it is a well-known fact that in flares the with low dispersion (about 200 A/mm in the emission of ionized Ca and Mg atoms prevails. blue region, and about 800 A/mm in the red For example, in our spectral material there are region of the spectrum) have been photo- photographs of two meteors that emit light graphed at the Ondfejov Observatory. It is only in the red region of the spectrum. But evident from these spectra that the emissions when a flare occurs, very strong Ca II emission of ionized Ca and Mg for some meteors are is observable. visible from the beginning of the meteor. Thomas and White (1953) reported that the For others, these emissions are absent, although Ca II emission is absent in the spectra of as is shown from rocket explorations, ionized Giacobinids, and connected this with the low Ca and Mg atoms of meteoric origin are pre- velocity of the shower. Let us suppose, as the sent at heights of about 100 km (Istomin, 1961). highest limit, the same meteoroid diameter for From direct photographs it was possible to Giacobinids as for our meteors, and let us use determine the fundamental dynamical and the end heights determined for Giacobinids. physical characteristics for three spectra of this Then we find that these meteors emit light collection. Using the mass determined from above the region of slip flow, where transition- the photometry and the decelerations, we flow conditions prevail, and thus according to calculated the diameter of the meteoroid on our results it is impossible to observe Ca II the assumption that the density of meteoric emissions. Similarly, for Super-Schmidt mete- material is 3.5 gm/cm3. Following Baker ors also emitting light above the slip flow, i.e., (1959), we determined the parameter B (i.e., in transition or free-molecule flow, it is possible the ratio of meteor-body diameter to the mean to admit from our point of view that the prin- free path of the emitted molecules with respect cipal blue emissions Ca II and Mg II must be to the oncoming molecules), and the flow con- absent and the faint meteors show redder. ditions of the interaction between meteoric This fact was recently pointed out by Ceplecha body and the surrounding atmosphere. The (1959), and was photoelectrically observed by Ca II and Mg II emissions appear when the J. Davis (private communication). It is inter- flow conditions change from transition flow to esting to point out that the height interval at slip flow. In the case when the slip-flow con- which ionized atoms begin to emit light is the ditions are already valid from the beginning of same as that found by Jacchia (1949) for a meteor light emission, then the ionized emissions slight increase of brightness shown by some occur from the beginning. This is fulfilled for photographic meteors, and explained by the the meteors we investigated in the height in- possible influence of the flayer. terval from 96 to 100 km. We may conclude that Baker's parameter B It is interesting to compare this result with is the more important one, governing the pro- my recent finding (Rajchl, 1960) that the slip- cesses of light emission and its efficiency T, flow conditions are also present when a sudden and perhaps is more important than the meteor increase in meteor brightness (flare) occurs; velocity itself. But it must be emphasized that our results are only preliminary and more 1 Details of this paper will be published in Bull. Astron. Inst., Czechoslovakia. observational material is required for con- 1 Astronomical Institute of the Czechoslovak Academy of Sciences, firmation. Ondfejov, Czechoslovakia. 155 156 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS References JACCHIA, L. G. BAKER, R. M. L., Jr. 1949. Photographic meteor phenomena and 1959. The transitional aerodynamic drag of mete- theory. Harvard Coll. Obs. Reprints, orites. Astrophys. Journ., vol. 129, ser. 2, no. 31. pp. 826-841. RAJCHL, J. CEPLECHA, Zd. 1960. A note to the problem of meteor flares. 1959. On the colour index of meteors. Bull. Bull. Astron. Inst. Czechoslovakia, vol. Astron. Inst. Czechoslovakia, vol. 10, 11, pp. 33-34. p. 39. ISTOMIN, V. G. THOMAS, R. N., and WHITE, W. C. 1961. Nitrogen ions in the earth's upper atmo- 1953. The physical theory of meteors. IV. sphere and nighttime ionization in the Inquiry into the radiation problem—a E layer. Dokl. Acad. Sci. U.S.S.R., vol. laboratory model. Astrophys. Journ., 137, p. 1102. vol. 118, pp. 555-567. Spectrographic Observations of Meteors in U.S.S.R. in 1957-1960 By E. N. Kramer,1 K. A. Liubarsky,2 and V. I. Ivanikov' The first photographs of meteor spectra mm; D:F=1:4.5, F=250 mm; D:F=1:4.5, were obtained at the Moscow Observatory by F=135 mm) with objective prisms, angles S. N. Blajko in 1904. After this only amateurs 16°, 18°, 30° and diameters 45 and 30 mm. All were engaged in meteor spectrography. How- the stations used panchromatic films. ever, some good spectra were obtained by In the period 1957-1960, 95 spectra were K. P. Staniukovich and G. O. Zateishchikov in obtained, as follows: 32 in Ashkhabad (Liubar- 1934, G. O. Zateishchikov in 1936, V. M. sky), 10 in Odessa (Kramer), 25 in Doushanbe Chernov in 1939, and K. A. Liubarsky in 1950. (formerly Stalinabad) (Ivanikov), and 28 in During the IGY comparatively powerful spec- Simferopol (Pushnoy). tral patrols came into operation at some The catalog of the spectra with the number observatories. of lines n > 5 is given in table 1 and is a partial In the Astrophysical Laboratory of the continuation of the list published by Asta- Physical-Technical Institute, Academy of Sci- povich (1958). ences of Turkmenian S.S.R., Ashkhabad, The dispersion on all the spectra is small. In U.S.S.R., the patrol consisted of 8 cameras the blue region of the spectrum it is equal to (D:F=1:2.5, F=250 mm) and 2 cameras 100-200 A/mm, and for the Simferopol spectra (D:F=1:5, F=500 mm) with objective as high as 1000 A/mm. Some of the meteors prisms made of flintglass, angle 20°, diameter have a small angle between the direction of 120 mm. In the Astrophysical Institute, the dispersion and the meteor trail. As a Academy of Sciences of Tadjik S.S.R., Dou- rule, the stars of spectral type A0-A5 do not shanbe, the patrol was designed under the appear on the negative near the meteor spec- direction of L. A. Katasev and was fully com- trum, and the irregular distortion of the wide- pleted in 1960. It consisted of 6 cameras angle objectives do not permit the use of these (D:F= 1:2.5, F=250 mm) with objective stars for purposes of comparison. This is the prisms made of flintglass, angle 25°, diameter reason why the stations in Ashkhabad and 250 mm. The calculation of the prisms was Odessa, independently of each other, used the published by Ivanikov (1957). The cameras method of the construction of the dispersion were equatorially mounted and clock driven. curve on the base of some line in the meteor The Odessa Astronomical Observatory ran a spectra, which may be easily identified on the patrol of four cameras (D:F= 1:4.5, F=125 negative. The fact that the identification is mm) with objective prisms, angle 17°, and correct has been checked by means of the diameter 45 mm. interval coordination of the identification of At the meteor station4 of the Simferopol the other lines (Liubarsky, 1961). For this section of VAGO, the patrol consisted of six purpose the list of the most widespread lines cameras of different types (D:F=1:2, F=125 in meteor spectra was employed. This method was used partially in Doushanbe. i Odessa Astronomical Observatory, U.S.S.R. 1 Astrophysical Laboratory, Ashkhabad, U.S.S.R. In Ashkhabad and Doushanbe the method * Astrophysical Institute, Doushanbe (formerly Btalinabad), of control based on the multiplets was used; U.S.8.R. • All Union Astronomical Geodetical Society, U.S.S.R. if we identify any line of a multiplet in a 638805—63 13 157 158 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Meteor spectra with the number of lines n >5 World No. List* Date Shower lines No. Observatory 1 1957 Aug. 6 Sporadic 10 255 Odessa 2 1958 Aug. 7 ? 10 264 Doushanbe 3 12 ? 54 279 Doushanbe 4 13 Perseid 11 281 Odessa 5 13 Sporadic 8 280 Odessa 6 Sept. 7 Sporadic 29 297 Ashkhabad 7 8 ? 12 Doushanbe 8 8 Sporadic 10 298 Odessa 9 9 Sporadic 18 299 Odessa 10 Oct. 21 Orionid 28 307 Ashkhabad 11 22 Orionid 8 308 Ashkhabad 12 1959 Aug. 6 Perseid 7 Ashkhabad 13 11 Perseid 5 Simferopol 14 12 ? 31 Doushanbe 15 12 Perseid 5 Simferopol 16 12 Sporadic 7 Simferopol 17 13 Perseid 9 Simferopol 18 13 Cassiopeid 8 Odessa 19 13 Perseid 7 Simferopol 20 13 Sporadic 7 Simferopol 21 1960 Mar. 28 Sporadic 5 Ashkhabad 22 June 17 Sporadic 6 Ashkhabad 23 18 Sporadic 33 Ashkhabad 24 26 Sporadic 10 Odessa 25 Aug. 3 Perseid 6 Ashkhabad 26 13 Cassiopeid 12 Odessa 27 Oct. 21 ? 25 Doushanbe • See P. Millman, Journ. Astron. Soc. Canada, vol. 53, pp. 271-276, 1959. spectrum, the other weaker lines of this mul- and K Ca II lines, but the ionization and ex- tiplet must also be in this spectrum. The citation potentials of silicon are very large Ashkhabad station used the method of selec- (about 20 ev). Neutral silicon is not identified. tion of spectral lines according to the order of Meanwhile any other identifications of this their appearance in comparison with the line were not found. potentials of excitation. In some spectra the presence of the line One spectrum in Odessa was reduced by near the well-known green line of oxygen obtaining the mercury spectrum under condi- (forbidden 5577 A) was noted. The identifi- tions that are closest to the condition for cation of this line as O I may be erroneous as obtaining the meteor spectrum (Kramer, 1959). in this region there is a fairly strong line of The Doushanbe station began the study of neutral calcium Ca I (23). the dispersion as a function of the point in Most of the spectra show a strong unresolved the field of view, with the purpose of estab- system N2 1 P.G. According to Liubarsky the lishing the standard dispersion curves for intensity of the nitrogen bands depends less every point. In all spectra the lines of neu- on the variations of the brightness of the tral and ionized metals were identified: iron, meteor than does the intensity of the metal calcium, magnesium, sodium, manganese, co- lines. The Ashkhabad station (1960, June 18) balt, and chromium. In the spectra of the obtained the meteor spectrum in which, except fast meteors the lines 6340-6350 A are frequent for N2 1 P.G., there is a strong system Na 2 and strong. They are identified by Cook and P.G. Fourteen bands are resolved in this Millman (1955) as Si II. This identification, system. however, is very strange as this line often The Doushanbe station (1961, February 9) appears earlier than the D Na I, and the H obtained a spectrum without atomic lines. METEOR SYMPOSIUM—KRAMER, LIUBARSKY, AND IVANIKOV 159 This spectrum consists of three unresolved For one spectrum (Odessa, 1958, August 12), systems of bands; one of these systems may be Komarov tried to find the relative intensities identified as N2 1 P.G., but the other two are of the lines on the assumption that the density unidentified. is proportional to the logarithm of the intensity The Odessa station (1956, October 8) ob- in the density interval 0.4-1.4. He found that tained the spectrum of a — 12m bolide (Kramer, in the flares the maximum intensity is displaced 1958) with cyclical pulsating brightness. The toward the blue region of the spectrum and pulsations continued through the bright bursts; hence the temperature rises. He also found X^, the bursts are independent of the pulsations. (the color of the meteor) by using the com- The cycles of the pulsations begin from sharp parison star a Lyrae. double maxima, which are followed by intervals Liubarsky noticed the strong asymmetry of of constant brightness. The brightness in- the profiles of certain lines (for example, H and creases and falls very rapidly. Toward the K Ca II). The asymmetry may perhaps be end of the trail the intervals between the caused by the presence of these lines in the components of the double maxima as well as meteor train and by drift of the train. It is between the maxima are shorter. A continuous unlikely that the blending of faint lines may spectrum appears in the maximum. cause such an asymmetry for the H and K lines, Kramer explains these pulsations by arrang- but for the weaker lines we must take this ing a cyclical regime at certain values of the possibility into account. velocity and sizes of a meteor body. Near the boiling point, when the vapor pressure References and external pressure are equal, a secondary ASTAPOVICH, I. S. shock wave arises from the surface of the body 1958. Meteor phenomena in the earth's atmos- and the melting layer instantly evaporates. phere. Publ. House for Physical and Mathematical Literature, Moscow. The bolide flashes up for a second time after COOK, A. F., AND MILLMAN, P. M. the departure of the secondary shock wave. 1955. Photometric analysis of a spectrogram of a The Ashkhabad station (1958, September 7) Perseid meteor. Astrophys. Journ., vol obtained the spectrum of the — 9m bolide. 121, pp. 250-270. According to the visual observations of Asta- IVANIKOV, V. I. povich, the height of the final point is 20 km 1957. The choice of an objective prism for meteor spectrophotometry. Bull. Stalinabad and the velocity is about 30 km/sec. The Astron. Obs., no. 20. bolide has a metallic spectrum only and the KRAMER, E. N. lines of the ionized elements are absent. N2 1958. The spectrum of a bolide from October 8, 1 P.G. is very strong. The bolide showed 14 1956. Astron. Circ, Acad. Sci. U.S.S.R., flares separated by regular intervals. As Liu- no. 195. 1959. Spectra of three bright meteors. Proc. barsky notes, there are no spectral changes in Odessa Astron. Obs., vol. 5, no. 1. the flares; the flares may perhaps be due to LIUBARSKY, K. A. rotation, but they are not due to the variations 1961. Spectral observations of meteors. Bull. in temperature. Comiss. Comets and Meteors, Acad. Sci. U.S.S.R., no. 5. Diffusion Effects Observed in the Wake Spectrum of a Geminid Meteor By Ian Halliday1 The spectrum of a spectacular Geminid in the first-order image of the Na D lines. A meteor was photographed on the night of small portion of the first- and second-order December 12/13, 1960, at the Meanook Meteor spectra of Vega appears in the lower left corner Observatory (^=54°37'N, X=113°21/W). of the photograph. This is one of two meteor observatories operated by the Dominion Observatory in northern Height and velocity measures Alberta. The spectrum exhibits interesting Since the meteor was photographed from only peculiarities. one station, no direct measures of height or The meteor was photographed with a con- velocity are possible. Reasonably accurate verted Kl9 aerial camera, focal length 12 inches, heights may be obtained, however, by assuming focal ratio f/2.5, using Kodak Spectroscopic a radiant position and velocity for the Geminid I-D emulsion on glass plates 8 by 10 inches. shower and then determining ranges and hence A Bausch & Lomb replica transmission grating heights from the observed angular velocity. with 300 lines per mm was mounted immedi- ately in front of the camera objective. The A standard Geminid radiant at a=113°, camera was occulted 12.5 times per second by a 8= +32°, was assumed (Millman, 1954) since rotating shutter with a closed-to-open ratio of the meteor appeared within an hour of the peak 2:1. of the shower (solar longitude 261°.l). This The meteor was not observed visually but assumed radiant was displaced to allow for the was recorded on an exposure that began at effect of zenith attraction. It was then found 3h32m U.T. and ended at 4b05m U.T., Decem- that the observed trail passed through the ber 13, 1960. The meteor was initially outside apparent radiant shortly before the end of the exposure, which gave a computed time for the the field of view of the camera, but the zero- h m order image entered the field before peak appearance of the meteor of 4 01 U.T. luminosity was attained, while the first- and A plot of the observed angular velocities second-order spectra entered the field at pro- along the trail indicated a marked deceleration gressively lower heights. in the latter half of the observed trail. The The spectrum is reproduced in plate 1. The angular velocity at the bottom of segment 4 was meteor moved from upper right to lower left. assumed to correspond to a meteoric velocity of The right-hand edge of the photograph is the 36.0 km/sec, and other heights and velocities edge of the original plate. The first segment were based on this assumption. The meteor of the zero order at the top of the photograph is was 90° from the radiant near the end of seg- the sixth segment that appeared on the plate, ment 10, at a minimum range of 106.0 km from while the lower end of the zero order (beyond the camera. Table 1 lists the computed heights segment 13) has been omitted at the left. The (corrected to sea level) and corresponding segment numbers are shown opposite each velocities at the lower ends of selected segments. segment of the spectrum along the right-hand Errors in the assumed meteoric velocity, radiant edge. Segments 20 and 21 are detectable only position, and speed of shutter rotation could i Dominion Observatory, Ottawa, Canada. result in the heights and velocities of table 1 161 162 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS being in error by as much as 3 to 5 percent. TABLE 2.—Measured emission lines in the Geminid There is no doubt, however, that the meteor spectrum—Continued penetrated to unusually low heights in the Xtoras Atom or ton Mutt. Xub atmosphere. 3797. 1 Fel 21 3795. 0 3798. 5 TABLE 1.—Heights and velocities 3799. 5 Height of 3815. 9 Fel 45 3815. 8 bottom of 3821. 0 Fe I 20 3820. 4 segment Veloeii Segment (km) (km/te 3825. 1 Fe I 20 3825. 9 1 72.9 36. 1 Fel 45 3827. 8 5 67.2 36.0 3832. 8 Mgl 3 3829. 4 9 61.5 35.4 3832. 3 13 55.9 33.7 Fel 20 3834. 2 15 53.3 32.3 3839. 1 Mgl 3 3838. 3 17 50.7 30.3 3856. 4 Fel 4 3856. 4 19 4& 5 26.7 Si II 1 3856. 0 21 46. 6 3859. 7 Fe I 4 3859. 9 3878. 2 Fel 4 3878. 6 The meteor spectrum Fel 20 3878. 0 3886. 5 Fe I 4 3886. 3 The dispersions of the first and second orders of 3900. 3 Fel 4 3899. 7 the meteor spectrum were found to be 95 and 3905. 5 Fel 4 3906. 5 46 A/mm, respectively. Atomic emission lines 3921. 4 Fe I 4 3920. 3 3922. 9 indentified in this spectrum are shown in table 3933. 6 Ca II 1 3933. 7 2. A total of 95 features are identified. The 3943. 6 All 1 3944. 0 spectrum appears normal for a meteor of this 3960. 6 All 1 3961. 5 velocity except that the lines of ionized mag- 3968. 8 Ca II 1 3968. 5 nesium, calcium, and silicon are relatively Fel 43 3969. 3 3997. 2 Fel 276 3998. 1 strong compared to those in the spectra of Fel 278 3997. 4 fainter meteors with the same velocity. Almost 4005. 0 Fel 43 4005.2 all the lines in table 2 were also identified in a 4030. 8 Mn I 2 4030. 8 recent study of Perseid spectra (Halliday, 1961). 4033. 5 Mn I 2 4033. 1 The only significant addition is the pair of lines 4034. 8 of Al I at 3944 and 3961 A. These are barely 4046. 1 Fel 43 4045. 8 detectable by sighting along the lines in a 4063. 4 Fel 43 4063. 6 30 X enlargement of the strongest segment of 4071. 7 Fel 43 4071. 7 the second order. The much greater relative 4132. 4 Sill 3 4130. 9 Fel 43 4132. 1 strength of the H and K lines for fast meteors 4144. 4 Fel 43 4143. 9 makes the detection of faint lines in this Fel 523 4143. 4 spectral region very difficult for Perseid spectra. 4155. 5 Fel 354 4156. 8 Fel 355 4154. 5 TABLE 2.—Measured emission lines in the Geminid 4200. 9 Fel 42 4202. 0 spectrum Fel 152 419a 3 Fel 522 4199. 1 Atom or ion Mutt. Xld> 4217. 3 Fe I 3 4216. 2 3683.1 Fel 5 3679. 9 4226. 6 Ca I 2 4226. 7 3683. 1 Fel 693 4227. 4 Fel 21 3687. 5 4250. 9 Fel 42 4250. 8 3705. 4 Fel 5 3705. 6 Fel 152 4250. 1 3707. 8 Fel 21 3709. 2 4253. 6 Cr I 1 4254. 3 3719. 9 Fel 5 3719. 9 4260. 6 Fel 152 4260. 5 3735. 6 Fel 5 3737. 1 4272. 1 Fe I 42 4271. 8 Fel 21 3734 9 Fel 152 4271. 2 3747. S Fel 5 3745. 6 4274. 2 Cr I 1 4274.8 3748. 3 4282. 2 Fel 71 4282. 4 Fel 21 3749. 5 Cal 5 4283.0 METEOR SYMPOSIUM—HALLIDAT 163 TABLE 2.—Measured emission lines in the Geminid TABLE 2.—Measured emission lines in the Geminid spectrum—Continued spectrum—Continued XmtM Atom or ion MuU. Xl.b Atom or ion MuU. Xl.b 4290. 2 Crl 1 4289. 7 5402. 8 Fel 15 5397. 1 Cal 5 4289. 4 5405.8 Fel 3,41 4291. 5 5429. 8 Fel 15 5429. 7 4307. 6 Fel 42 4307. 9 5448. 8 Fel 15 5446. 9 Cal 5 4307. 7 5456. 9 Fel 15 5455. 6 4314. 2 Fel 71 4315. 1 5499. 0 Fel 15 5497. 5 4325. 8 Fel 42 4325. 8 5501. 5 4353. 1 Mgl 14 4351. 9 5506. 8 Fel 71 4352. 7 5528. 3 Mgl 9 5528. 4 4376. 1 Fel 2 4375. 9 5573. 2 Fel 686 5572. 9 4383. 6 Fel 41 4383. 5 5587. 2 Fel 686 5586. 8 4404 8 Fel 41 4404. 8 5615. 5 Fel 686 5615. 5 4415. 9 Fel 41 4415. 1 5685. 8 Nal 6 5682. 6 4427. 1 Fe I 2 4427. 3 5688. 2 4461. 6 Fe I 2 4461. 7 5712 Fel 686 5709. 4 4481. 4 Mgll 4 4481. 1 Mgl 8 5711. 1 4481. 3 5892. 9 Nal 1 5890. 0 Fel 2 4482. 2 5895. 9 4530 Fel 68 4528. 6 5980. 2 Sill 4 5979. 0 4920. 4 Fe I 318 4919. 0 6063 Fel 207 6065. 5 4920. 5 6105 Cal 3 6102. 7 4955. 1 Fel 318 4957. 3 6121. 9 Cal 3 6122. 2 4957. 6 6139 Fel 169 6136. 6 5008. 9 » Fel 16 5012. 1 Fel 207 6137. 7 Fel 318 5006. 1 6157. 0 Nal 5 6154.2 5041. 3 Fe I 16 5041. 1 6160. 7 Fel 36 5041. 8 6162. 7 Cal 3 6162. 2 5108. 5 Fel 1 5110. 4 6229 Fel 207 6230. 7 Fel 16 5107. 5 6251 Fel 169 6252. 6 Fe I 36 5107. 6 6347. 1 Sill 2 6347. 1 5146. 1 » Fel 16 5142. 9 6369 Sill 2 6371.4 5150. 8 6397. 8 Fel 168 6393. 6 Fel 383 5139. 5 Fel 816 6400.0 Nal 8 5148. 8 6438 Cal 18 6439. 1 5153. 4 6461 Fel 168 6462. 7 5169. 4 Mgl 2 5167. 3 Cal 18 6462.6 5172. 7 6494.4 Fel 168 6495. 0 Fe I 1 5166. 3 5168. 9 Variation in luminosity Fel 36 5171. 6 Fel 37 5167. 5 The photograph in plate 1 indicates that this 5183.6 Mgl 2 5183. 6 meteor maintained approximately peak lumi- 5207. 2 Crl 7 5204. 5 nosity from segments 6 to 16, inclusive, an 5206. 0 atmospheric path length of 27 km. Before 5208. 4 segment 6, the zero order is considerably 5227. 2 Fel 37 5227. 2 Fel 383 5226. 9 weaker, while the spectral lines are fading 5232. 9 noticeably by segment 17 and drop abruptly 5269. 5 Fel 15 5269. 5 in intensity at segment 19. Peak luminosity Fel 37 5270. 4 appears to occur at about segment 11, near a Fel 383 5266. 6 height of 59 km. The lines in the blue region 532a 5 Fel 15 5328. 0 of the first-order spectrum appear stronger here Fel 37 5328. 5 5341. 8 Fel 37 5341. 0 than in subsequent segments, although the 5370. 5 Fel 15 5371. 5 variation between segments 11 and 14 is quite • Feature noticeably broad. small. 164 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS To estimate the peak luminosity the zero- quency curve occurs about two segments order image of the meteor was compared with after the peak of the light curve. From seg- zero-order star trails. The meteor trail is ments 9 to 12 the frequency is fairly constant definitely weaker than the trail of Vega and near 150 cps, but jumps suddenly at segment appears to match the zero order of a Cygni. 13 to about 320 cps and then declines, only The meteor image trails across the plate at slowly, for another two segments. The ampli- a rate 5,820 times greater than does the image tude of an individual fluctuation reaches a of a Cygni, corresponding to a difference of maximum near segments 10 to 12, resulting 9.4 magnitudes. The apparent magnitude of in a more pronounced fluting of the spectrum a Cygni is +1.3, from which the meteor magni- here than at other places. tude is estimated at —8. This is a "panchro- The periodic flaring might be interpreted matic" magnitude, which may be influenced by as indicating the repeated crumbling or frag- such factors as emulsion sensitivity and the mentation of minor amounts of material from blaze of the grating. In other cases (Cook the main meteoroid in the form of solid frag- and Millman, 1956; Milhnan and Cook, 1959), ments. The increased surface area exposed it proved to be only slightly different from to ablation would then account for the increased photographic magnitudes. luminosity at each minor flare. The meteor exhibits a rapid flickering super- Alternatively, the phenomenon might be imposed on the gradual rise and fall in overall closely associated with rotation of the meteor- intensity. The frequency of the flickering was oid. If the meteoroid is quite irregular in estimated in each segment by measuring the shape and rotates in such a manner that it fraction of a complete segment occupied by the alternately presents a large side and a smaller largest number of complete cycles observed end to the oncoming atmosphere, then the within the segment. The results were con- amount of material ablated, in gaseous form, verted to cycles per second and are shown would vary in a periodic manner. The light, plotted in figure 1. of course, is produced by the hot ablated gas Over most of the observed trail the flicker and not by the solid surface. is within the range of 100 to 300 cycles per A third possibility, suggested by Dr. L. G. second. In a sense, it follows the general Jacchia, is based on laboratory studies of light curve with a gentle rise to maximum ultraspeed pellets by Rinehart (Rinehart, and a steeper decline. The peak of the fre- Allen, and White, 1952). A flashing phenom- enon was observed for the pellets and ex- 360 r plained by an oscillation of the pellet during its flight. The luminosity will be low if there 320 is a leading edge of the object, and high when 280 a whole face of the object is exposed to ablation. For a meteor at relatively low heights where ,240- continuum flow occurs, as is expected for this Geminid, an oscillation may be set up that "200- would lead to a flickering phenomenon. Sud- den jumps in the frequency of the flicker, such 160- as are observed near segment 8 and again just i20r before segment 13, could indicate a tumbling of the meteoroid to oscillate about a new axis. 801- On the rotation hypothesis these changes might be interpreted as a fracturing of the 40- object into two or more pieces with an accom- panying increase in rotational velocity for 8 10 12 16 18 SEGMENT each piece. It is noted, however, that no FIGURE 1.—Frequency of the flickering plotted against general increase in luminosity accompanies position along the trail as indicated by segment numbers. these changes in the period of the flicker, HALLIDAY PLATE 1 10 I I 13 16 17 18 A Geminid meteor spectrum showing parts of the zero order (at left), and first- and second-order spectra. Segment numbers are indicated at right. HALLIDAY PLATE 2 Greatly enlarged portion of the meteor spectrum showing the split wake-lines of H and K in segment IS of the second-order spectrum. METEOR SYMPOSIUM—HALLIDAY 165 which suggests that the effective exposed nents; i.e., displaced to longer or shorter wave- area is not greatly changed. This would lengths compared to the position of the main appear to favor the oscillation hypothesis. spectral line in the adjacent segment of the normal meteor spectrum. The flickering effect The wake spectrum can be detected in the wake, although in some The zero-order image shows a slight wake in cases only one component appears to strengthen. the gap between the segments, beginning at The individual components show distinct curva- segment 6. The D lines of Na I show a strong ture in places with little correlation between wake where they enter the field of the camera the two components. In some instances the at segment 11, and the wake is observable as wake components can be traced down the trail far as segment 19. Peak intensity of the wake well into the following segment of the meteor occurs at segment 15, where the following mul- spectrum. tiplets can be detected in the wake: Na I, 1; Individual descriptions of the components Mg I, 2, 3; Ca I, 2; Ca II, 1; Mn I, 2; Fe I, 1, 2, of the H and K lines, in the second order, are 4, 5, 15, 20, 21, 41, 42, 43. given below for segments 14 to 18. Most of these multiplets have been listed in Segment 14-—The red component is the earlier studies of wake spectra (Halliday,1958), stronger, shows some curvature, and appears although the manganese lines near 4030 A to fluctuate more due to flickering than does have not been listed previously. The wake the violet component. The images are within spectrum shows the usual preference for lines of 1 mm of the extreme edge of the plate, which low excitation potential, but it is less pro- limits the value of this segment. nounced than in most other cases. The line of Segment 15.—A large-scale reproduction of Mg I, 2 at 5183.6 A is moderately intense in this region is shown in plate 2. The red com- the wake, whereas its presence was quite doubt- ponent is strong near the main segment and ful in most earlier wake spectra. The H and K can be traced back farther than the violet one. lines of Ca II are strong in the wake and persist The effects of flickering are more evident in the essentially as far back into the gap as the other violet component. Both components show wake lines, in contrast to some spectra in which some minor kinks, particularly near the main the H and K lines decay much faster in the wake segment. than the low-excitation Na I and Fe I lines. The most significant feature of this spectrum Segment 16.—The red component is very appears to be a clear splitting of the stronger strong in the lower portion of the wake, while wake lines into two components. It is observ- the violet component is weak. The violet able in the second-order spectrum from segments line strengthens considerably at an earlier flare 14 to 18. The H and K lines show the splitting in this segment of the wake. The red compo- most distinctly because of their strength, but nent may be detected higher than the violet other lines in which it is also observable are: one, and also continues about two-thirds of the X 4045 of Fe I, 43; X 4226 of Ca I, 2; and X 4383 way through the main part of segment 16. of Fe I, 41. For the Na D lines, in the first- With decreasing height within the segment, the order spectrum, the situation is complicated by red component approaches the main line but the fact that the splitting is in general compa- fades out before actually joining the main rable to the separation of the two D lines. In spectral line. segments 18 and 19, however, the separation due Segment 17.—The violet component is much to the splitting effect is large enough to show in the stronger in this segment. Just above the spite of the duplicity of the spectral feature. main segment the splitting is quite large, while For the remainder of the wake lines the inten- the violet line bends over toward the red com- sity in the wake is either too low or the resolu- ponent a little farther up in the wake. tion of individual lines is too poor to identify Segment 18.—The wake is now quite weak, both components of the wake lines. with the violet component stronger than the The two components of one wake line may red component. The splitting of the two com- be considered as "red" and "violet" compo- ponents is larger than in earlier segments. 638805—63 14 166 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS In the first-order Na D lines the two compo- Figure 2 shows three computed curves for the nents are not clearly resolved until segment 18. intensity profilo plotted along a radius of the Faint double wake-lines are also detectable in column from x=0 at the center to a;=1.0 at the segment 19 for Na D, with an even wider outer edge. The three curves correspond to an separation than in segment 18. effective thickness t of the luminous portion of the column of 0.05, 0.15, and 0.40, in units of the Interpretation of the double lines column radius. The resolution of the plate is One obvious explanation to consider for the insufficient to rule out very small values of t, doubling of the spectral lines in the wake is a but the components are sharp enough to indicate physical splitting of the meteoroid into two that the effective value of t is not much greater pieces. This has been observed for some than 0.15. meteors; for example, in spectrum 132 (Millman One complete segment corresponds to 0?080, and Cook, 1959). A splitting of the meteoroid made up of an exposure of 0*027 and an occulted was suggested as one possibility to ac- interval of 0?053. The wake is observable count for the sudden jump in the fre- approximately halfway across an occulted por- quency of the flicker at segment 13. While this tion, from which it follows that the effective dura- may have occurred it is not the explanation tion of the luminosity is considerably less than for the double wake-lines, since the wake-lines one shutter break. For all points within the are generally converging near their lower ends gap, the total exposure is the same, 0?027, but the in each segment, while a splitting of the mete- "age" of the column when the exposure starts is oroid would lead to diverging fragments. In different for each point. A point at the lower segment 16 the red component can be traced end of a normal wake, where the main meteor for a total distance equivalent to half a com- spectrum segment is starting, receives an expo- plete segment, whereas the exposure time is sure from age t=QH) to tf=0?027. For a point in only one-third of a segment; i.e., the red com- the middle of the gap, the exposure runs from ponent is not an exposure on a trailing fragment. <=0?027 to *=0?054. This is about the last The two components are not due to Doppler observable point in most segments; i.e., there is effects of an expanding column. The observed insufficient luminosity emitted after t—0?03 to splitting would correspond to a relative expan- record on the plate. sion, transverse to the direction of meteor motion, of about 600 km/sec. This is a most unreasonable value when the initial velocity of the meteor was only 36 km/sec. The best interpretation is that the wake represents a time exposure of the expanding column of meteoric gas, taken while the lumi- nosity is decaying. Meteor spectrographs are slitless spectrographs, which produce a broad image of a broad source even in monochromatic light. The two wake components apparently represent a side view of a hollow luminous column that shows two well-defined edges because the effective thickness of the luminous portion is small compared to the thickness of the entire column. Note that in comparing the two components at a particular point on the trail, one should compare two points on a line per- pendicular to the direction of meteor motion, rather than two points on a line parallel to the FIGURE 2.—Computed intensity profiles for three values of t, spectral dispersion. the effective thickness of the luminous column. METEOR SYMPOSIUM—HALLIDAY 167 As noted earlier, some of the wake lines, par- In segments 15 and 16, the red components ticularly the red components, can be traced can be traced into the main segment. They down into the following segment. For different begin with separations from the main spectral points within the segment itself, all exposures line (corresponding to a radius of the column) start at t=0 but have different durations. The of about 24 meters. One-third of the way down durations decrease linearly with distance down through the segment the separation is about 16 the segment from a maximum exposure of 0?027 meters, and this is maintained for another third at the top to 0?000 at the bottom (where the of the dash, until the line fades out as noted shutter occults the camera for the next 0?05). previously. At the lowest observable points In segment 16, it was noted that the red com- these red components are displaced about 16 ponent was observable two-thirds of the way meters from the main column, recorded in an through the segment; i.e., for any exposure exposure which lasted for only 0?009. duration longer than 0*009. In all cases where separations for lines of Ca I or Fe I could be measured, the observed Magnitude of the splitting points fell quite well on the corresponding curve Once it is known that the split lines correspond derived for the H and K lines of Ca II. to a physical separation of material in the The splittings within one segment of the wake meteor column, it becomes of interest to deter- are fairly constant; i.e., the red and violet com- mine the magnitude of the splitting. The ponents are nearly parallel. The mean values separation of the two components was measured for the five segments show a curious dependence at frequent intervals in each segment of the on height: the splitting increases with decreas- wake, and the results were plotted. The scales ing height and may be represented quite well by were converted to diameters of the column, in the linear relation meters, by allowing for the plate scale and the t Discussion 15 16 n 18 (tee) H 0.000 31 49 69 78 At the low heights of this meteor one should 0.004 37 34 49 73 82 expect a shock front to form around the mov- 0.008 42 39 44 61 85 ing meteoroid. The observed wake luminosity, 0.012 45 49 54 83 however, is produced by meteor atoms of cal- 0.016 53 77 0.020 77 cium, sodium, and iron, not by atmospheric Mean 40 37 49 64 80 particles. An effective front of expanding me- 168 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS teoric material appears to exist inside the heated very large compared with thermal velocities. volume produced by the shock front. Violent Manning mentions that the ion column for a collisions at the surface of this meteoric front bright meteor may be larger than for a faint continue to excite the meteoric atoms for inter- one if ionization can still be accomplished during vals up to about 0903. As shown earlier, the the expansion of the atom column. For cal- sharpness of the two components of the strong cium, 6.09 ev are required for ionization, which wake lines indicates that the effective thickness is higher than the excitation of any wake line. of the luminous portion of the meteor column This makes it more likely that an ion of Ca II, is not more than about 15 percent of its radius, which emits H or K radiation 0?03 after the or else the lines would appear more diffuse. passage of the meteor, survived nearly all of The wake components indicate a diffusion of that interval as an ion, and was then excited meteoric material into the atmosphere, but the with an additional 3.1 ev to radiate, rather than effective diffusion coefficients are so large as to that the ion was formed from a Ca I atom and excited to radiate by any collision occurring suggest the phenomenon is essentially explo- ? sive. During the initial stages the diameter of later than perhaps 0 002 after the atom left the the column is observed to expand at a rate in meteoroid. excess of 3 km/sec. It is not surprising that Rapid rotation or oscillation for this me- the expansion should be very rapid compared teoroid is suggested by the regularity of the to normal diffusion since the amount of mate- flickering phenomenon and is supported to some rial deposited by such a bright meteor is so extent by the split wake-lines. The variation great as to constitute a major change in local in relative intensity between the red and violet density along the meteor trajectory. The rate components could arise from either mechanism, of diffusion of meteoric material in this case is since varying amounts of material would be so great that it seems reasonable to infer that ejected from the meteoroid in a manner which much fainter meteors, perhaps any meteor is not radially symmetrical. This could also brighter than 0 magnitude, will also give rise account for some of the curvature observed in to appreciably more rapid diffusion than is ob- the wake lines. served by radar for faint meteors. Great care One Perseid spectrum is known which shows should be taken, then, that normal diffusion a similar but much less pronounced splitting coefficients are not considered as a known quan- of the wake-lines. It is spectrum 223 in Mill- tity in calculations involving the gas columns man's (1958) list of meteor spectra. The split- left by bright meteors. ting is evident in one segment of the H and K It is of considerable interest to note that the lines, but it is confined to a very short path ion column of Ca II ions expands at the same length near a bright flare. The fact that the rate as the atom column of Ca I or Fe I. This phenomenon is so pronounced for the Geminid is not normally considered to be the case, at meteor may be associated with its low height least for radar observations, which, of course, in the atmosphere. The much smaller mean are for meteors about 10 magnitudes fainter than free path may tend to localize the zone where this Geminid. Manning (1958), for example, appropriate collisions may occur and effectively quotes Loeb (1934) to support a statement that shield the inner portion of the wake column, the ion column will diffuse only one-fifth as giving rise to the hollow-tube effect. fast as the atom column. Loeb's results are References based on experiments in which ions were allowed to diffuse through an electric field. He states COOK, A. F., AND MILLMAN, P. M. 1956. Photometric analysis of a spectrogram of a (p. 548): "the velocity gained from the field Perseid meteor. Astrophys. Journ., vol. between impacts is further supposed small 124, pp. 476-477. compared with the thermal velocities of agita- HALLIDAT, I. tion." It is not clear that these results are 1958. Meteor wakes and their spectra. Astro- applicable to the meteor case, where the pres- phys. Journ., vol. 127, pp. 245-252. 1961. A study of spectral line identifications in ence of an electric field is unlikely and the Perseid meteor spectra. Publ. Dominion systematic velocity of the meteor particles is Obs. Ottawa, vol. 25, pp. 1-16. METEOR SYMPOSIUM—HALLIDAY 169 LOEB, L. B. MlLLMAN, P. M., AND COOK, A. F. 1934. Kinetic theory of gases. Ed. 2. McGraw- 1959. Photometric analysis of a spectrogram of a Hill Book Co., New York. very slow meteor. Astrophys. Journ., MANNING, L. A. vol. 130, pp. 648-662. 1958. The initial radius of meteoric ionization trails. Journ. Geophys. Res., vol. 63, OPIK, E. J. pp. 181-196. 1958. Physics of meteor flight in the atmosphere. MILKMAN, P. M. Interscience, New York. 1954. A provisional list of the major meteor showers. Journ. Roy. Astron. Soc. RINEHART, J. S.; ALLEN, W. A.; AND WHITE, W. C. Canada, vol. 48, pp. 193-195. 1952. Phenomena associated with the flight of 1958. Photographic meteor spectra (appendix 6). ultra-speed pellets. Part 3. General fea- Journ. Roy. Astron. Soc. Canada, vol. 52, tures of luminosity. Journ. Appl. Phys., pp. 275-276. vol. 23, pp. 297-299. Abstract A study is made of an unusual Geminid meteor spectrum, photographed between heights of 73 and 47 km. Peak luminosity, estimated at magnitude —8, occurred near 59 km. The meteor exhibited a pronounced flicker, with a frequency varying from 50 to 300 cycles per second. The flicker may be associated with rotation or oscilla- tion of the meteoroid. A prominent wake spectrum was photographed between 50 and 55 km, consisting of lines of Na I, Mg I, Ca I, Ca II, Mn I, and Fe I. The stronger lines are split into two distinct components attributed to a hollow luminous column rather than splitting of the meteoroid or Doppler effects. The diameter of the column varies from 40 to 80 meters, but shows no variation from one atom or ion to another. Diffusion coefficients computed from the ob- served rate of expansion exceed the normal values at these heights by a factor of more than 10s. The expansion in this case might be considered as explosive. On the Frequency of Occurrence of the Auroral Green Line (5577 A) in Perseid Spectra By John A. Russell * In a note announcing his identification of the appeared briefly at the beginning of the trail. forbidden line of O I at 5577 A in the spectra Spectrum No. 119 was reproduced in a general of three meteors, Halliday (1958) pointed out paper on Perseid spectra (Russell, 1959) as an the difficulty of explaining why the line had not example in which the H and K lines, though been detected previously. In a subsequent present, were not dominant. When this spec- study of old spectra, Millman (1960) located a trum was treated photographically by the proc- number of appearances of the auroral line, ess used on No. 120, the auroral green line also including one in a spectrum from the 1933 became evident. Harvard Leonid campaign and one obtained in More recently the presence of the auroral 1950 by the author. This latter spectrum, No. line in two additional spectra, Nos. 103 and 121 on the World List (Millman, 1952), had not 107, obtained in 1948 and 1949 with the same been reduced at the time of Millman's identifi- instrument, was detected by careful reexamina- cation because of its mediocre quality. tion of the negatives. This brought the green Spectrum No. 121 is one of five Perseid line identifications to five out of the seven spectra (World Nos. 118-122) obtained in a spectra photographed with the same equipment 31-hour interval with the same instrument: a during three consecutive Perseid showers. 30° prism, f/5.6 lens of 130-mm focal length, We note that the auroral line was not found and Super XX roll film. It is evident that in the spectrum that showed only the H and K these spectra should provide some information lines, or in the one which showed little besides on the consistency with which the green line strong H and K lines. A correlation between appears under similar instrumental and atmos- the strength of the green line and that of the pheric conditions. Spectrum No. 118 contains H and K lines was originally suggested by a only the H and K lines of ionized calcium. In Perseid spectrum, No. 198, obtained with dif- spectrum No. 122 the H and K lines were ferent apparatus (Russell, 1959). This was the strong, but the remainder of the spectrum was first Perseid in the author's experience that very weak. In neither of these spectra could failed to show H and K lines or that exhibited the 5577 line be detected. Spectrum No. 120 strongly the O I line, the density of which in was reproduced in an earlier paper (Russell, this spectrum was second only to that of the 1953). It was well exposed, but suffered from sodium D-lines. The extent of this possible a poor orientation of the path with respect to correlation was investigated in the following the prism edge, from foreshortening as the result manner. From estimates made visually out- of the proximity of the meteor to the radiant, side of noticeable bursts, the plates were and from the location of the spectrum in the arranged in order of increasing density of the very corner of the photographic film. When K line relative to the blend at 3830 A, and the original negative was enlarged and the con- numbered from 1 to 7. The plates were then trast increased by two successive printings on arranged in order of increasing density of the contrast lantern slide plates, the green line O I line relative to the D lines and again num- 1 Department of Astronomy, University of Southern Cali- bered consecutively. The results are illustrated fornia, Los Angeles, Calif. in figure 1. 171 172 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS To the extent that so small a sample can be taken from the "Summaries of Mount Wilson relied upon, the correlation appears to be real. Magnetic Observations of Sunspots," which ap- Spectrum No. 198 (photographed 1956) and peared regularly in Publications of the Astro- Halliday's spectrum No. 302 (photographed nomical Society of the Pacific until 1959. No 1958), which consists solely of the green line, consistent pattern is evident relating the green would extend the graph linearly toward the line to solar activity, although the absence of lower right. Why a correlation such as this extreme solar conditions during the period should exist is not clear. As most of the data 1948-1950 makes the negative result inconclu- plotted in figure 1 are from meteors observed sive. The atmospheric depth to which the green within a period of 31 hours, it may be necessary line penetrates relative to the other lines is re- to explain some of the differences in the spectra markably uniform. It generally appears con- as the result of differences in the meteoroids currently with the blend of iron lines at 3830 A rather than as exclusively the effect of variable and fades about the time that the iron blend atmospheric conditions. Of possible signifi- at 3740 A becomes visible. As all of these cance is the observation that the strength of spectra are of meteors with a common high the green line is almost entirely unaffected by velocity, they can provide only partial support bursts (Halliday, 1960), whereas the strength of Halliday's evidence that the green line is of the H and K lines is most strongly influenced exclusively a high-velocity feature. by them (Russell, 1950). TABLE 1.—Relation of auroral line emission to geo- magnetic and solar activity Aug. 7- e World Number number num- tunspot suntpot ber of groups groupt spec- Auroral Geomagnetic that that 6- e trum Date line activity day month 103 8/12/48 Weak Great 10 12.4 5- • 107 8/13/49 Weak None 6 8.6 ORDER 118 8/12/50 Missing Moderate 6 7.0 OF 4- 0 119 8/12/50 Weak K 120 8/12/50 Weak 3830 121 3- • 8/13/50 Weak None 122 8/13/50 Missing 198 8/11/56 Strong Great 13.5 2 e 9 274 8/12/58 Missing Not great 14 14.8 1- 0 When spectra taken with other instruments and emulsions are considered, a broader though 12 3 4 5 6 7 less homogeneous base is provided for studying ORDER OF 5577/ D correlations. In 1951, by which time solar FIGURE 1.—K-line density versus density of 5577 in seven activity was below average, a shutter-chopped Perseid spectra. Perseid, No. 124 (Russell, 1959), which strongly showed H and K lines and the first positive Halliday (1960) finds some evidence for a group of nitrogen bands, failed to display the correlation between the occurrence and depth green line. A low-dispersion but fairly well- of penetration of the green line, on the one exposed spectrum of a Geminid, No. 140, hand, and geomagnetic and solar activity, on obtained in 1953 shows no green line. This the other. Milhnan (1960), however, points absence fits both the low-velocity and the solar- out that the Leonid spectrum in which he activity correlation. A sporadic meteor of found the green line was photographed at the unknown velocity, No. 148, photographed in depth of the sunspot minimum of 1933. The 1954, showed very strong H and K lines but no geomagnetic activity and solar activity during green line. Both spectrum No. 198 (1956) the Perseid showers of 1948-1950, 1956, and and a similar but much fainter spectrum 1958 are indicated in table 1. The data are obtained the next night with the same instru- METEOR SYMPOSIUM—RUSSELL 173 ment show a relatively strong green line but no inverse correlation between the strength of the H and K lines. Table 1 indicates that a great H and K lines and the strength of the auroral magnetic storm was in progress at the time and green line at 5577 A; (2) the spectra consist- that general solar activity was high. ently show the green line except where its If general solar activity is a factor in produc- absence would be expected in the light of state- ing the green line, no spectrum obtained by the ment (1); (3) the position of the green line author should show it better than the Perseid relative to the rest of the spectrum was consist- photographed in 1958 with the 18-inch Schmidt ently similar. A consideration of several at the Mount Palomar Observatory (Russell, other spectra indicates that (1) any correlation 1960). This spectrum, No. 274, shows very between general solar activity and the occur- strong N2 bands in the red and 40 measurable rence of the green line appears weak, at best; (2) features. In the region of the spectrum where there is no violation of Halliday's findings that the green line would normally appear, there is the green line is not found in the spectra of low- a broad, amorphous darkening that is a little velocity meteors. sharper on one side than on the other. The sharper boundary is close to 5577 A. This References feature was not included in the published dis- HALLIDAT, I. cussion as no lines are seen in it under the 1958. Forbidden line of O I observed in meteor microscope. It differs from the green line's spectra. Astrophys. Journ., vol. 128, pp. customary behavior in that it does not appear 441-442. 1960. Auroral green line in meteor wakes. sooner than the other lines; actually, the N2 bands appear first, much farther back along Astrophys. Journ., vol. 131, pp. 25-33. MILLMAN, P. M. the trail. If this feature were produced by the 1952. Photographic meteor spectra. Journ. Roy. forbidden O I line, blurred by winds during the Astron. Soc. Canada, vol. 46, pp. 121- brief period of radiation, a minimum wind 125. velocity of 5 km/sec would be required to ac- 1960. The auroral green line in a Leonid spectrum. count for the breadth of the line. A. F. Cook Journ. Roy. Astron. Soc. Canada, vol. (private communication) suggests that this 54, pp. 189-192. feature is produced by additional, unresolved RUSSELL, J. A. N2 bands. The conclusion that the green line 1950. Intensity variations in the spectrum of a is not involved in this spectrum seems strongly Perseid meteor. Contr. Meteorit. Soc, justified. Although general solar activity was vol. 4, pp. 250-254; Pop. Astron., vol. 58, pp. 134-138. higher in 1958 than in 1956, no great magnetic 1953. Asymmetrical lines in the spectra of storm was in progress at the time of the 1958 meteors. Meteoritics, vol. 1, pp. 103- shower. 105. 1959. Some Perseid meteor spectra of the past Conclusions decade. Sky and Tel., vol. 18, pp. 549- 551. The seven Perseid spectra obtained with the 1960. Molecular nitrogen in the spectra of two same instrument over the 3-year period 1948- Perseid meteors. Publ. Astron. Soc. 1950 indicate that (1) there appears to be an Pacific, vol. 72, pp. 291-295. Abstract A reexamination of seven Perseid spectra, obtained with the same instrument, has revealed five instances in which the auroral green line of O I at 5577 A is visible. All of the seven spectra were photographed during the showers of 1948-50, five of them within a 31-hour period. An inverse correlation appears to exist in these spectra between the strength of the auroral line and that of the H and K lines of ionized calcium. The points of appearance and disappearance of the auroral line relative to those of other lines were consistently similar. A consideration of several other meteor spectra indicated little if any correlation between the occurrence of the green line and general solar activity. There is, however, no evidence contrary to Halliday's findings that the green line appears only in the spectra of high-velocity meteors. 638805—63 15 Negative Ions and Luminosity in Meteor Trains By T. R. Kaiser x The work reported here forms part of a theo- of electrons, negative and positive ions respec- retical study of the processes of dissipation of tively. The diffusion coefficient D is treated ionization in meteor trails and, as it is hoped as a constant, identical for all species. While to report this more fully elsewhere, the relevant this will be essentially true in the absence of parts of the theory are given only in outline. heavy negative ions, it is only an approxima- The main causes for decay of volume elec- tion in the general case. However, from a gen- tron density in a meteor trail are diffusion eralization of the theory of ambipolar diffusion (molecular and eddy), electron-ion recombina- (Kaiser, 1953), it seems that it is not likely to tion, and negative ion formation. The latter vary by more than a factor of two. The terms is complicated somewhat through the compet- ai, a?, 03, and a4 are the rate coefficients for ing processes of electron detachment and mutual attachment, detachment, recombination, and neutralization of positive and negative ions. mutual neutralization, respectively; ax will be proportional either to molecular density n cm"3 Both recombination and mutual neutralization 2 can contribute to the meteor train luminescence for radiative attachment, or to n for three- through the processes body attachment in which the excited negative ion is stabilized by collision; a^ will be propor- A+-\-e-*A+hv (radiative recombination), (1) tional to n for collisional detachment or con- stant for photo-detachment; a3 will be constant (2) since recombination appears to be radiative (Kaiser and Greenhow, 1953; Kaiser, 1953). where the primes denote excited states. The The contribution of recombination to train positive ions will be the products of collisional luminosity has been discussed by Kaiser (1955). ionization and the negative ions will be O2 and A reassessment of this process suggests that it 0", produced by electron attachment. In the may be significant for the brightest meteors meteor height region molecular oxygen is pre- but that generally it is likely to be less impor- dominant. tant than mutual neutralization. In the pres- Formulation of the problem ent analysis, therefore, we will neglect it and put 03=0. The expressions for the rate of change of volume The most important quantity for radio me- density of the various ionic species are: 1 teor studies is the electron line density, ae cm" . (3) The corresponding quantities for the other spe- cies are a+ and ^ —a3nen+— (5) L~ 2 dt (7) (6) If we take g as the mean number of photons where nt, n_ and n+ are the volume densities produced in reaction (2), / will be measured 1 University of Sheffield, England. in photons per second per cm length of train. 175 176 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Solutions to equations 3 to 6 ( a3=0) The form of the electron and negative ion We will assume the trail is formed with negli- decay is illustrated in figure 1. gible radius at time t—0. (iii) a^Q. (i) O2=0. Approximate solutions have been found for Equation (3) becomes extreme cases of slow and rapid neutralization. X = (a) 4 7r-ri Here at and a_ are given ap- (8) proximately by solutions (11) and (12), with a+ decreasing with time: with the solution (Kaiser, 1953) (9) 1 (13) where n<=£m°xv(-TTu) ao) a,ia (l—e~r) A 0 (14) and OQ is the initial electron and positive ion dt line density. This solution has been used to a1 derive values of ai from long enduring radio where echoes (Kaiser, 1953; Davis, Greenhow, and Hall, 1959a, 1959b). and Equations (4) and (5) contain the non-linear neutralization terms and are not amenable to exact analytic solution. (ii) at=0. The train luminosity will follow equation This is not directly helpful for the train (14) and will have decayed significantly when 1 problem since it gives 7=0; however, it is T~1, i.e., tf~ (ai+e^)" . capable of exact solution which may be of The initial intensity 70 is, from equations (7) value in radio-echo studies and is helpful in and (14): obtaining an approximate solution when a« is (15) sufficiently small (see below). The solutions are: (b) a?4»l. In this case the negative ion volume density rises rapidly (instantaneously ct.—:— -(a,+o,)0j, (11) for zero initial radius) to n-^di/di and main- tains this value throughout the region of the column of ionization where w«/w_ exceeds a+=a0= constant ^*- — "--•" For all ionic species \ 4*Dt*xp\~WiJ- 1 When t becomes large compared with(a1-f-o2)" , /X _ / a,*«2 FIGURE 1.—Variation of electron and negative ion line density for a«=0. Solid line, a«/ao; broken line, a-fa^. METEOR SYMPOSIUM—KAISER 177 If also camera, over the atmospheric height range 84 we get to 93 km. They find the decay to be expo- I — da+ a nential with a duration which increases with ~di Oil0e- ih (16) height (i.e., with decreasing atmospheric density). They have also estimated the rate (17) of train decay for a long enduring train observed where equation (16) remains valid to by Liller and Whipple (1954), which suggests < that the duration is a maximum at about If x^>\ but z4a1/(a1+a2) ^l (which requires a^ai) the initial intensity is still approxi- 92.5 km and decreases above this height. mately as given by equation (17) but the de- They infer that the decay of light from both cay is more rapid and it is estimated that normal and long-enduring trains follows an the intensity will fall to a fraction of its ini- exponential law and that the rate of decay tial value in a time is a well-defined function of height. The decay constants obtained by Howard and tar1- (18) Hawkins are plotted in figure 3 as a function of height and atmospheric molecular density 3 The significance of the above is that, provided ncm~ , the latter as given by Nicolet (1959). x^>l, negative ions are initially mopped up The decay constant is defined as dM/dt, where by neutralization before detachment can op- M is the train magnitude. erate and, unless a% is extremely large com- These results strongly suggest that below 92.5 km the decay obeys equation (16) (equa- pared with au this will continue to £>af \ In the limiting case, a?=0, it is possible tion (15) with xt hence ax can be deduced directly from experi- ment. The theoretical initial decay rate for a finite resolving power has been evaluated on the assumption that the camera records a line source as having a brightness profile exp (—TP/XI2), where x is measured along the normal from the axis of the image. It is further assumed that D (b) Three-body attachment, n FIGURE 2.—Decay of train intensity for various values of dM m M_ IY n V , o *"»T (21) i./ n J 178 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS • HAWKINS t HOWARD O LILLER t WHIPPLE IOO I 3O 31 32 33 In fit cm3) FIGURE 3.—Train decay rate, dM/dt magnitude sec"1, as a function of atmos- pheric density and height. The theoretical curves, which are discussed in the text, are derived for three-body (a) and two-body (b) attachment. where the subscript "min" refers to the values Substituting the above values we get a;i~9 corresponding to minimum dM/dt. The result meters, which implies a resolvable train diame- of fitting the formulas (20) and (21) to the ter of the order of 20 meters. This agrees experimental results is illustrated by the curves well with Whipple's estimate of the order of plotted in figure 3. It is clear that the experi- 30 meters for the Baker Super-Schmidt cameras mental data strongly support three-body at- (Liller and Whipple, 1954). tachment, at least below 92.5 km. In fitting In spite of the above, the explanation of 2 formula (21) we obtain at min=6X10" the increasing decay constant above 92.5 km and hence as due to diffusion is not completely convincing, n2sec~1 (22) particularly if one considers in more detail the form of the decay in surface brightness with for the relation between attachment coefficient time. However, although one could specu- and total molecular density. The result also latively suggest additional explanations, this implies that a2 > at and xi ^ 1 up to 92.5 is hardly justified until more experimental data km. If we take Z>=5X10* at this level are available. (Greenhow and Lovell, 1960) and note that The relation between meteor and initial trains have been observed for meteors as faint as magnitude +2.5 (Hughes, 1959) for which train magnitude 13 1 «0~10 cm~ (Kaiser, 1953) we obtain a limit If the meteor produces p photons through for the mutual neutralization coefficient, namely, collisional excitation for each electron and we ignore any difference between meteor and 7 3 1 train spectral distribution, then, for at > 10~ cm sec" * (23) In deriving expression (21) we obtain a ~2.5flog,o(p/2)-logio(aiT)], (25) relation for the camera resolution, characterized by*,: provided ai!T GBEENHOW, J. S., AND LOVELL, A. C. B. 1955. Meteors. Spec. Suppl. Journ. Atmosph. 1060. The upper atmosphere and meteors. In Terr. Physics, vol. 2. Pergamon Press, Physics of the upper atmosphere, J. A. New York. Ratcliffe, ed., pp. 513-549. Academic KAISER, T. R., AND GREENHOW, J. S. Press, New York. 1953. On the decay of radio echoes from meteor HAWKINS, G. S., AND HOWARD, W. E. trails. Proc. Phys. Soc. London, vol. 1959. Decay of light from a meteor train. 66B, pp. 150-151. Astrophys. Journ., vol. 130, pp. 1003- LlLLER, W., AND WHIPPLE, F. L. 1007. 1954. High-altitude winds by meteor-train pho- tography. In Rocket exploration of HUGHES, R. F. the upper atmosphere, R. L. F. Boyd 1959. Meteor trains. Smithsonian Contr. As- and M. J. Seaton, eds., pp. 112-130. trophys., vol. 3, no. 8, pp. 79-94. Interscience, New York. KAISER, T. R. NICOLET, M. 1953. Radio echo studies of meteor ionization. 1959. The constitution and composition of the Advances in Physics (Phil. Mag. upper atmosphere. Proc. Inst. Radio Suppl.), vol. 2, pp. 495-544. Eng., vol. 47, pp. 142-147. Abstract The decay of meteoric ionization due to the combined effects of diffusion, electron attachment and detachment, and mutual neutralization of positive and negative ions, has been studied theoretically. It is proposed that the excited neutral atoms resulting from mutual neutralization produce meteor train luminosity. Interpretation of experimental data, in the light of the theory, leads to values for the electron attachment coefficient which agree reasonably with those obtained from radio observations and suggests that attachment occurs through a three-body process. Luminosity from Large Meteoric Bodies By H. Julian Allenl and Kenneth K. Yoshikawa l It is usual in meteor theory to ascribe the Gas-cap radiation origin of the luminosity of meteors to collisions During the development of long-range ballistic of air molecules with, the surface molecules of missiles and, more recently, of space vehicles the meteoric body. In this case the rate of intended to enter the atmosphere of earth or removal of mass from the surface is directly other planets, the problem of gas-cap radiation proportional to the luminosity, which affords a was readily recognized (see, e.g., Allen and Eggers, most convenient means for analysis of meteor 1958) by aerodynamicists as one that would flight. It seems not to be generally appre- require considerable study if satisfactory heat ciated that the above procedure must be shields were to be designed for such vehicles. restricted to meteoric bodies which experience An intensive effort, both theoretical and free-molecule flow—that is, to those bodies experimental, has been made to understand the which are small compared to the mean free phenomenon (Kivel and Bailey, 1957; Kivel, path of the air molecules. 1959; Armstrong, 1959; Hochstim, 1957; For meteoric bodies in continuum flow, Ziemer, 1960; Meyerott, Sokoloff, and Nicholls, wherein the body is large compared to the air 1960; Yoshikawa and Wick, 1961; Camm, mean free path, the air molecules approaching Kivel, Taylor, and Teare, 1959). The phe- the body enter into a highly compressed layer nomenon is a complex one and far from well of gas which precedes the body itself. Linde- understood. The experimental studies, more- mann and Dobson (1923) aptly called this layer over, have not been carried into the speed the "protective gas cap." The most energetic regimes of most meteors. Thus, in the follow- collisions in the case of such flow occur between ing paragraphs concerning gas radiation it must the air molecules in the atmosphere and in this be constantly borne in mind that the description gas cap. The resulting luminosity from the of the phenomenon is probably incomplete, and gas cap is not a consequence of mass-loss rate the magnitudes discussed are extrapolations of of surface molecules, but the radiation from it our present knowledge. may be a principal source of aerodynamic When continuum air is compressed on enter- heating of the meteoric body and thus cause ing a gas cap of a meteoric body, it is drastically removal of mass by ablation. Plate 1 gives an slowed down relative to the body. The high illustration of gas-cap radiation from a body as kinetic energy of the stream is then almost photographed in a shock-tube wind tunnel at entirely converted to heat. The translational, Ames Kesearch Center. The airspeed is only rotational, and vibrational modes of the mole- 4.3 km/sec with the ideal gas temperature in the cules are first excited. At meteor speed the shock layer at about 4500° K, yet the luminous energy is sufficient, moreover, to dissociate and gas cap near the bow of the body and near the ionize a large fraction of the air in the cap. control surface for this simulated space vehicle These atomic and molecular species become the is clearly evident. It should be noted that at sources of the radiation. In addition, time is the instant this photograph was taken, no required to establish equilibrium between the ablation of the surface of this metal model was gas species during and following the compres- occurring. sion transient so that the amount of the radia- tion is a function of this time of relaxation. i National Aeronautics and Space Administration, Ames Research The whole process then can be visualized as Center, Moflett Field, Calif. 181 182 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS the gas cap is low, the time scale is such that the body surface may be in the position rela- tively close to the start of the original com- pression, as at "B" (fig. 1). The radiation will then be principally from "out-of-equilibrium" air. On the contrary, if the gas-cap density is high, the whole of the transient behavior will occur in a much shorter time and hence effectively farther from the body. In this case "C" may represent the position of the body so that the radiation received may be well approxi- mated by the equilibrium value as the inte- grated mean, provided the density is not so high that reabsorption occurs within the layer. Where reabsorption occurs, the radiation, of course, becomes limited. Using the results of Yoshikawa and Wick (1961), we have calculated the radiative heat- transfer coefficient in equilibrium flow \aB, which the body receives from the gas cap. FIGURE 1.—Variation of pressure, temperature, and radiation For this purpose it was assumed that the total in air following shock compression. volume of the radiating cap is the product of the frontal area of a sphere of radius r and the shown in figure 1. During the initial com- thickness of the gas cap. The results are shown pression ("A") the temperature attains values in figure 2. approaching that corresponding to an ideal The out-of-equilibrium effects were obtained molecular gas; it then decreases as energy is by extrapolating data from AVCO (Camm, diverted to dissociate and ionize the molecules Kivel, Taylor, and Teare, 1959) and Ames and to form interim molecules until chemical Research Center (e.g., Allen, 1961) with the equilibrium is established. The level of radia- assumption that the total gas-cap radiation is tion takes somewhat longer to subside to an limited to about 15 percent of the total energy equilibrium value. The radiation intensity as a maximum since body size or reabsorption received by the body surface therefore depends within the gas would limit further losses. upon the position of the surface relative to this (Note that the "black body" limit was far time-dependent process. If the air density in above this.) The results are shown in figure 3. 100 - 0 2 4 6 8 K) 12 M 16 18 20 Velocity, v, km sec"' FIGURE 2.—Radiative heat-transfer from the gas cap of a sphere for air in chemical equilibrium METEOR SYMPOSIUM—ALLEN AND YOSHIKAWA 183 FIGURE 3.—Radiative heat-transfer coefficient for air out of chemical equilibrium. Thereon \o is the radiative heat-transfer co- input to the body surface must appear as efficient for the air out of chemical equilibrium, ablation of the surface. Thus and p is the ratio of the ambient air density p at altitude h to that at sea level, po. The dm dt~~ (2) ratio XG/XGB will be less than unity for large values of ~j>r when reabsorption becomes where f is heat of ablation (thejheat energy important. absorbed by a unit mass of the body during ablation), and X is the heat-transfer coefficient. Theory for meteors, including continuum On the other hand, it is well known that the flow ablated mass from meteoric bodies in free- The equations for acceleration and mass loss molecule flow emits radiation (see, e.g., Opik, rate for all meteor air flow regimes are given 1958). For continuum flow it is again to be in the following. The notation used is that of expected that a fraction of the total radiation Jacchia (1948). resulting from ablation can be related to the The equation of motion, which relates the time rate of change of mass by the equation acceleration of the meteoric body to its mass dm and to the drag it experiences in flight, is the (3) familiar one dt dv where / is the total radiation, Ia is the gas- 1/3 (1) cap radiation, and r is the luminous effi- m 0 ciency factor. where v is velocity, m is mass, p is air density, In the determination of mass loss for meteoric y is drag coefficient, and Ami/3 is the effective bodies in free-molecule flow (which constitute cross-sectional area. nearly all of the observed meteors), it has The time rate of change of mass may, in been customary to put all of the reliance on principle, be obtained in two ways: On the equation (3) rather than equation (2). The one hand, since for meteoric bodies of all but reasons for accepting equation (3) are: the smallest in size, (1) the heat radiated from (i) The gas-cap radiation (/ Gas-cap radiation from an entry body. METEOR SYMPOSIUM—ALLEN AND YOSHIKAWA 185 ,.- i __ 3 iff" FIGURE 4.—Meanook meteor luminous intensity and comparison with gas-cap radia* tion for spherical bodies ing reasons. If the body on entering the at- For a sphere in continuum flow mosphere is tumbling fairly rapidly, the mean drag force experienced will be that of a sphere. 7=2- (5) When the initial tumbling rate is low, it can be shown that the body will generally stop its rotation and, instead, oscillate about an attitude thus equation (1) can be solved for the radius for which it is aerodynamically stable. Now it should be noted that stable attitudes are not opo / pV \ ,a\ those for which streamline flow would result. r=—^— I -7-777 I * \y) Thus rodlike or platelike shapes will assume an attitude for which the largest dimensions are Of course the meteor density is not known, perpendicular to the direction of motion. Such a priori, and must be determined. Table 1 bodies in their high-drag attitude have approxi- gives six values of v (not counting the extra- mately twice the drag coefficient of a sphere. polated value for vm given by Millman and Cook) As an opposite extreme, a right circular cone and five of acceleration as a function of k. will fly with its apex forward only if the total These data can be used to obtain a general cone angle is greater than about 40° of arc. series expression for v in the following way: In this case the drag is about one-fourth that for Allen and Eggers (1958) have shown that when a sphere, but it should be noted that for such a no mass is lost shape the conical point will promptly ablate and v*=t?m(F, (7) hence the drag will be increased. On the aver- age, then, we expect that the assumption of a where k is a constant (negative). When mass spherical shape would give an error in the drag is lost during descent, as is the case for meteors, coefficient most probably less than a factor of 2. then k must be a function of altitude—or, Under the assumption that the meteor is spheri- more conveniently, p. Thus we may express cal, then the velocity as In v8= (8) i so that m=-irr Pm o , dtf/dh 2 where pm is the meteor density. v '' - ••>!• « 638805—63 16 186 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Observed data for Meanook Meteor 132 * Epoch (t) Altitude (A) Speed (v) Luminous Acceleration seconds meters m sec"1 intensity (dv/dt) m sec~2 0 67. 59X103 17. 42X103 253. 7 .317 363. 1 .538 478.5 .980 824.0 1.0 52. 71X 10s 16. 58X103 847.5 -1. 88X10» 1.157 1038. 0 1.4 47. 12 X 10s 15. 45X103 1087. 0 -4. 02X103 1.423 1057. 0 1.732 602.5 1.8 42. 11X103 13. 10 X 10s 501.5 -8. 78X103 1.909 353.2 2.0 39. 98X103 11. 28X103 240.0 -9. 97X103 2.042 310.4 2.086 405.5 2.2 38. 21X 10s 9. 16 X 10s 37.35 -11.21X10* •Additional observations: Two pronounced flares observed at f=0.8 sec and i=2.08 sec (fig. 4); meteoric body observed to break in two pieces at t —1.75 sec (A=42.63X 10s m); velocity at entrance to atmosphere from extrapola- I 1 tion by Millman and Cook, »oo = 17.59X10 m sec" . Deduced data: entry angle Zfi=29.8° (cos ZR=0.868). Radiant power conversion: one zero-magnitude lumi- nous intensity equals 2.73 X10"" kilowatts-m-J. But since the density is very nearly an Thus between equations (8) and (15) one exponential function of h, that is, may find the values of as many of the C co- (10) efficients as one has values for v and dvjdt, given the variation of density with altitude. In the where /3 varies slightly with h, then calculations of this report, the 1959 AKDC atmosphere (AFCRC, 1959) was used for p and p+h(dp/dh). Our first attempt to find the C values using an IBM-704 computer was un- and equation (9) becomes satisfactory because severe oscillations of the functions became evident for low altitudes. It was surmised that this difficulty arose because (12) we had required equations (8) and (15) to have continuity through the altitude 42,630 meters Now we note that since the flight path is (corresponding to Epoch 1.75 sec) where the essentially straight so that the vertical velocity body was observed to break into parts. Ac- IB cordingly, we tried employing two sets of these dh „ (13) equations: one above the altitude of the break —=—v cos Zfi, and one below. These results are reasonably then satisfactory, but it was necessary to reduce v* dv_dv dh_ 1 dv2 (14) from 17.59 km per second (Millman and Cook dt dh dt 2 dh extrapolation) to 17.54 km per second to avoid so that having the radius decrease with increasing alti- tude at the highest altitudes. The coefficients in equations (8) and (15) in the two altitude (15) regimes are: METEOR SYMPOSIUM—ALLEN AND TOSHIKAWA 187 for h >42,630 m (70= 0.9772322 X 10 O, =-0.5223564 X 10s 5 O2=-0.8908289 X 10 8 <73= 0.6952320 X 10 (16) Ct= -0.3252341 X 10" 13 C6= 0.8007179 X 10 lfi C6= -0.8640033 X 10 for &<42,630 m Oscillations occur also with these functions, and it would probably be best to have allowed Velocity , v , km-sec"1 similar discontinuities at the altitudes corre- sponding to the two observed flares (at Epochs Deceleration,-dv/dt , km-sec'1 t—O.S sec and <=2.08 sec). However, the re- O IO 20 30 40 50 60 70 80 f sults obtained with the above coefficients are Meteor density - rodius product, fmr , kg m" certainly adequate for yielding the variation of FIGURE 5.—Observed and calculated characteristics for radius with altitude to evaluate gas-cap radia- Meanook meteor. tion Io and heat transfer X to be given later. The calculated values of v, dv/dt, and the mass loss was calculated for the various assumed product pmr are given in table 2 as a function densities and the luminous intensity, from the of altitude (along with the values of (3+h(df}/dh) ablated mass evaluated from the equation, used), and are shown in figure 5. The observa- tional data of table 1 are also shown as the dm circled points for comparison. The calculated (18) product pmr shows a decrease of about 20 per- cent at the observed altitude of breakup. The sum of IQ-\-IA is plotted along with ob- The data of table 2 were used to determine the served luminous intensity / in figure 6. This radius as a function of altitude for values of limit indicates the meteor density to lie between meteor density from 10,000 to 50 kg m~8. The 2,000 and 1,000 kgm~3. Comparing these two gas-cap radiation was then evaluated in accord- limits we chose water density (1,000 kgm~8) as ance with the method indicated earlier. The the most appropriate value, and evaluated the calculated values of Io are plotted on figure 4 radius and mass as a function of altitude along along with the total luminous intensity observed with the corresponding Knudsen numbers for the meteor. If T0 were zero, then Ia would (ratio of air mean-free-path to body radius) constitute all the radiation, and the indicated and the Reynolds numbers (the product of the density for the meteor would He between 1,000 diameter, density, and velocity divided by the and 500 kgm~3 representing one limit. To absolute viscosity). From the calculated Knud- evaluate the other limit, it was assumed that sen numbers (table 2) it is clear that the meteor experienced continuum flow throughout Jacchia's (1948) value of log TO= —18.19 (in his units) was applicable to the ablated mass-loss the observed part of the descent. The Rey- rate for this meteor. From the calculated nolds numbers are so low that the flow in the boundary layer must be laminar. With this variation of the product pmr, the time rate of 188 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS I I I H I I I O on Q H t-.* w o\vo Jt NS o tn en d f^ nof- HHOOHHCVIr?> tn B1? •NO ONOO IAiAIArnrj o o\c-vo t-w jo-* A ONC-VO IAIAIPINO t-ON-* evi c-rj oooooo ooooooooooooooo 8 8 88888 88 88 8888^8 88888 888 8888888 •a i li IfNlA lf\NO VQNO t-t^ao ONO\O\ ONOO ON ON00 lf\ 1 T3 :8888888888888888888 8: 9 C\ t^ ON d oo iTiSjo rn en 1 *-? !£!5ScH^S OO t— t^NO NO ITwt Q>oo cow co rH O\ t— NO t—oioaocooiAOeo IACOCO-* ON NO NO H CO.* CO H o\co IA CTvOO CO OJ ON CVJ -* -* CU CV1 CK IA CVli?\qNOHON>HCOHpNlAr-|IAQONHCOCOH CV1 IA ON-* ON-* HCONO tAlTNlAI^p IA3; O ON H CO t- Q P-^t H CO IA CO IA Oi ON O cvl cvj e\i co co^t IA IAN2 r-co ON O OJ COIACO O-* t-HvooNOcoHCUON t—co V- w HHHHHOICMdlro (nj -* irsNO I—00 oo ON O f S- CO t— ONCO CO IA CO t— 1ACO -=»• co^t- t— H -* ON t~-NO CO IA O NO CO C—-* 1ANO ON CO ON < tA-4; CM Q CO NO CO O NO CU C- CM VO ON >-{ CO CO r-l C— H CO CU Q> H CO Q\-* p iH O O 5-N. -*-*-*.* COCOCOCOCVlOJi-lrHO ONONOO t—NO 4 rtH ON NO -=t OvDCVlF-OCUcOHO METEOR SYMPOSIUM—ALLEN AND TOSHIKAWA 189 Time, seconds FIGURE 6.—Meanook meteor luminous intensity and comparison with sum of gas-cap and maximum ablative radiation for spherical bodies. knowledge it is possible to estimate two limiting When ablation by vaporization occurs, the values for the ratio vapor which is pushed into the boundary layer partially fends off the hot air from the surface, thereby reducing the shear and hence the heat- transfer coefficient. The ratio of convective heat transfer for vapor ablation to that for corresponding to ablation by vaporization on liquid ablation may be obtained (Adams, 1959) the one hand, and fusion on the other. The heats of ablation used were 8X106 m2 sec"2 as 9 2 2 and 1.4X10 m sec~ (usual values for stone), (20) respectively, and again y of % was employed. The value of the heat-transfer coefficient has two contributions, one due to radiative heating where for the constant K a value of 0.3 is a good from the gas cap and the other due to the con- average. vective heating resulting from the shear in the In table 2 values for Xcy, \?F» and X© are given, laminar boundary layer. The radiative com- and are appropriately combined to give the ponent, X<7, was calculated from the total gas- expected limits for X/yf for ablation by vapor cap radiation as previously discussed assuming run-off and liquid run-off. The latter quanti- an absorptivity of unity for the surface. This ties are plotted in figure 7. value applies if either fusion or vaporization is Finally, the value of X/yf can be inferred the process of ablation. This is in contrast to directly from the experimental data by use of the convective heating, since the convective equations (2), (4), (5), and (13): component for ablation by fusion, \, , differs r 16p cos Z /dr markedly from that for ablation by vaporiza- w R (21) pv2 \dh tion, \,y, as discussed by Adams (1959). For ablation by fusion, the convective heat- Since r varies somewhat erratically with h, the transfer coefficient is that which would occur derivative was evaluated at any altitude k by for a solid body. If Lees' (1957) estimate is dividing the value of r at A+1000 m less r at used (or see Chapman, 1959), k—1000 m by the AA=2000 m. The resulting computed values from equation (21) are given in table 2 and plotted in figure 7. Even with 0.000216 (19) the smoothing the computed values do not pro- 190 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS tib O .5 1.0 1.5 2.0 25 3.0 3.5 Heot transfer factof.-jj xlO7, sec' m'* FIGURE 7.—Calculated and experimental values of the heat-transfer factor for Meanook meteor. duce a smooth variation, but clearly it is evi- speeds. A possible reason for this is that T0 dent that the values are reasonably close to the might vary as a negative power of velocity, but computed limit for vapor ablation. It should the required variation appears far more than be noted, however, that this does not mean that can be considered reasonable, particularly since a substantial loss of surface material in the the ablation radiation in the present case must liquid state or even the solid state does not occur, for, as has been shown by Dean R. Chapman in a yet unpublished work on the ablation of tektites, even a small fraction of ablation by vaporization serves to depress the heat transfer by convection. Discussion Light.—It is clear, in the present case, that 1=0 equation (3) for the calculation of mass cannot be used since the appropriate value of r0 and its variation to velocity probably bear no rela- tion to the values and variations suggested by Opik and Jacchia. In any event, T0 would be expected to be much less than that for the 2 A .6 free-molecule regime and, since even using Gas cap radiation fraction. Jacchia's value gives the relative proportion of FIGURE 8.—Gas-cap radiation as a function of maximum ablation light to gas-cap light as low as 35 per- possible total radiation. cent as shown in figure 8, the difference between the total radiation and the gas-cap radiation be much smaller than we have assumed. There must be so small that no precision from equa- is no doubt that there can be a considerable tion (3) could be expected. error in our estimates of the gas-cap radiation, There is a disturbing trend of the gas-cap for, as noted earlier, the actual meteor body light as compared with the observed luminous may not be near spherical in shape and no intensity shown on figure 6, since, if we match doubt changes in shape in the course of descent. the two at any given point, the calculated gas- Also, the gas cap is well out of chemical equi- cap light exceeds the observed lumination at librium over most of the range, and the meteor the higher speeds, but is less than at the lower velocities here are beyond those for which we METEOR SYMPOSIUM—ALLEN AND YOSHIKAWA 191 I si i FIGURE 9.—Equilibrium gas temperature as a function of altitude and velocity. 1 J 1 1.23,000 •» 1 —i— ^T. 12,000 V — n -1 n ! —i J r n ! ,- ft/ jjh H 11 i i / J 1 .4 .5 Wow lt*9U>, p. , o FIGURE 10.—Power spectral density of gas-cap radiation for equilibrium air. have obtained data for gas radiation, so that ables. It is seen that even for equilibrium flow the ^results given had to be extrapolated the gas temperature would vary from about extensively. 17,000° K early in the entry down to about However, it should be noted that errors in 8,000° K or less near the end of the flight. assessment of the gas-cap radiation may not be For the actual gas flow, which is well out of the principal cause for the observed differences, chemical equilibrium, the maximum tempera- but rather the observed intensity interpreted as ture would be considerably higher. Figure the total radiation emitted may be in consid- 10 shows the theoretical estimates of the erable error itself. This can be seen from the power spectral density for the gas-cap radiation following: In figure 9 are shown the gas-cap in equilibrium for 8,000° K and 12,000° K temperature and density as a function of (Meyerott et al., 1960), and for 23,000° K velocity and altitude for equilibrium flow. The (Armstrong, 1959) for a gas-cap density appro- heavy dashed curve shows the trajectory of priate for the Meanook meteor. This figure the Meanook meteor expressed in these vari- shows that for wavelengths well into the 192 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS ultraviolet important contributions exist, par- many points if it is called for at the altitudes ticularly when the velocity is high. However, corresponding to the flares. the observed intensity for the meteor was The relation of the computed vapor ablation obtained from measurements essentially in the and liquid ablation values to the experimental range from 0.3 to above 0.6 micron. Thus the values of the heat-transfer factor looks reason- observed luminous intensity approximates the able. The actual heat-transfer factor is ex- total radiation emitted only near the end of the pected to be more nearly allied with the vapor flight when the speed is low. In the early case even though a sizable fraction of the portion of the flight, when the speeds are high, ablated material is liquid, or even solid, since the total radiation emitted must be much the vapor blowing in the boundary layer larger than is indicated by the plot in figure 4. greatly reduces the convective-heat transfer; In consequence, the disagreement may not be thus the heat-transfer coefficient is well approxi- as real as the present evidence indicates. In mated by the calculated vapor value. any event, the density of the meteor is probably less than our analysis has indicated. Future prospects Data reduction.—It was noted earlier that a It is to be expected that future research in our better analysis of the series solution for v and ground laboratories will prove very useful to dvjdt would have been to break the solution the meteor astronomer by providing better val- (to avoid the necessity for continuity) at the ues for the total gas-cap radiation and its altitudes corresponding to the two observed power spectral density. Moreover, it should flares. The data, as presented by Millman be possible to better understand the nature of and Cook, are insufficient to allow such a the radiation from the ablated mass in the procedure. It would clearly be valuable to future. Future work should also permit much attempt to reanalyze this meteor using the better estimates for the effect of blowing due original unfaired data from as many points as to vaporization in the boundary layer on the there were shutter breaks on the original films. convective-heat transfer. The principal limi- However, the analysis presented gives results tations of the ground laboratory work in the that, in general, agree with the observed past have resulted from the fact that it was behavior of the meteor. In particular, as not possible to reach even the lowest of meteor noted on figure 5, the computed values of the speeds in the laboratory. Recently, however, radius-density product show an increase in we have managed to enter the meteor speed negative slope at the altitudes corresponding regime, and have made measurements of gas- to the flares,a s would be expected. Moreover, cap radiation and ablation radiation up to it was noted by Millman and Cook that when speeds of 13 kilometers per second. We do the breakup of a meteor occurred, a small not expect ever to be able to reach the high- piece was kicked forward ahead of the main est meteor speeds, but we will probably improve body. This could account for the unexpected the present situation materially. slope observed for the density-radius product Dr. McCrosky recently suggested the con- at altitudes just below the altitude for breakup. struction of a series of optical stations in the Heat-transfer factor, X/7f.—It is of interest Mississippi Valley to track fireballs. The pur- to note that the experimental values for X/7f pose of this system would be primarily to expe- are consistent with the observations of the dite finding the meteorites so that their radio- flares in that an increase in X/yf is noted near activity could be analyzed soon after their the altitudes corresponding to the flares. How- arrival on earth, thus improving our knowl- ever, there are several other increases noted at edge of cosmic origin and environment. The altitudes where no flares were observed, but system McCrosky proposes could also, in our these may result from the fact that in the opinion, provide useful tracking data to obtain analysis the velocity function was forced to be a better understanding of the heating of mete- continuous at the flare points. The required ors. The analysis of Meanook Meteor 132 is continuity promotes residual oscillations at very encouraging in this respect. Clearly, the METEOR SYMPOSIUM—ALLEN AND TOSHIKAWA 193 present methods are capable of giving the veloc- CAMM, J. C; KIVEL, B.; TAYLOR, R. L.; AND TEARE, ities and altitudes and the flight path with J. D. about the precision required for such analyses. 1959. Absolute intensity of non-equilibrium radi- ation in air and stagnation heating at Examination of the meteorite shape, mass, and high altitudes. Research Report 93, material properties would play a very important Avco-Everett Research Laboratory. part in such heat-transfer analyses. However, CHAPMAN, D. R. it is evident that not only should McCrosky's 1959. An approximate analytical method for system be designed to give the best coverage studying entry into planetary atmos- pheres. Report 11, National Aeronau- for trajectory evaluation, but also the cameras tics and Space Administration. themselves should be designed to give radia- HOCHSTIM, A. R. tion response on the plates for a wide enough 1957. Gas properties behind shocks at hypersonic range of wavelengths. Also, adequate spectro- velocities. I. Normal shocks in air. graphic equipment should be provided. In Convair Zph(GP)-002. JACCHIA, L. G. this way very valuable information on gas-cap 1948. Ballistics of the upper atmosphere. Har- and ablation radiation would be obtained for vard Coll. Obs. and Center of Analysis, speeds far higher than we will probably be Massachusetts Inst. Tech., Tech. Rep. able to obtain in the laboratory. No. 2. KIVEL, B. References 1959. Radiation from hot air and stagnation heating. Research Report 79, Avco- ADAMS, M.C. Everett Research Laboratory. 1959. Recent advances in ablation. Journ. Amer. KIVEL, B., AND BAILEY, K. Rocket Soc, vol. 29, pp. 625-632. 1957. Tables of radiation from high temperature air. Research Report 21, Avco-Ever- AIR FORCE CAMBRIDGE RESEARCH CENTER ett Research Laboratory. 1959. Abbreviated English tables of the ARDC LEES, L. model atmosphere. Geophysics Re- 1957. Recent developments in hypersonic flow. search Directorate, Air Force Cambridge California Inst. Tech. Publ. 428; also, Research Center, ARDC. Jet Propulsion, vol. 27, pp. 1162-1178. LlNDEMANN, F. A., AND DOBSON, G. M. B. ALLEN, H. J. 1923. A theory of meteors and the density and 1960. On the motion and ablation of meteoric temperature of the outer atmosphere to bodies. In N. J. Hoff and W. G. which it leads. Proc. Roy. Soc. Lon- Vincenti, eds., Aeronautics and astro- don, ser. A, vol. 102, pp. 411-437. nautics, Proc Durand Centennial Con- MEYEROTT, R. E.; SOKOLOFF, J.; AND NICHOLLS, R. W. ference, pp. 378-416. Pergamon Press, 1960. Absorption coefficients of air. Report New York. LMSD 288052, Lockheed Aircraft Corp. 1961. Problems in atmospheric entry from para- MlLLMAN, P. M., AND COOK, A. F. bolic orbits. Journ. Japanese Soc. 1959. Photometric analysis of a spectrogram of a Aeron. Space Sci., vol. 9, no. 85, pp. very slow meteor. Astrophys. Journ., 43-50. vol. 130, pp. 648-662. OPIK, E. J. ALLEN, H. J., AND EGGERS, A. J., JR. 1958. Physics of meteor flight in the atmosphere. 1958. A study of the motion and aerodynamic Interscience Publ., New York. heating of ballistic missiles entering the YOSHIKAWA, K. K., AND Wick, B. H. earth's atmosphere at high supersonic 1961. Radiative heat transfer during atmosphere speeds. Report 1381, NACA. entry at parabolic velocity. Technical Note D-1074, National Aeronautics and ARMSTRONG, B. H. Space Administration. 1959. Mean absorption coefficients of air, nitro- ZlEMER, R. W. gen, and oxygen from 22,000 to 220,000° 1960. Extended hypervelocity gas dynamic charts K. Report LMSD 49759 (July 15), for equilibrium air. Space Technology Lockheed Aircraft Corp. Labs., Inc., TR-60-0000-09093. Preliminary Notes on Some Results of Photographic Multiple Meteorite Fall of Pfibram By Zd. Ceplecha1 This paper is based on the double-station face is negligible in the range of temperatures photographs of the meteorite fall of Pfibram. under consideration. The details of this first photographic meteorite We measure the distance x from the boundary fall have been published (Ceplecha, 1961). plane into the body, and make x=0 on the Ten plates are available of the fall, which was boundary plane. Let T be the temperature of photographed from two stations 40 km apart; the body in the atmosphere as a function of the four stony meteorites that were recovered the time t and the distance x. Let To be the belong to different photographed trajectories temperature of the whole body after a very of individual fragments. long time; To is, of course, equal to the tem- The photographs allow the determination of perature of the undisturbed air. We can then the end heights of the individual fragments into define the temperature T according to the which the meteorite was split. The velocity at equation the end point can also be found by computing T=T0+r. (1) the invisible ("dark") part of the trajectory. Thus it seems to be of great importance to If X is the heat conductivity, 8 the density, and derive theoretically the connections among c the specific heat of the body, we can define different parameters at the end of the visible the coefficient /8 by the equation ("light") trajectory of a large body. This is simpler than studying the visible trajectory (2) itself, because its end can be taken as the begin- V ning point of the invisible trajectory. Then we can formulate our problem by the well-known equation of heat conduction: The end height of the visible trajectory The thickness of the meteorite crust is usually I2 (3) very small relative to the dimensions of the dz body itself. Thus we can solve the problem by assuming a semi-infinite space full of meteorite It is useful to choose the zero point of the mass, the boundary of which is a plane. Air time at the moment when the light trajectory of density p moves with velocity v perpendicu- ends; then larly to the surface of this boundary plane. tB=0. (4) Let the fraction A of the total kinetic energy The initial condition for r can then be written in of the moving air be transferred through the the form boundary plane into the body. Thus A is the r , (5) heat-transfer coefficient. The energy in the jr body itself can be transferred to the air only by where rB is the temperature at the surface when heat conduction. The radiation from the sur- the light trajectory ends. The end condition, from equation (1), is then 1 Astronomical Institute of the Czechoslovak Academy of Sciences, Ondfejov, Czechoslovakia. T(«,x)=0. (6) 195 196 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS The boundary condition is given by the merical factor 2, which originated from equa- equation tion (7), under the assumption of a plane bound- ary. The energy in this case goes in only one direction. In reality, the energy is distributed over all directions in space. We shall assume This is the complete mathematical formula- uniform distribution of the energy. This tion of the problem. The complete solution assumption is supported, for the jEJ-point, by of equation (3) with the additional conditions the fact that there was no difference in the in equations (5), (6), and (7) is not possible in thickness of the crust between the front and the analytical form. But it is possible to solve back of the Velka meteorite (one of the Pfi- these equations exactly in analytical form if bram multiple fall). Then the numerical we are satisfied with a solution at one point only. factor 2 in equation (11) will be four times This point will be the point E, where the light greater and we arrive at the equation trajectory ends. The value pi? in equation (7) is an unknown _|_ (12) transcendental function of time. An analytical 8\(-k)1/2' function of time can be substituted on the assumption that both these functions are which is valid if we can neglect the curvature of identical in their values and first derivatives. the meteorite surface as a factor in the concen- Then the whole problem can be solved by ana- tration of the heat conducted inwards. This lytical functions. is possible, because the dimensions of the crust Let us take of the Pfibram meteorites are smaller by three (8) orders of magnitude than the dimensions of the bodies. Numerical computations of the heat- where the subscript E denotes values on the transfer coefficient presented in the following surface of the body at the end of the light section are based on equation (12). trajectory, and A: is a coefficient which must be The solution of the problem contains the determined. If we use the exponential law of term (—k)m, which of course must be real. the air density This is possible only if the equation p=Poe~M, (9) b cos z (13) then the condition of equal values and deriva- P*> 3r- tives together with the drag equation will yield m for k the equation is valid. Equation (13) is a very good cri- k=bvB cos zB—3T — pBvB. (10) terion for testing whether a meteorite of mass m m can follow a fireball that ends at a height Here m is the constant mass and s the constant hB. The numerical values obtained from equa- area of the cross-section; T is the drag coefficient, tion (13) are presented in table 1. and zB the zenith distance of the radiant. Now, if we replace the right-hand side of equation (7) TABLE 1.—Critical and end heights (km) of a fireball. by equations (8) and (10), then equations (3), (5), When hE of the fireball is less than the presented value, (6), and (7) can be solved. Without going into a stony meteorite of mass m can follow (for iron mete- details of derivation, I write the solution directly orites the heights are S or 4 km lower) mass m _ 1/2 (11) COS ZB 2\(-Ar) ' 100 gr lkg 10 kg 100 kg 1,000 kg where k is given by equation (10). This is the exact solution, but it is valid for one point only. 1 37 32 27 23 17 0.5 42 36 31 27 22 However, we are nor interested in solutions at 0.1 55 48 42 37 32 other points. Equation (11) contains the nu- METEOR SYMPOSIUM—CEPLECHA 197 The heat-transfer coefficient Equation (12) was used to compute the end Equation (12) can be used for the computation velocity vB, with different values of A, shown of the heat-transfer coefficient A, if other in table 2. The end velocities of the same values are known. In the case of the Pfibram meteorites can be estimated from the computed and the observed dark-flight distances. The meteorites, we have pB and b from the end height Us which was directly measured from end velocity for the Luhy, Velkfi, and Hojgfn the double-station plates. The zenith distance meteorites was provisionally estimated to be 7 of the radiant z was also measured directly. km/sec, and for the DraJkov meteorite to be 11 R km/sec. Thus it is evident from table 2 that The end velocity vB can be computed from the dark-flight distance of meteorites, but the value the heat-transfer coefficient at the end of the is only provisional at present, because the light trajectory was about 0.002 or 0.003. wind-tunnel measurements have not yet been The thickness of the meteorite crust completed. The same holds true for the prod- The solution of equations (3), (5), (6), and (7) uct r— (Ceplecha, 1961). The surface tern- can be presented in a more general form if we TYh perature at the beginning of the light trajectory are not interested in the surface temperature of a meteor was computed by Levin (1956). alone. If r(x) is the temperature in the body It is evident that this value, giving the limiting at the distance x from the surface, then the conditions for the rapid evaporation, is the equation same for the end of the light trajectory. The 1/2 T(x)=r exp[-(-A:) x/ 8] (14) density of the meteorites was measured, but Jf i the measurements of the heat conductivity are is valid. If we use the temperature of the not yet finished. The heat conductivity X is of melting point rm, the thickness d of the meteor- course the main factor in equation (12), and ite crust follows from equation (14): strongly determines the numerical results. We have not yet obtained values for the heat con- ductivity from direct measurements, and there- 0 log I*j d= 1/2 (15) fore use the mean value for stony meteorites 0.434 (-£) ' as given by Opik (1958). Thus the numerical values given in tables 2 and 3 must be taken as The theoretical thickness of the crust for provisional, and are accurate to an order of A=0.002 is about 0.1 mm for all the meteorites magnitude only. of the Pffbram fall. The numerical data are presented in table 3. The measured thickness TABLE 2.—Computed end-point velocity VE (kmfsec). of the crust for the Velka meteorite is between From equation (12), with the following constants: T =2.1X10S; X=2X105; c=lX107; 5 = 3.6; cos 0.18 mm and 0.29 mm—about twice or three E times the theoretical value. This agreement zR=0.683 is quite good, as many provisional values were Heat-transfer coefficient A Meteorite used in the computations. 0.01 0.005 0.002 0.001 References Luhy 4.0 5.3 7.6 10.1 CEPLECHA, Zd. Velkii 5.2 6.9 9.9 13.1 1961. Multiple fall of Pffbram meteorites photo- HojSin 5.8 7.7 11.1 14 6 graphed. 1. Double-station photographs Draikov 6.5 8.5 12.3 16.3 of the fireball and their relations to the found meteorites. Bull. Astron. Inst. Czechoslovakia, vol. 12, pp. 21-47. TABLE 3.—Computed thickness d of the meteorite crust LEVIN, B. J. (m=1.5X10s, A = 0.002) 1956. The physical theory of meteors and Meteorite d(cm) meteoric matter in the solar system. Luhy 0. 011 Publ. House Acad. Sci. U.S.S.R., Moscow. Velkii . 009 OPIK, E. J. HojSin . 006 1958. Physics of meteor flight in the atmosphere. DraJkov .015 Interscience, New York. 198 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Abstract The theory presented in this paper can be applied to meteorite falls. The heat-transfer coefficient at the end point of the light trajectory and thickness of the meteorite crust can be computed. The preliminary value of the heat-transfer coefficient at the end points of four Pfibram meteorites was estimated to be about 0.002 or 0.003. The definitive numerical values will be published in the Bulletin of the Astronomical Institute of Czechoslovakia in a series of papers entitled "Multiple Fall of Pffbram Meteorites Photographed." Results from an Artificial Iron Meteoroid at 10 km/sec By R. E. McCrosky* and R. K. Soberman2 The results described in this paper are derived mass loss. Presumably the velocity of the from a successful attempt to produce a meteor meteor will also influence the rate of luminous in the upper atmosphere with a body of known energy production. As a minimum statement mass and composition. This was accomplished of the problem one can write: by accelerating a small pellet of stainless steel by an air cavity charge attached to the nose of dm (1) a high-speed re-entry rocket. The experi- ment was made possible only through the cooperation of a number of organizations. where / is the instantaneous intensity, m the The test vehicle was a Trailblazer I rocket, meteoric mass, and /(») some function of the developed by the Applied Material and Physics velocity. The data obtained from observa- Division of the National Aeronautics and tions of natural meteors are consistent with an Space Administration at Langley Field, Va., expression of the form for use in a re-entry research program conducted by the Massachusetts Institute of Technology /=TU ~ (2a) Lincoln Laboratory. The optical observations described were made by the Harvard Meteor or T0 dm ~ T (26) Project with considerable technical assistance I=-^-dt*' from Lincoln Laboratory. The air cavity charge was constructed and tested by the Air where T0 is the luminosity coefficient and T=T0 V Force Cambridge Kesearch Laboratories. is the ratio of the observable luminous energy The primary purpose of the experiment was to the kinetic energy. the determination, to within narrower limits Certainly other factors besides the mass loss than previously available, of the efficiency with and velocity also influence the luminosity, which the kinetic energy of a meteoroid is notably the composition of the meteoroid and transformed to luminous energy in the photo- its aerodynamical properties. But at the graphic region. From a simple inspection of present time one must be content, generally, meteor spectra, one can ascertain that a great with the simplified theory implicit in equation majority of the meteor luminosity is produced (2). The considerations of additional factors by the excitation of the meteoric atoms, partic- must await a determination of such basic ulary for the luminosity in the photographic parameters as the meteor mass and density. (3500 to 5000 A) region. Whether the abla- Equation (2b) may be integrated to give tion is by direct vaporization, melting, or fragmentation, the meteoritic material must be m=mp=— I -5 dt- (3) vaporized before appreciable luminosity is produced. The vaporized atoms are then Since / and v are observable quantities, the excited by collision. The instantaneous inten- mass of the meteor can be determined if T is sity of a meteor must be related to the rate of 0 known. Masses determined from the observed • Harvard College Observatory, and Smithsonian Astrophyslcal Observ- luminosity are referred to as photometric atory, Cambridge, Mass. masses (mp). A number of estimates of this > Geophysics Eesearch Directorate of the Air Force Cambridge Re- search Laboratories, Bedford, Mass. or related quantities are available (Whipple, 199 200 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 1943; Cook, 1955; Opik, 1955; McCrosky, 1961; partly of solid or liquid fragments, these two Cook, Jacchia, and McCrosky, in this symposium masses should not agree and are not inter- (p. 209); Lazarus and Hawkins, in this sympo- changeable. If fragmentation were the only sium (p. 221). However, the scatter in the important factor, the photometric mass would values is extreme; a factor of more than 10a never be smaller than the dynamic mass and separates the highest from the lowest estimate equation (5) would yield a lower limit on the and this uncertainty, or more, enters essen- density. Unfortunately, however, mp may tially into all mass determinations of meteors. be underestimated. We have assumed in The mass of a meteoroid can also be deter- equation (3) that the entire mass has been mined, in principle, from the observed decelera- vaporized by the end of the trajectory. This tion. The drag equation, condition is reasonably well met for high- velocity meteors, but for low-velocity meteors de_ rAd~2/3pv2 there may be a substantial terminal mass 1 3 (4) dt~ m ' ' that is not accounted for in mp. For this reason, md may exceed trip if fragmentation includes as unknowns the drag coefficient T, is absent, and the derived value of 8 would be the mass, and a shaped factor A. Here an upper limit. AS~2/3=Am~m, where A is the presentation To summarize, observations of natural me- area and 8 the density of the meteoroid. The teors have not yet yielded reliable data on the atmospheric density p is known to sufficient precision at meteor heights. In view of the luminous efficiency, meteor masses, or meteor large uncertainties of m and 5, one is prompted densities. The present experiment was in- to assume reasonable values for T and A and tended to supply a reliable value of the luminous solve equation (4) for the unknown: efficiency of one meteoric element, iron. The experiment The Trailblazer I rocket (see pi. 1) is normally a 6-stage vehicle; the first three stages hit a If the meteoroid density were known, one velocity package (the final three stages) to an could determine a "dynamic mass" (md) from altitude of approximately 300 kilometers and equation (5). There is considerable evidence these three stages fire downward in sequence. (e.g., see Whipple, 1955; McCrosky, 1955; The last, or sixth stage, is a 5-inch spherical and Jacchia, 1955) that the majority of rocket motor which, after burnout, enters the meteors, those of cometary origin, are extremely atmosphere at 6 or 7 km/sec and serves as a fragile. The weak structure may be associated test object for re-entry studies. In the case with bodies of low density and one cannot of the Trailblazer I used for the meteor ex- safely assume that the density of these meteor- periment, an air cavity charge was mounted on oids is that of meteoritic stone. Until inde- the 5-inch rocket motor as a seventh stage of pendent measures of meteor densities are propulsion (see pi. 1). This explosive pro- made, and this is not immediately likely, jectile accelerator is shown in plate 2. It equation (5) is of limited value in making esti- consists of a chemical delay fuse, a "tetryl" mates of masses. booster and a cylindrical charge of plastic If the mass can be determined from equation bonded RDX molding explosive. The air (3), it would appear that the density can be cavity in the RDX is sealed by a 5.7-gram derived from equation (5). However, some pellet of Stainless Steel 304. When ignited, caution is required here. The dynamical mass a plane detonation wave—utilizing the air represents the mass of the major meteoroid in the cavity as a buffer—propels a portion of body, whereas the photometric mass refers the steel pellet forward with a velocity of about to that portion of the entire meteoroid that has 4 km/sec. These projectiles have been re- not yet produced appreciable luminosity (i.e., covered intact in the laboratory to determine that part which is not yet vaporized). If the their mass. Such a recovered projectile is entire mass consists partly of a parent body and shown, next to an unfired pellet, in plate 2. McCROSKY AND SOBERMAN PLATE 1 Left, Trailblazer I-G 7th-stage rocket. Right, 6th-stage spherical rocket engine and 7th-stage gun. Top, 7th-stage pellet before and after detonation. Bottom, Disassembled 7th-stage gun showing test pellet. METEOR SYMPOSIUM—McCROSKY AND SOBERMAN 201 BEFORE 8EF0RE EXPERIMENT EXPERIMENT AFTER EXPERIMENT AFTER EXPERIMENT I 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4 7 4.8 4 9 2.1 2.2 2.3 VELOCITY (KM/SEC) MASS (GRAMS) FIGURE 1.—Velocities achieved by air cavity charge during FIGURE 2.—Post-detonation masses of air cavity charge pellet ground tests. determined from ground tests. sidered. The submeteoric velocity (see table The resultant velocity and mass of these test 1) of the pellet, although not intentional, is projectiles are shown in figures 1 and 2. All virtual proof of the origin of the body. Fur- the charges referred to in these figures, as well thermore, an upward extrapolation of the as the charge utilized in the experiment, were trajectories of the sixth and seventh stages fabricated in an identical manner from the indicates a coincidence of these objects in space same batch of materials and in time that agrees, to within the accuracy In producing the artificial meteoroid, the of the observation, with the position and time air cavity charge was to be fired shortly after at which the seventh stage was intended to burnout of the 5-inch rocket motor at an alti- fire. The photographs certainly refer to the tude of 195 kilometers. h seventh stage pellet. The seven-stage Trailblazer was fired at 00 56m 20' E.S.T. on April 21, 1961. The re-entry The photographs were reduced by the stand- of the steel pellet was observed visually at ard precision method described by Whipple and 01h 02m 00B by many observers. It was photo- Jacchia (1957). The summary of the results graphed by the Super-Schmidt meteor cameras are given in table 1. We followed the usual at the Arbuckle Neck, Va., and the Eastville, meteor analysis procedure and, by least squares, Va., stations of the M.I.T. Lincoln Labora- fitted the observed distances D along the trail tory; and also by smaller cameras operated by by an equation of the form the National Aeronautics and Space Adminis- D=a+bt+cekt. (6) tration at Eastville, Va., and at Coquina Beach, N.C. The re-entry of the sixth stage The values of the constants, and their probable was also observed visually and photographic- errors, are included in table 1. D and t take ally. the value zero at the beginning point. Since it was conceivable that the meteor Photometry.—The same photometric tech- would be near or below the observable thresh- niques described by Whipple and Jacchia have old, duplicate observations were made when- been used to determine the meteor light curve. ever possible. For this reason we relinquished Each dash (shutter opening) of each Super- the chance of spectrographic observations Schmidt trail was compared with trailed images with the Super-Schmidt equipped with a prism. of stars in the Harvard Standard region at 5= + 15°, a=75°. The comparison film had Observational results been obtained earlier in an excellent (New Trajectory.—The probability that the observed Mexico) sky. In addition to the usual correc- meteor was of natural origin and not the result tions for reciprocity failure, the relative trailing of the experiment is small but must be con- rates of star and meteor, and the distance of 638805—63 17 202 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Trajectory data derived from Super-Schmidt between the comparison and the calibration photographs of re-entry pellet films. However, the value of the air mass Duration 0. 77 sec coefficient is surprisingly low for a sky at sea Maximum Mv, -0.4 level on the east coast. Perhaps this should Velocity at beginning 9. 8 km/sec be attributed, along with the nonzero value of Beginning height 69. 4 km a, to a nonuniform sky that was relatively poor End height 62. 6 km Beginning range 254 km near the zenith because of local haze but End range 254 km degrading less rapidly than normal at increas- Zenith distance at midpoint 76° ing zenith angle. The cause of this slightly Radiant (1950) a=204?5 anomalous behavior, whatever it might be, 5= 37?5 does not concern us. The re-entry film and Zenith distance of radiant, ZR 15° Constants of equation (6) comparison film were identical emulsions ex- a= 0.012±0.020 km posed on an identical sky. The photometric 6= 9.787 ±0.046 km/sec measures of both re-entry and calibration c= 2.8X10-«±0.3X10-*km 1 films were made against the same comparison k= 12.5 sec" film. Thus our "absorption" measures repre- the meteor, very considerable absorption cor- sent corrections that must be applied to the rections were also indicated. The sky condi- re-entry film whether these departures result tions at the time of the experiment were from actual absorption or from changes in particularly good for the locale, but the obser- system sensitivity. vations were made at a low elevation. Cali- The corrected measures are plotted in figure bration photographs of the re-entry region were 3. The agreement between the results from made, with each Super-Schmidt camera, within the two stations is generally excellent. The minutes after the event to provide for these free-hand curve drawn to the points has been absorption corrections. The cameras were taken to be the true light curve for the re- kept stationary for these exposures. About mainder of this analysis. 20 star trails on each Super-Schmidt calibration film were compared with star trails on the Luminous efficiency comparison film. The calibration stars were Equation (3) gives the luminous efficiency as chosen at about the re-entry azimuth and at zenith distances from 48 to 85 degrees. True T JL dt> (8) apparent photographic magnitudes for these stars were taken from the Henry Draper catalogue, and the apparent absorption values where T is the duration of the meteor. We were fitted by least squares to the equation will assume at the outset the mass to be 2.2 grams, realizing that the terminal mass may (7) not be negligible for this pellet and that, therefore, the luminous efficiency derived will be where Z is the air mass above sea-level at the a lower limit. The intensity / is given by the zenith distance of the star. The observed light curve. We have included in the inte- values fit the equation over the entire range of grated intensity those extrapolated regions at the zenith distance with a mean scatter of 0.3 the beginning and the end of the light curve magnitude. Some of this scatter is certainly when the meteor was fainter than film limit. due to errors in the Henry Draper magnitudes. The intensity as used here is defined by: We obtained the following values for the two stations: MM=-2.5log/. (9) a b Arbuckle Neck 0. 17 0. 30 The velocity at the end of the trajectory is Eastville 0.77 0.24 rapidly decreasing and not particularly well determined. The formal solution of the tra- The nonzero values of the constant a can be jectory suggests a velocity of 5 km/sec at T= interpreted as a difference in the filmsensitivitie s 0.75 second. If this velocity is accepted, METEOR SYMPOSIUM—McCROSKY AND SOBERMAN 203 the contribution to the integral in the region + SS from 0.7 to 0.9 second becomes inordinately -0.5 - o SK large and unrealistic because of the uncertain o extrapolation of the light curve and the poorly 0 + - / \ determined velocity. If one arbitrarily sets c the velocity at 6 km/sec whenever the formal solution yields a lower value, one obtains 19 0.5 / \ To=8.6X10~ (cgs and MPg). If instead of a limiting velocity of 6 km/sec, one chooses \ 8.5 km/sec (the point on the descending branch 1.0 \ of the light curve for which the contribution of \ I/i? first exceeds that of the previous inte- S 1 I 1 19 gration step), then TO=6.6X1O" (cgs and 0 .1 .2 .3 .4 .5 .6 .7 8 .9 Mpg). We will take t (tec) FIGURE 3.—Light curve of the artificial meteor. 16 T0=8X10- (cgs and Mpt) (10) The contribution of both Ni and Mn to the as the observed value in our discussion. luminosity is certainly small and, in any case, One must clearly recognize the limitations is probably comparable to the contribution of a attached to this value. It is a lower limit to similar mass of iron. However, the gross dis- the efficiency of stainless steel at velocities of crepancy in the relative abundance of Cr is a about 10 km/sec. In attempting to derive matter for concern. To derive a value for the other values from this, one must enter areas of luminous efficiency of iron alone, one must speculation. In the following sections we will first know the ratio of the efficiency of Fe to investigate three corrections that are necessary that of Cr. A few Cr I lines are seen in the to convert the value in equation (10) to a lumi- best spectrum of a low velocity meteor. They nosity coefficient valid for meteoric material at are generally blended with the iron lines which 10 km(sec. The corrections cannot be derived are far more prevalent. If one accepts that the without assumptions. We offer the following Cr/Fe ratio found in chondritic meteorites or analysis as a "best estimate" for those who may in cosmic material generally (^1%) is also require it. If the mass-luminosity problem valid for cometary meteoroids, then the avail- of meteor physics is to be solved by the present able data on meteor spectra and on the abso- experimental techniques, observation of a lute oscillator strengths of the important number of materials at several velocities will spectral lines of the two elements suggests that be necessary. the luminous efficiency of Cr does not differ Correction jor the composition of the pellet. enormously from that of Fe. For the present, —The choice of pellet material was dictated by we prefer the assumption that T0(Cr)=T0(Fe), the physical properties rather than by the and that the value given in equation (10) chemical composition of the metal, and for for the steel pellet is representative of pure this reason the pellet composition does not iron. If, instead, we assume as probable match either meteoritic iron or the metallic extremes either that T0(Cr)=l/10r0(Fe) or components of meteoritic stone. A comparison To(Cr) = 10T0(Fe), then the corresponding val- 18 19 of the major constituents of these materials ues for ro(Fe) become 1.0X10" and 2.9X10" , relative to the iron content is given in table 2. respectively. Correction jor terminal mass.—The rate of TABLE 2.—Composition of various materials relative to mass loss from a nonfragmenting meteoroid is their Fe content proportional to the energy flux and inversely Stainless Met. Met 304 stone iron proportional to some energy of ablation f (heat Fe 100 100 100 of vaporization or fusion), and is given by Cr 28 1 trace Ni 15 5 9 dm Mn 3 1 (ID 204 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS where A is the heat-transfer coefficient. Divid- Levin (1956) suggests that intense vaporiza- ing equation (11) by equation (4) and integrat- tion will occur at temperatures above 2000°— ing one obtains 2800° K. We will assume the higher figure as a conservative value for our purpose. Hand- %=m0 exp I —| (fljj—vtyl (12) book values for the emissivity of liquid steel are about 0.28. From equation (14), vf=3.5 where It acquires about 10 percent of the energy major axes are of the order of 12 mm and the needed to melt entirely. Conduction prevents semi-minor axes about 3 mm. The body may any melting of the front surface. Radiative also be reasonably well characterized as a cooling is negligible. cylinder 12 mm in diameter and 2.5 rcym thick. (4) The major heating occurs in the con- Consider a hemispherical drop, of radius r, tinuum flow regime below 85 kilometers. Some on the edge of a cylinder, of radius a, spinning melting must occur before the onset of lumi- at a rate «. This drop will be thrown off when nosity. This melt may be spun off, blown off, the centrifugal force exceeds the surface tension collect at the edge, or creep to the rear surface. force, or when Complete melting is conceivable if sufficient heat is transferred to the rear surface by the (16) flow. It is more likely that melt collecting on the edge begins to be removed by drag and where S is the surface tension. For steel centrifugal forces. The mass-area ratio of (8=1.2Xl& dynes/cm and 5=7.9 gm/cm3) the these nearly spherical drops will be similar to limiting value of r is about 1 mm. That is, the major body, and little differential decelera- the droplet need have a diameter only com- tion will occur. parable to the thickness of the pellet to be (5) Eventually complete melting must occur. unstable against centrifugal force. Spraying Whether the remaining liquid mass is stable or will certainly occur whenever molten material not depends on the drag load, the angular reaches the edge. In addition, the drag force momentum, and the restoring force of surface on the drops will enhance the spraying process. tension. The drag load increases through the The conduction of heat inward from the first half of the trajectory but, in compensation, leading surface will delay melting for a time. the angular momentum decreases and the sur- The solution for the heat transfer within a face tension increases as the body sheds droplets. cylindrical meteor entering an exponential at- (6) If the body is stable, mass loss by melting mosphere has been given by Levin (1956), and will cease and the body, after an interval of is applicable here. For a thin cylinder of thick- further heating, will begin to vaporize strongly ness D and conductivity X, with the major until the velocity is sufficiently reduced. If area presented to the air stream, one finds the the body is unstable, it will break into droplets temperature to be which will continue to vaporize until the e2x/x 1 velocity vf is reached by each droplet. °+ The beginning of luminosity must be accom- panied by either (a) vaporization from the surface of the pellet, (b) vaporization from where W(t)——£— is the heat flux per unit area small droplets thrown off from the pellet, or (c) fragmentation of the pellet into some very at time t, and is a characteristic heating small particles. We discard (c) as improbable. depth given as If (b) is important, (a) does not occur appreci- ably; for if the pellet is losing mass efficiently by melting, its temperature cannot approach 8cv cos that necessary for vaporization. To demon- strate that (b) is a possible mechanism, we /3 is the scale height of the atmosphere, c is the must show both that the surface can have specific heat of the material, and ZR is the melted and that there was sufficient force to zenith distance of the radiant. For Stainless remove droplets against surface tension. The Steel 304, X=2XlO6 ergs/cm sec degree and geometry of the pellet is a determining factor c=5XlO8 ergs/gm degree. for each of these conditions. Using these values, one finds that if A=0.3, The shapes of pellets recovered on the firing sufficient energy to raise the surface to the range are roughly that of one-half an oblate melting temperature will have been introduced spheroid with the plane surface leading. The by the time the pellet becomes visible. Since 206 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS A is almost certainly considerably larger than that the droplet disintegrates into about 102 0.3 (equation (15) gives a value of A«0.5 at pieces of a=l mm. Assuming that these par- 70 km), some melting must have occurred. We ticles act independently from the incoming deduce that initial mass loss is due to melting air stream, equation (15) gives A«0.8. In and spraying. The droplets are small enough all probability the effective body size for each to be nearly in free-molecular flow and are sub- of the particles is larger than its physical ject to a heat-transfer coefficient of near unity. dimension since the cluster of droplets will This form of spraying may continue until act on the air stream in concert to some extent. the pellet is completely melted, or until suf- But since this value of A is an upper limit, it ficient angular momentum has been shed with will yield a lower limit in the correction to the the droplets. Various attempts to describe luminous efficiency, which is what we seek here. the mass-loss phenomena in sufficient detail From the above discussion we arrive at the to distinguish between these two events have, following best estimate for the terminal mass: not unexpectedly, been unsuccessful. However, (1) The first half of the mass is introduced the observed light-curve offers a clue, which as small droplets by spraying at essentially although indefinite, cannot be ignored. At the initial velocity of the meteor and ablates t=OA second (see fig. 3) the intensity subsides with A«l (o-=7X10-12) until the velocity temporarily. This is seen on Super-Schmidt of vr is reached. For this computation we films taken from stations separated by 50 have used the heat of vaporization of iron, kilometers. The decrease is certainly real and 6.4X1010 ergs/gram plus the energy required not the effect of uneven atmospheric absorption. to raise the melt from 1800° to 3500° K, Compared to the light curves of natural meteors, 1.1X10"10 ergs/gram. We have presumed free it is a most unusual event. We suspect that molecular flow and therefore have used F=l. this represents the time when spraying ceased Less than 1 percent of the mass remains at or decreased markedly, and that the delay in 4 km/sec. the light-curve represents the time required (2) For the droplets formed by the disruption to melt the remaining solids. At this time, of the liquid body we have used A=0.8 and at least half of the initial mass remains in the F=0.5 as characteristic values for continuum pellet since half of the luminosity is yet to be flow. Then We shall give a detailed discussion of theory constant. For a rough unpolished surface and observation for three meteors, one iron and we may expect this constant to approach unity two stone. The first is known to be a pure iron (Opik, 1958, pp. 52-54). The second possibility from its spectrum. The other two yield results is that the meteor may have been in viscous or consistent with those of the first if they are laminar continuum flow. In that case we expect assumed to be composed of meteoritic stone. that (Probstein and Kemp, 1960) Finally, a satisfactory comparison is made with 1/2 the observed re-entry of an artificial iron meteor. A~(1.6Z/r) , (2) The result is a determination of the luminous where I is a mean free path in air and r is the efficiency law for iron and solid stone meteors radius of the meteroid, assumed to be a sphere. valid in the range of velocity from 9 to 21 km Since sec"1. Iron meteors (3) 3 It is convenient to approach the problem of the for I in cm, pa in gm cm" (Minzner, Champion, luminous efficiency of meteors by starting with and Pond, 1959), we have objects made of meteoric iron or pure iron. 1/2 One meteor showing the spectrum of an iron A~1.14X10-*(rPa)- , (4) without stone inclusions has been reported by and F~5.7 X10"V" "WF3. (5) Halliday (1960). We expect that initially the meteoroid con- Let us make the further approximation that ducted all the heat received at its surface to the interior. The surface temperature must then pa~Pa exp [- (h-ho)/H], (6) have been determined by the rate at which heat was received earlier along the trajectory. If where His the scale-height defined by the meteroid were not rotating, the part of the l surface normal to the direction of flight would HJ%rlnpX (7) have received a heat flux and h denotes height above sea-level. The rF=- 2 (1) approximation of equation (6) amounts to the assumption that the scale-height of equation where A is the heat-transfer coefficient, pa (7) is a constant independent of h. We now the atmospheric density, and V the velocity. note that Since very little change in velocity was observed h—A«= — («—s0) cos Zs, (8) over the trajectory, we may assume that V is constant in equation (1). where s is the distance along the trajectory Two possibilities will be considered. The from an arbitrary zero point, s0 is the value first is that the meteor may have been near or in of s corresponding to the height h$, and ZB is free molecular flow; in both cases A will be a the zenith distance of the radiant. Further- 1 Smithsonian Astropbyslcal Observatory, and Harvard College Ob- more, servatory, Cambridge, Mass. s— (9) 638805—63 lg 209 210 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS according to the approximate assumption of Flow of molten iron constant velocity. Substitution of equation We consider a convex meteoroid with axial (9) into equation (8), substitution of the result symmetry about the direction of flight. Let y into equation (6), and substitution of that be perpendicular to the surface outward, r result into equation (1) (with A~l) and into the radius of curvature in the xy-plane, and equation (5) yields r2 the radius of curvature in the plane normal to the local x-direction. Let 6 denote the exp [(VIH) cos ZR(t-Q], (10) angle between the y-direction and the direction of flight. The equation of continuity takes the form or - £- v(r2u sin 0)+^ (18) r2 sin 9 bx by X exp [i (V/H) cos Z«(*-*o)]» (H) where u, v are the x and y components of velocity of the iron. The equation for the conduction where we adopt the smaller value as correct. of heat takes the form Equations (10) and (11) can be written as bT. dT , b2T ox oy Oy2 (19) F=Foexp[(t-t0)/T], (12) where for equation (10) we have where a steady state has been assumed and only the dominant term in the Laplacian of the temperature has been kept on the right-hand cos ZB), (13) side. The equation of the transport of the u- component of momentum is and for equation (11) we have bu bu d / bu\ dp /onN i 1/2 i F0=5.7XW- r- f&V ,T=2H/(y cos ZR). (14) Levin (1956) has shown that for a flux of the The dynamic viscosity is taken to have the form form given in equation (12) extending back to of a step function of temperature: t=—», the following temperature distribution 7>1800°K, M=«, is expected in a semi-infinite medium at negli- (21) 1 1 gible initial temperature bounded by a plane T<1800° K, M=2X10~2 gm cm" sec" , surface at x=0, as given by Opik (1958, pp. 156-158). T(x,t)=T0(t) exp (-x/xj, (15) We begin by linearizing equation (20) in u, v, whence (16) (17) where x is the distance from the surface, F(t) and we denote the rate of flow of shearing is given by equation (12), K is the conductivity, momentum across the boundary y=0 by T0, cv the specific heat, pm the density, and k the whence thermometric conductivity of meteoric iron. (22) As soon as To= 1800 °1£ is reached (the melting temperature, Opik, 1958, pp. 156-158), molten ron will start to flow. To form an idea of y> the conditions required to remove molten iron by n nbx from the meteoroid, we must consider the problem of flow of the melt. u(v)= ^p_V. (23) METEOR SYMPOSIUM—COOK, JACCHIA, AND McCROSKT 211 We next assume that u(bT/bx) T =A' p V2 sin 6 cos 6, (34) In (bT/by) =lim In (bT/by) +^ (y-ym). (26) O a v where we take for the deceleration the standard we find form 2 dV . PaV (42) (48) a I pm 7ii where r is the drag coefficient, A is a shape factor (1.208 for a sphere), and m is the mass of the meteoroid. It then follows that the X inertial component (41) of the pressure gradi- ent may be written sin d. (43) t3 t s X(y -5ymy"-5y!ny -5y m) ]}*'• (49) The sum of equations (40) and (43) is the total pressure gradient to be used in equation We assume that all the terms in the exponential (33), are sufficiently small so that equation (49) may be replaced by 2 /3 QP=-PaV sin 0 [- cos e-TA(pJmy ~\ ox t_rx J — ^0+^ (Zf+CpTjpnVny, where *' \ ^2 >\~ I Z3 ~>TZ )' (44) (50) As x approaches zero, 6 approaches zero and, by axial symmetry, r2 approaches r, and the Thus we have z-derivatives of ru r2, and ym approach zero. In this region equation (33) may be integrated 0>y>ym,T^Tm+^Ur+cpTm)Pmvw{y-ym) (51) to yield and r i m ym>y>- ,T^=Tm exp I y (y—yTO) • ' (y-ymy-tfm L J Equation (47) may be expressed in the form ;lim ^snm r2. x-»0 x->0 (46) ^y (53) By substituting y—ym we obtain the fl-com- We now assume that vaporization is negligi- ponent of velocity at great depths: ble, i.e., #0=0, and we obtain Hm p^ r-X> 6 r (47) ^y (54) METEOR SYMPOSIUM—COOK, JACCHIA, AND McCROSKY 213 This is a cubic equation in yjr. The solution of this equation must then be substituted into (58) equation (50) to find To. Next, the assump- tion that vaporization is always negligible must and by retaining the first root as the only non- be verified by computing the rates of loss of extraneous value, we have the result that mass and heat by vaporization at To. Finally, 0~54°7. Melt will begin to flow past here as the assumption that the argument of the expo- soon as To, by conduction alone according to nential in equation (49) is small must also be equation (16), reaches Tm. If r is sufficiently verified. large the melt will blow off as droplets, other- Before the computation near x=0 can be wise the surface temperature will rise until undertaken we must return to equation (33) vaporization removes the iron. The criterion and ask where on the surface the second term for the former process is (Opik, 1958, pp. 83-86) on the right-hand side might vanish. At such 7.6s, a point melt would presumably flow past, as T V (59) it is neither removed nor accumulated by the Pa flow field. At smaller z the flow field would In free molecular flow r^l. In continuum remove melt, at greater x it would accumulate flow with negligible shear transfer the dynamic melt. On our assumptions the excess accumu- pressure (eq. 38) alone contributes to the drag, lation plus the local ablation would be removed and if we integrate over the exposed hemisphere, 2 by vaporization. If flow of melt without abla- we find from the local contribution pd cos 0, tion continues past this critical value of x, it f*r/2 can only do so if u(bT/bx) at y=0 dominates sin 0 cos3 0d6=~- over v(bT/by). We ignore the trivial case of T~21 sii Jo T0 From the computation for free molecular flow where cp$ is the specific heat of the solid, fr this flux, with A^l, gives an upper limit for the heat of fusion, and cPi the specific heat of the beginning height of 72.2 km, only 3.0 km the liquid. Opik (1958, pp. 156-158) gives cP,= above the photographed beginning height. 6.91 X106 ergs gm"1 deg-1, f,=2.69X109 ergs 1 6 1 1 This model is valid only for r<0.53 cm (from gin" , cp,= 6.66X10 ergs gm" deg- . We A=l in eq. (2)). We also require that r>x0 assume that the beginning of ablation occurred by equation (17) otherwise the surface temper- at A=69.2 km. Then the result is yjr~—3.2 X 3 0 ature will be higher than by equation (16) and 10~ for 7o^TOT=1800 K, and the final re- 3 T ablation will begin at a greater height. We sults are yjr=— 3.1X10" , 7 0=1874°K. It have k=0.004 cm2 sec"1 for meteoric iron from may easily be shown from the vaporization law given by Opik (1958, pp. 156-158) that k=- (66) rates of removal of both heat and mass by CpPm vaporization are smaller by three orders of magnitude compared with the rates due to whence, in viscous or continuum flow, x <0.39 0 flow of molten iron. Also the characteristic cm, and in free molecular flow with rapid rota- depth of heating is tion XQ=0.27 cm. r must lie in the range of about 0.3 to 0.5 cm in order that this latter model should be a valid physical picture. —=0.15 cm, For the nonrotating model, r must be suffi- ciently large so that A<1 at a height of 74.1 and the depth of the molten layer is —ym=4.8X km in order to prevent the onset of ablation 10~8 cm, which is about l/20th of the depth of METEOR SYMPOSIUM—COOK, JACCHIA, AND McCROSKT 215 heating. This circumstance makes the expo- equation (72) into the above expression we nential in equation (49) nearly unity, and vali- find dates the use of equation (50). (73) Photometric mass and end point from the 27 / light curve The light curve has the appearance of a standard A series of values of IB.BI/max may be selected and PB!P« and pshe found by solving equation light curve (Opik, 1958, pp. 63-65) collapsed in (73). For these values we obtain duration. In the approximation where the scale height is constant, equation (6) may be Pma* PE __r> PE (74) used. If constant velocity is also assumed, the PB 3 p Pmax standard light curve takes the form B From equation (71) these ratios yield (70) == hB—hm*z H In (Pmax/PB)> (75) where pe is the ambient atmospheric density at l-hB=H In (76) the end point of the trail. Then pa is given by equation (6), h by equation (8), and s by equa- where hB, Am,,, hs are the observed height of tion (9), which completes the conversion of the beginning, maximum intensity, and end, respec- independent variable to time. From equation tively. We now eliminate the scale height H (6) we have by obtaining the ratio ^from equations (75) and (76): Pa =exp[-(h-he)/H]. (71) (77) = We now propose to alter H from its true value x—hit In to one that will fit the observed light curve. The first form is an observed quantity, the We do so for two purposes: to measure the total second a computed quantity from equations photographic light emitted by the meteor; (73) and (74). Thus the computed value and to extrapolate from the observed end point must be matched with the observed value to to the true end point of the meteor. find IB,B/IIMLX- Then equation (75) or equa- Halliday (1960) gives beginning and end tion (76) yields the effective scale-height for heights determined by the threshold of the the light curve, emulsion. On such a short trail we assume the threshold to be unchanged from beginning to (78) end. Let us denote the intensity of the meteor (Pmax/ps) In at these points by IB,E- Then the density at maximum intensity Pm is given by where the ratios have been computed from ax equation (74). Then from equations (71) and (72) the true end height (complete abla- Pmax 1 (72) Pe tion of the meteoroid) is and from equation (71) we have he=hmtix—H In (p«/pmax)=^max—H In 3. (79) The total light emitted up to the disappearance of the meteor at te is =_<*. [Yl-£* Pmax LA Pe where p , PB are the densities at the observed B . (80) beginning and end points. By substituting x 216 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS The photometric mass introduced by Millman Fdt=\ exp jj COS ZR(t — te) dt and Cook (1959) is (81) 2 cos ZR F4 cos Z, where we have substituted equations (6), (8), where the terminal mass is neglected. and (9), and integrated. For an upper limit From the spectrogram, Halliday (1960) gives on the diameter we neglect conduction of heat ^s=68.3 km, ^«=64.9 km, and from the direct in directions other than the direction of flight, photograph, & =69.2 km, ^=64.7 km. From B and use T =T in equation (69) for f. We his figure 1, the ninth of thirteen visible seg- 0 m then obtain ments of the spectrogram is seen to be the brightest. The corresponding value of Amax is d< (82) 66.0 km. Then the two values of & are 2 and 2£pm cos 2%, respectively, which in turn yield H=1.56 and 1.62 km, and A«=64.3 km in both cases. since all the heat input must ultimately be used Finally, he gives the maximum brightness in for melting. The limit from equation (82) is panchromatic light as absolute magnitude —4.2. 2.3 cm, so that the upper limit on the radius is The spectrum of a slow meteor is overwhelm- 1.2 cm. A lower limit on the radius can be ingly due to iron, except for the D-line of established by assuming equation (1) to be sodium and the first positive group of nitrogen valid when A=l until equation (5) predicts a bands (Milhnan and Cook, 1959); hence, except lower value, and then by using equation (5) for these two emissions, the reduction from with r fixed therein at its initial value. For a panchromatic to photographic magnitude given value of r the two fluxes are equal when should be the same. The contribution of the 8 1.297X10" D-line may safely be neglected, but the first pa=- (83) positive group of N2 is not seen in Halliday's whence we have spectrogram, and hence the reduction found by Millman and Cook for a slow cometary meteor 3 dt of +0.2 magnitude should probably be altered =^PeV r explj cos by an unknown negative quantity. It seems best to make no correction at all, hence the Jr exp intensity at maximum brightness was COsZ (t-t )~jdt, [2H ' t e where in units of a zero magnitude star. From H equation (81) the photometric mass was ^=1.2X 10~17 0 Mag cm~3 sec4, We now find, with d=2r, that 1 V2H 1.297X10"8 where H=1.59 km is adopted as the effective Pe scale-height of the light curve. This photo- 2 cos ZR i 1/2T/2Z7 4 metric mass and the end height of 64.3 km are +1.14X10" 1/2 {l 1 our two fundamental observed quantities. r cos ZR 8 -[1.297X10- /(rp.)]^}=2fPmr, Limits on mass from end height then For the nonrotating model we can set an upper 1.297.297 XX10~8 2.28X10"4 I /„ \ 1 /O limit on the mass by putting A= 1 in equation 4fP(BcosZfi { rpe (1). Then the total heat input to the end 114X1Q-4 point of the trajectory will be given by METEOR SYMPOSIUM—COOK, JACCHIA, AND McCROSKY 217 18 18 and if we put whence we find TO=6X1O~ to 10~ 0 Mag 22 1 8 1.14X10-* /1.297X1Q-V1Q-V fQA. gm" cm" sec* as r varies from 0.5 to 0.9 cm. (84) 18 %m^"m . )) ' For r=0.7 cm, we have TO=2X1O~ 0 Mag gm"1 cm"3 sec4. we obtain On the other hand, McCrosky and Soberman rpw g (see this symposium, p. 199) find that, from a re- tt«-2it»+5.19X10-« 2T?rf =0- (85) entering iron pellet of known mass, TO=8X1O~W 0 Mag gm"1 cm"3 sec*. If we now use the The last term in equation (85) has the value logarithmic mean, we obtain 0.0526, and we find r>0.61 cm. This is, however, a closer limit than the upper limit log T0= —17.70 for spectrum no. 210, previously found. log TO= —18.10 for re-entering pellet, Iron For the rotating model, an upper limit can log TO= —17.90 adopted pro tempore. be found as above by putting A=l, but using equation (65) instead of equation (1). The For low velocity meteors the spectrum in the limit obtained is a limit on r rather than on d photographic region is almost exclusively dom- and is expressed by inated by iron lines (Millman and Cook, 1959). (86) Hence we should be able to derive the luminous -< Pe -2ir{pm COS ZR efficiency of meteoric stone in photographic light from that of iron by multiplying the which establishes the limit r<0.74 cm. luminous efficiency of iron by the fraction of A lower limit can be established as above but the mass of meteoric stone due to iron. Ac- with the left-hand side divided by v for averag- cording to Opik (1958, pp. 159-162) this frac- ing the flux over a rotation and the right-hand tion is 0.154, from which we deduce that side divided by 2, since the radius rather than log TO= 18.71 for stone. the diameter sets the limit. The result is Coefficient in deceleration equation for stone 1-297X1Q-8 The deceleration equation usually appears in 2ir COS ZR p, r. the form , 1.14X10"* 2 dV_ 1 PaV m l (89) r cos ZR dt~ K^? <*' I" /1.297X10"8V/ n_ where K is an observable quantity, and then (90) =0. (87) In terms of physical parameters, it takes the The last term in equation (87) has the value form 0.083, whence we find r>0.44 cm. (91) From the above analysis we may estimate 0.5 Coefficient in formula for luminous efficiency If we insert into equation (91) this value of A The coefficient in the law for luminous effi- and the value of T0 for stone found above, then 6 ciency, the mass, and the photometric mass i£=3.4X10« for r=l, and 8.1X10 for r=l/2. are related as follows (Millman and Cook, The corresponding values of the logarithm are 1959): log isT=6.61 and 6.91, respectively. These 9 9 J/ expected values for stony meteors may be con- mf =^ (88) m trasted with log .K~6.29 observed for cometary 638805—63 -19 218 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Observed data for meteor No. 1242 (Beginning height, 74.9 km on FA 8808; maximum light, 47.1 km on FA 8808 and AI 39336; end height, 38.0 km on FA 8808, 36.6 km on RH 12890; cos ZR=0.632.) Epoch Height Velocity Acceleration Air density* I I (smoothed) 2 3 Jt-JL3 4 t(sec) i(km) F(km/sec) dV/dt (km/sec ) pa(gm/cm ) O Mag cm" sec O Mag O Mag 18 1. 0526 62.56 11.93 -0. 185 2. 58X 10~7 7. 56X10- 3.0 3.0 8 1.6842 57.78 11.81 -0. 389 4. 49X10-' 6. 69X10-1 3.3 3.3 7 8 2. 3158 53. 10 11.48 -0. 660 7. 49X10- 5.64X10-1 4.8 4.0 2. 9474 48.61 10.94 -1. 024 1. 278X l0-« 4.07X10-18 4.3 4.0 8 3.5790 44.38 10. 13 -1. 505 2. 195X10-' 2. 432X10-1 4.0 3.3 4. 2105 40.57 8.87 -2. 23 3. 698 X 10-« 0. 885X10-i» 2. 1 2. 1 • According to Minzner, Champion, and Pond (1959). TABLE 2.—Reduced values of parameters on assumption that 1/K and a are constant for last three epochs Epoch t Photometric mass ^/if 1/K a/K sec 0 Mag cm"3 sec 4 0 Mag I/3 gm"1 cm sec4/3 O Mag I/3 gm"1 cm3sec-J/3 I. 0526 10. 09 X 10-18 1.087X10-7 8. 41X10-i» 1.6842 9.22X10-18 1.303X10-7 6. I6XIO-19 2. 3158 8. 17X 10-" 1. 347X lO-7 5. 75X10-i» 2. 9474 6. 60 X 10-" 1. 266X10-* 5. 19X10-i» 3. 5790 4. 96X 10-18 1. 140X10-7 4. 78 X 10-i» 4. 2105 3.42X10-18 1. 155X10-7 5. 15X10-i» TABLE 3.—Deduced values of parameters on assumption that 1/K is constant over trajectory (The values of a are computed using the mean VK deduced for computation of the masses.) a/K Epoch t Photometric mass ^' 1/K 0 Mag xt% gm"1 cm3 a sec O Mg cm"3 sec * 0 Mag m gm~i cm sec4/3 sec -2/3 cm2 sec"2 1. 0526 1.710X10-17 1.285X10-7 5. 92X10-i» 3.6X10-12 1. 6842 1.623X10-17 1.573X10-7 4.23X10-1' 2. 6XIO-12 2. 3158 1.518X10-17 1.656X10-7 3.81X10-1' 2.3X10-12 2. 9474 1.361X10-" 1.598X10-7 3. 20 X 10-i» 2. 0X10-12 3. 5790 1. 197X10-'7 1. 529X10-7 2. 66X 10"i» 1.6X10-12 4. 2105 1.042X10-17 1.676X10-7 2. 45X10-" 1. 5X10-12 meteors. Two such meteors with anomalously We next assume that 1/K alone is constant large log K have been well photographed in the From the reciprocal of the cube of equation (90) Harvard Program, and will be discussed in turn. we have Meteor No. 1242 Meteor No. 1242 was photographed on Feb- ruary 6, 1945, from Harvard College Observa- tory in Cambridge, Mass., on plate no. AI •D J (92) 39336, and at the Oak Kidge Station (now Agassiz Station) in Harvard, Mass., on plates The coefficient of ^//e and the entire last term are observed quantities. The equation is to be nos. FA 8808 and RH 12890. The meteor 3 appeared at 5h 12m 17" UT. Table 1 lists the solved for ^t and 1/K as unknowns. We do pertinent data from the reductions. so by grouping the first three epochs and the last three epochs and using average values for The method of Millman and Cook (1959) is 7 18 each. The results are 1/K= 1.628 X10" 0 Mag used to determine that~^=2.5X10~ from the gin"1 cm sec,logii:= 6.79, and last three points. The results are given in 0 Mag gm~3 sec* (see table 3). We estimate table 2. These results are not satisfactory that r~l/2 since we are presumably far from since neither 1/if nor a is constant. free molecular flow, and hence log TO= — 18.36. METEOR SYMPOSIUM—COOK, JACCHIA, AND McCROSKY 219 TABLE 4.—Observed data for Meteor No. 19816 (1/K(obs)=value computed from eq. (90); 1/K(corr)-1/K(obs) corrected for presence of two equal fragments; beginning height, 91.1 km (SK); minimum light, 60.1 km; end height, 47.1 km(SL); cos Z«=0.716.) .Jt-Jt, I 1/K(obs) 3 1/3 Camera t h V dV/dt Pa O Mag cm- O Mag OMag gm-' 1/K(corr) 2 3 4/s 2 2 sec km km/sec km/sec gm/cm sec4 cm sec cm sec" 12 SK 0.764 80. 13 20.67 -0. 158 2. 117X10-8 6. 60X10-" 1.2 1. 52X10-' 1. 52X10-' 6. 0X10- 2 SL 1. 135 74.02 20.57 -0. 416 5.739X10-8 6. 07X10-" 1.3 1. 45X10-' 1. 45X10-' 2. 9X10-' SK 1.860 64. 16 19.87 -2.02 4. 43X10-" 2.5 1. 84X10-' 1. 84X10-' 1. 8X10-" 2. 123X10-' 2 SL 2.038 61.06 19.27 -2.75 7 3. 87X10-19 2.5 1. 74X10-' 1. 38X10-' 1. 7X10-' 3. 105 X10- 12 SK 2.425 56.49 17.85 -5.92 7 2. 71X10-19 1.9 2. 32X10-' 1. 84X10-' 1. 2X10" 5. 178 X10" 12 SL 2.623 53. 75 16.36 -8.45 6. 935X10"7 2. 05X10-" 1.9 2. 68X10-' 2. 13X10"' 1. 5X10" SK 2.824 51.87 14.63 -9.67 1. 14X10-" 1.3 2. 53X10"' 2. 01X10-' 2. 6X10-12 8.662X10-' 12 SL 3. 124 48.92 10.77 -8.90 1. 23lX10-« 0. 32X10-" 0.23 1. 98X10-' 1. 57X10"' 6. 0X10" Smoothed magnitudes are used at some and SK 9821) at the Sacramento Peak and epochs, primarily because oscillations appear on Organ Pass meteor stations in New Mexico. the FA and AI plates near maximum light. The meteor appeared at 0Sh 50ra 13s UT on A very short period flicker is found on the RH December 8, 1958. The entire trajectory of plate. The following model may be used to 62 km was photographed on at least one of the represent the observed behavior. The meteor- cameras, but the initial 12 km and the final oid is assumed to have been in rapid rotation 4 km were photographed at only one station. upon entering the atmosphere. After free The light curve is particularly smooth, ex- molecular flow was left behind, a torque de- hibiting only one brief flare of 0.2 magnitude. pending upon orientation developed and became During the final portion of the trajectory, two stronger. First, the angular velocity of rota- particles are in evidence, and it is likely that the tion was diminished at all orientations except flare represents the time of splitting of the those for which no torque occurred, and re- original body. The deceleration history of the storing torques were met on either side of these separate particles indicates that they have null orientations. The period lengthened until nearly comparable masses. The lifetimes and rotation ceased and oscillation began about an the total luminosity of each particle are similar. equilibrium orientation. This change occurred In the following analysis we have made the around maximum light. Thereafter the period approximation that the masses and area of the of oscillation became progressively shorter. particles are exactly identical although, in fact, Next, we consider the shape bounded by two some difference between them must exist for right circular cylinders of equal diameter with differential deceleration to be observed. Apart axes intersecting perpendicularly. We then from this single fracturing, there is no evidence make the axis of rotation perpendicular to the of unusual wake or terminal blending that axes of the cylinders. Finally, we orient the frequently is visible in meteor trails formed by axis of rotation perpendicular to the direction fragmenting cometary bodies. of flight. Such a model does not really deviate The data for this meteor is given in table 4. much from a sphere, and yet it can be shown Four deceleration solutions were made for each that the torques produced upon it are about trail. The values for K'1 are generally lower equal to those required to fit the observed for the later solutions, presumably due to the periods. This model can also be made to fit presence of the two fragments. In the penul- an observed proportionality of the period of timate column of table 4, we increased K~l by oscillation to the inverse square root of the air 21/3 for those solutions where two fragments density along the lower part of the trail. existed. The corrected values show no signifi- cant trend with time. It appears that the ter- Meteor No. 19816 minal mass is negligible, as would be expected This sporadic meteor was photographed with for a meteor with an initial velocity of 21 km/ the Super-Schmidt meteor cameras (SL 12537 sec. Assuming the same values for A, T, andpm 220 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 5.—Orbital elements of asteroidal meteors Corr. rad. Meteor a e Si i r-u+Q, sec) a S Iron Meteor 332T0 + 73!2 13.4 1.05 .117 .928 1.173 247?5 219?1 13!3 106?6 Meteor No. 1242 3?0 + 59?2 12.2 1.33 .262 .984 1.68 172? 1 317?1 6?9 129?2 Meteor No. 19816 70!5 -0?2 20.7 2.24 .662 .756 3.72 65?4 75!7 12?6 141?1 Pribram Meteorite 191!5 + 17!7 20.9 2.42 .674 .790 4,05 241?6 17?1 10?4 258?7 as were used for meteor no. 1242, we find log been published previously (Halliday, 1960; TO= —18.24. These two values are in excel- Whipple, 1954; and Ceplecha, 1961). lent agreement. It is apparent in the table that these four meteors represent two distinct classes of orbits. Optimum values of the luminous efficiency However, an inspection of the orbits of the factor for iron and stone asteroids with small perihelion distance indicates Spectrum no. 210 yields, on reduction for that this is probably a chance grouping. the abundance fraction of iron in stone, log TO= —18.51 for stone. For the above two References meteors log ro= —18.36 and —18.24. The CEPLECHA, Zd. mean of the three values is 1961. Multiple fall of Pffbram meteorites photo- graphed. Bull. Astron. Inst. Czecho- log T = —18.37 for stone. slovakia, vol. 12, pp. 21-47. O HALLIDAT, I. 1960. The spectrum of an asteroidal meteor Application of the reduction for the abundance fragment. Astrophys. Journ., vol. 132, fraction of iron yields log TO= — 17.56 for iron, pp. 482-485. while McCrosky and Soberman (see this sym- LEVIN, B. J. posium, p. 199) find log ro= —18.10 for a re- 1956. The physical theory of meteors and mete- entering iron pellet of known mass. This dis- oric matter in the solar system. Publ. House Acad. Sci. U.S.S.R., Moscow. crepancy corresponds to a deviation in the shape (German translation published by Aka- factor (larger than for a sphere) by the antilog- demie-Verlag, Berlin, 1961, pp. 47-49.) arithm of 0.18, or of about 1.51. If we adopt MlLLMAN, P. M., AND CoOK, A. F. this new corresponding value of the shape 1959. Photometric analysis of a spectrogram of factor we have a very slow meteor. Astrophys. Journ., vol. 130, pp. 648-662. MINZNER, R. A.; CHAMPION, K. S.; AND POND, H. L. .4=1.83, log r0= — 18.10 for iron, and 1959. The ARDC model atmosphere. Air Force — 18.91 for stone. (93) Surveys in Geophysics, No. 115, Geo- physical Research Directorate, Air Force Orbits of asteroidal meteors Cambridge Research Laboratories. OPIK, E. J. The meteors discussed in this paper are of 1958. Physics of meteor flight in the atmosphere. unique value because their physical character- Interscience, New York. istics indicate that the objects are of asteroidal PBOB8TEIN, R. F., AND KEMP, N. H. origin. These meteors, together with the 1960. Viscous aerodynamic characteristics in Pffbram meteorite that was photographed in hypersonic rarefied gas flow. Journ. flight (Ceplecha, 1961), represent a new source Aerospace Sci., vol. 27, pp. 174-192. of data on the orbits of asteroidal material with WHIPPLE, F. L. small perihelion distance. The information is 1954. Photographic meteor orbits and their distribution in space. Astron. Journ., collected in table 5. Three of the orbits have vol. 59, pp. 201-217. Meteor Ionization and the Mass of Meteoroids By Donald M. Lazarus1 and Gerald S. Hawkins: The excitation and ionization observed in where / is the atomic radius in units of a^; AEis meteor trails have long been assumed to result the energy of the transition in ev; Mx is the from inelastic collisions of meteor atoms with mass of the incident atom relative to the proton those of the atmosphere after their ablation from mass; and E the impact energy in ev. For themeteoroid surface. This assumption is based fast collisions, T/T (7) where rx is the internuclear separation, rt 01 the position of the i electron of the sodium If we define K'=K—k', equation (7) becomes atom relative to its nucleus, and a and A are interaction parameters. We assume that the electron wave-functions can be separated, and that removal of the 3s electron does not appre- ciably change the wave function of the inner The probability for the ejection of the valence electrons. With these approximations, the electron with wave number k' in the direction matrix element for the transition to the con- 9', H'=L~ ,(r2)—< sin sin r\2 (4) where ^*'(r2) and fa (r2) are the ionized and 3 f initial state functions, respectively, and L is where ok>(6 for the ejected electron; with k'=mv'l%, the electron momentum divided by Planck's con- so that the total cross section for ionization stant over 2x, from the ground state is given by the expression, METEOR SYMPOSIUM—LAZARUS AND HAWKINS 223 contribution to the integral is for small values O/W = olK. sin Odtkk sin O'dO'dv'dk' (11) For smaU valueg of ^ sin edddip sin e'dd'd If we retain the first two terms in equation (18) (13) k sin OdO=^ dK; and the leading term in equation (17) and introduce the notation, and K'2=K2+k'2-2Kk' cos B', ^<4- 1 ~i I ZiZAlC I ~T (19) (14) Kk'sw0'dB'=K'dK'. the cross section becomes Substituting equation (8) in equation (12) and „ , 9X2nirmn2b2A2B2 carrying out the integrations over (20) r*'"" d¥ r°+k dK The coefficients S, R, Q, and P are functions of a, 6, k0) and A:'. The interaction parameters A and a were chosen to give the best possible fit to the oxygen potential determined by the The first integral is easily evaluated: numerical integration of the radial charge distribution obtained from the Hartree (1957) numerical wave-functions. The value of b was take from the Slater (1951, app. 13, p. 476) _2b*f 1 1 screening constants. ~7 \[62+(.K:-fc')2]7 [&2+(#+fc')2]7 The numerical integration of equation (20) was performed for several values of k0 and the resulting cross section is plotted in figure 1. 3 \{b2+{K-k')2]* I The ionization probability The probability of ionization of an atom with wave number kt in one collision is The lower limit ko—k for the second integration is determined by applying the principle of the (21) conservation of energy: where 2 [cos 0<±(M!/M?— sin (27) where MI is the mass of a meteor atom, 0, the angle of scattering of the meteor atom, and /z2 VELOCITY v km/sec the mass of the atmospheric atom. The average of t} over the normalized angular FIGURE 1.—The dependence of ionizing cross section on t velocity. distribution in the approximation k^a^l is first Born approximation (Schiff, 1955, chap. (28) 7, sec. 26), M1+M2 If I <72-4 _4 (22) In the range of meteor velocities, the ioniza- tion cross section given in figure 1 can be h2 approximated by the expression, where a——271/3' e represents the electronic, and Ze the nuclear charge. In the energy (29) range under consideration, If we let Pt be the probability that ionization will result from the Ith collision, it is easily seen that (24) .*{: i-1 "^ f>k 4=1 if the wtb collision is the last one in which The possibility of ionization from an excited sufficient energy is available for ionization. state has been neglected because the lifetimes However, of excited states are much less than the average time between encounters in the upper (31) atmosphere. Since /3<<^1, it is possible to simplify the last so that of equations (24) to the expression Pi«ft, (25) so that the total probability for the ionization of an evaporated meteor atom becomes The elastic scattering cross section used in (26) the calculation of /3 is subject to the same criti- METEOR SYMPOSIUM—LAZARUS AND HAWKINS 225 cism as the ionization cross section; i.e., it gives meteor atom and the atmospheric constituent too large a cross section at low energies because occur during a single collision have been neg- distortion is not taken into account. Since /3 lected because such events are much less involves the ratio of the cross sections, it is probable than single-excitation collisions. likely that the effect of neglecting distortion will be reduced. The effect on the inelastic Application to meteoroids cross section is generally greater than in the The number of ions produced in an infinitesimal elastic collision so that /3 will remain an over- volume of a meteor trail is related to the mass estimate. of ablated meteor material in that volume by From the cross section given by equation (2), the equation which results from the near crossing of potential a energy curves, we obtain for this process an dQ—-=- dmo> (34) ionization probability of Ml where m is the average mass of a meteor atom and /3 is the ionization probability. The calcu- V 1-v ) °J lation of /3 enables one to determine the mass (33) of a meteor by the integration of equation (34). If the velocity is constant, the number of ions The velocity dependence of the probability per unit length of trail is related to the visual resulting from curve crossing appears to be in magnitude M at that point by the equation better agreement with radio-meteor measure- ments than with the Born approximation result. (35) Nevertheless, it is doubtful whether this type of collision would be prevalent in the majority which is derived in the appendix (p. 227). of encounters between meteor and atmospheric Making use of the relation between trail atoms. length x and height H, where f is the zenith The Born approximation cross section ap- angle, we find that pears to be more compatible with the great majority of experimental results (Massey and dx= dH, (36) Burhop, 1952, ch. 7). Moreover, the manner in which the Born approximation is expected to cos f ' fail in low-energy collisions (Massey and Smith, and equation (35) can be written in the form 1933; Mott and Massey, 1949, ch. 11; Mott, 1931) is well known, and has been discussed 6 J—l—dH. (37) above. It has been used with considerable cos f success to derive the range-energy relations for The variation of magnitude M along the alpha particles and protons. trail is given by the normalized light-curve One of the assumptions made in applying (Whipple and Hawkins, 1958, sec. 3, p. 525) these conclusions to meteor phenomena is that which, for 360 measured meteors, is closely the upper atmosphere is monatomic. This as- approximated by the parabolic form sumption, of course, is not precisely correct although it is true for the majority of atmos- (38) pheric atoms at 100 km. Those diatomic mole- H cules present in the upper atmosphere will tend o to reduce /3 because of the large cross section where Ho is a two-value parameter. We take for rotational excitation. This competing proc- Hn=Hb for the portion of the light curve be- ess may lead to a term 2aB(kt) in equation (21) tween the beginning and the maximum, and which is comparable with (39) ET is the ionization energy (Bates and Griffing, > 1953). In addition, it is necessary to multiply the cross sections by n, the number of valence which is readily integrated to give for equation electrons available for ionization. Thus the (34) the value relative ionization probabilities are related by In r_ IT i TT PWo—Afnj.il (40) the ratios of n(^-6\ Table 1 gives these factors and the abundances. Let us assume for the moment that the mete- oroid consists en tirely of sodium. For a sodium- TABLE 1.—Total ionizing probability oxygen collision ^=0.42, and the following Number of Estimated Weighted contribution Element atoms per per atom values of /3 are obtained: for a Perseid of veloc- 100 0(Na=l) ity 60 km/sec, /3=1.6X10"2; and for a Geminid of velocity 35 km/sec, 0=2.7 X10~8. Hawkins Na 0.8 1.000 0.0080 (1956) has shown that M0=40.75 for a Perseid Mg 14.6 . 198 .0290 and 38.89 for a Geminid. The mass m0 is ob- Al 1.4 1.518 .0213 tained with these radar values for meteors of Si 16.2 .288 .0468 Ca maximum magnitude, Mmax =0. Thus for the 1. 1 .972 .0106 Perseid, m =0.212(l/cos f) gm, and for the Fe 6.2 .384 .0238 0 Inert 59.6 .000 .0000 Geminid, Wo=0.620(1/cos f) gm. These results do not, of course, represent the Mean /3 per atom . 1323/S(Na) mass of a meteoroid since the calculations were performed for sodium-oxygen collisions. The oxygen may be considered a reasonable approx- The average ionization probability 0 is 0.1323 imation to the average atmospheric atom, but /Sjva, and the average mass of a meteor atom the properties of the other meteor constituents is 3.72 X10"28 gm. The insertion of these val- may lead to different masses than those obtained ues in equation (40) gives mo=1.66 (1/cosf) for the above "sodium meteor." The proper gm for the Perseid, and 4.85(l/cos f) gm approach (determining an average J8 and £i) is for the Geminid. There are additional uncer- to evaluate the ionization cross sections and the tainties in these figures due to the assumption respective ionization probabilities for each con- of the cosmic mix, and the sodium-like be- stituent, and weight each by its abundance in havior of all atoms. Unfortunately, this un- meteoroids. Since the calculations of the cross certainty cannot be adequately assessed. sections for all meteor constituents is not fea- Using the radio meteor values for Mo given sible, and since their velocity dependence is by Hawkins, one readily obtains an upper expected to be similar to that of sodium, the limit for the luminous efficiency from equation following rough approximation will be used to (7) of the appendix. For a meteor of visual obtain some idea of the effects of these elements magnitude 0, on the calculation of meteor masses. 3 The average atomic mass is easily evaluated W,8=3.194X10- for «=60 km/sec, from the cosmic abundances of the elements and (Brown, 1949). If the ionization probabilities T//3=3.295X10~2 for r=35 km/sec, for the various elements have the same form, they will differ only in magnitude. The mag- so that the upper limits are nitude of the cross section is inversely propor- tional to a high power of the energy required T=8.9X10~6 for »=35 km/sec METEOR SYMPOSIUM—LAZARUS AND HAWKINS 227 and References T=5.1X10"5 for 0=60 km/sec, BATES, D. R., AND GRIFFING, G. 1953. Inelastic collisions between heavy particles. with the computed values of /3. I. Excitation and ionization of hydrogen atoms in fast encounters with protons Appendix and with other hydrogen atoms. Proc. Phys. Soc. London, vol. 66A, pp. 961-971. The luminous efficiency T is defined by the BATES, D. R., AND MASSEY, H. S. W. relation between 7, the visible energy produced 1954. Slow inelastic collisions between atomic per second, and the rate of ablation: systems. Phil. Mag., ser. 7, vol. 45, pp. 111-122. 1 dm0 BROWN, H. 2 dt (1) 1949. A table of relative abundances of nuclear species. Rev. Mod. Phys., vol. 21, pp. By dividing equation (34) by dx=—vdt, we 625-634. obtain HARTEEE, D. R. 1957. The calculation of atomic structures. John dm 0 (2) Wiley & Sons, New York. dx v HAWKINS, G. S. 1956. Meteor ionization and its dependence on and eliminating /dt between equations (1) velocity. Astrophys. Journ., vol. 124, and (2) gives pp. 311-313. LANDAU, L. D. (3) 1932. Theory of superconductivity. Phys. Zeit- schr. Sowjetunion, vol. 2, p. 46. Opik (1937) obtained the luminosity relation MASSEY, H. S. W. 1949. Collisions between atoms and molecules at M=7.1-2.5log 7 (4) ordinary temperatures. Phys. Soc., Rep. 10 Progr. Phys., vol. 12, pp. 248-269. from the standard fluxes of established stars. MASSEY, H. S. W., AND BURHOP, E. H. S. 1952. Electronic and ionic impact phenomena. Combining equations (3) and (4), we find for Clarendon Press, Oxford. the visual magnitude MASSEY, H. S. W., AND SMITH, R. A. 1933. The passage of positive ions through gases. Proc. Roy. Soc. London, ser. A, vol. 142, M=7.1-2.5 log10 Q UIOBU^. pp. 142-172. MOTT, N. F. (5) 1931. On the theory of excitation by collision with heavy particles. Proc. Cambridge Finally, for constant velocity, the second term Phil. Soc, vol. 27, pp. 553-560. MOTT, N. F., AND MASSEY, H. S. W. in equation (5) is a constant and can be com- 1949. Theory of atomic collisions. Ed. 2. Clar- bined with the first term to yield endon Press, Oxford. OPIK, E. J. 1937. Researches on the physical theory of meteor (6) phenomena. III. Basis of the physical theory of meteor phenomena. Publ. where Obs. Astron. Univ. Tartu, vol. 29, no. 5. SCHIFF, L. I. (7) 1955. Quantum mechanics. Ed. 2. McGraw- Hill, New York. SLATER, J. C. Acknowledgment 1951. Quantum theory of matter. Ed. 1. The research in this paper was supported McGraw-Hill, New York. STUECKELBERG, E. C. G. jointly by the U.S. Army, Navy, and Air Force, 1932. Theorie der unelastischen Stosse zwischen under contract with the Massachusetts In- Atomen. Helvetica Phys. Acta, vol. 5, stitute of Technology. pp. 369-422. 228 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS WHIPPLE, F. L., AND HAWKINS, G. S. ZENER, C. 1958. Meteors. In Handbuch der Physik, S. 1932. Non-adiabatic crossing of energy levels. FlQgge, ed., vol. 52, pp. 519-564. Proc. Roy. Soc. London, ser. A, vol. 137, Springer-Verlag, Berlin. pp. 696-702. Abstract An upper limit for the ionizing probability /3 is obtained for sodium atoms that have evaporated from the surface of a meteoroid entering the atmosphere, and the result is generalized to apply to the other meteor constituents. Com- parison of this result with radio meteor measurements leads to a meteor mass which suggests an upward revision of the meteor mass scale. An upper limit for the luminous efficiency is also obtained. The lower limit for the mass of a zero-magnitude Geminid meteor incident at 45° is found to be 6.86 gm, and the mass of a zero magnitude Perseid is 2.35 gm. Statistical Verification of the Physical Theory of Meteors By B. J. Levin1 and S. V. Majeva1 Hawkins and Southworth (1958), using the ing values for M=K for the same value of photographs of 360 meteors obtained with Am=mnm—mML. Only 13 meteors (3.5 percent) Baker Super-Schmidt cameras, compared their have anomalous position of both points, i.e., a light curves (for three points—beginning, maxi- very short path; this means that only a flare mum light, end) with the theoretical curve was photographed. For seven meteors (2 per- for nonfraginenting meteor bodies. They cent) the approximately normal position of the showed (see fig. 5 of their paper) that almost beginning point is accompanied by the all measured points lie within the theoretical anomalous position of the end point; this means light curve. The same is true if, instead of that the meteor ended in a terminal flare or taking a curve for n= % as was done by Hawkins shortly after a flare. and Southworth, we take a curve for ju— % In spite of a very liberal classification of which gives a shorter path length. The beginning and end points as "normal," only 117 parameter n characterizes the relation between meteors (33 percent) have approximately nor- the changes of the frontal cross-sectional area mal positions of both points. At the same time, S and of the mass Mot the meteor body according for 221 meteors (62 percent) the position of the to the formula S^M" (Levin, 1956, formula beginning points is anomalous, i.e., they lie too 3.24). The rising branches of these curves close to the points of maximum light, while the are similar, but the descending branch for end points He more or less normal. Such n=% is much steeper than for fi=%, and even deviations from theoretical light curves indi- steeper than the curve drawn by Ananthakrish- cate the fragmentation of meteor bodies. If nan (1960) according to the hypothesis that we are to regard as normal only those beginning the luminosity factor r~vp (instead of the points which lie in the region of theoretical conventional r~t;). light curves for different values of n, i.e., those Jacchia (1955) noted that the observed path which lie to the upper left from the straight length of meteors is shorter than the theoretical line corresponding to M=0, then 87 percent of one, and he ascribed this to fragmentation of meteors become anomalous. For them the meteor bodies. More convincing evidence in mean increase of light intensity is greater than favor of the fragmentation theory can be the maximum theoretical value for meteors obtained by comparing the positions of begin- produced by nonfragmenting bodies. ning and end points (relative to the point of Levin (1956) used the data on 140 meteors maximum light) for the same meteor. This photographed at Harvard by conventional was done by using the data of Hawkins and cameras to obtain a statistical verification of Southworth. Moderate deviations of points two basic formulas for the physical theory of from the theoretical curves were not taken into meteors. Below are given the results of an account, and only those beginning or end points analogous verification of these formulas using were regarded as anomalous for which HB—HML the data for 337 meteors photographed by or HML—HB were less than a half of correspond- Super-Schmidt cameras. Hawkins and South- worth kindly sent us their unpublished data on 1 O. Schmidt Institute of Physics of the Earth, U.S.S.R. Academy of Sciences, Moscow. the masses of meteor bodies. These masses 229 230 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS were obtained by the photometric method (by The reduction of the observed maximum integrating the light curve) and are expressed magnitudes mML, with theoretical values for in the Opik-Jacchia scale. They are averages a, 0, 7, gave for "normal" meteors a dispersion of values found for each film of a pair. South- e=0*29, and for fragmenting meteors e=0*64. worth informed us that there is frequently a The least-squares solution is given in table 2. large difference between these individual values. According to nonfragmentation theory, the TABLE 2.—Observed parameters in the equation for air density at the height of maximum light p "Normal" Fragmenting ML meteors meteors can be represented by the formula (Levin, a 0.86±0.08 0. 73±0.03 1956, formula 3.40) 0 3. 67±0.08 2. 67±0. 10 7 0. 72±0.09 0. 46±0. 10 c cos Z, ±0*49 and if the drag coefficient and the heat-transfer For meteors produced by obviously frag- coefficient are constant, a=0.33, b=2, and menting bodies, the values for a, /3, 7 are e=l. The data for 116 "normal" meteors and substantially smaller than the theoretical ones. for 221 meteors produced by obviously frag- This probably shows that the fragmentation menting bodies were analyzed separately with masks the dependence of mML on Mo, v0, cos Z. the theoretical values of a, b, c given above, the For "normal" meteors, a, 0,7 have intermediate dispersion e of the reduced H is 4.7 km for values, probably also because of fragmentation ML 2 "normal" and 4.8 km for fragmenting meteors. in a more moderate form. The least-squares solution leads to a marked The least-squares solution also gives values reduction of dispersion and gives the values of of (mML)i0, the maximum absolute magnitude the parameters shown in table 1. It remains of a meteor produced by a body with Mo= 1 gm, flying vertically into the atmosphere at a TABLE 1.—Observed parameters in the equation pML velocity of »0=10 km/sec. We obtain "Normal" Fragmenting for "normal" meteors: (m f )io= + lf?2 meteors meteors A i a 0. 32 ±0. 05 0. 15 ±0.05 and b 0. 89 ±0. 18 1. 9 ±0. 14 for fragmenting meteors: (mafI,)i0=-|-0'74. c 0.85 ±0.21 0. 55 ±0. 15 t ± 3. 6 km ±3.9 km These values are expressed in photographic magnitudes. If we convert them to a visual unclear why the value for b for "normal" scale and, in spite of some doubts, use the same meteors is about half the theoretical value, correction +1*8 as for meteors photographed while the values for a and c are close to theory, by conventional cameras, we obtain the values and why for fragmenting meteors the reverse +3*0 and +2*2, respectively. The first is true. value (for "normal" meteors) is close to the The analogous verification was done for the previously obtained value +2*86 (Levin, 1956). maximum light formula Accordingly, meteor bodies that produce a meteor of zero visual magnitude during vertical /ar-A/gflg COS* Z motion with initial velocity po=4O km/sec must have initial masses 0.062 gm and 0.030 gm, (Levin, 1956, formula 3.60). The values a=l, respectively (instead of M =0.055 gm obtained 0=4, and 7=1 correspond to the theoretical o in 1956). It is clear that fragmentation de- basis of the photometric method for the deter- creases the path length of a meteor and at the mination of masses of meteor bodies. However, same time increases its brightness; therefore since the masses are obtained by integration of the light curve which for real meteors deviates > Note added in March 1962. Further Investigation showed that after from the theoretical one, the least-squares solu- a simple allowance for fractlonation, the theoretical values of a, 0, y give tion gives somewhat different values for a a good representation of all observations together with a dispersion +0»16; i.e., even better than a least-squares solution for "normal" meteors. (Levin, in press.) METEOR SYMPOSIUM—LEVIN AND MAJEVA 231 a smaller initial mass is needed to produce a HAWKINS, G. S., AND SOUTHWOBTH, R. B. meteor with a given maximum magnitude. 1958. The statistics of meteors in the earth's atmosphere. Smithsonian Contr. Astro- The manner in which the fragmentation in- phys., vol. 2, no. 11, pp. 349-364. fluences the estimates of the masses of meteor JACCHIA, L. G. bodies and of the amount of meteoric material 1955. The physical theory of meteors. VIII. accreted by the earth deserves further Fragmentation as a cause of the faint- investigation. meteor anomaly. Astrophys. Journ., vol. 121, pp. 521-527. References LEVIN, B. J. ANANTHAKBISHNAN, R. 1956. The physical theory of meteors and mete- i960. Light curves of meteors. Nature, vol. 187, oric matter in the solar system. Publ. pp. 675-676. House Acad. Sci. U.S.S.R., Moscow. On the Color Index of Meteors By John Davis 1 A comparison of photographic and visual spectral transmission curves. The author has meteor magnitudes reveals an interesting re- reduced his data in the same way as Jacchia, lationship. Jacchia (1957) has compared the but only 53 meteor observations were available, meteor color index, defined as the difference and the larger errors associated with the mean between the photographic magnitude and the points in figure 1 can be accounted for by the corresponding visual magnitude, with the maxi- difference in the number of observations. mum intrinsic brightness of a meteor as The variation of color index with brightness measured with a photographic system. The has been explained by Jacchia in terms of the photographic system used was a Super-Schmidt PurkynS effect. He suggested that the varia- camera (Whipple and Hawkins, 1958) with tion is a spurious physiological effect due to the blue-sensitive film corresponding approximately difference in the spectral sensitivity of the eye to the International Photographic system for bright and faint sources (photoptic and (Allen, 1955), and the visual magnitudes were scotoptic vision). the direct estimates of an observer. Jacchia Ceplecha (1959) has made a similar study found that the color index showed very little with the difference that he used panchromatic variation with meteor velocity but a large film, sensitive to approximately 6700 A, to variation with maximum photographic bright- obtain his "photographic" magnitudes. He ness, the ratio of "photographic" to "visual" thus obtained a panchromatic index which light decreasing with maximum meteor bright- he found to increase as the maximum brightness ness. The author has confirmed Jacchia's of the meteor decreased. This is shown in results during the combined photographic and figure 1 for comparison with the color index. radio-echo meteor program at Jodrell Bank The quite different results obtained with blue (Davis, Greenhow, and Hall, 1959; Davis, 1959). and with panchromatic systems have been The results of both Jacchia and the author suggested by Ceplecha to originate in real are shown in figure 1, where color index has spectral changes of the meteor light with a been plotted against visual magnitude in order previously unsuspected high emission in the that comparison may be made with results red becoming relatively predominant in fainter obtained with a photographic system of dif- meteors. ferent spectral sensitivity. Jacchia has divided All the observations described have been the observations of 294 meteors into four based on subjective estimates of visual mag- groups according to absolute photographic nitudes, and these are suspect. Systematic magnitude, and deduced the mean color index differences up to 0.5 magnitude between ex- for each group. The two points given by perienced observers have been found, and it Jacchia for the brightest meteors were de- appears that a single visual magnitude estimate rived from the observations of 45 meteors has a standard deviation of the order of ±0.8 made with the Harvard small cameras. It magnitude. Inspection of the data included is possible that color indices derived from small in figure 1 shows that the standard deviation cameras may differ from those from Schmidt of a single visual estimate is ±0.8 from the cameras due to differences in their relative author's observations. Jacchia's results give a standard deviation of ±0.6 magnitude, but < Nuffield Radio Astronomy Laboratories, Jodrell Bank, England. Now at School at Physics, University of Sydney, Australia. since the majority of his results are based on 233 234 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS have been taken into account in the calcula- tions. The results have been combined with Colour Indei the spectral sensitivity curve of the Kodak photographic emulsion (V6010) to give the relative spectral sensitivity curve of the camera- emulsion combination. Spectral sensitivity Ponchromolic curves for the normal and dark-adapted eye Index (Allen, 1955) are also included for comparison with the green photoelectric system. Common -4 -2 O +2 +4 filter glasses have been used; nevertheless, a Visual Magnitude reasonably good match has been obtained, FIGURE 1.—The variation of meteor color index with meteor particularly in the case of the blue system. brightness. X, The results of Jacchia obtained with a The varying components of the photomultiplier blue photographic system; Q, results of Davis obtained outputs over a bandwidth of 1/10 to 20 c.p.s. with a blue photographic system; O, results of Ceplecha have been recorded with a Siemens high-speed obtained with a panchromatic photographic system. pen recorder. The amplitudes of meteor signals from the the mean of two visual-magnitude estimates, blue (OBlO) and green (OGr2) channels have these figures are in good agreement. They been converted to arbitrary magnitude scales suggest that the main cause of scatter in color which correspond to the photographic and index can be ascribed to the inaccuracy of the visual magnitude scales respectively, except estimates of visual brightness. Objective for the addition of constants which at this measurements of visual magnitudes are there- stage have not been determined. In figure 3, fore very desirable in this kind of work. green magnitudes have been plotted against Two photomultipliers with Tri-alkali cath- the author's direct visual estimates of meteor odes of 5-inch diameter have been used in a study of this problem. They have been set up magnitude. The broken line represents a with a screen above them to define a zenithal linear, 1:1 relationship between the two magni- field of 50° diameter and with Chance glass tude scales, whereas the full line has been filters to provide the required spectral re- derived from figure 1 for the hypothetical case sponses. One tube has been fitted with an of a constant color index with the observed OBlO filter to give a spectral sensitivity cor- responding approximately to a blue photo- Meniscus Schmidt graphic system, and the other has an OGr2 Pnologrophic filter to give a response corresponding to a Blue Photoelectric mean visual observer (mean photoptic and scotoptic). The relative spectral responses of the photocathode-filter combinations are shown in figure 2. The relative spectral sensitivities of the photocathodes and the ^ Nornol Eye spectral transmission of the filters have been A\\ .-.Dork Adopted Eye measured at the National Physical Laboratory, Teddington, England. The relative spectral sensitivity curve of the photographic system (Davis, in preparation) is shown in figure 2 for comparison with the blue photoelectric system. Photoelectric The spectral transmission curve for the Royal Society Meniscus Schmidt camera has been O3 OS O7 Ot computed from data supplied by Chance WAVELENGTH IN MICRONS Brothers Ltd. The transmission of the lenses, reflection at mirror and at alTglass-air surfaces FIGURE 2.—The relative spectral sensitivity curves of the three photocathode-filter combinations. METEOR SYMPOSIUM—DAVIS 235 been derived (2? is the amplitude of the blue photoelectric signal on an arbitrary scale). + 1 An absolute calibration of the blue magnitude Constant Colour Index—observed scale has been made from fifteen meteors +2 variation due to the Purkyne effect photographed and recorded photoelectrically. The photographic magnitude scale is indicated in figure 4 alongside the blue scale, the uncer- tainty in the calibration being ±0.1 magnitude. u +4 The curve representing the variation in color O 2 4 6 index with meteor brightness determined from 2-5 Loo G visual and photographic observations, which is FIGURE 3.—A comparison between visual magnitudes and shown in figure 1, has been transposed to the green photoelectric brightness. figure 4. It must be noted that the color index given in figure 4 is not absolute but variation being due to the Purkyne" effect. gives only the relative change in index with A total of 24 meteors has been used in this meteor brightness. Owing to the uncertainty diagram. Since the major error lies in the in the calibration of the blue and the green estimate of visual magnitude, the data have magnitude scales against the photographic been divided into four groups according to and visual scales, it is not possible to fix the their value of 2.5 log G, and the mean values position of the curve relative to the ordinate of visual magnitude have been derived (G is of figure 4 with sufficient accuracy for compari- the amplitude of the "visual" photoelectric son with the photoelectric color indices. There- signal on an arbitrary scale and is proportional fore the position of the curve relative to the to the incident light intensity). The diagram ordinate scale has been chosen arbitrarily to shows clearly that it is not possible to explain give the best fit to the points. It will be seen the variation in terms of the Purkyne" effect that a good fit is obtained, thus supporting the alone. This conclusion is further supported conclusion drawn from figure 3. The implica- by figure 4 in which the photoelectrically tion is that the variation is due to a real change determined color index is plotted against blue in the meteor spectrum. The Purkyne" effect magnitude. Some 40 meteors have been used has a much smaller influence on the results in this diagram and, since the error in 2.5 log than was envisaged by Jacchia. B/G is the greater, the data have been divided Another series of observations has been into five groups according to their values of made in which the OGr2 filter was replaced by 2.5 log B, and mean values of 2.5 log B/G have an OR2 filter to give a response from approxi- mately 6000 to 7000 A with the peak sensitivity -O-5 from 6100 to 6500 A. The relative sensitivity curve of this photocathode-filter combination O is shown in figure 2. A red index, which is J-t-OS defined as the difference between blue (OB 10) magnitude and red (OR2) magnitude, has '+ IO been plotted against blue magnitude in figure +• 1-5 5, in which 44 meteors have been used. The data have been treated in the same manner as •I-2-O for figure 3, being divided into five 2.5 log B IO 6 4 2-5 Log B groups, and mean values of 2.5 log B/R have been derived. Again the index is only relative, -6 -4 -2 O +2 Photographic Magnitude but it shows clearly that faint meteors are relatively redder than bright meteors. A FIGURE 4.—The variation of photoelectric color index (2.5 smooth curve has been drawn through the log BIG) with meteor brightness. O, Points representing the mean of 8 meteors; -f, point representing a single points in figure 5 to indicate the trend of the bright meteor. results. 236 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS -OS- to red light as the meteor passes from the faint O- beginning, through maximum brightness, to the s lower end of the trail. Millman suggests that • +O-5 faint meteors may exhibit a spectrum similar to the faint early parts of bright meteor trails. If + I-O this is so, changes in color index along the light curve of a bright meteor should be observable + 1-5 with the photoelectric system. No evidence to +2-O 6 4 support this has been found so far, but the 2-5 Log B problem is being investigated further. -6 -4 -2 O +2 I am grateful to Sir Bernard Lovell for his Photographic Magnitude encouragement in this work and for the re- FIGURE 5.—The variation of the photoelectric red index (2.5 search facilities at Jodrell Bank. I also thank log BfR) determined photoelectrically with meteor bright- Messrs. Turner and Newall for a research ness. Each point represents the mean of 8 meteors. fellowship. Approximately half of the meteors observed References in the photoelectric program have been Gem- ALLEN, C. W. inids, and the remainder sporadics. No signif- 1955. Astrophysical quantities. Athlone Press, icant difference has been found between meteors London. of the Geminid shower and sporadic meteors. CEPLECHA, Zd. The preliminary examination has shown that 1959. On the colour index of meteors. Bull. Ceplecha's proposed explanation of the ob- Astron. Inst. Czechoslovakia, vol. 10, p. 39. served variation of color and panchromatic DAVIS, J. indices with meteor brightness in terms of a 1959. Simultaneous photographic and radio echo changing spectrum is correct. If the total observations of meteors. Thesis, Uni- light from a meteor is considered, faint meteors versity of Manchester. have relatively greater emission in the green DAVIS, J.; GREENHOW, J. S.; AND HALL, J. E. 1959. Combined photographic and radio echo and red regions of the spectrum than bright observations of meteors. Proc. Roy. meteors. The transition in the emission spec- Soc. London, ser. A, vol. 253, pp. 121- trum over the range of visual magnitudes from 129. —2 to +3 is apparently a gradual one. JACCHIA, L. G. Dr. Peter Milhnan (personal communication) 1957. On the "color index" of meteors. Astron. has pointed out that the spectra of bright Journ., vol. 62, pp. 358-362. meteors exhibit changes along the meteor trail. WHIPPLE, F. L., AND HAWKINS, G. S. The great majority of all spectra containing 1958. Meteors. In S. Flugge, ed., Handbuch der Physik, vol. 52, pp. 519-564. sufficient detail show an increasing ratio of blue Springer-Verlag, Berlin. Abstract Objective measures of meteor color indices have been made with a two-channel photoelectric apparatus. Photo- cathode-filter combinations giving spectral responses corresponding to a Super-Schmidt photographic system and a mean visual observer have been used. The results show a variation of color index with meteor brightness in agree- ment with the subjective results of Jacchia and Davis. However, the variation can no longer be accounted for in terms of a physiological effect alone as proposed by Jacchia. It must arise from a real change in the meteor spectrum with meteor brightness as Ceplecha has suggested. This interpretation is further supported by the observations made with blue and red sensitive photoelectric channels. SPALDING, COLTER, ET AL. PLATE 1 a, The electronic equipment, b, Three motion-picture sequences of Perseid meteors. I I! i SPALDING, COLTER, ET AL. PLATE 2 a, Perseid meteor photographed in August 1960. b, Photograph of a bright meteor. Image-Orthicon Photographs of Meteors By J. Spalding,1 G. Colter,2 C. L. Hemenway,3 J. A. Cole,3 and J. F. Dugan3 The function of this brief communication is Equivalent speed: ASA 1,000,000 Operating mode: 500 lines, non-interlaced threefold: First, to show some image-orthicon OT photographs of meteors; second, to make sure Limiting magnitude: ^>6 .O for 1/30 sec exposure. that the characteristics of image-orthicon tech- Plate la is a photograph of the electronic niques are understood by the scientists interested equipment employed. The image-orthicon in the study of meteors; and third, to seek camera is not shown and is outside the building advice for the best possible means of exploiting directed toward the sky. This system has the the capabilities of these techniques for meteor advantage that the observer is in a comfortable, research. heated room. The recording technique that The most important characteristic of image- we have used is direct photography of the orthicon techniques for the study of meteors monitor screen. The video tape technique is their very high sensitivity, which ranges suggested by D. W. R. McKinley * offers a from 200 to 400 times that of the fastest number of advantages. photographic films. This high sensitivity per- Plate 2a is a picture of a 3d-magnitude mits good time resolution at the expense of some Perseid meteor that we obtained in August loss in spatial resolution. Typical resolution for 1960; the meteor motion is effectively "frozen," meteor studies is of the order of 10 lines per milli- and the meteor wake is clearly separated from meter. The photo cathodes presently available the meteoroid. Thus it is possible to study are approximately 1.75 inches in diameter. The separately the time and spatial development of readout from such equipment is extremely the meteor, and its wake and train phenomena. rapid, and 30 pictures per second can be Plate 26 shows what happens when a bright recorded with little difficulty. Image orthicons meteor comes into view. Although this —1 are available with different spectral sensitivities. magnitude meteor overloaded the system, the There are tubes sensitive in the ultraviolet, in meteor wake can still be distinguished. Note the infrared, and in the visual. Images can be that this system has sufficient speed to photo- stored on an image-orthicon target if desired. graph a weak aurora (below the meteor) in In fact, we have stored an image of stars for a l/30th of a second. half-hour without refrigeration and found that Plate 16 shows three motion-picture se- the image brightness decayed by less than a quences of meteors taken at l/8th of a second magnitude. By pulsing the cathode to target intervals during the last Perseid shower. The potential, one can take exposures as short as a first sequence shows a faint meteor with a millisecond without difficulty. short-duration train. The second and third The operational characteristics of the system, sequences show meteors that fragmented, and as used this past summer for motion picture the spatial developments of the associated photography of meteors, are as follows: flares are seen along with longer-duration Objective: 50 mm. fl, f/1.4 trains. Tube: GE Z5396 (visual sensitivity) In summary, faint meteors, wakes, and > General Electric Company, Schenectady, N.Y., and Dudley Ob- trains can readily be studied by image-orthicon servatory, Albany, N.Y. 1 General Electric Company, Schenectady, N.Y. 1 Dudley Observatory, Albany, N.Y. • Meteor Science and Engineering, McGraw-Hill, New York, 1961. 638805—63 20 237 238 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS techniques. We note that the magnitudes of We welcome comments and suggestions for meteors can be measured by direct comparison the exploitation of this equipment. We have with the stars since short-duration exposures already obtained low-resolution spectra of are used. The equipment is suitable for taking meteors and trains with an objective prism motion pictures, which permit study of time coupled to this equipment. A transmission and spatial development of meteors and their grating spectrograph has been constructed and spectra. will be in use shortly. Meteoritic Erosion in Space By Fred L. Whipplex Recently developed techniques in meteoritic tion of exposure age leads to a determination astronomy now enable us to determine upper of an upper limit to the rate of erosion acting on limits, or rough evaluations, of the erosion rates the meteorite while it was in space. The actual on the surfaces of meteoroids moving in orbits erosion rate experienced by the meteorite about the sun. may well have been smaller, depending upon Successive sections of this paper discuss the the unknown time at which the meteorite was observational data on erosion rates, the theory separated from a larger parent body. of high-velocity cratering, the deduced mean Suppose, as an example, that in a recently density of the interplanetary dust component fallen meteorite the quantity of the stable and evidence for a dust cloud about the earth, argon isotope A38 is measured, as well as that of general physical effects on meteoroids in space, the isotope A39, which has a half-life of some 325 and effects that meteoroids may have on space years. Since the relative cross sections for vehicles. average cosmic-ray spallation are known for these two isotopes, the A39 measurement Evidence for erosion of meteoroids in space provides a determination of the average rate When cosmic rays penetrate solids in space, of production of A38 over some 100 revolu- they produce a number of transmutation or tions of the meteoroid in orbit about the sun spallation products in the material of the solid- and over a large number of sunspot cycles (the Some of the atomic species so produced are latter in case the cosmic-ray activity is depen- stable against radioactive decay, while others dent upon solar activity). have a half-life that may be short or long. A Let us assume that the measurements are measurement of a stable isotope so produced made somewhere near what was the center of establishes the integrated effect of cosmic rays the meteoroid before it was ablated and broken over the period of exposure, while a measure- by atmospheric passage, and that the meteor- ment of a short-lived isotope provides the oid was roughly spherical, of radius R. Let cosmic-ray spallation rate over a recent, short L be the distance through which the cosmic- interval of time. After correction for spallation ray spallation effect is reduced by a factor of e. cross sections, the ratio of the observed abun- Then the measured A39 content leads to the dances of two such isotopes in a recovered me- rate of production of A38 given by the equation teorite provides a determination of the exposure age. (1) The penetrating powers of cosmic rays are definitely limited. One may assume that the where C is the constant determined by the cosmic-ray rate has been statistically constant measurement. in the past and conclude that the exposure age If we postulate a statistically linear erosion represents the interval of time since the indi- rate of the external surfaces by all causes while vidually recovered meteorite was separated the meteoroid was in space, the accumulated from a larger body. Whipple and Fireman A38 will amount to (1959) showed that an alternative interpreta- 1 Smithsonian Astrophystcal Observatory, and Harvard College Ob- 3*=C fT exp [-{R+tt)IL]df (2) servatory, Cambridge, Mass. Jo 239 240 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 8 where T' is the interval of time since the five oldest average about 6 X10 years. I shall, meteoroid was exposed to cosmic rays, therefore, adopt as a significant estimate, presumably by separation from a parent body. probably correct to within a factor of 2, an The usually defined exposure age T follows upper limit to the erosion rate of nickel-iron 7 1 from a constant rate of production by cosmic meteorites in space, e=1.2X10~ cm yr" , to be rays producing the A38; from equation (1) the used in the present discussion. accumulated A38 is given by The fact that the exposure ages by the argon method and by similar methods have con- 3 A »=CTexp(-R/L). (3) sistently come out much smaller for stony meteorites than for iron meteorites has been a On integrating equation (2) and equating 38 subject of considerable interest for several the values of A to that given by equation (3), years. Fireman and DeFelice (1960) suggested we find that the difference results from the physical weakness of stones as compared to irons, result- ing in a more rapid etching of the stones, a major assumption of this paper. In a summary If we assume that T is very great compared of argon measurements for various stony mete- to T, we obtain an upper limit to the erosion orites, Fireman and DeFelice (1961) find that rate by the simple expression, for nine chondrites the A39 decay rate ranges 1 e=L/T. (5) only from 5 to 12 kg~* min" . The exposure ages and the A38 contents are similarly confined Eventually our knowledge of the breakup in range (e.g., A38 only from 1 to 2X10~8 sec of asteroids in the formation of meteorites may gm"1). Typical exposure ages by the argon be adequate to enable us to put in a significant method and also by similar methods for chon- value of T' in equation (4). At present, the drites generally are less than 10s years. Until upper limit to the erosion rate given by equa- more definitive results are available, I shall tion (5) represents an adequate statement of adopt the typical exposure age of 4X107 years our knowledge. determined for the Bruderheim meteorite by For the Sikhote-Alin nickel-iron meteorite Fireman and DeFelice, and a shielding constant Fireman (1958) found T=5X108 years by a Z=70 cm, suggested by these authors. The measurement of the decay rate of A39 and upper limit to the erosion rate for typical stony an evaluation of the A38 content from the He3 meteorites then becomes e=1.7X10~8 cm yr"1. measured in that meteorite. Whipple and Again the uncertainty may be comparable to a Fireman (1959) adopted a shielding constant factor of 2. Z=72 cm on the basis of spoliation measures For photographic meteoroids, Jacchia and in a high-energy accelerator, and thus derived Whipple (1961) find semi-definitive evidence an upper limit to the erosion rate of a nickel- that the lifetime of photographic meteoroids iron meteorite as «=1.5X10~7 cm yr"1. Fisher in space is relatively short, of the order of (1961) has criticized this determination; he 10,000 years or perhaps less. Some 60 percent prefers a value based upon Ne21 measurements of the precisely reduced photographic meteors in the Grant meteorite leading to a value can be assigned to streams or associations; also «=1.1X1O~8 cm yr"1. A detailed critical dis- there is no tendency for meteors having aphelion cussion of the pros and cons concerning mete- distances within the perihelion distance of orite exposure ages by various isotopic measure- Jupiter to have their lines of apsides aligned ments would become too involved for the with those of Jupiter, as is so conspicuous for present paper and probably would not result the asteroids. in any definitive conclusions in the present Since so many of the meteor orbits cross that state of the art. Measurements of 17 iron of Jupiter and are subject to relatively great meteorites giving exposure ages by A39 or Cl3a perturbations by that planet, the tendency to rates are listed by Anders (private communica- streaming among photographic meteors in- tion) and average about 3.4 X108 years. The dicates a relatively short lifetime. Over METEOR SYMPOSIUM—WHIPPLE 241 periods of time adequate for the Poynting- third-magnitude meteor as representative of Kobertson effect to reduce the aphelia ap- the fainter limit of the photographic determina- preciably, Jupiter perturbations should cause tions. This assumption would reduce the value most of the stream identities to become un- in equation (7) to a rate of approximately 10~5/P recognizable. The lack of systematic align- cm yr"1. However, the determinations for ment of the apsides among the meteors of iron and stony meteorites are both upper limits, smaller aphelion distance again indicates that and tentative indications suggest that the Jupiter perturbations have not had time to photographic meteoroids may, indeed, be de- play an appreciable role for these meteoroids. stroyed in a shorter mean time than 10* years More precise evaluation of the time scale in- in the inner region of the solar system typical volved is under study by S. Hamid at the of the orbital paths expected for recovered Smithsonian Astrophysical Observatory, but meteorites. No clearcut information is yet for the moment, an estimate of 10* years ap- available concerning the density p to be applied pears to be reasonable, although perhaps too in equation (7). The lowest mean value is long. p=0.05 gm cm"3 suggested by Cook and Destructive forces are thus clearly at work Whipple (Whipple, 1955), but based upon a among the cometary meteoroids. Barring rather unreliable method of mass determination totally destructive collisions, an erosion rate that depends upon the forward motion in the should fairly describe the destructive processes. incoming train of a bright photographic meteor. Thus, to determine the erosion rate for the There is little question but that some stream photographic meteoroids, we need a determina- meteoroids may average densities even smaller tion of the meteoroid dimensions. Again, to than this, for example, the Giacobinid stream provide an upper limit, let us assume that the of 1946, but average values for more typical 104-year period of destruction applies to a zero- photographic meteoroids may more probably magnitude meteor with an entry velocity of 30 lie near 0.3 gm cm"3. This problem of mete- km sec"1. Until the luminous efficiency factor oroid density is under continuous investigation; or the ionizing efficiency factor has been theo- better values will be obtainable when more retically or experimentally determined for definitive examples of stony meteors have been photographic meteoroids, we have available for investigated by photographic meteor tech- dimensional determination primarily the drag niques. We have clearcut evidence, partic- measurements in the Harvard photographic ularly that by McCrosky (1955), that photo- meteor program. These measurements provide graphic meteoroids are weak structures and the following relation between meteoroid radius that many of them crush under dynamic pres- R and density p, for a 0th magnitude meteor at sure of the order of one-third atmospheric velocity 30 km sec"1: pressure, approximately 10~* dyn cm"2. Opik (1955) previously deduced such a crushing i?p=0.23 (cgs). (6) strength from less definitive observational data. I am indebted to Allan F. Cook and Luigi G. The basic assumption of this paper, that the Jacchia for this evaluation, which is probably erosion rates calculated above for iron, stony, significant to 20 or 30 percent. The largest and cometary meteoroids are actually measures uncertainty rests in the assumed shape factor of space erosion, is an ad hoc assumption for of the meteoroid, which was taken to be 1.66, which no great affirmatory evidence can be and the drag coefficient, taken as (7d=2.0; both amassed. Its chief raison d'etre rests on the enter linearly on the right side of the equation. fact that the values increase sequentially as Accepting a time of 10* years for total we go from materials of stronger to weaker erosion, we estimate for the erosion rate in structural strength, or better, from greater to equation (6), lesser crushing strengths. Also there seems 5 1 little reason to believe that etching processes c(phot. met.)=2.3X10" /p (cm yr" ). (7) such as ionic sputtering would be highly sensi- This value may be taken as a rather high tive to strength of materials. The binding in estimate since we might well have chosen a crystalline structures is not so vastly variable 242 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS from one material to another (see, for example, samples are made, the problem does not appear Reiffel, 1960). Crater formation, however, damning to the erosion concept. where there is no gravitational or other mecha- It is quite clear, as Anders points out, that nism for retaining crushed material, should pure compressive strength is not the vital increase markedly as the target material quantity that should measure the cratering becomes weaker structurally. efficiency of high-velocity impact. Some nu- Anders (private communication) carefully merical measures of malleability or of brittleness discusses the pros and cons of the erosion con- should enter the cratering formula. Everyone cept for irons and stones, and questions the knows that violent shock on a meteorite will conclusion severely on the following grounds: break more stone than iron. The volume of (a) Some of the harder stones have short ex- displaced material is mostly determined by posure ages, while the "friable" Norton County whether the material breaks when compressed has an age of 2 to 5X108 years, (b) Macro- violently as the exposure shock wave spreads scopic strength cannot be the major factor in and attenuates. The iron in most chondrites cratering by dust, only the microstructure; the is usually not adequate by volume (5 to 10 mechanical strength of iron and stone crystals percent), and is too poorly distributed to offer do not differ much; also from Buddhue's (1942) major protection to the stony material against measures of crushing strength even the macro- crushing and shattering. Undoubtedly it does scopic strengths of some stones are comparable offer appreciable protection in some chondrites, to that of the Descubridora, Mexico, iron (Om), but the stronger crystals in the stone will fall 3.7X109 dyn cm"2, (c) Strong clustering of apart when the softer matrix is worn away by ages occurs for the chondrites near 22X106 cratering impact. years versus much shorter ages; clustering also The questions raised by Anders cannot be occurs for the irons near 3X108 years and ignored and suggest a number of experiments 6X108 years; the hexahedrites seem to show and measurements that must be made to shorter life times than the medium octahedrites. finally clarify the situation. I prefer for the Some of these objections are also voiced by moment, however, to adopt the erosion concept Eberhardt and Hess (1960). as a working hypothesis applying to the older The shorter derived exposure ages, whether exposure ages, and to pursue the consequences for irons or stones, whether for hard or soft of this hypothesis. material, whether or not clustered in time, really do not bear on the erosion problem. Cratering by high-velocity impact There is every reason to believe that collisions No definitive theory or experiment yet defines and breakage among meteoroids must occur. the laws of crater formation by high-velocity Thus the shorter exposure ages (when not due impact. Indeed, the broad dependence of the to leakage of gas) are reasonably interpreted phenomena upon velocity or nature of the in the now classical fashion as measuring some materials is not yet established. Rapid prog- mean time since a breakup or breakups. ress, however, is now being made in this area The great age for the Norton County of research, both in theory and in experiment. achondrite presents a problem. Edward L. The only theory that appears broad enough to Fireman has permitted me to express his opinion cover the types of target materials of interest as follows: "The exposure age of Norton County in meteoroids is by Opik (1958). He calculates is more uncertain than that of other meteorites that a projectile (or meteoroid) of mass ft, because its Fe content (~1 percent) is the striking a large solid surface of density p, and lowest of any meteorite. Thus argon is pro- compressive strength S, at a meteoric velocity duced mainly from Ca and K; the He may be v, will crush and displace a mass M, according produced by a high Li6 content. Some Li and to the equation target measurements in Norton County are M necessary to put its exposure age on a firmer basis." Until these experiments have been done, and until complete measures on single where k is a dimensionless constant that depends METEOR SYMPOSIUM—WH1PPLE 243 upon the shape and nature of the impinging For hemispherical craters in aluminum or body. He derives values of k=4.7 for iron stone, his formula (1961a) is impinging on stone, and k=4.3 for stone on iron. Fireman and DeFelice (1960) apply his Penetration=2.35X10"2(M»)1/3 (cgsunits). (9) theory to the problem of erosion rates and meteoritic collision in space. Thus, Bjork's equation for crater formation From effects of surface nuclear bursts, S. agrees with Opik's insofar as velocity is con- Glasstone (1957, par. 5.8) adopts a scaling law cerned, but carries no explicit term for the strength or ductility of the material. For in which the linear dimensions of the displaced 3 earth vary as the cube root of the energy E. stone (p=2.64 gm cm" ) his equation for ratio Thus, the volume or mass of the earth displaced of crater mass to projectile mass is closely and brecciated varies as the energy of the M explosion (with dependence upon the scaled —=7.0X10-5y (cgsunits). (10) depth of the explosion). His constants lead to an energy per unit mass of dry earth displaced The corresponding energy per gram displaced 1 10 1 equal to about 1X109 erg gm"1. at 0=15 km sec" becomes 1.0X10 ergs gm" , Shoemaker (1960), in his calculations of the a relatively high value. mass that produced the Arizona Meteorite I can find little experimental evidence con- Crater (Barringer Crater), adopts Glasstone's cerning the dependence of cratering upon the scaling law and uses constants equivalent to density, strength, or nature of materials. 2.3 X108 erg gm"1 for displaced earth at an Birkhoff et al. (1948) Jfind that the depth of impact velocity of 15 km sec"1. He also uses penetration of a high-velocity jet varies as the scaling law (E)1/3i, which would further p-1/2 Thus, the volume of material displaced in a jet-produced crater should vary more reduce the required ergs per gram for micro- 1 3/2 scopic particles. I used the energy-scaling law rapidly than p" , perhaps as much as p~ , and 10 1 the mass as p° to p~1/2, in contrast to Opik's and a constant of 1.2 X10 erg gm" for alumi- +1/2 num (Whipple, 1957). conclusion, p . Opik's relation (eq. 8) produces a volume of Summers and Charters (1959), from relatively 1 displaced material proportional to the momen- low-velocity impact (to 3 km sec" ) of copper tum rather than to the energy of impact. Thus, and lead projectiles on both copper and lead at #=15 km sec"1, he calculates the impacting targets, find a relation which becomes closely mass of the Arizona iron meteoroid to be some 2 12 M o /v\ 20 times greater than does Shoemaker (1.5 X10 —=5.93—(-) (cgs units), (11) gm vs. 6.3X1010 gm). Opik adopts, however, a 15 crushed mass of 10 gm, about three times the where pm is the density of the projectile, and c is value adopted by Shoemaker, reducing the the velocity of sound in the target material. theoretical discrepancy to a factor of eight Equation (11) follows the energy scaling law; times. From equation (8), if «=15 km sec"1, the calculated energy per unit mass displaced and I have compared the various theories and Let us then adopt k=4.0 in equation (8), formulas for crater formation with the Charters since the meteoroid will probably be less dense photographs of the crater formed in granite by a than stone (£=4.3 for stone striking iron). At nylon ball (see table 1). The theories or formulas a relative velocity v, interplanetary material of by Opik, Glasstone, and Shoemaker fit the mean space density p, will, upon encountering a results from the photographs fairly well, while randomly oriented surface, contribute a mass the predictions by Summers and Charters, vp,/4 per unit area per unit time. The linear Whipple, and Bjork are less satisfactory, pre- erosion rate dR/dt will vary as p"1 times the dieting too small a crater volume by one to two mass erosion rate; hence by equation (8), the orders of magnitude. linear erosion rate becomes TABLE 1.—Crater by nylon ball on granite at dR__ p,v* v=5.S4kmsec-1 ~dt=e==(pS)1/2' ' ' Energy of Mass displaced displaced mass Mass projectile _ _. . . , , . . Source era gm-i M/M Mean space density of dust and comparisons Opi-(eq. )• We can now apply the observed etching rates to Glasstone l X 1O» determine a mean space density of particles by Shoemaker 2.3X108 means of equation (12); hence Bjork 38 Summers and Charters e(pS)112 b (eq. ll) 7.0X10»° 2.0 P.=—~2 (13) Photographs by Charters 6.6X108 220 • Assumed crushing strength of granite, S—1.5X10»dyn cm-'. _,, , - , . , * Assumed velocity of sound in granite, c-6.oxio» cm sec->. 1 he space density and tne velocity of en- counter both vary in some inverse fashion with I shall adopt Opik's solution and constants in solar distance (~r~1/2?), while the time of passage this paper, since they appear to apply fairly well varies in a direct fashion (^r372?). Hence the in an experimental example that simulates the mean value of p, will be somewhat weighted actual situation in space. There can be no towards the outer parts of the meteoroid orbits, doubt that some other quantity than compres- A more refined treatment is scarcely justified at sive strength, perhaps some function of ductility this stage of the study. or brittleness, must more appropriately measure The last column of table 2 gives calculated the material displaced in crater formation by values of the mean space density based on high-velocity impact. Also density seemingly equation (13) and the adopted values of the should enter the equation in a different manner. erosion rates for irons, stones, and photographic More theoretical and experimental results are meteoroids. For irons the value of crushing sorely needed. Fortunately, the velocities of strength is probably too great, but is intended to interest are near enough to the observed velocity include some effect of ductility or malleability, so that the still unresolved question of mo- as compared to the value for stones, checked mentum or energy scaling does not greatly generally by the experiment of the nylon ball affect the calculated results. In both cases on granite. For photographic meteoroids the the scaling is proportional to mass, and no value of £ is probably too small, representing theory suggests much deviation from mass the most fragile 20 percent observed. The scaling so long as the projectile contains a mean velocity is taken at 10 km sec"', as very large number of atoms. representing a value corresponding more to the TABLE 2.—Calculations of space density from erosion rates Compressive Meteoric Mean Erosion rate Density strength velocity space density t(cm sec-1) p(gm cm,-1) S{dyn cm,-2) v(km sec~l) p.igm cm,-') Ir°ns 3. 8X10"15 7.8 2X1O'<> 10 1. 5X10"21 Stones 5.4X10-" 2.6 2X10» 10 3.9X10"" Phot, met, 7. SXIO-'V1 P 2X104 10 0. 10X10-lV"I/1 METEOR SYMPOSIUM—WHIPPLE 245 distance of the earth or Mars from the sun 3.5X10-*° 2 (15) than to the value in the asteroid belt (~5 km N(R) = R * ' sec"1)- The value y=10 km sec"1 is small for meteors, even after neglecting earth gravity. where N(R)dR represents the number of The mean space density calculated from data particles per cm3 in the particle-radius range on photographic meteoroids is more than an dR. In his integration he cuts off the upper order of magnitude smaller than those based on limit at R=0.035 cm; the lower limit-, does not erosion of irons and stones, if we adopt p~l much affect the sum. gin cm"3. A smaller value of p is possible and The mean space density derived by Mc- the crushing strength is probably underesti- Cracken, Alexander, and Dubin (1961) from mated; hence the fit is satisfactory in view of measures on space vehicles near the earth is the uncertainties in all the critical data. A considerably greater than van de Hulst's value. somewhat higher value of v than 10 km sec"1 Comparison, however, is specious since their might be more applicable to cometary meteor- particle-size distribution law is oids than to asteroidal material, because of the 6.0X10"28 greater orbital inclinations and eccentricities. N(R): 1 (cgs units). (16) This would increase the discrepancy. On the other hand, the greater mean applicable solar The mass integral depends almost solely distance and inclination would be expected to 10 reduce the average space density of material upon the lower limit in R, which at mass ~10~ gm leads to a mean space density near the earth encountered by cometary meteoroids. Hence 20 3 the discrepancy is in the expected direction and of 3.2 X10" gm cm" . The space density is doubled if the mean velocity of impact is taken may represent reality. 1 Comparison with other data is difficult. For as 15 instead of 30 km sec" , and increases the Zodiacal Cloud in the neighborhood of the rapidly as the limiting particle radius is reduced. earth's path, van de Hulst (1947) derives But the real difficulty in comparison lies in 1.0 X10"21 p gm cm"3 as the density near the the area provided for diffraction of light by earth's orbit. If we accept a fraction / of this small particles, diffraction being proportional density as applying to the mean region of to the total area. From McCracken et al. cometary asteroids, and assume that the density (1961) the particle area per unit volume is 5.7X10 -17 .-2/3 cm2 per cm3, while van de of interplanetary particles equals that of the 20 2 3 cometary meteoroids, we can solve for p and p, Hulst requires only 6 X10" cm per cm near by this relation and by table 2, leading to the earth, or about one thousandth as much. The newer measures by Blackwell, as indicated pJ/3=0.1// (cgs units). (14) by the calculations of Ingham (1961), require even a lower scattering area, roughly 10"21 cm2 3 If we arbitrarily adopt some reasonable value per cm at the earth's distance from the sun. for/, say 1/4, allowing for a decrease in p, with The great diffracting area observed by Mc- Cracken et al. is provided by particles largely solar distance and away from the ecliptic 4 3 plane, we solve equation (14) to derive p~0.5 in the dimension range from 10" to 10~ cm gm cm"3 and p,= 1.2X10~22 gm cm"3. The that are most responsible for scattering at corresponding value of the mean density near angles of from 1° to 10° from the sun. Hence the earth's orbit is then 5X10"22 gm cm"1. their deduced particle density would increase Other analyses of the Zodiacal Cloud obtain the intensity of the Fraunhofer Corona a lower values of p, than van de Hulst by factors thousand times its observed intensity, unless of 10 or more times (see, for example, the list the particles are confined to a small volume by Nazarova, 1961). The derived density is near the earth, essentially within the moon's very sensitive to the upper limit of the cutoff distance. This evidence seems to confirm the in the distribution law applicable to the smallest existence of a dust cloud about the earth particles. Typical is van de Hulst's derived (Whipple, 1961a, 1961b; Hibbs, 1961a, 1961b; law questioned by Dubin, 1961). 638805- 246 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Consequences of erosion to the Poynting- particles can be neglected. The Poynting- Robertson effect Robertson effect becomes operative when only 10~2 to 10~4 of the total space density remains Let us again accept the postulate made above 2 that cometary meteoroids capable of producing in particles the order of 10" times the mass of the particle eroded. If, for example, van de photographic meteors are eliminated by space 4 erosion of some kind in the order of 10* years. Hulst's law (eq. 15) is valid to Jf2=10~ cm, 1 percent of the mass would then be contained in The average lifetime t of any such meteoroid 4 will then be R/e, or by equation (7), the range 1 to 13X10" cm. Hence, cometary particles less than approximately 60/x in radius 4 *=#PX4.3X10 years, (17) and initially in very short-period orbits, or in longer orbits of high inclination to the ecliptic, where R and p, the radius and density, respec- might survive long enough to spiral into the sun. tively, of the meteoroid, are measured in cgs Note that high-inclination orbits are always units. favored because of the observed concentration of Now the Poynting-Robertson effect (Wyatt dust to the ecliptic (really Jupiter's orbit and Whipple, 1950) will cause a particle to plane). Note also, however, that these smaller spiral from an initial orbit into the sun in a particles are already counted in the diffraction time theory for the origin of the Fraunhofer Corona. 7 t=RPKX10 years, (18) As the exponent a of R in the relation N(R)~R~a becomes greater than 4, where the where K depends upon the initial eccentricity increase in mass per logarithmic step in R or n and perihelion distance of the orbit, and is of is constant, only the smallest particles allowed the order of unity for short-period comet orbits. by light pressure survive long enough to spiral Typical values of .EL are: 0.14 for the Geminid into the sun. The total fraction of the mass in meteor shower; 0.6 for the Taurids and Encke's these small particles, however, becomes larger comet; 2.8 for the Bielids; 7.2 for the Leonids; as a increases. If high-energy ions contribute and 18.6 for the Lyrids. We see that the appreciably to space erosion, however, small 2 Poynting-Robertson effect acts at a rate 10~ particles may fare no better than large ones 4 to 10~ times as fast as the estimated erosion for with regard to the Poynting-Robertson effect typical meteor showers. Thus, if we take the (see, for example, Whipple, 1959; and Reiffel, erosion rate as fixed for particles of all sizes, the 1960). Poynting-Robertson effect can be neglected for On the other hand, if the small cometary all cometary meteoroids, but could possibly be meteoroids are stronger and tougher than are effective for some of the asteroidal particles (see the larger ones, erosion may be less serious for table 2), such as irons in highly inclined orbits. them. An increase by a factor of 100 in com- Even for stones, however, the effect could pressive strength would increase the limiting probably be neglected. particle size by a factor of 10 and the limiting At least two other factors must be considered mass by a factor of 1,000. as we apply the above reasoning to smaller We may conclude, therefore, that the data particles: (a) erosion changes to destructive and assumptions of the present paper would collision when the eroding mass is the order of 3 preclude most of the small particles from 10~ times the mass of the particle eroded or attaining low inclination, nearly circular orbits, greater; and (6) the physical properties of unless they are much stronger physically than cometary meteoroids may depend upon their the average photographic meteoroids. Higher size. inclination orbits would somewhat favor Poynt- Thus, as we consider erosion of smaller ing-Robertson reduction in orbit dimensions particles, the upper limit of mass for the for all particles. Unfortunately, we have no eroding particles must be reduced proportionally satisfactory information about the physical or to the mass of the particle eroded. The cross chemical nature of the small particles (below 2 section for destructive collision by larger M~10- gm). METEOR SYMPOSIUM—WHIPPLE 247 Dust as a hazard to space vehicles should be protected on the moon, and "tough" Penetration rates for exposed surfaces in space material should be chosen insofar as is practica- near the ecliptic between Venus and Mars well ble for exposed surfaces whose function could away from the earth should probably be con- be impaired by etching. siderably smaller than the values given by Note that the continued operation of Van- the author in 1957. The mean space density guard I for nearly three years does not guarantee near the earth was taken as 4X10"20 gm cm"3 that the protecting windows for the solar cells as compared to the values derived here, in the are not severely etched. Even severe etching range from 4X10"22 to ~1XK)-20 gm cm"3 of a glass surface does not necessarily reduce near the earth's orbit. Bjork (1961a) finds much of its transparency to light. slightly smaller penetrations for given masses Acknowledgment and velocities, while possible low densities of the particles themselves may reduce their pen- I am indebted to Dr. Edward Anders for pre- etration powers. Lower average velocities also publication material, and especially to Dr. will apply, especially towards Mars, again Edward L. Fireman for helpful advice and reducing the penetrating power. discussion. In the immediate neighborhood of the earth References we still have no reliable information on the distribution of particles in the mass range from BIRKHOFF, G.; MACDOUGALL, D. P.; PUGH, E. M.: 5 3 AND TAYLOR, G. 10~ to 10~ gm. Extrapolations from Mc- 1948. Explosives with lined cavities. Journ. Cracken et al. suggest that my 1957 estimates Appl. Phys., vol. 19, pp. 563-582. may be too large. On the other hand, the BJORK, R. L. rough agreement of values of mean space 1961a. Meteoroids vs. space vehicles. Journ. Amer. Rocket Soc, vol. 31, pp. 803-807. density derived from etching rates and from 1961b. Analysis of the formation of Meteor Cra- the Zodiacal Cloud calculations suggests that ter, Arizona: A preliminary report. most of the mass (or mean density) in space is Journ. Geophys. Res., vol. 66, pp. 3379- carried by particles intermediate in size between 3387. the finest dust and optical meteoroids; i.e., in BuDDHUE, J. D. 7 3 1942. The compressive strength of meteorites. the range 10~ to 10~ gm. Hence my 1957 Pop. Astron., vol. 50, pp. 390-391. estimates may, in fact, be realistic values rather CHARTERS, A. C. than safe upper limits near the earth. 1960. High-speed impact. Sci. American, vol. The etching rates for exposed surfaces can 203, no. 4, pp. 128-140. be determined with some confidence from table DUBIN, M. 1961. Remarks on the article by A. R. Hibbs, 2 as safe limits near the ecliptic between Venus "The distribution of micrometeorites and Mars, but not in the immediate vicinity of near the earth." Journ. Geophys. Res., the earth. McCracken et al. have produced vol. 66, pp. 2592-2594. almost incontestable evidence that a high con- EBERHARDT, P., AND HESS, D. C. centration of fine dust exists near the earth. 1960. Helium in stone meteorites. Astrophys. Journ., vol. 131, pp. 38-46. Thus etching rates within the moon's distance FIREMAN, E. L. may be a few times greater, perhaps even orders 1958. Argon-39 in the Sikhote-Alin meteorite fall. of magnitude greater, than at much larger Nature, vol. 181, pp. 1613-1614. distances from the earth, but still negligible FIREMAN, E. L., AND DEFELICE, J. for metals as tough as aluminum. 1960. Argon-39 and tritium in meteorites. Geo- chim. et Cosmochim. Acta, vol. 18, Exposed surfaces of weak and brittle mate- pp. 183-192. rials should be avoided in space vehicles where 1961. Tritium, argon 37, and argon 39 in the an etching rate of 1 micron a year would be Bruderheim meteorite. Journ. Geophys. undesirable. On the moon, the etching (and Res., vol. 66, pp. 3547-3551. possibly even penetration) may be aggravated FISHER, D. E. 1961. Space erosion of the Grant meteorite. by material ejected from meteoritic impacts. Journ. Geophys. Res., vol. 66, pp. When not in use, transparent or delicate surfaces 1509-1511. 248 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS GLASSTONE, S., ED. 1958. Meteor impact on solid surface. Irish 1957. The effects of nuclear weapons. U.S. Astron. Journ., vol. 5, pp. 14-33. Atomic Energy Commission, Washing- REIFTEL, L. ton, D.C. 1960. Erosion of meteorites in space and the HIBBS, A. R. density of interplanetary gas. Nature, 1961a. The distribution of micrometeorites near vol. 185, pp. 821-823. the earth. Journ. Geophys. Res., vol. SHOEMAKER, E. M. 66, pp. 371-377. 1960. Penetration mechanics of high velocity 1961b. Author's reply to the preceding discussion meteorites, illustrated by Meteor Crater, on the article, "The distribution of Arizona. In Report Int. Geol. Congr., micrometeorites near the earth." Journ. 21st Session, Norden, 1960, Part 18, Geophys. Res., vol. 66, pp. 2595-2596. pp. 418-434. VAN DE HULST, H. C. SUMMERS, J. L., AND CHARTERS, A. C. 1947. Zodiacal light in the solar corona. Astro- 1959. High-speed impact of metal projectiles in phys. Journ., vol. 105, pp. 471-488. targets of various materials. In Proc. INGBAM, M. F. Third Symposium on Hypervelocity Im- 1961. Observations of the zodiacal light from a pact, F. Genevese, ed., pp. 101-110. very high altitude station. IV. The Armour Research Foundation, Chicago. nature and distribution of the inter- WHIPPLE, F. L. planetary dust. Monthly Notices Roy. 1955. On the mass-luminosity relation for Astron. Soc, vol. 122, pp. 157-176. meteors. Astron. Journ., vol. 60, pp. JACCHIA, L. G., AND WHIPPLE, F. L. 182-183. 1961. Precision orbits of 413 photographic 1957. The meteoritic risk to space vehicles. In meteors. Smithsonian Contr. Astrophys., Proc. 8th Int. Astronautical Congr., vol. 4, no. 4, pp. 97-129. Barcelona, 1957, pp. 418-428. Springer- MCCBACKEN, C. W.; ALEXANDER, W. M.; AND DUBIN, Verlag, Vienna, 1958. M. 1959. Solid particles in the solar system. Journ. 1961. Direct measurement of interplanetary dust Geophys. Res., vol. 64, pp. 1653-1664. particles in the vicinity of Earth. Na- 1961a. The dust cloud about the earth. Nature, ture, vol. 192, pp. 441-442. vol. 189, pp. 127-128. 1961b. Particulate contents of space. In Medical MCCBOSKT, R. E. and biological aspects of the energies of 1955. Fragmentation of faint meteors. Astron. space, P. Campbell, ed., pp. 49-70. Journ., vol. 60, p. 170. Columbia University Press, New York. NAZAKOVA, T. N. WHIPPLE, F. L., AND FIREMAN, E. L. 1961. Investigation of meteoric particles on the 1959. Calculation of erosion in space from the third Soviet artificial earth satellite. cosmic-ray exposure ages of meteorites. Planetary Space Sci., vol. 8, pp. 82-85. Nature, vol. 183, p. 1315. OPIK, E. J. WYATT, S. P., JR., AND WHIPPLE, F. L. 1955. The masses and structure of meteors. In 1950. The Poynting-Robertson effect on meteor Meteors, T. R. Kaiser, ed., pp. 33-35. orbits. Astrophys. Journ., vol. Ill, pp. Pergamon Press, New York. 134-141. Abstract Photographic meteor studies and spallation products of cosmic rays in meteorites provide upper limits or evalua- tions of erosion rates for meteoroids in orbit. The rate increases by several orders of magnitude in the sequence: irons, stones, and photographic meteoroids. The sequence suggests that the erosion rate depends in some inverse fashion upon the strength or brittleness of the materials, a natural expectation if erosion is produced by crater-forming impacts with interplanetary dust. On this assumption and by applying Opik's high-velocity impact theory, the author derives a mean space density for the dust at roughly 10~21 gm cm~*. Only the very smallest cometary meteoroids can spiral into the sun by the Poynting-Robertson effect if the as- sumptions of this paper are true. Strong evidence for a high concentration of dust in the earth's immediate vicinity suggests that erosion rates may be greater near the earth than in deep space from Venus to Mars and even beyond. The surface of space vehicles should be carefully chosen or protected to minimize etching by interplanetary dust, if long exposure times are anticipated. The hazard may be greatest at the moon's surface. The Absence of Magnetic Micropulsations of Meteor Origin By C. Ellyett1 and G. B. Gillion' Kalashnikov (1949) claimed to have found only in very limited parts of the year that a an increase in the occurrence of micropulsations second peak of shower activity occurs around of the earth's magnetic field during the dates 8 to 10 a.m. Second, the association of the two of certain northern hemisphere meteor showers. phenomena, which was claimed to exist, was Pulsations of Amy amplitude could be recorded. based on low-rate meteor observations (Whipple The measured band of micropulsation frequen- and Hawkins, 1958). If the percentage occur- cies unfortunately extended upwards beyond rence of micropulsation activity is to increase 2 c/s, and hence into the region of sferics of appreciably, an accompanying large increase thunderstorm origin. Campbell (1960) has of subvisual meteors is implied. Observations reviewed this work, and concludes that the at high radar-meteor rates, however, have now validity of the results is increased by the fact shown that major showers comprise a relatively that the observations were made in November, small proportion of the total count (Whipple December, and January—a period of mini- and Hawkins, 1958; Ellyett and Keay, 1956). mum activity in the annual cycle of thunder- Finally, Jenkins, Phillips, and Maple (1960) storm activity. observed micropulsations at 1.5 c/s, with an Hawkins (1958) repeated Kalashnikov's in- improved sensitivity of 0.3m7. They claim vestigations using apparatus of about the same an occasional big increase in the occurrence sensitivity, and found no significant increase. of these micropulsations on some nights, His equipment, however, covered the band 0.5 correlating with meteor showers. The graphical to 10.0 c/s, and hence could also have recorded correlation does not look convincing, and in sferics. any case the meteor data used are the Harvard Further work was done by Campbell (1960), visual meteor observations taken during the who carried out a 7-month series of micropulsa- first half of the century. These are again tion measurements in California in 1958, with low-rate observations. a limiting sensitivity of 20mY. The periods On general grounds, there is little in the recorded lay between 0.04 and 0.4 c/s. An diurnal distribution of micropulsations that average amplitude was obtained for each hour. fits the picture of meteor occurrence. Con- Normalizing factors ranging from 1.02 to 33.3 tinuous micropulsations (Pc), trains (Pt), giant for various hours of the day, compared with the pulsations (Pg), and long-period fluctuations rate at 10 a.m. L.T. as reference, were applied to (LPc) all increase in amplitude toward the determine an average daily value. Departures auroral regions. Observations of Pc show a in micropulsation amplitude from this daily 27-day solar period, and both Pt and Pg are average were then compared with the occur- essentially nighttime phenomena. The possible rence of the meteor showers. This procedure exception is "pearls" at about 1 c/s, but these would appear to be doubtful in two respects. on occasion last fairly regularly for many hours. First, undue weight in scaling within the day If a single meteor produces a magnetic appears to be given to the 10 a.m. micropulsa- pulse or train of pulses—as it must if the tion amplitudes. Meteor shower occurrence suggestion of meteor origin is correct—then normally peaks well before 6 a.m. L.T., and it is it is obviously desirable that both sets of phenomena be observed at the same site. 1 Physics Department, University of Canterbury, Christchureh, New Zealand. This has not previously been done. 249 250 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Meteor equipment minimum readable input level of 0.04 /*V, corresponding to a magnetic field charge of A radar transmitter maintained a signal of 8 81 kw peak power and 26 M pulse width at 0.3mY (=0.3X10" gauss). The sensitivity, 69.5 Mc/s, with a pulse recurrence frequency and the frequency filtering at 1.5 c/s, was thus of 150 c/s. Single crossed folded dipoles purposely made the same as that of Jenkins, above wire-mesh artificial earth screens were Phillips, and Maple (1960). (There is no used for transmitting and receiving, and reason why the particular micropulsation fre- resulted in all-sky viewing. The receiver, quency of 1.5 c/s should be associated with followed by normal oscillograph and camera meteors. Two advantages of this frequency recording, contained design features that elimi- are that it is below sferics interference, and is nated false meteor rate variations due to both in a region of high loop sensitivity.) The slow variations of the background noise recording paper passed through the pen recorder (MacLauchlan, 1960) and pulse-type inter- at a rate of 3 inches per minute, which just ference. Meteors were recorded down to a allowed 1.5 c/s to be resolved. magnitude of +9.1. Results Micropulsation equipment A period of 24 hours of continuous observing Magnetic flux variations were detected by a commenced at 6 p.m. L.T. on 29 July 1961. two-turn buried loop, with an effective area Both equipments were looking at approximately of 1.6X10* square meters, leading to a gal- the same sky area. vanometer. Light from the undisturbed gal- During this period the 5-Aquarid shower vanometer mirror fell centrally on two photocells (a=340°, 8= —19°), which is the strongest placed side by side. Deflection of the mirror known meteor shower to occur in either hemi- by an incoming signal altered the photocell sphere (Ellyett, Keay, Roth, and Bennett, current, which was then amplified by a triode 1961; Keay and Ellyett, 1961), exerted its that formed one arm of a bridge. A twin-T peak effect at about 3 a.m. L.T. and was above rejection filter, tuned to 1.5 c/s, supplied the horizon for several hours centered on this negative feedback. The overall gain of the time. detection circuit, including an amplifier in The time of occurrence of each individual the pen-recording unit, was 4X108, giving a meteor was read to an accuracy of ±2 seconds. The micropulsation and the meteor time scales 3OO . were locked together, and a correlation was considered positive if the two phenomena 25O _ . DO commenced within 4 seconds of each other. A total of 2,750 meteors was recorded during the 24 hours. aoo . _ 80 First, large-amplitude low-range meteors, with durations in many cases of up to 30 ISO . seconds, were individually compared with the micropulsation trace. There was an obvious absence of any correlation. The number of half-minutes in each hour when micropulsations at 1.5 c/s were present was then plotted against the total hourly meteor count. The micropulsation distribution was as expected, with a noon dip. As shown I I I I I I I I I I I I I in figure 1, there is no correlation between the 00 two rates of occurrence. FIGURE 1.—Number of meteors/hr compared with the number of %-mins/hr containing 1.5 c/s micropulsations. Observa- It would thus appear fairly conclusive that tions commenced at 6 p.m., local time, on 29 July 1961. no correlation exists within the present limits Heavy line, meteors; dotted line, micropulsations. of sensitivity. Further work is proceeding. METEOR SYMPOSIUM—ELLYETT AND GH.LION 251 Acknowledgments HAWKINS, G. S. 1958. A search for magnetic effects from meteors. This work has received contract assistance Journ. Geophys. Res., vol. 63, pp. from the Electronics Research Directorate of 467-475. the Air Research and Development Command, JENKINS, A. W.; PHILLIPS, C. A.; AND MAPLE, E. U.S. Air Force, under contract AF64(500)-6 1960. Observed magnetic effects from meteors. (Meteors); and from the Office of Naval Re- Journ. Geophys. Res., vol. 65, pp. search, U.S. Navy, under contract Nonr(G)- 1617-1619. KALASHNIKOV, A. G. 00027-62 (Micropulsations). 1949. On observations of the magnetic effect of meteors by the inductive method. Dokl. References Acad. Sci. U.S.S.R., vol. 66, pp. 373-376. CAMPBELL, W. H. (Translated by E. R. Hope, Canadian 1960. Magnetic micropulsations accompanying Def. Res. Board, T-130-R, 1954.) meteor activity. Joum. Geophys. Res., KEAY, C. S. L., AND ELLYETT, C. D. vol. 65, pp. 2241-2245. 1961. The latitude dependence of radar meteor ELLTETT, C. D., AND KEAT, C. S. L. shower observations. Journ. Geophys. 1956. Radio echo observations of meteor activity Res., vol. 66, pp. 2337-2343. in the southern hemisphere. Australian MACLAUCHLAN, E. C. Journ. Phys., vol. 9, pp. 471-480. 1960. Noise suppression in pulse receivers. ELLYETT, C. D; KEAT, C. S. L.; ROTH, K. W.; AND Australian Journ. Phys., vol. 13, pp. BENNETT, R. G. T. 750-752. 1961. The identification of meteor showers with WHIPPLE, F. L., AND HAWKINS, G. S. application to southern hemisphere re- 1958. Meteors. In Handbuch der Physik, S. sults. Monthly Notices Roy. Astron. Flugge, ed., vol. 52, pp. 519-564. Soc, vol. 123, pp. 37-50. Springer-Verlag, Berlin. Abstract Simultaneous observations of micropulsations greater than 0.3m-y at 1.5 c/s, and of meteors greater than magni- tude + 9.1, at the same site during the peak of occurrence of the 5-Aquarid meteor shower, failed to show any corre- lation, either with individual meteors, or on an hourly rate basis. The Density Distribution of Telescopic Meteors Around the Earth's Orbit By L. Kresak and M. Kresakova * Recent studies have provided substantial vehicle is greater for smaller meteors, approxi- evidence that the rate of meteoric particles en- mately by a factor of ten for each two magni- countered by the earth varies considerably, tudes. For the telescopic meteors recorded owing both to the occurrence of recognized during the Skalnate Pleso program, the average meteor showers and to the nonuniform density probability of collision is estimated at about distribution of the sporadic background. These 10~8 m~2 day"1; this is not a negligible figure even irregularities, appearing periodically in the form for the space vehicles and durations of flight of annual variation of meteor rates, are of par- expected in the near future. ticular interest in connection with the constitu- Supplementing the results of earlier studies tion of minor meteor streams at advanced (Kresakova and Kresak, 1955), we report here stages of dispersion, and with the meteoric risk on a total of 4,573 meteors recorded during to space vehicles, especially those moving in 1,364 hours of observation more or less uni- satellite orbits around the earth or launched to formly distributed over the various dates and the moon. The evaluation of the annual varia- hours of the night. All observations were made tion is made difficult by the variability of the geometrical conditions depending on geographic with the same type of instrument, 4-inch binoc- latitude and by the additional effect of diurnal ulars with a magnification of 25. variation produced by the rotation of the earth. The net annual variation produced by the Another difficulty in the interpretation of changes of meteor density around the orbit of observed rates lies in the variability of the the earth may be established only after elimi- velocity distribution; the different sensitivities nating the effect of nonuniform radiant distri- of individual observing techniques to meteors bution on the visible and invisible hemisphere. of different velocities complicate this problem. This may be done by separating the annual and diurnal variation with the use of a reference The main purpose of this paper is to analyze model of apparent radiant distribution. In the the distribution of meteor orbits around that of present work we used three such models: the earth, on the basis of telescopic observations made at the Skalnate" Pleso Observatory for the Model A, the uniform apparent radiant distri- period 1946-1959. These include faint tele- bution, giving constant hourly rates; Model scopic meteors down to the 10th apparent B, an isotropic distribution with respect to the magnitude; more than half of the data are con- apex, with a continuous decline of radiant den- cerned with meteors of 8th and 9th magnitude. sity towards the antapex (Hoffmeister, 1931) These faint meteors are of particular signifi- (the "equivalent velocity" constant (c=3.9) cance for two reasons: (1) In the constitution of was computed directly from our observations); meteor streams they may reveal evolutionary and Model C, the standard distribution derived changes caused by the effects of solar radiation, by Hawkins (1956) from the radio-echo data. which are substantially greater on smaller parti- This distribution is characterized by a pro- cles; (2) the likelihood of a collision with a space nounced ecliptical concentration, two main maxima near the sun and antihelion point, and 1 Astronomical Institute of the Slovak Academy of Sciences, Bratislava, a secondary maximum at the apex. Czechoslovakia. 253 254 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS For each of the models the expected hourly tice, 1957), is, surprisingly, less consistent with rates for any instant may be predicted and the telescopic observations than is model B. compared with the observation. An overall It appears as if there were an overcorrection comparison is shown in figure 1, where the of the observed variation (a mirror-like image abscissae indicate the solar longitudes and the of model A); the radiant distribution suggested ordinates the time in M.E.T. In this network by Hawkins would fit better with a diminished of coordinates we plotted curves of constant density gradient. This may be attributed either reduced hourly rates obtained by combining to the selective effect of angular velocities the observations into normal places, interpo- applying to telescopic observations, or to an lating, and smoothing the results. For a per- increasing preponderance of direct orbits with fect representation of the diurnal variation, the decreasing size of particles. It may be noted curves should change into vertical straight lines; that for model B this effect was already indi- for a perfect representation of the annual varia- rectly taken into account by determining the tion, they should change into horizontal straight basic constant ("equivalent velocity") from the lines; and for an entire agreement between telescopic data; a lower degree of apical con- theory and observation they should vanish. centration than indicated by the naked-eye Obviously, sj7stematic departures may be an- observations was also found. ticipated in the annual rather than in the diur- It is clear that none of the models of radiant nal variation, and the vanishing of the diurnal distribution can account for the annual vari- variation alone would offer convincing proof of ation in terms of the earth's rotation combined a nonuniform distribution of meteors around with the obliquity of the ecliptic. Evidently, the orbit of the earth. such features as a rapid increase of meteor Figure 1 indicates that model A is distinctly rates in June and July, or a decrease in Decem- inferior to models B and C. A systematic in- ber and January cannot be explained by any crease of meteor rates from the evening to the other acceptable model of radiant distribution. morning hours and a sharp maximum in the Although the trend of diurnal variation in autumn reveal the main features of diurnal and model A is entirely reversed in model C, never- annual variations, established long ago in the theless the annual variation preserves its main naked-eye range. In model B the total ampli- features unchanged. Hence the actual varia- tude is essentially reduced, and the curves, as tions of meteor density around the earth's orbit expected, run mainly in the vertical direction. are established beyond doubt. Model C, which represents satisfactorily the The course of the net annual variation, freed radio observations (Hawkins, 1956) and the from the effects of the earth's orientation ac- naked-eye observations (Hawkins and Pren- cording to the three models, is shown in figure 180 270 360 \(X60 0.60/ 360 90 180 FIGURE 1.—The combined pattern of annual and diurnal variation referred to models A, B, C. The first solid line corresponds to the median hourly rate, the next solid lines to 1.20, 1.40, etc., and the broken lines to 0.80, 0.60 of the median value. Abscissae, solar longitudes; ordinates, time M.E.T. METEOR SYMPOSIUM—KRESAK AND KRESAKOVA 255 2. Here the amplitudes of the annual variation groups according as their periods were less for models B and C are 1:1.8 and 1:2.2, respec- than or greater than 500 years. With regard tively. Despite considerable differences be- to the periods of recognized meteor showers, a tween these two models, the curves are similar correlation appears possible for the former enough, indicating in each case a lower level of group but highly improbable for the latter. activity in the first half of the year (©=280° The stream associated with each of the comets to 100°) and a higher level in the second half was considered as possibly active within 30° (0 = 100° to 280°). Even the three maxima of solar longitude before and after the closest near © = 130°, 190°, and 250° are consistently approach for the former group of comets reproduced in both types of representation. (regularly of low inclination), and within 15° for the latter group. The results are shown in the lower part of figure 2, where we see that the density distribution Dl of short-periodic com- etary orbits agrees well with the density distribution of telescopic meteors. As expected, the comets moving in very prolonged orbits (D2) exhibit no correlation at all, and this excludes any alternative explanation of the agreement in terms of seasonal selective effects of the discoveries of comets on the distribution of their nodes. Consequently, there is strong evidence that at least some of the irregularities of the annual variation are due to minor meteor streams of cometary origin, with the parent comets known, but too dispersed to be dis- tinguished separately. Analyses of the annual variation determined by radio techniques have been published by Hawkins (1956) for Jodrell Bank, and by Weiss (1957) for Adelaide. The net variation determined by Hawkins is based on model C, and is directly comparable with our results. The model chosen by Weiss is different. It assumes a line source round the ecliptic with an apparent strength up to the distance of 90° FIGURE 2.—The net annual variation referred to models A, from the apex four times higher than in the B, C, compared with the distribution of comet approaches Dl (P<500years) and D2 (P>500years). Abscissae- remaining part of the ecliptic. Experience solar longitudes. with different models, however, including those used by Weiss, indicates that the result will We have attempted to test the source of the not differ greatly from those based on model C. irregularities of the annual variation by means A comparison of the annual variation deter- of cometary statistics. For the reasons pointed mined at Adelaide, Jodrell Bank, and Skalnate" out by Levin (1961), Hoffmeister's (1948, p. Pleso is shown in figure 3. A previous study 100) cometary indices are not suitable for this (Kresakova and Kresak, 1955) demonstrated purpose, i.e., for taking the periodic comets that the annual variation of the telescopic together with the non-periodic comets that meteors obtained at Skalnate" Pleso agrees well appeared in a relatively short time-interval. with the results of Hoffmeister (1937), derived Instead, we used as a reference catalogue from the naked-eye observations of Schmidt, Kramer's (1953) extensive list of cometary Heybrock, and himself. Hence it may be orbits approaching the earth within 0.3 A.U. concluded that the difference in the magnitude The comets of this list were divided into two range covered by the different methods of 256 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS striking feature of the first of these curves is its symmetry with respect to the summer sol- stice. This feature suggests some seasonal effect producing a correlation between the dura- tion of the daily illumination by the sun and the ratio of ionizing to luminous efficiency of meteors. Such an explanation would require an inverse trend of the annual variation in the southern hemisphere, where the seasonal shift is in six-month periods. Some indications of such inversion actually appear in the figure, but they are not sufficiently conclusive. If, however, we take into account the annual variation of naked-eye meteors, determined for FIGURE 3.—The comparison of the net annual variation obtained at Adelaide (dotted line), Jodrell Bank (broken the southern latitudes by Mclntosh (1934) and line) and Skalnate Pleso (solid line). Abscissae—solar by Hoffmeister (1937, p. 60), with a deep longitudes. minimum in January and February, the inver- sion of the curve becomes more distinct. There observation cannot represent the essential is a considerable difference between the ampli- source of the discrepancies. These discrepancies tudes of the two curves in figure 4. In fact, may result either from the difference in geo- the explanation suggested requires a difference graphic latitudes or from the difference in the of this kind, since the conditions of illumination type of technique used for the determination of are much more variable during the year at meteor rates. Jodrell Bank (<£=+53°.4) than at Adelaide It is surprising that our telescopic observa- (0= —34°.9). At Jodrell Bank the atmos- tions agree much more closely with the obser- pheric levels in which meteors appear are almost vations at Adelaide, in the southern hemisphere, continuously illuminated at the time of the than with those at Jodrell Bank, located in summer solstice. If this explanation is valid, approximately the same geographic latitude as it would partially account for the abundance Skalnate Pleso. From this fact it follows that of the summer daytime meteor showers detected the main discrepancies are not due to the actual only by radio techniques. Since all these distribution of the radiants in declination. streams encounter the earth at about the time The difference between the annual variation of the summer solstice, they should be relatively of radio-echo and that of photographic rates was pointed out by Hawkins and Prentice -2.0 (1957), who attribute it to the fact that the sensitivities of the two methods of observation vary with the geocentric velocities of the -1.5 meteors. This interpretation requires a con- centration of high-velocity meteors in the sector of the earth's orbit traversed in the summer months. However, direct measure- ments of meteor velocities, e.g., those by McKinley (1951), show, on the contrary, that there is only an average number of high- velocity meteors in June and July, and a maximum in September and October, consistent with the extreme declination of the apex. Figure 4 shows the annual variation of the FIGURE 4.—The annual variation of the ratio of reduced hourly rates Adelaide: Skalnate Pleso (broken line), and ratios of the hourly rates at Jodrell Bank and Jodrell Bank: Skalnate Pleso (solid line). Abscissae— at Adelaide to those at Skalnate Pleso. A solar longitudes. METEOR SYMPOSIUM—KRESAK AND KRESAKOVA 257 much weaker if they could be observed optically. radiation on the evolution of meteor streams It is important to emphasize that the basic (Kresak, 1960). data of Hawkins and Weiss were obtained at In addition to the showers known from other substantially different wavelengths (4.2 and techniques of observation, eight showers of 11.2 m, respectively), but the effect of this telescopic meteors have been established with difference is difficult to determine. a sufficient degree of probability. They ap- From the similarity of the curves of annual peared at solar longitudes 25°, 131°, 158°, variation at Adelaide and Skalnate Pleso, 184°, 194°, 248°, and 316°, most of them with despite their difference of 84° in geographic out any indication of periodicity. Only rarely latitude, we deduce that the mam contribution did their hourly rates exceed those of the spo- of the annual variation is due to meteor radiants radic background by a factor of two; striking of low declination. This conclusion offers an phenomena, comparable to the Perseids or additional proof of the ecliptical concentration Geminids in the naked-eye range, are absent at of radiants produced by a significant prevalence 7th to 9th magnitudes. In general, it may be of direct meteor orbits of low inclination and concluded that the phenomenon of meteor short period. An earlier discussion of the streaming becomes gradually less prominent telescopic meteor directions (Kresak, 1955) as we pass from photographic through naked- led to the same conclusion. The correlation eye to telescopic meteors. This phenomenon between the annual variation and the distribu- probably reappears again, to a much greater tion of the points of approach of short-periodic extent, in the range of much smaller particles cometary orbits (fig.2) requires an ecliptical for which rapid variations in frequency have concentration as well. Thus three entirely been established by the impact measurements independent types of evidence suggest that in on high-altitude rockets and artificial satellites. the range of faint telescopic meteors the orbits Among faint telescopic meteors there is rather a are distinctly concentrated toward the plane system of minor, considerably dispersed streams of the ecliptic. that do not produce abrupt changes in meteor We shall now comment on the occurrence of rates over a period of a few days, but that do shower meteors in the telescopic observations. influence the total annual variation. The permanent showers, which at their maxima As for the meteoric risk to space vehicles moving in the vicinity of the earth's orbit, we contribute a majority of the meteors observed, conclude primarily that: are hardly detectable in the range of tele- (1) At the level of the probability of en- scopic meteors of 7th to 9th magnitudes. In counter 10~6 m~2 day"1, the risk from the the total for the whole year they were at least sporadic background varies with a maximum seven times less frequent with respect to the seasonal amplitude of about 1:2, the safest sporadic background than in the naked-eye period of permanently low activity lasting from observations. This lack of faint shower- January till May. The periods of highest meteors is consistent with the recent results activity occur in August, September, the first obtained by Kresakova (unpublished), who, half of October, and in December. from naked-eye observations at SkalnatS Pleso, (2) The particles producing this activity are found a systematic difference between the very likely of cometary origin and constitution, magnitude functions of shower meteors and of as indicated by the relation of meteor rates to sporadic meteors, with the constant K=2.6 the distribution of nodes of short periodic and 3.5, respectively. An extrapolation would coraetary orbits. require the ratio of shower to sporadic meteors (3) The permanent meteor showers represent to be four times lower in the present telescopic a risk that is negligible in most cases, and is observations than in those made by the naked generally much lower than might be anticipated eye. The actual relative decrease in the num- from photographic and naked-eye observations ber of shower meteors appears even sharper, of larger bodies, or from impact measurements presumably because of the effect of corpuscular of smaller bodies. 258 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS (4) Showers rich in telescopic meteors are KRAMER, E. N. scarce and without distinct periodicity. Only 1953. Cometary radiants and the association of quite exceptionally do they increase the risk meteor streams with comets. Izv. by a factor of two or three. It must be Astron. Obs. Odessa, pp. 163-247. remembered, however, that this result is de- KRESAK, L. rived from night-time observations, in which 1955. The direction distribution of telescopic meteor showers with apparent radiants near meteors. Contr. Astron. Obs. Skalnat6 the sun could not be detected. Pleso, vol. 1, pp. 9-35. 1960. The effect of solar corpuscular emission on the magnitude distribution of meteors. Acknowledgments Bull. Astron. Inst. Czechoslovakia, vol. The authors are greatly indebted to all mem- 11, pp. 1-9. bers of the staff of the Skalnate Pleso Observa- KRESAKOVI, M., and KRESAK, L. tory who took part in the observations analyzed 1955. On the activity of telescopic meteors and here, especially to the most active observers: some related problems. Contr. Astron. Dr. L. Padjusakova, Mr. A. Mrkos, Dr. A. Obs. Skalnate' Pleso, vol. 1, pp. 40-77. Becvaf, Mr. M. Dzubak, and Mrs. A. Antalova. LEVIN, B. J. Our thanks are due to Mrs. L. Durkovidova 1961. The distribution of true radiants of meteor bodies down to a definite limit of mass. and Mr. A. Aldor for their aid in numerical Symposia at the 5th meeting of CSAGI, computations. Moscow, 1958. In Annals of the Inter- national Geophysical Year, vol. 11, References pp. 187-193. Published by Pergamon HAWKINS, G. S. Press. 1956. Variations in the occurrence rate of meteors. MCINTOSH, R. A. Astron. Journ., vol. 61, pp. 386-391. 1934. Report of the section for the observation HAWKINS, G. S., and PRENTICE, J. P. M. of meteors, 1929-1931. New Zealand 1957. Visual determination of the radiant distri- Astron. Soc. Bull., no. 21. bution of sporadic meteors. Astron. MCKINLET, D. W. R. Journ., vol. 62, pp. 234-240. 1951. Meteor velocities determined by radio ob- HOFFMEISTER, C. servations. Astrophys. Journ., vol. 113, 1931. Zur Theorie der Variation der Sternschnup- pp.225-267. penhaufigkeit. Veroff. Univ. Berlin- Babclsberg, vol. 9, no. 1, p. 17. WEISS, A. A. 1937. Die Meteore. Akademische Verlagsgesell- 1957. The distribution of the orbits of sporadic schaft, Leipzig. meteors. Australian Journ. Phys., vol. 1948. Meteorstrome. J. Barth Verlag, Leipzig. 10, pp. 77-102. Atmospheric Trajectories of Telescopic Meteors Lubos Kohoutek* and Jiri Grygar* Double-station observations of telescopic teors. Next it was found that increase of the meteors were organized at the amateur meteor mean height with magnitude, if it exists at all, expedition to Mount Bezovec (\=17O58' E. is very slight in the magnitude interval from Gr., Statistics of Meteor Streams By R. B. Southworth 1 and G. S. Hawkins In the period 1954 to 1957 a random sample and dates of occurrence—was satisfactory. comprising 360 of the meteors photographed It became clear that a comprehensive quantita- from two stations by the Baker Super-Schmidt tive criterion was required that would embrace meteor cameras were reduced under the "Short all the elements of the orbit. Trail" program (Hawkins and Southworth, 1958). The radiants and velocities of 359 A criterion for stream membership meteors were sufficiently accurate to permit We may regard a meteor stream as composed computation of dependable orbits (Hawkins of those meteors originating from a single and Southworth, 1961). Since this is the only comet during a single ejection or over an random sample of photographic meteors with interval of time. As yet, we cannot reliably individually reliable orbits, it provides valuable identify stream membership by physical or material for a study of meteor distribution in chemical characteristics, although some physi- space. The degree to which the meteors are cal differences have been observed between grouped into streams is a basic feature of this different streams (Jacchia, 1960). Meteors distribution. arising from a comet will move in orbits similar The major meteor streams are conspicuous in to the comet's orbit, at least for a while. If any random sample of photographic meteors planetary perturbations or other forces dis- and virtually all of their members are readily perse the meteors so that the similarity of identifiable. Since in the sample there are their orbits can no longer be seen, the stream many meteors that belong to each such stream, is effectively destroyed. Thus, in practice, a the characteristics of the stream are well de- meteor stream is recognized and defined as a fined, as is the range of each characteristic group of meteors in similar orbits. within the stream. Furthermore, these streams Occasionally meteors from two or more are well-known from other studies. comets may have similar orbits. If the dif- Members of a stream not previously known, ferent origins cannot be recognized from the or not well represented in the sample under orbits, then the meteors form a single stream study, are not as easy to identify. Early in in practice, although arising from more than the Short Trail program it became apparent one comet. that some of the meteors were associated into A quantitative criterion for stream member- other groups or streams, apart from the pre- ship therefore requires a quantitative measure viously known major streams. Numerous of orbit similarity (or difference). We shall other meteors were doubtfully associated with consider the difference between orbits to be the streams. Several attempts, based upon a difference between meteors. Let D(A, B) be variety of principles, were made to classify such a quantitative measure of the difference the meteors into streams and a "sporadic" between meteors A and B. Similarly, let remainder. None of the classifications based D(M,N) be of the same mathematical form, but on geocentric quantities—radiants, velocities, let it measure the difference between a mean orbit M of a stream and an individual meteor N. 1 Smithsonian Astrophysical Observatory, and Harvard College The two measures D(A, B) and D(M, N) corre- Observatory, Cambridge, Mass. spond to two possible ways of defining a stream. 'Harvard College Observatory, Cambridge, Mass., and Boston University, Boston, Mass. (1) A stream may be defined, for example, as 261 262 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS all meteors N such that D(M, N) does not The fact that all the observed meteors pass exceed a given value DM. close to the orbit of the earth imposes one (2) We may define a stream by serial association constraint on the observed five elements, and between members. We may state that two mete- thus reduces the number of independent ele- ors A and B are associated if D(A, B) does not ments to four. The earth is not important in exceed a standard value Ds. A stream may the evolution of meteor streams, however, so then be defined as a group of meteors in which that this constraint is not relevant to the dif- every member is associated with one or more ferences within streams. Accordingly, we other members, and all members are associated disregard it in formulating D. together directly or indirectly. In order to construct a single measure com- As a third possibility we may define a phase prising five independent differences, we draw an space in which each meteor is represented by a analogy with a five-dimensional orthogonal point and the distance between points is the coordinate system and consider each element as difference between meteors. Then a stream can a coordinate. Any orbit then corresponds to a be defined as the meteors in a region of high point in this conceptual space, and the distance density in the phase space. This form has between two points is a natural measure of the theoretical advantages but requires far more difference between the two corresponding orbits. data than we possess. We shall therefore use Although orthogonal, the coordinates need not the first two alternatives. be rectangular; usually these will be curvilinear No particular measure of orbit difference is coordinates. We may approximate them by specified a priori. Indeed, we have found no rectangular coordinates in any sufficiently small way of determining a "best" or "optimum" region, however, and this approximation will measure. The magnitude of the cause of orbit prove to be adequate for our purpose. The difference, when known, is often the best meas- difference between meteors A, B is then of the ure of the difference. We do not know the form causes of the dispersion within meteor streams in any quantitative manner, however, and it D(A, B)= (1) would be quite wrong to assume the causes. Consequently, we have looked for a measure where Cj are functions of the elements which are that is as simple as possible and is expressed in at our disposal and which determine the geom- terms of orbital elements as far as possible, etry of the space. since these are the most familiar indices of orbit The value of D as a measure of difference differences. Because differences within streams depends on how well the Cj are chosen or can be are small, our measure need not be accurate for chosen. The c}, for our purpose, are scale factors large differences in order to be useful. determining the relative importance of the dif- One orbit may differ from another in any of ferent Cj in D. Each c} should be inversely six independent ways. One of these, the posi- proportional to the expected standard devia- tion of the meteor in the orbit at a given time, tion of the corresponding element Cj in a stream. is not significant for our purpose because the An attempt to base the scale factors only meteors of a stream rapidly spread out all along on the streams would be an attempt to lift the orbit. Apart from the instantaneous posi- ourselves by our own bootstraps, because we tion of the meteor, the orbit may be specified by are still trying to determine which meteors five elements. These may be the usual elements: are stream members and which are not. We a, e, w, Q,, i; or they may be any set, out of an have checked the reasonability of the scale infinite number of possible sets, of five inde- factors that were ultimately chosen by com- pendent functions of the usual elements: Cu puting the mean standard deviation of each where; = 1,2,3,4,5. Each of the five differences element on the assumption of an idealized between two orbits A, B may be expressed as dispersing mechanism. This is discussed below. the difference between the values of an element: In accordance with all the foregoing, we have selected the elements and scale factors METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 263 2 described below. Angles, and thus the differ- [D(A,B)] =[eB-eA}> ences of angular elements, are measured by their chords, i.e., by twice the sine of half the angle. This is convenient; for small angles it is a close approximation to radian measure, +sin i sin i [2 sin and for all angles from 0 to 2* it is a smooth A B continuous measure of the absolute value, without difficulties of sign or definition. We measure a difference in orbital plane by the angle IAB between the orbital planes. IAB comprises two elements and their scale factors; this appears explicitly in the identity at low inclination, or, after substituting equa- tion (3a) for (3b) in the last term, at any in- clination. This expression is readily computed with a desk calculator and a special table of chords. +sin iA sin iB (2 sin (2) A model of stream dispersion Three elements describe an orbit within the The scale factors c} of the various elements in orbital plane. We measure a difference of D, as shown in equation (4), were selected on a shape by the difference in eccentricity e, and variety of grounds of simplicity, reasonability, of size by the difference in perihelion distance q. and previous experience. By qualitative argu- In contrast to the major semi-axis a, q is well ments we expected that, after scaling, the ele- determined by observation and has a small ments would make roughly equal contributions range of values. A difference in orientation to D. The utility of D is empirically established of the orbit within the plane is measured by below on known streams; nevertheless, a the angle TLAB between the major axes, theoretical basis for D is clearly desirable if it weighted by the scale factor e. Thus UAB can be found. Although we made this the- is the difference between the longitudes of oretical study after D had already been applied perihelion. When the orbits are in different to the meteors, the validity of D is confirmed planes, HAB is the difference between the in the main. longitudes of perihelion measured from the An excellent measure of the difference be- intersection of the orbits. In terms of familiar tween orbits, if it can be found, is the amount elements, of perturbation that will transform one of the orbits into the other. A perturbation of an arc sin orbit arises from an accelerating force applied for a period of time; hence the velocity change ^], (3a) so arising is the natural measure of the per- turbation. In general, no single perturbation can suffice because the two orbits will not or, to a sufficient approximation when iA and is axe small, intersect; we may expect that an appreciable difference between a pair of orbits that were (3b) once the same will normally be the result of many perturbations. Since D comprehends By combining equations (1), (2), and (3a), we all orbital elements, it appears possible that D find is indeed a measure of total perturbation; we 264 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS must determine, however, whether the different elements may be so combined. We consider an idealized model for orbit change, postulating random vectorial incre- ments 5V to the vectorial orbital velocity yH of a meteor. The natural unit of orbital speed at any particular distance r from the sun is the speed XJ of a body in a circular orbit of that radius; U varies as r~1/2. We take 8V to be proportional to U because, broadly speaking, we would expect most kinds of perturbation to be more intense nearer the sun. The test ap- plied to D consisted in finding the sensitivities of the different elements to perturbations 5V; these sensitivities should be comparable. To measure the sensitivities, we computed the time averages, denoted "Grad," of the magni- tudes of the gradients of the various elements in D with respect to perturbations 5V. They are shown in figure 1, plotted as functions of eccentricity. The computation of these grad- ients is described in the appendix; they are approximate, except for Grad IAB which is FIGURE 1.—Time average gradients of elements with respect exact. The value of Grad UAB on the graph to velocity perturbations which are proportional to the has already been multiplied by the scale factor circular speed. e. To obtain Grad q, the value in the figure must be multiplied by a. Broken lines in the orbits into the other. Estimating the factor figure indicate the extreme values of Grad q of proportionality from figure 1, we find that possible for a meteor observed on earth. D is about 3/2 of the mean velocity increment. Grad e, e Grad H , and Grad I are quite AB AB Application to known streams satisfactorily similar for values of e<0.85. Grad q is a function of both a and e, but is com- We applied D first to the known streams that parable to the others for most observed values were sufficiently represented in the sample, of a when e<0.85. For eccentricities near using both forms of the stream criterion dis- unity the model implies that the different cussed above. A list of the members of each elements in D have very different sensitivities stream had already been prepared by use of to perturbations of the kind postulated. At conventional criteria. The difference D(M,N) these high eccentricities, the time averages between the mean orbit M of the stream and primarily reflect the conditions near aphelion; the orbit of each member N was found, as were the effect of this is not clear. Possibly the the differences D(A, B) between pairs of meteors apparent practical success of D for high eccen- in each stream. The values of D(A,B) were tricity depends upon averaging overweighted examined to find the minimum value Ds mto of elements with underweighted elements. Ds such that all the members were associated. If this computation had been available Table 1 summarizes the values of D(M, N) and earlier in the analysis, we could no doubt have also includes D3 mln. There is obviously a close modified the form of D at the cost of simplicity. correlation between Ds miD and the mean value Except when the eccentricity is nearly one, of D{M,N). however, we may conclude that D in its present Both Ds mia and the mean value of D{M, N) form is approximately proportional to the are also correlated with the average spread of a average velocity increment, in units of the cir- stream radiant, as found by Wright, Jacchia, cular speed, which will perturb one of the and Whipple (see Whipple, 1947; Wright and METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 265 TABLE 1.—Distribution of values of D(M,N) for previously identified members of known streams, and values of As mia and O—C Range of D(Af,i\T Streams Total Mean DSmia O—C .00 .05 .09 . 13 . 17 .21 .25 .29 to to to to to to to to .04 .08 . 12 . 16 .20 .24 .28 .32 Quadrantids 3 2 5 0.04 0.05 Orionids 4 3 3 1 1 12 .08 . 15 0.38 Pereeids* 1 3 1 3 1 9 * 11 * 12 .68 Geminids 14 2 16 .02 .04 .22 S. t Aquarids t 2 2 . 10 .20 1.8 S. S Aquarids 2 1 3 .03 .07 .80 N. 5 Aquarids 1 1 1 3 .10 . 15 1.2 a Capricornid 2 2 .08 . 16 2. 1 S. Taurid-Arietid 1 2 2 1 1 2 1 10 . 18 . 19 .83 N. Taurids 1 2 3 1 7 . 13 . 19 1.02 Andromedids 2 2 .08 . 16 Total 24 18 11 10 3 2 2 1 71 .09 'Incomplete observations; shower peak missed. tOne meteor, characterized as a "very doubtful member" by Wright, Jacchla, and Whipple (1957), has been omitted. Whipple, 1948, 1950, 1953; Wright, Jacchia, between radiant spread and shower duration and Whipple, 1954, 1956, 1957). Their value (Wright, Jacchia, and Whipple, 1957), and a of the average deviation 10— C\ of an extended correlation of either D3mln or mean D{M,N) trail from the least-squares moving radiant with shower duration is easily shown. Thus all appears in table 1 for nine streams. They these measures of meteor orbit difference within have not determined a value of \0—C\ for a stream are intercorrelated. Quadrantids or Andromedids. Figure 2 shows It will be seen in table 1 that D(M, N) ex- ceeds 0.20 for only four meteors, one Northern the correlations with Ds mla and the average D(M,N). There is a well-known correlation Taurid and three Southern Taurid-Arietids. Thus we may use the condition D(M,N) a e i Q> IT Corr. Rad. Foo D(M,N) R.A. dec. (km/ Stream Greenwich date (a.u.) (deg) (deg) (deg) (deg) (deg) sec) KNOWN STREAMS South 52 June 29.43208 1.50 0.852 7.4 97.6 51.8 299.3 -16. 1 32. 1 0.38 i Aquarids 53 July 21.35918 1.37 0.921 2.9 298.2 87.9 32a 7 -13.9 35.9 0.24 *53 Aug 6.17192 1. 71 0.827 1.9 313.4 77.8 331.9 -13.1 30.9 0. 11 53 Aug 8.19278 3.32 0.927 7.5 315.3 80.8 335. 1 -15. 1 36. 1 0. 16 Mean values 1.98 0.882 1.2 301. 1 74.6 323.8 -14.6 S3. 8 0.22 South 52 July 25.44425 2.85 0.978 27.3 302.4 96.2 335.9 -17. 1 43.0 0.05 5 Aquarids 52 July 28.31172 2. 36 0.973 30.0 305. 1 99.3 339.6 -16.5 43.7 0.03 52 July 29.40913 2.40 0. 970 29.4 306.2 98.8 340.2 -16.6 42.2 0.02 Mean values 2.54 0.974 28.9 304-6 98.1 338.6 -16.7 42.6 0.03 North 52 Aug 25.35970 2.38 0.947 17.0 152.0 114.9 354.3 + 4.7 39.0 0. 15 8 Aquarids 53 Aug 10.43442 2.41 0.960 22.5 137.4 105.2 342.8 +0.5 40.5 0.06 53 Aug 13.19186 2.50 0.949 24. 1 140. 1 102.8 343.0 + 2.5 40.0 0.09 Mean values 2.43 0.952 21.2 143.2 107.6 346. 7 +2.6 39.8 0.10 a Capri- 52 Aug 22.20417 3. 19 0.793 8.3 149.0 46.6 321.2 -2.2 24. 3 0.08 cornids 53 Aug 8.36042 2.65 0.753 7.7 135.4 35.5 309.8 -6.3 23.5 0.08 Mean values 2.92 0. 773 8.0 142.2 41.0 315.6 -4.2 23. 9 0.08 S. Taurid- 52 Sept 17.31692 2.47 0.894 4.5 354.3 118. 7 10.6 + 1.4 33.8 0.28 Arietids 52 Sept 19.37056 1. 17 0.736 6. 1 356.3 125.7 16.8 + 1.2 26. 5 0.17 52 Sept 20.28367 1.37 0.741 3.0 357.2 118.4 12.5 + 2.2 29. 9 0.26 52 Sept 25.36076 1.52 0.846 7.9 2.2 134.5 22.8 + 4.0 31.8 0. 10 52 Oct 13.28086 1.85 0.840 5.9 19.8 142.6 34.8 + 9.2 31.4 0. 10 52 Oct 14.35000 1.47 0.844 6.2 20.9 154. 1 40.6 + 11.5 31.8 0.27 52 Oct 19.26487 1.95 0.806 4.3 25.8 138.6 36.0 + 10. 1 29. 2 0.08 52 Oct 22.47095 2. 76 0.829 5.3 29.0 128. 1 33.6 + 7.4 27.9 0.12 52 Oct 27.35625 1.79 0.869 9.9 33.8 163.8 52.5 + 12.5 33.3 0.42 *53 Oct 9.24410 1. 73 0.791 5.4 195.6 131.9 24.7 + 15.6 28. 8 0.20 53 Oct 9.24725 1.45 0.813 7.5 15.6 144. 1 34. 1 + 7.7 30.4 0. 12 Mean values 1. 78 0.819 5.0 13. 7 136.4 29.0 + 7.5 so. 4 0.19 North 52 Oct 16.33975 2.26 0.912 3.4 202.9 155.6 39.7 + 17.4 35.7 0. 12 Taurids 52 Oct 19.38343 1.44 0.782 5.0 205.9 149.9 38.8 + 19.6 28. 6 0. 12 52 Oct 27.35799 2.40 0.856 3.0 213.8 148.4 42.6 + 18.9 31. 1 0. 13 52 Nov 12.19596 2.25 0.839 3. 1 229.7 162.8 58.4 + 23.0 30.7 0. 10 52 Nov 12.37045 2.35 0.836 2.3 229.9 159.7 57.1 +22. 1 29.8 0.10 52 Nov 12.37333 1.98 0.855 5.3 229.9 172.5 53.0 +25. 1 32.0 0.23 *53 Oct 10.21343 2. 18 0.929 4.7 196.6 155.5 36.0 + 16.6 37.4 0. 17 53 Nov 3.29681 1.97 0.800 3.6 220.5 151.4 47.6 + 21.3 2a 8 0. 13 Mean values 2. 10 0.851 3.8 216. 1 157.0 47.9 + 20.5 31.9 0.14 METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 269 TABLE 2.—Minor streams—Continued a e i a Corr. Rad. Fc D(M,N) R.A. dec. (km/ Stream Greenwich date (a.u.) (deg) (deg) (deg) (deg) (deg) sec) KNOWN STREAMS—Continued Androme- *52 Sept 20.28438 2.80 0.764 0.8 357.2 75.0 351.8 -4.9 23.0 0.33 dids *52 Sept 25.36001 2.80 0.775 4.7 182.2 83.6 355.2 +5.5 23.8 0.25 (Bielids) *52 Oct 9.19792 2.50 0.773 2. 1 195. 8 105.2 13.3 +8.7 25.2 0. 16 *52 Oct 13.27778 2.38 0.787 8. 1 199.8 116. 6 18.5 + 18. 1 26.8 0.32 *52 Oct 21.32940 1.75 0.558 8.0 27.8 95.4 22.6 — 10.4 18. 1 0.22 *52 Oct 24.35000 2.48 0.707 4. 1 30.8 100.4 21.4 +0.6 20.7 0.07 52 Nov 7.10870 2.59 0.708 8.4 224.6 10&9 22.9 +27.6 20.9 0.20 *52 Dec 9.26872 2.68 0.666 3.5 77.1 116.8 44.8 +6.0 16.5 0.29 *53 Oct 2.31090 2.54 0.773 3.7 188.8 97.0 5.4 +7.7 24.8 0. 16 *53 Oct 6.28447 1.81 0.703 3.6 12.7 109.6 16.9 +2.0 24. 1 0.20 *53 Oct 7.39307 1.92 0.649 1.9 13.8 93.9 9.3 +0.3 20.6 0.07 *53 Oct 10.31215 2.39 0.700 7.5 16. 6 88.3 12.3 -9.6 21.2 0. 16 53 Nov 7.46139 2.29 0.679 0.0 224.7 113.5 31.1 + 12.7 19.8 0. 17 *53 Nov 13.24583 2.02 0.577 3.8 50.5 102.0 29.5 +0.6 16.6 0.25 Mean values 2.85 0.701 0.4 25.9 100.4 18.6 + 4-6 21.6 0.20 KNOWN STREAMS HAVING ONE MEMBER IN OUR SAMPLE North 52 July 27.22705 3.48 0.947 15.0 124. 1 76.9 323.8 -6.8 38.5 _ i Aquarids Draconids 53 Oct 9.19222 3.37 0.704 24.6 195.5 12.5 270.9 +47.2 20.2 (Giacobi- nids) Monoce- 53 Dec 10.51604 20.16 0.992 39.8 78. 1 209.4 101.6 +7.7 43.9 — rotids NEWLY FOUND POSSIBLE STREAMS X Geminids 53 Jan 13. 29493 2.01 0.739 1.0 292.8 208.3 118.4 + 22. 1 25.6 0. 11 53 Jan 13. 45089 2.58 0. 779 17.2 292.9 200.6 123.0 + 41.8 27.0 0. 19 53 Jan 13.45411 2.52 0.798 11.3 292.9 208. 1 122.4 + 33.2 27.6 0. 13 53 Jan 23. 38956 2.41 0.725 4.2 123. 1 200.0 117.9 + 14.0 22.7 0.20 Mean values 2.88. 0. 760 6.8 295.4 204-2 120.4 + 27.8 25. 7 0. 16 p Geminids 53 Jan 15.40959 2.41 0.716 7.4 294.9 189.0 111.6 + 34.6 22.3 0. 11 53 Jan 19.19278 2.55 0.694 4.4 298.8 179.5 105.5 + 31.9 20. 1 0.06 53 Jan 21.37852 2.81 0.723 1. 1 301.0 180.6 105.9 +24.9 20.0 0.08 54 Jan 28.17159 3.34 0.762 7.9 307.6 184. 1 114. 6 + 37.8 21.0 0.07 Mean values 2.78 0. 724 5.2 300.6 183.3 109.4 +32.3 20.8 0.08 638805—63 22 270 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Minor streams—Continued a e i Q, Corr. Rad. Vao D(M,N) R.A. dec. (km/ Stream Greenwich date (a.u.) (deg) (deg) (deg) (deg) (deg) sec) NEWLY 1^OTJND POSSIBLE STEEAMS—Continued £ Orionids 53 Mar 14.16204 1.94 0.488 4.2 173.2 169. 1 80.6 + 0.2 12. 9 0.09 54 Feb 11.42659 1.80 0.479 3.4 322. 1 173.9 106.4 + 36.5 13.9 0.09 54 Mar 5.29565 1.84 0.472 1.0 164. 1 185.6 113.6 + 12.6 13.2 0.08 Mean values 1.86 0.480 0.9 159.8 176.2 100.2 + 16.4 IS. 3 0.09 *Ursa 53 Mar 18.33889 1.80 0.497 6.9 357.4 220.9 161.8 + 31.8 15. 6 0.08 Majorids 53 Apr 10.24583 1.63 0.398 12.0 20.0 222.5 185.4 + 53.7 14.6 0.09 Mean values 1. 72 0.448 9.4 8. 7 221. 7 173.6 + 42.8 15. 1 0.08 a Bootids 53 Apr 9.35927 1.90 0.599 21.0 19.2 26a 3 209.8 +27.6 22.4 0.10 53 Apr 16.28881 2. 13 0.674 16.0 26.0 282.3 210.4 + 14.8 23.4 0.10 Mean values 2.02 0.636 18.5 22. 6 275. S 210. 1 + 21.2 22.9 0.10 y Bootids 53 Apr 13.25625 3.76 0.786 26.6 23.0 259.5 213.0 + 33.4 25.9 0.08 53 Apr 21.36280 3.34 0.733 29. 1 30.9 253.3 222.3 +42.7 24.5 0.08 Mean values 3.55 0. 760 27.8 27. 0 256.4 217.6 + 38.0 25.2 0.08 £ Librids 53 Apr 13.25916 1.71 0.914 2.8 23.0 344.9 224.4 -15.6 37.0 0.09 53 May 5.28417 2.95 0.925 3.2 224.4 353.3 239. 1 -22.2 36. 8 0.09 Mean values 2. S3 0.920 0.2 213. 7 349.0 231.8 -18.9 36. 9 0.09 METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 271 TABLE 2.—Minor streams—Continued a e i a ,x Corr. Rad. Foo D(M,N) R.A. dec. (km/ Stream Greenwich date (a.u.) (deg) (deg) (deg) (deg) (deg) sec) NEWLY FOUND POSSIBLE STREAMS—Continued p Draconids 53 Apr 16.42532 4.32 0.770 43.6 26. 1 218.3 256.6 + 59.8 29.6 0.09 53 Apr 21.36042 15.72 0.936 47.1 30.9 217.9 263.4 + 61.0 32.4 0.09 Mean values 10.02 0.853 45.4 28.5 218. 1 260.0 +60.4 81.0 0.09 4> Bootids 53 May 7.33958 1.50 0.379 19.9 46.4 270.3 231.6 +40.4 17.6 0.06 53 May 8.38998 1.39 0.321 17.9 47.4 270. 1 231.8 +41.5 16.2 0.08 54 June 4.24583 1.70 0.417 22.7 73.0 275.0 245.0 +50.5 18.2 0.13 Mean values 1.5S 0.S72 20.2 55.6 271.8 286.1 +44.1 17.8 0.09 £ Cepheids 53 May 9.42072 1.00 0. 157 16.4 48.4 127.4 341.8 +63.9 14.7 0.10 54 Apr 28.30873 1. 15 0. 172 22. 1 37.4 164.7 313.8 + 6a3 16.6 0.10 Mean values 1.08 0.164 19.2 42.9 146.0 827.8 + 66.1 15.6 0.10 a Ursa 53 May 13.16348 2.04 0.512 2.8 52.0 24a 3 17a 0 + 15.1 13.6 0.16 Majorids 53 June 5.17987 2.48 0.596 9.4 74.2 240.5 161.0 +49.0 14.8 0.03 53 June 6.19792 3.55 0.725 13.3 75. 1 230.6 143.5 +57.9 17.2 0. 18 Mean values 2.69 0.611 8.5 67. 1 289.8 160.8 +40.7 15.1 0.12 Cylids 52 May 19.21518 1.03 0. 131 0.8 58. 1 148.7 57.5 +27.2 11.5 0.06 52 Oct 19.44228 0.95 0.120 2.7 26.0 144.7 50.2 -7.4 11.4 0. 10 53 Apr 11.14238 1.00 0.017 0.0 — 101.2 — — 10.6 0.10 53 Apr 15.24742 1.00 0.017 0.0 — 101.2 — — 11.0 0. 10 54 June 3.25929 1.08 0.148 6.2 72. 1 179.9 76.9 + 62.8 12.2 0. 14 Mean values 1.01 0.087 1.9 52.1 185.1 60.5 +27.6 11.8 0.10 $ Ophiucids 52 May 21.36274 2.62 0.915 4.5 240.2 9.4 256.3 -25.4 36.0 0. 19 53 June 4.20432 2.34 0.838 3.0 253.2 5.2 262.4 -25.9 30.8 0.03 53 June 8.37832 3.42 0.893 3.1 257.2 8.2 266.3 -25.8 32.4 0.06 53 June 16.30529 3. 11 0.827 4.5 264.8 357.0 265.0 -2a 6 27.4 0.19 53 June 16.31098 2.86 0.837 4.9 264.8 5.7 269.7 -2a 4 29.0 0.08 Mean values 2.87 0.862 £0 256.0 5.1 264.0 -26.8 81.1 0.11 M Draconids 52 June 17.36898 2.83 0.641 26.7 86. 1 269.4 23a 4 +60.3 20.6 0. 18 53 June 8.28750 3.28 0.720 31.8 77. 1 296.2 255.5 +37.9 25.0 0.24 53 June 13.26806 3.70 0.736 31.3 81.9 286.4 253. 1 +46.8 24.1 0. 12 54 June 9.37083 2.81 0.644 39.6 77.9 273.5 267.7 +56.6 26.7 0. 13 54 June 11.38442 3.27 0.697 33.6 79.8 280.1 255.9 +51.7 24 4 0.04 54 June 23.21042 2.74 0.629 38.2 91. 1 270. 1 267.8 + 66.9 25.9 0. 18 Mean values S.10 0.678 S3. 6 82.8 279.8 256.4 +68. 4 24-4 0.15 r Herculids 53 June 20. 35000 2.07 0.521 28.9 sa 7 291.0 25a 6 + 50.4 21.2 0.10 54 June 25. 24461 2.81 0.642 21.3 93. 1 286.2 237.8 +46.4 las 0. 10 Mean Values 2.44 0.582 25.1 90.9 288.6 248.2 +48.4 20.0 0.10 53 Aug 6. 16985 2.53 0.624 1.2 313.4 347.0 272.3 -27. 4 15.5 0. 12 M Ophiucids 53 Aug 7. 39845 2.06 0.515 7.7 134.5 330.7 254.0 + 11.6 13.7 0. 13 53 Aug 10. 22697 2.36 0.587 4.3 137.2 343. 1 266.9 -6.6 14.5 0. 10 Mean values 2.82 0.575 S.6 185.0 840.8 264-4 -7.5 14.6 0. 10 272 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Minor streams—Continued a e i a X Corr. Rad D(M,N) R.A. dec. (km/ Stream Greenwich date (a.u.) (deg) (deg) (deg) (deg) (deg) sec) NEWLY FOUND POSSIBLE STBEAMS—Continued 53 Aug 9. 23255 2.59 0.611 ia6 136.3 327.5 257. 1 + 38.8 17.4 0.04 t» Herculids 53 Aug 13. 42525 2.50 0.617 17.3 140.3 352. 1 280.1 + 26.6 17.7 0.13 53 Aug 15. 42275 2.06 0.518 18.7 142.2 341.2 270.2 + 38.8 16. 9 0.06 Mean values 2.38 0.582 i8.se 139.6 840.3 269.1 +84-7 17.2 0.08 52 Sept 14.37389 1.93 0.720 12.0 351.4 87.4 2.5 -15.7 25.0 0.09 0 Cetids 52 Sept 20.29047 1.40 0.592 15.5 357.2 96.7 14.0 -18.8 22.5 0.09 Mean values 1. 66 0.656 13.8 354.3 92.0 8.2 -17.2 23.8 0.09 52 Sept 16. 33958 1.57 0.422 1.0 173.4 41.6 327.2 -9.0 13.9 0.13 Capri- 52 Oct 9. 19742 1.46 0.314 &9 195.8 17.6 278.6 + 30.2 12.7 0.10 cornids 53 Aug 13.35836 1.37 0.348 0.9 320.2 17.1 302.4 -23.7 13. 6 0. 14 Mean values 1. 47 0.361 3.0 169.8 25.4 302.7 -0.8 13.5 0.12 52 Sept 19. 23462 2.70 0.713 0.7 176.2 59.4 341.4 -6.3 20.0 0.20 52 Sept 27.30972 1.79 0.599 10.9 184 1 79.8 349.8 +ia5 20.0 0.22 52 Oct 16. 21458 2.61 0.670 2.6 22.8 71.3 359.5 -7.7 17.5 0. 13 52 Nov 12. 19033 3.41 0.717 7.6 229.7 70. 1 342.2 +22.1 15. 6 0.14 a Pegasids 52 Nov 12. 19349 2.46 0.605 6.9 229.7 68.7 339.8 + 22.4 14 4 0. 12 52 Nov 12. 26391 3.42 0.719 7.4 229.8 70.6 343.0 +21.9 15.5 0.14 53 Oct 3. 25367 1.46 0.432 11.5 189.7 74.5 345.0 +2a5 16.5 0.24 53 Oct 3. 29383 1.73 0.501 5.4 189.7 64.0 343.1 + 10.4 15. 7 0.15 53 Nov 6. 33041 3.06 0.704 7.0 43.6 81.2 15.8 -15.5 16.8 0.25 Mean values 2.52 0.629 4-5 206.1 71.1 848.8 + 10.5 16.9 0.18 X Draconids 52 Oct 12. 33141 5.49 0.818 72. 1 198.9 1&4 181.9 + 81.4 43. 5 0. 10 52 Oct 19. 44602 3. 14 0.683 77.4 206. 0 24.3 167.3 + 76.0 45.0 0. 10 Mean values 4.82 0.750 74.8 202. 4 21.4 174.6 + 78.7 44.2 0. 10 52 Oct 17. 40034 10.37 0.919 173.8 203.9 71.4 102.4 + 26.3 70.6 0. 11 c Geminids 52 Oct 22. 26372 26.77 0.971 173.0 208.8 85.6 104 2 + 26.5 70. 6 0. 14 53 Oct 16. 42579 22. 87 0.965 173.3 202.7 76.3 9a 6 +26.8 70.5 0.03 Mean values 20. 00 0.952 173. 4 205.1 77.8 101. 7 + 26.5 70. 6 0.09 I Taurids 52 Dec 9. 26560 2.60 0.682 2. 1 257. 1 130.3 51.0 + 24.0 18.3 0. 10 53 Jan 15. 21244 3.98 0.763 0.7 114.7 140.7 65.3 + 19.3 16. 0 0. 10 Mean values 3.29 0.722 0.7 275.9 185.5 58.2 +21.6 17.2 0. 10 METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 273 TABLE 3.—Spurious streams found in the artificial sample Stream Greenwich Date a (a.u.) e i (deg) G (deg) x (deg) Original Clas- D(M,N) sification (A) 52 Mar 21. 36111 3.23 0.713 4.3 43.6 257.2 Sporadic 0. 13 52 Sept 25. 17361 1.79 0.510 11.5 25. 1 256.4 Sporadic 0. 14 53 Aug 13. 42525 2. 50 0.617 4.8 46.4 25& 2 # Herculid 0.06 53 Aug 15. 42275 2.06 0.518 4.6 47.4 246.4 Sporadic 0. 12 Mean values 2.40 0.590 6.S 40.6 254-6 (B) 52 Apr 2. 38125 2.80 0.644 1.9 21. 1 210.4 Sporadic 0.07 53 May 13. 16348 2.04 0.512 4.9 0.6 196.8 o) Ursa Majorid 0. 13 54 June 25.24461 2.81 0.642 1.2 16.6 209.7 T Herculid 0.07 Mean values 2.55 0.599 2.7 12.8 205.6 (C) 52 June 24. 20301 2. 79 0.720 8.6 39.6 283.2 Sporadic 0. 15 52 Sept 16. 33958 1.57 0.422 4.4 48.4 276.7 £ Capricornid 0.20 52 Sept 19. 23462 2. 70 0.713 7.6 30.0 273.2 a Pegasid 0. 14 53 Oct 3. 29383 1.73 0.501 4.2 46.3 280.5 a Pegasid 0. 12 54 Mar 1. 36968 2. 01 0.639 4.8 25. 1 276.8 (D) 52 Sept 16. 33723 3. 22 0.937 8.9 63. 1 13.6 Sporadic 0.07 53 Apr 13. 25916 1. 71 0.914 3.9 46.5 8.4 f Librid 0.07 Mean values 2.96 0.926 6.4 64.8 11.0 (E) 52 Sept 25. 17579 15.94 0.947 7.0 48.4 275.9 Sporadic 0.08 53 Nov 2. 35675 6.87 0.872 13.3 51.2 271.7 Sporadic 0.07 Mean values 11.41 0.910 10.2 49.8 273.8 (F) 52 Aug ia 25543 20.92 0.952 0.2 24.0 207.4 Sporadic 0.08 53 Oct 9. 31647 19.45 0.949 9.7 30.9 207.9 Sporadic 0.08 Mean values 20. 18 0.950 5.0 27.4 207. 6 (G) 53 Jan 13. 45411 2.52 0.798 1.9 23.2 298. 4 x Geminid 0.06 54 Feb 6. 38914 2.63 0.810 2.4 32. 1 30a i a Leonid 0.07 Mean values 2.58 0.804 2.2 27.6 SOS. 2 (H) 53 Apr 7. 28372 1.98 0.809 4.2 53. 0 346. 0 Sporadic 0.09 54 Feb 10. 40343 3.69 0.871 0.9 77.9 354.2 Sporadic 0.09 Mean values 2. 84 0.840 2.6 65.4 350. 1 274 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 4.—Further data on streams Stream Hourly Number of Inclina- Mean Radiant rate members tion D(M,N) X—X© MAJOR STREAMS Quadrantids 4.30 5 71?6 0.04 278° 64° Orionids 3.69 12 164.9 .08 247 -7 Perseids* *1. 67 9 113.8 *. 11 283 37 Geminids 5.92 16 23.3 .02 208 10 KNOWN MINOR STBEAMS S. t Aquarids 0.46 4 1?2 0.22 201° -1° S. & Aquarids 1.80 3 28.9 .03 210 -7 N. 5 Aquarids 0.37 3 21.2 .10 207 7 a Capricornids 0.29 2 8.0 .08 175 12 S. Taurid-Arietids 0.71 11 5.0 . 19 197 -5 N. Taurids 0.54 8 3.8 . 14 195 2 Andromedids 0.67 14 0.4 .20 172 — 1 KNOWN MINOR STREAMS HAVING ONE MEMBER m OUR SAMPLE N. Aquarids _ 1 15?0 202° 7° Draconids — 1 24.6 — 78 70 Monocerotids — 1 39.8 — 205 -14 NEWLY FOUND POSSIBLE MINOB STREAMS X Geminids 0.63 4 6! 3 0. 16 182° 8° p Geminids 0.89 4 5.2 .08 166 10 a Leonids 0.74 27 2.2 .29 171 2 £ Orionids 0.43 3 0.9 .09 119 -2 $ Ursae Majorids 0.20 2 9.4 .08 144 36 a Bootids 0.44 2 18?5 . 10 176° 31° 7 Bootids 0.67 2 27.8 .08 171 49 8 Librids 0.24 2 0.2 .09 201 0 v Draconids 1.24 2 45.4 .09 204 82 £ Cepheids 0.32 2 19?2 .10 344° 68° w Ursae Majorids 0.38 3 8.5 .12 79 31 Cyclids — 5 1.9 .10 0 Ophiucids 0.37 5 4.0 . 11 189 —4 fi Draconids** 1.31 6 33.5 .15 148 74 T Herculids 0.84 2 25! 1 .10 140° 68° M Ophiucids 1.76 3 3.6 .12 128 15 f> Herculids** 0.92 3 18.2 .08 129 58 0 Cetids 0.71 2 13.8 .09 186 -20 ( Capricornids 0.24 3 3.0 . 12 134 18 o Pegasids 0.56 9 4?5 .18 148° 15° X Draconids 0.57 2 74.8 .10 273 68 « Geminids** 1.09 3 173.4 .09 256 3 7 Taurids 0. 15 2 0.7 .10 145 2 METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 275 TABLE 4.—Further data on streams—Continued Stream Hourly Number of Inclina- Mean Radiant rate members tion D(M,N) X—\Q /3 SPURIOUS STREAM (A) _ 4 6?3 0. 11 136° 22° (B) — 3 2.7 .09 98 13 (C) — 5 5.9 . 14 161 14 (D) — 2 6.4 .07 202 6 (E) — 2 10.2 .08 152 22 (F) — 2 5.0 .08 90 16 (G) — 2 2.2 .06 183 2 (H) 2 2.6 .09 192 2 'Incomplete observations; shower peak missed. ••The most probable streams. Streams with a photographic rate less than **- Geminids three per hour we shall call the minor streams. Known streams. The Southern 5 Aquarids are the only minor • Newly found possible streams.! stream in our sample with a mean D(M,N) less than 0.08. Because the duration of a shower is correlated with its mean D(M,N), the total number of meteors in the sample is correlated k— Quadrantids with the product of ordinate and abscissa in figure 3; thus a stream in the lower left corner is comparatively unlikely to be detected. Nevertheless, the sharp boundary of the points at Z?=0.08 seems meaningful as an approximate lower bound to the mean dispersion within the majority of minor streams. On the right edge of figure 3, the " S. < Aquarids streams. The Southern Taurid-Arietid stream Perseids is known to be composite, as its name indicates. Either of these composite streams, if it could be divided into its originally independent parts, » Leonidnids .. _ S. Taurids-Arietids \ would be represented by two or more points Andromediddidss nearer to the majority of streams in the v 6. > Aquarids N. I Aquarid 0.10 0.20 0.30 Nature of the newly found minor streams MeanD(M.N) Figure 4 shows histograms of the normalized FIGURE 3.—Relation between hourly rate and internal distribution of values of D(M,N) in each of dispersion of streams. four classes of streams. The data for major Thus neither our value of the rate nor of the streams and previously known minor streams mean dispersion within the stream is represent- are drawn from table 1, Perseids being omitted; ative of the stream. Accordingly, no further for newly found minor streams from table 2; use is made of values of D from the Perseid and for spurious streams from table 3. For stream. each class of stream the ordinate is the percent 276 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 1 1 1 1 1 Previously known streams 60% 7 8 9 10 11 12 13 14 15 16 - Major streams \ 50% 1 "^ Spurious streams 40% \ 23456789 30% j Spurious streams 20% K \ 2 3 4 5 / \ *^\^^ Newly found minor streams Number of meteors 10% \ V \ FIGURE 6.—Distribution of number of meteors in a stream. 1 \ V /C~~\^— Previously known minor streams Major streams are unshaded. 0.1 \-V0.\2 0.x3 . 0.4 0.5 0.6 D(M.N) space from the mean elements of the stream. A stream with members uniformly distributed FIGURE 4.—Normalized distribution of D(M,N) in different through a volume in this phase space would classes of streams. have a histogram in which the ordinate varied of values of D within an interval 0.04 unit as the cube of the abscissa. Thus, the histo- wide. gram for newly found minor streams indicates, All these curves indicate high central con- so far as the numbers are significant, that the densation in the streams. D may be inter- density of points at D(M,N)=0.1 has actually preted as distance in a four-dimensional phase decreased from the density at lower values of D. Figure 5 shows the distribution of values of D(A,B) within streams. It is clear in both 45- figures 4 and 5 that the newly found streams are less compact than the previously known minor 40 streams. Figure 6 shows the distribution of the number of meteors in a stream. The 35 comparatively low values of D(M,N) for spurious streams occur because such a large proportion of the spurious streams have two members only. We may also conclude from figure 6 that it is the newly found minor streams with two or few members that are more likely 20- to be chance groupings. Figure 7 shows the distribution of the inclina- tions of the mean orbits of streams. From the concentration of all the spurious stream inclina- - Previously known minor streams tions to low values we conclude that the newly found minor streams with low inclinations, like those with few members, are less likely to be Spurious streams genuine streams. 0.0 The distribution of radiants of meteors and streams in our sample is shown in figures 8 and 9, FIGURE 5—Distribution of D(A,B) in different classes of referred to a system of coordinates defined by streams. the ecliptic plane and the direction to the sun. METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 277 Previously known streams n n n Newly found minor streams FIGURE 7.—Distribution of inclinations, Major streams are unshaded. Spurious streams These radiants have been corrected for zenith selected four of the provisional streams as the attraction and diurnal aberration. We found most probable, the ^Bootids, yuDraconids, the ecliptic latitude /3 of each radiant and also 0Herculids, and eGeminids. We have indicated the ecliptic longitude X—Xo measured from the above that the Cyclids are quite probably not sun's direction as origin. The radiants of a stream, and that the a Leonids are certainly major streams and of all other meteors in the not a single stream. More definite conclusions sample are so plotted in figure 8. The apex can hardly be drawn from our sample. of the earth's motion is in the center of the diagram, the antapex at the right and left ends. Comparison with other results The large cluster of points at the right is nearly McCrosky and Posen (1959) have examined centered on the anti-sun at X—Xo = 180°, 2538 meteors photographed by the Super- /S=0°; while the direction to the sun at left Schmidt cameras, and by comparing radiants, center is X—Xo=/3=0°. velocities, and dates, have found seven new The last two columns of table 4 list the mean streams. Substantially all of our meteor sample values of X—Xo and (J for each stream radiant. is included in theirs. Four of the streams in Figure 9 repeats figure 8, substituting mean each list are associated. The e-Geminid stream radiants of minor streams for the individual bears the same name in both lists. Their h radiants of their members. The "Cyclid" Arietids are the same as our ijTaurids, and radiant spread is too large to permit plotting. their ^Pegasids are included in our TPiscids. In figure 9, the minor streams are seen to Finally, their aVirginids are included within cluster principally around the ecliptic in our overgrown 70' \-\ \\\x # /// /7TT7 :\ \ \A \ v\\ u • . . .\ V.\ --\\\'x:\\ V.\A..\ \ \ \ \\\\\\\ \ \ \ \ J'- hi \\\\ \ \ \.\ L/XZV//7 x\\\\\ • RESIDUAL METEOR -70' -70° -80° -80° -90' 0 MAJOR STREAM FIGURE 8.—Distribution of radiants of major streams and residual meteors. 80° 90° 80° _—•— 70" .—— V —-—-*P 60° **— y 60' .—• m / IIS ?=>- 50°^^%^ x: y 50° y y// 7/ . \v _ ^^ >< ^^ y / \ *c '/• I. X" \ / / rt\v\\ N 30°X// /, // /. / i r ./. I- \ \'\ \ \\ \ \XX30° •\ > / / uYZL / / / i /pa —• \- \ • i —•— \20° i . \- \° \\ / / / r .\ \XX / /' s io / / 1 • '\ rtl—LLL 1 LJTT I / / / t • • i \ • \ .\. .\ \ \°\ \ vA ° —r • — 1 f1 l c \ \ \ l_ 0' /3o° (J V • OR! • • • a •• 1 • 1 1 l 1 1 / / l ./ \ \ 7-20° \ / /.- -2>rWv \xx\ \ \ \ • •/ / / /: / / •> //'/Tt-, / \ 0 \ \ \ \ / / / / \ \ \ 1 -J-Z2 /, 7 /XAo 3 "o? \ X v \ \ \ 1 / / / —v—*c s *~ -40^ \ 5^40° X \\\ \ Y77Z/ y^ >< 'A \ *y \. 1 -\44-LL y §§§^ ^- >< \ S>X A ^A 1th ///^ <<& ^50° -^- •—. ^^ —. t SPORADIC -60°^ > >^ ^> ^60°' METEOR •~~-—. /, ^^: ^\ \ —— O MINOR STREAM -70° . >\ ^^ — ^70° -80° -q -80° n MAJOR STREAM FIGURE 9.—Distribution of radiants of major streams, minor streams, and sporadic meteors. CO 280 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 5.—The closest correspondences between newly found possible streams and known comets Stream a e q i SI CO Equinox D £ Orionids 1.86 .480 .947 0?9 159?8 16?4 176?2 1950 Comet 1928 III 3.43 .71 .995 1.4 196.8 345.2 182.0 1928 .24 T Herculids 2.44 .582 .998 25! 1 90?9 197?7 288?6 1950 Comet 1930 VI 3.09 .672 1.011 17.4 76.8 192.3 269.1 1930 .28 n Ophiucids 2.32 .575 .975 3?6 135?0 205?3 340?3 1950 Comet 1 1678 3.07 .627 1. 145 2.9 163.3 159.5 322.8 1678 .26 De Vico-f 1844 I 3. 10 .617 1. 186 2.9 63.8 278.7 342.5 1844 .23 Swift J 1894 IV 3.25 .572 1.392 3.0 48.8 296.6 345.4 1900 .43 a Pegasids 2.52 .629 .872 4?5 206? 1 225? 0 71?1 1950 Comet 1819 IV 2.96 .699 .892 9.1 77.4 350.1 67.5 1820 .23 17 Taurids 3.29 . 722 .885 0?7 275?9 219?6 135?5 1950 Comet 1894 I 3.80 .698 1. 147 5.5 84.4 46.2 130.6 1894 .29 Only the connection between the a Pegasids stream passing within any given small distance and Comet 1819 IV, which was previously h from the earth's orbit per unit time is found by McCrosky and Posen (1959), is fully N2=2hwpVB. It is well known (see, for convincing. We should expect a parent comet example, Schiaparelli, 1871, p. 116) that to differ from the mean of an observed stream X=T&VllVl. From figure 10 we see that by more than the average observed meteor VB cosec 7= Va cosec L. Thus the ratio of the because the observed meteors, all of which number of meteors striking the earth to those strike the earth, are a restricted sample of the crossing its orbit, stream. Since all these comets except 1819 IV differ from the streams by more than 0.2, iV, TR2 VI however, and since some of the streams them- N2~2hVB VI sin L selves may not exist, anything more than the is independent of w and p. Following Whipple possibility of association between these remain- (1954), we approximate by Vt the velocity ing streams and comets remains to be shown. dependence of the ratio of the number of Space density of minor streams meteors brighter than a given magnitude on photographs to the number above a given In this discussion we shall use the symbols mass, but omit his special factor for velocities defined as follows: Va: Heliocentric velocity of meteors. Va: Geocentric velocity of the meteors, neglecting the earth's attraction. Vt: Accelerated geocentric velocity of the meteors, at the top of the atmosphere. Vm: Velocity of the meteors with respect to the earth's surface, at the top of the atmosphere. L: Angular elongation of the corrected radiant from the apex of the earth's way. y: Angle between Vo and Vg. ID: Width of meteor stream in the plane of the figure. p: Density of meteors in the stream. X: Earth's collisional cross-section for meteors. R: Earth's radius including the atmosphere. VM: Earth's orbital velocity. The number of meteors swept up by the earth in one pass through the stream is iV"i= FIGURE 10.—Geometry of the intersection of the earth with Xpw cosec 7. The number of meteors in the a meteor stream. METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 281 below 19 km/sec. The qualitative result is streams. The absence of closely associated not sensitive to a fractional change in the comets corroborates their relatively great age. power of Va* We approximate Vm by the There are some interesting relationships closely similar Ve. Thus (neglecting constants) among certain of these streams. The n Ophiu- we find that the ratio of the number of meteors cids and # Herculids form a pair of very similar passing near the earth's orbit to those photo- streams differing only in inclination. In con- graphed is given by trast to known pairs of streams on either side of the ecliptic, the juOphiucids are nearly in w V% sin L W= T/5 > (5) the ecliptic while the t? Herculids have an Vi inclination of 18°. The eGeminids bear a which may be interpreted as a weight of a somewhat similar relation to the Orionids, but photographic meteor in determining the present in this case the longitudes of perihelion also strength of its stream at the earth's orbit. differ. Figure 11 shows ifas a function of L for In figure 12 the orbits of all the newly found various values of the stream major semi-axis a. minor streams with inclinations less than 15° For all except nearly circular orbits, W has a are projected onto the plane of the ecliptic. strong maximum between Z=90° and L=130°. Although the majority of the streams plotted Figure 9 contains, in addition to the cluster here doubtless are chance associations, some of stream radiants already discussed, a smaller conclusions can be drawn. The tendency of group in high latitudes which includes three of these orbits to be approximately tangent to the the most probable streams. All of this group, earth's orbit is clearly a selection effect. The and most of the larger group fall within the tendency to have aphelia at or near Jupiter's range of high W, 90° FIGURE 11.—Weight W of a photographic meteor in determining the present strength of its stream at the earth's orbit, as a function of the angular elongation L of the radiant from the apex, and the major semi-axis a. 8 to /3 CETIDS PEGASIDS L 17 TAURIDS CAPRICORNIDS CYC LIDS p GEMINIOS CLIBRIDS > 3 CD -/iOPHIUCIDS X GEMINIOS URSA MAJORIDS \O"LE0NIDS FIGURE 12.—Orbits of all the newly found minor streams with inclinations less than 15° are projected onto the plane of the ecliptic. METEOR SYMPOSIUM—SOUTHWORTH AND HAWKINS 283 which no doubt do not exist, but the number of spurious streams may be estimated. Streams too small to have at least two meteors in our sample, streams with brief showers during a gap in the observing, variable and temporary streams: all may or will be missed. Of the 359 meteors in the sample, 42 or 12 percent belong to major streams, i.e., those with mean photographic hourly rates exceeding three; 48 or 13 percent belong to previously known minor streams, and approximately 39 or 11 percent to newly found streams. About 64 percent of the sample is sporadic. In the immediate neighborhood of the earth's orbit, the minor streams, compared to other meteors, are roughly twice as numerous as in the photo- graphic sample. The newly found streams are in fact old streams; they have been much perturbed, and some show obvious relationships with Jupiter. In overall orbital distribution, the minor streams do not differ, in any way that has been conspicuous to us without special scrutiny, svR from the sporadic meteors with which they are FIGURE 13.—Components of velocity and of perturbations intermingled. in velocity. Appendix Let the perturbations be expressed in units of To compute the average gradients of orbital the circular speed U=r~l/2; the magnitude of the elements with respect to perturbations, let gradient of G with respect to such perturbations the units be the astronomical unit, the solar is mass, and s~ year. Consider perturbations at (A2) a point P in a meteor orbit, at a distance r from Let Grad Q denote the average of \VG\e over the sun. Resolve the meteor's heliocentric the orbit with respect to time. Let t be the velocity VH into components as in figure 13: time of the meteor's arrival at P, counted from VR directed away from the sun, and Vs directed perihelion passage. Then, since the period of 3/2 forward in the orbit plane normal to VR. Re- revolution is 2jra , solve the perturbations in velocity into com- ponents: bVR and 8VS along VR and Vs, re- Grad G= (A3) spectively, and dVT normal to the orbit plane forming a right-handed set with SVR and SVS. Let G denote any orbital element, and \VG\ the magnitude of the gradient of G with respect to perturbations in velocity. Then 1/2 (Al) This is the desired measure of the importance of GinD. It is evident that differentiation with respect Of the components of D, differences within to dVR and 8VS may be replaced by differenti- the orbital plane are measured by 8e, Sq, UAB. ation with respect to VR and Vs, respectively. Obviously, Grad 8g=Grad e, and Grad Sq— 284 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Grad q. Let v be the true anomaly at P. Any (A14) perturbation hv in true anomaly implies a per- turbation — 8v in the direction of perihelion; then thus (Y2+Z2y<2 2 2 bv/bVs=rVs(rVs -l)/e , (A8) 2 = -rVB(rVs where Since e, v, q are functions of r, VB, Vs only, =O. (A9) and \jTr~) is similarly defined. With the aid of a particular set of elliptic Expressions for Grad e, Grad v, Grad g integrals, may now be found by substitution of equations (A6), (A8), (A9) into equation (A4). We = r[l- (A18) have not found it possible to evaluate those Jo expressions in closed form. Since an approxi- mate result is sufficient for the purpose, we )= C*[i- -1 sin2 u\l'2du, (A19) take advantage of the following lemma. Jo Lemma: If y and z are functions of x which are finite throughout a -4=2 arc cos{[— (A22) HAWKINS, G. S., AND SOUTHWORTH, R. B. 1958. The statistics of meteors in the earth's Also, atmosphere. Smithsonian Contr. Astrophys., vol. 2, no. 11, pp. 349-364. 1961. Orbital elements of meteors. Smithsonian Contr. Astrophys., vol. 4, no. 3, pp. 85-95. 2 , (A23) HAWKINS, G. S.; SOUTHWORTH, R. B.; AND STIENON, F. 1959. Recovery of the Andromedids. Astron. Journ., vol. 64, pp. 183-188. 2 , (A24) JACCHIA, L. G. 1960. Individual characteristics of meteor families (abstract). Astron. Journ., vol. 60, pp. 53-54. (A25) MCCROSKY, R. E., AND POSEN, A. 1/2 (dqJbVs)=4a(l-e) [2(l+e)E'(Tr/2) 1959. New photographic meteor showers. Astron. Journ., vol. 64, pp. 25-27. -(2-e)(l-e)F'M2)}/3ire. (A26) SCHIAPARELLI, G. V. 1871. Entwurf einer Astronomischen Theorie der The values of Grad e, e Grad UAB, and of I/a Sternschnuppen. G. von Boguslawski, Grad q in figure 1 were computed from equa- transl. Stettin. tions (A5) and (A16) through (A26). The WHIPPLE, F. L. largest possible error, computed by equation 1947. Photographic meteor studies. IV. The (A15), is 15 percent. Geminid shower. Proc. Amer. Phil. Grad IAB remains to be discussed. Since Soc, vol. 91, pp. 189-200. I measures the change of orbital plane, it 1954. Photographic meteor orbits and their AB distribution in space. Astron. Journ., is independent of velocity changes within the vol. 59, pp. 201-217. plane; hence WRIGHT, F. W.; JACCHIA, L. G.; AND WHIPPLE, F. L. 0. (A27) 1954. Photographic 5-Aquarid meteors. Astron. Journ., vol. 59, pp. 400-406. From figure 13 we see that 1956. Photographic a-Capricornid meteors. Astron. Journ., vol. 61, pp. 61-69. (A28) 1957. Photographic i-Aquarid meteors and evidence for the northern 5 Aquarids. Astron. Journ., vol. 62, pp. 225-233. Thus Grad IAB may be found exactly from equations (A4), (A6), (A18), (A19), (A27), WRIGHT, F. W., AND WHIPPLE, F. L. (A28), as 1948. Photographic meteor results for the month of October (abstract). Astron. Journ., 1/2 vol. 54, p. 54. Grad IAB=2(l-e)- 1950. The photographic Taurid meteors. Har- -(1-e) F'(X/2))/ZT, (A29) vard Coll. Obs. Reprints, ser. 2, no. 35. 1953. The photographic Perseid meteors. Har- and the values in figure 1 were so computed. vard Coll. Obs. Reprints, ser. 2, no. 47. Orbital Elements of Photographic Meteors By P. Babadjanov1 During the period June 1957 to December meteors were photographed simultane- 1959, 592 meteors were photographed at two ously on one camera, they are distinguished stations of the Institute of Astrophysics of the by an added Roman letter. Academy of Sciences of Tadjikistan, 359 of Date: Day and month. them from both stations; 185 doubly photo- U.T.: Hour and minute in Universal Time. graphed meteors were selected for accurate Shower: Name of the shower. reduction on the basis of trial length and Corr. rad.: Radiant (right ascension a and quality of image. The photographs were declination S) after correction for the taken with cameras NAFA 3C/25 (D=100 mm, earth's attraction and diurnal aberration, F=250 mm); a panchromatic emulsion with in degrees. Equinox 1950.0. sensitivity 1,100-1,300 units of GOST 0.85 Sin Q: Q is the angle between the great circles (approx.=350-400 ASA) was used. traced by the meteor, as seen from two The details of the cameras and the methods stations. of reduction have been described elsewhere e: Elongation of the corrected radiant from (Babadjanov and Sosnova, 1960). The veloc- the apex of the earth's motion, in degrees. ity of the meteor outside the atmosphere was va: Pre-atmospheric velocity, km/sec. computed on the basis of the formula: Vg' Geocentric velocity, km/sec. vB: Heliocentric velocity, km/sec. a: Semimajor axis, in a.u. The parameter k was determined by a grapho- e: Eccentricity. analytical method, and va and c were evaluated q: Perihelion distance, in a.u. by the method of least squares. The instants q': Aphelion distance, in a.u. of the meteors were determined by using a &: Longitude of the ascending node, in degrees. shutter with a gradually changing width. The Equinox 1950.0. error of a time observation did not exceed 1 &>: Argument of perhelion, in degrees. Equinox minute. 1950.0. Table 1 presents orbital data for 132 meteors i: Inclination of the orbit plane to the ecliptic, that have been reduced up to the present. in degrees. Equinox 1950.0. Following are definitions of the column heads r: Longitude of perihelion, in degrees. of the table: k: Whipple's comet-asteroid criterion. aCAP: a Capricornids. No: Meteor number, of the year of observa- IAQU: i Aquarids. tion minus 1900 (the first two digits); 5AQU: S Aquarids. number of the exposure in the current year PER: Perseids. (the next three digits); number of the S.T.: Southern Taurids. camera (the last digit). If two or more LMIN: Leo Minorids. N.T.: Northern Taurids. ' Astrophysics! Institute of the Academy of Sciences of Tadjikistan, U.S.S.R. GEM: Geminids. 287 288 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Table 1 contains many meteors belonging 63; No. 570233 (association 61 or 62); Nos. to well-known showers: 30 Perseids, 4 Genri- 591774, 570556, and 570582 (association 72 or nids, 3 6 Aquarids, 3 Northern and Southern 73); No. 570354 (association 69); and No. Taurids, and 6 a Capricornids. The sporadic 573162 (association 88). meteors also form some groups with analogous Thus, we see that an increase in the observa- radiants, velocities, and orbital elements. One tional material results in a decrease in the group, designated as "I," is formed by the number of the so-called sporadic meteors. meteors Nos. 570673, 570706, 570774, and Evidently the majority of sporadic meteors 582565b, and posesses the following mean belong to various not-yet-known showers. characteristics: Application of Whipplo's K criterion to the a=278?5 fi=127?6 elliptic orbits of sporadic meteors shows that 8= 48?8 e=0.86 a>=202?2 73 percent aro connected with comets (K>0), o.= 27.0 i = 36?2 the rest being of anteroidal origin (K<0). Tablo 1 contains 26 hyperbolic orbits, but The mean coordinates of the radiant are close almost all of them am perhaps due to observa- to that of Maltsev's (1930) radiant No. 76 tional errors. Only one hyperbolic orbit, that near c Draconis (a=279°, 5=55°); moreover, of meteor No. 570707, is probably real. This the Harvard meteor No. 8089 has almost the indicates that the actually hyperbolic orbits same characteristics. constitute less than 1 percent of the whole The second group, designated as "II," material. consists of three meteors, Nos. 572181, 572182, and 572384, and has the following mean References characteristics: BABADJANOV, P. B., AND SOSNOVA, A. K. a= 0?9 a-'=0.33 G= 2?7 1960. Photographic observation* of meteore in 5=_ 8?9 e=0.77 co=79?3 1957. Bull. Astrophys. Inst. Acad. Sci. Tadjikistan, U.S.S.R., no. 29, pp. 3-15. va= 23.7 9=0.65 i= 5?6 JACCHIA, L. G., AND WHIPPLE, F. L. No known visual radiant corresponds to the 1961. Precision orbits of 413 photographic me- mean coordinates of the radiant of this group. teors. Smithsonian Contr. Astrophys., Some meteors can be related to the meteor vol. 4, no. 4, pp. 97-129. MALTSEV, V. A. associations discovered recently by Jacchia 1930. Radiants of the most important meteor and Whipple (1961). Such meteors are Nos. streams. Russian Astron. Calendar, ed. 591444 and 570214, related to the association 4-e (permanent section), pp. 448-453. METEOR SYMPOSIUM—BABADJAXOV 289 E S; 81312 8 5 2 8 8 3 2 8 S S 3 S 8 K 8 3 3 & S S 9 8 8 3 8,. 55 R 2 3IS 2 8 £ * i* f • rj r i" -- i • r • r i -* **J * r•* r r . / .* : ; -* - c « c .- N * i. * — cccr. i ~ ^- s<^*iri &« 3 a88Rft5SSI3.R2 RS,h 6 » 3 8 8 E R « R « R S » « » 2 5 8 3 6 8 „ S « 8 2 «. R o S •• -. - i - ~ » ^ « ' a » « - * n » - -. v- ' i - • .-. « •» « m s r. -' _ o c. r - r;».-«.-- « o V - » ri - t-j « » '7 o =SS888853SIS„5S38««R388582 8e8S88S38S88S3SS85SS3«S 3 * --. - ri -r r. -.-..-.--: V -. ci r _-.-,-,— , — .- -. c, ci ci ri o <£ ,' i; r: ci a ii i fi - -' -r ri r, «' fi d ci pi n s •: f" = li|- iM-| SS • i 'l i ,2 8 ^ — r >•*. s ?: c c c o t- •- / N fi c tt v: N ^ - - r. * at c rt x r- x ?. •; >•*. c t- n - K c T « a a •* 00 y cc f N (3 CO • ci 111 ' 1 ' ' ' 1 1 1 35 '— c; r- %2 — 290 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS 38 8 S 8 9 3 8 2 S S3 8 « S 8 S 8 g 2 S S 8 8 8 S S 8 S S 8 8 S 5 S 8 3 8 8 3 8 " -»' I "J ' ... pj „• » p, • • ^ ' ' ei ' ( I ' ( ' I ft * •< xt |* " ' " * * *" 8«S2«^S^liS22'«^8fi««2Ji^«rt2 2^rt22 22e5«M'*8SfSt2p''°''ic'e'32SS " g o' «j » »' •J « N «»' O' «'» N - 9 • O' 8 M *' 9' • ^ «'«' « C3 V& SSi SSi S8 S 2! 55 aJ 2 2i Si Si S5 Si 8i SI 2 S CO CO 6Q CO CO CO CO *"* CO CO CO CO P3 C3 5*3 1* ^» *T"7" "*' • T rt 7 i"3 7" ~ "TT ' ' 1 3SRS&SSRSS3SS!iS3 O « « o o r- « - to a o> 00 - t» to-. r» i CO jj* • ' g S 5 S § Si 3 5i S S S S S g gj S S S S S" S S 8 S 3 S S S S S g 2 2 S 2 S 9' 9 8' 9 8* S 8* § 8 5 3 2 S2SS £* .5 f * 3"" II « QSCSPSfitOS DOSBSQSKPSOH PHBSPS. CHOSOHOSOSOSOSOSQSBS 03 HHMHH 7HHNHNN WWW H W W H K W W H W w ^ &. i, ft ^ cc««icio«ic«»»ce»ao««««»»««B«Boo»«i»i»»i»»««cooao>o»aooooiooioo«ao>oi I -r* £ <^ c; £ ^ > 2 ^JflQ METEOR SYMPOSIUM—BABADJANOV 291 P3 W H t^ M O Q « • co H^' IQ IH* to V d ri H n ^ Q d n* *f _ id e e e ^ N 10 N a a' H d « d H H « h CS oooo oooo ooo>ntiHHCicot(mn «' N N s i< n H ri « id ^ e a d N' «' ri ci H ci od a N N a n n ^ O O CO g N S N NOJN o> * * n -i ' H M i-i e< oo fi * i-i ti « o - H * ri F< M •<> H M M M S I II ? M **— b id id H N «' o o d «' cc ei § S 2 2 S SS 5 S3 S 9 fe Or oaQHaoo^Hiq 8 So§8S*oeS ^ *"* ^ ® ^ C% ^ 's" ^ ^* *^ S o o oo OJO>OS»CJOOOOOOIOO~-;<^^^.^;CJ ^JJSJSJSJ — "2" On the Structure of the 5-Aquarid Meteor Stream By A. K. Terentjevax The 5-Aquarid meteor stream has some The general study of the radiants shows the exclusive and peculiar properties: complexity of the S Aquarids and their large (1) The perihelion distance 2=0.06 a.u. is area with some "centers" (Pokrovsky, 1922; shorter than that of all other streams (having Astapovich, 1950). A study of the improved g>0.1 a.u.). The theoretical temperature Tof geocentric radiants and all the orbital data the meteor bodies near the perihelion should be gives the following results: T~1100o K (melting point of silicates). It is (1) The S Aquarids show two groups of possible that this high temperature accounts for radiants: one north and one south (main group) the peculiar general appearance of the shower of the ecliptic; the symmetry line is parallel to meteors which are sharp, show no wakes, and the ecliptic and lies 1?3 to the south. The north give off no sparks. The 5 Aquarids are also and south branches are symmetrical relative to observed at the ascending nodes of their orbits the plane of Jupiter's orbit. This may be the as a daytime meteor shower, the Arietids. effect of perturbations. The visually deter- (2) The radiants of this shower cover a large mined third "ecliptic branch" of the 5 Aquarids area, and many other radiants of some minor is, apparently, a group of different radiants of ecliptical streams occur in the same region. many minor streams and of sporadic meteors. (3) The 5 Aquarids are the richest shower of Simultaneous visual and photographic observa- the southern hemisphere, but the number of tions of this group are needed. The new good photographic and visual observations is ephemeris (daily motion) of the improved insufficient. There have been some attempts geocentric radiants is given in table 1. Mc- (Wright, Jacchia, and Whipple, 1956; 1957) to Intosh's (1935) ephemeris is in good agreement obtain the structure of S Aquarids and neighbor- with these data for the southern 5 Aquarids. ing streams (i Aquarids, aCapricornids), based Derbeneva (Stalinabad) in 1960 obtained the upon an insufficient number of double-station following ephemeris for the radiant from photographs; only some of the single-station telescopic meteors (Equinox 1950.0): meteors are genuine S Aquarids. a«=338?8+0°89 (Long, of o-122?4), Much reliable data concerning a great number of the 5-Aquarids' visual radiants have been 5«=-15?4-|-0o39 (Long, of O-122?4). compiled in the U.S.S.R., some with a determi- (2) The period of visibility is apparently from nation of geocentric velocities. These data long, of o = 120° (July 23) to long, of © = 148° were published in Russian, and have remained (August 22), with maximum at long, of insufficiently known abroad. A new paper O = 125°-126° (maximum of southern 8 Aqua- (Terentjeva, to appear) tabulates all double- rids only, about July 29), Equinox 1950.0. station photographs, visual radiants, and radar The 600 fireballs observed in China (10th to observations of the 5Aquarids since their 12th centuries) give for the 5 Aquarids a shift discovery in 1871. The total number of radiants of 4 to 5 days in the maximum date over 1000 is about 600, and they include radiants of the years; the radiant area then was also very large. neighboring showers (i Aquarids, aCapri- Undoubtedly, Jupiter disturbed this stream. cornids, etc). During a period of 28 days the earth passes 'Kiev Astronomical Observatory, V S.8.R. through the stream, which passes very close to 293 294 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS the sun; therefore a very small variation in cannot be explored because of the absence of radiant position gives a great difference in exact observations of the later shower. It is inclination, perihelion distance, and other possible that the structure of the Capricornids elements. The perturbations alone are insuffi- is identical with that of the 8Aquarids: cient to explain all the peculiarities of the Mclntosh's (1935) radiant No. 250 being the observed orbit which may be affected by physi- southern branch, the a Capricornids the cal agents as well. northern one. Mclntosh's radiants Nos 240 The velocities of the photographic 5-Aquarid and 264 apparently fit the ecliptical branch of the Capricornids, but not that of the "south- meteors show a systematic difference AVg of about +4 km/sec relative to the velocities ern lAquarids," as suggested by Wright, obtained by radar (Almond, Bullough, and Jacchia, and Whipple (1957). The ecliptical Hawkins, 1952; Almond, 1952). The semimajor group of "August a Capricornids" (Wright, axes of photographic orbits are thus enlarged Jacchia, and Whipple, 1956) is doubtful. If by a factor of 2. The mass distribution and the the "August a Capricornids" were real, the space density in the stream were studied in photographic error of the radiant would be 6° detail (Terentjeva, 1961); the mass distribution to 8°. The preliminary ephemeris of the a j {M)~Mr%0, the mean space density D(M) = Capricornids (Equinox 1950.0) is: 12X10"9 km"3, in accord with the later determination by Moscow Section of VAGO, ttB=303? 14-0.63 (Long, of o-124?0) Crimea Expedition of 1960. A fine shower of i Aquarids and more than 5B=_10?4+0.17 (Long. ofo-124?0) three other showers are acting simultaneously (photographic obs.) in the same region; to separate them we need double-station photographic and exact visual The association between the a Capricornids observations. The data obtained by Wright, and Comet Lexell 1770 I is more probable than Jacchia, and Whipple (1957) concerning i that with Comet 1948n (=1948 XII) as sug- Aquarids are doubtful: of nine double-station gested by Wright, Jacchia, and Whipple (1956). meteors, some are not i Aquarids. The radiants Comet 1770 I is also connected with some of the i Aquarids are also near the ecliptic, in a ecliptical streams and may be related to Comet very confusing region. There are many 1948 XII (Tisserrand's criterion). The appli- radiants of meteor streams near the ecliptic, cation of Tisserrand's criterion to the study acting successively; they correspond to the of the connection between comets and streams orbits of short period (Astapovich, 1950). gives a considerable dispersion of the C- The study of these streams (i Aquarids, a constant (for a single meteor orbit). Ap- Capricornids, 8Aquarids) therefore demands parently the meteor bodies not only suffer great care. gravitational perturbations, but are possibly The neighboring shower of the a Capricornids influenced by other physical agents. In these has two groups of radiants: the proper a cases Tisserand's criterion cannot be applied. Capricormds (an old meteor stream observed in A lengthy paper on this subject with a the 19th century), and a possibly new group of bibliography of sources (over 100 references) ecliptical a Capricornids. Their correlation will be published in Russian. TABLE 1.—New ephemeria of geocentric S-Aquarid radiants (Equinox 1950.0) Stream Radiants Daily motion Long, of© Type of observation Northern 5 Aquarids as = 334!5 + 0?85 125!0 All photographic «s=-5?4 + 0?35 125T0 and visual Southern 5 Aquarids as=338?5 + 0?88 125T0 «s=-16?9 +0?36 125T0 All photographic METEOR SYMPOSIUM—TERENTJEVA 295 References POKROVSKY, K. D. ALMOND, M. 1922. Meteor showers of the southern hemisphere. 1952. The orbit of the 8 Aquarid meteor stream. In Trans. 1st Meeting of the Russian Jodrell Bank Ann., vol. 1, pp. 22-28. Society of Amateur Astronomers (Mir- ALMOND, M.; BTXLLOUGH, K.; AND HAWKINS, G. S. ove'de'nie'), pp. 172-173. Petrograd. 1952. Radio echo observations of the daytime TEBENTJEVA, A. K. meteor streams in 1952. Jodrell Bank 1961. Results of the 5-Aquarid and Perseid Ann., vol. 1, pp. 13-21. meteor showers observations in 1959. ASTAPOVICH, I. S. Bull. Comiss. Comets and Meteors, 1950. The ecliptic orbits of meteor bodies in the Acad. Sci. U.S.S.R., no. 5, pp. 29-35. solar system. Izv. Turkmen Sec. Acad. WEIGHT, F. W.; JACCHIA, L. G.; AND WHIPPLE, F. L. Sci. U.S.S.R., no. 4, pp. 95-96. 1956. Photographic a-Capricornid meteors. MCINTOSH, R. A. Astron. Journ., vol. 61, pp. 61-69. 1935. An index to southern meteor showers. 1957. Photographic i-Aquarid meteors and evi- Monthly Notices Roy. Astron. Soc, vol. dence for the northern S Aquarids. 95, pp. 709-718. Astron. Journ., vol. 62, pp. 225-233. Masses of Comet Giacobini-Zinner and the Draconid Meteor Stream By Y. V. Yevdokimov1 Comet Giacobini-Zinner is of great interest in 1940, with a tangential component of velocity because of the many valuable observations not greater than 13.7 meters/second. made of the related Draconid meteor showers. We have recently worked at combining obser- A precise study of the comet's motions leads vations of the comet in 1933, 1939, and 1946. to some striking conclusions about the manner The perturbations from five planets, in the in which meteor particles are ejected from the period from 1933 to 1946, have been computed, comet's nucleus. and we have obtained two systems of elements: An earlier paper (Yevdokimov, 1955) on the relation between the comet and the Draconid (1) For observations in 1933 and 1939, meteor stream gave the exact value of the osculation 1939, Oct. 14.0: longitude of the ascending node of the meteor o M0=341 06'57!45±0r22 stream in 1946 at its maximum as a=3.5138522±16 J2 =196°14'3ir6±3!2 (1946.0). e=0.7166606±7 max co=171°47/27?62±0!93 This paper also established the displacement of z=30o44'28T67±0?62 (1950.0) the plane of the central part of the stream's G=196°14'59T75±ir64 orbit. The longitude of the ascending node of the stream was greater than that of the comet. (2) For observations in 1939 and 1946, In 1898 the comet approached Jupiter and osculation 1939, Oct. 14.0: o markedly changed its orbit. Table 1 shows M0=341 06'55!46±3?71 the changes in the perihelion distances since a=3.5139365±63 1897. e=0.7166601±9 In view of these data we may conclude that w=171o47'3l!49±0!93 the earth could encounter only the meteoritic (1950.0) bodies of the stream ejected since 1900. Hence, G=196o14'56T71±l!73 the tangential component of velocity could not be less than 14.5 meters/second. A comparison of these elements shows the The minor distance along the orbit between changes from one revolution (1933-1939) to the comet and the meteors that encountered another (1939-1946), as follows: the earth in 1946 (15.4 days) indicates that / the tangential component of velocity must not AM0=-l r99 be greater than 13.7 meters/second. Aa =+0.0000843 If we suppose that the meteors were ejected earlier (in 1913, 1926, or 1933), the ejection Ae =-0.0000005 velocities would be lower than those of the meteors observed in 1933. Thus we conclude Aco =+3?87 that the main mass of meteors observed in M =—0715 1946 at maximum was ejected by the comet i Kazan University, U.S.S.B. A£2 =—3*04 297 298 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 1.—Changes in perihelion distance Penhelion Year distance (a.u.) 3.04 : 1897 1.20 Afin 16.8 1900 .93 1913 .98 Hence we conclude that during the period of 1926 .99 one revolution, from 1939 to 1946, Comet 1933 1.00 Giacobini-Zinner ejected 17.5 percent of its 1940 1.00 1946 1.00 mass. 1959 .94 This result is consistent with that obtained by Dubyago (1948), who concluded that the velocity of the ejected meteor particles multi- In contrast to the change in the longitude of plied by the ratio of their mass to the mass of the ascending node of the meteor stream et (+16? 8), that of the comet diminished the comet (vt /" ' in cm/sec) has the follow- (— 3? 04) and became equal to ing values: for Comet Encke, 19; for Comet 196°14'll!82±l!73 (1946.0). Biela, 14; for Comet Brooks, 17. If we assume the velocity of ejected particles It is well known that a change in the longi- to be 10 meters/second, we obtain for Comet tude of the ascending node depends only on the Encke 2 percent, and for Comets Biela and force perpendicular to the plane of an orbit. Brooks 1.5 percent of mass lost. Thus, our Hence we may conclude that the changes in the conclusions that a considerable loss of mass longitude of the ascending node of both comet occurs during each cometary revolution are and meteor stream resulted from the same force proved in various ways. at the time of ejection in 1940, and that some component of that force was perpendicular to References the plane of the orbit. This may be expressed DUBYAGO, A. D. as follows: 1948. On secular acceleration in the motion of periodic comets. Astron. Journ., U.S.S.R., vol. 25, pp. 361-368. YEVDOKIMOV, Y. V. AflCom.=r sin (»+«) cosec i „- * 1955. Concerning the association of the Gia- cobini-Zinner comet with the Draconid Ai2 t.=7" sin (»+«) cosec i „ * stream. Astron. Circular, Acad. Sci., me U.S.S.R., no. 159, pp. 21-24. Dynamical Evolution of the Perseids and Orionids By Richard B. Southworth x This report sketches the methods and the products of the differences; in that sense, it conclusions to date in a study of the effects is a first-order computation. of planetary perturbations on meteor streams. The secular perturbations of an orbit by a The research is continuing, and the material planet are the average of the total perturba- will be published in greater detail in another tions by that planet, where the average is volume of these Contributions. taken over all positions of the planet in its own orbit. The effect of secular perturbations Method on a meteor stream is a gradual shift of the Chiefly to be considered here are the differ- whole stream; differential secular perturba- ences among the orbits of different meteors in tions between different parts of the stream may a stream, or equivalently the distribution of gradually further distort the stream. The meteors in a stream. Since these differences secular perturbations of the Perseids and can be directly observed in photographs, the Orionids were computed by Smart's (1953, principal task here is to predict them. Neither pp. 218-231) formulation of Gauss's method. the observation nor the prediction would have The results, especially the differential secular been possible a few years ago; now, however, perturbations, were found to be negligible the Harvard Meteor Program has made compared to the other perturbations to be sufficiently accurate observations, and the discussed next. IBM 704 and 7090 computing machines at The average perturbation of a planet over the Smithsonian Observatory have been able a few revolutions does not differ much from to perform the computing. the secular perturbation. This is not true, The only streams treated thus far in the however, for meteor streams. In particular, study are the Perseids and Orionids. Suffi- when as in the case of the Perseids and Orionids ciently accurate observations have been pub- the stream passes close to the orbit of Jupiter lished for 30 Perseids and 19 Orionids (Wright or Saturn and has a long period, the meteors and Whipple, 1953; Whipple, 1954; Jacchia undergo large perturbations at irregular inter- and Whipple, 1961; Hawkins and Southworth, vals. The effects of these perturbations on 1961). Every meteor within about 10 km/sec the stream are treated here by a numerical in speed, 10 degrees in radiant, and 10 days in method, partly due to Hamid (1950). This date from the shower mean has been included. method, which involves a good deal of com- The observational errors are all probably less puting, is described as follows. than 0.2 km/sec or 0°.2. First, tables are prepared of the perturbation For each stream a reference orbit was com- of a single meteor during one complete revolu- puted for a hypothetical meteor coming from tion from aphelion to aphelion. The values the observed mean radiant at the observed are found by numerical integration along the mean speed on the observed mean date. The orbit. A table is prepared for every perturbing analysis deals primarily with differences from planet that may be considered, and a set of this reference orbit, but neglects squares and initial positions of the planet is chosen, uni- formly spaced around its orbit. The different values in each table are the perturbations i Harvard College Observatory, and Smithsonian Astropby steal Observatory, Cambridge, Mass. that would occur if the planet were at these 299 300 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS different initial positions when the meteor acter. Saturn's perturbations upset the peri- was at aphelion. The chief approximation odicity of Jupiter's, however, at least for the in this entire study is that these perturbations Perseids and Orionids. It is convenient to are computed only for meteors moving in the express the average difference between meteors reference orbit. From another point of view, in the stream by the variances and covariances this is the limitation of the first-order formula- of the elements. The computed variances tion. This approximation should be adequate increase approximately linearly with the time, for the Perseids and Orionids, but will be as they would in the case of a true random discontinued in future studies of other streams. walk. The second part of the perturbation compu- The last part of the perturbation computa- tation is a Monte Carlo computation. This tion is the calculation of the effects of the consists in tracing out the individual perturba- perturbations on the observable part of the tions of a set of meteors chosen at random stream. The observable meteors—those whose from an initial Gaussian distribution. Each orbits intersect the earth's orbit—form only a chosen meteor is treated one revolution at a four-dimensional cross section of the five- time. The initial elements and the initial dimensional distribution of meteors in the positions of the perturbing planets are given. stream. It is convenient to define a set of The meteor's perturbations in its first revolu- descriptive parameters called "local quanti- tion are found by interpolation in the tables ties"; these comprise the (unobservable) mini- previously computed from the initial planet mum distance between a meteor's orbit and positions. Thus the elements at the end of the earth's orbit, and four observable quanti- the first revolution are found, and then, from ties: the right ascension and declination of the the new value of the period, the positions of true radiant, the geocentric speed, and the the planets at the end of the revolution. It is date of observation. To the first order, there then possible to compute the perturbations in is a linear matrix transformation between the second revolution, and so on. This process orbital elements and local quantities, and the is performed about a hundred times on about two modes of description are equivalent. This a hundred meteors. transformation is applied, in fact, to the tables One can theoretically predict the general of perturbations per revolution, and the Monte behavior of the perturbed meteors, but it will Carlo computations are actually performed in be sufficient here to examine the computed terms of local quantities. Thus the time rates results, which agree with the theory. The first of increase of the variances of the local quan- feature is the well-known rapid speed of the tities are computed directly, and the rates of meteors along the orbit of the stream. Of increase of the variances in the observable course, the time required to spread out depends sample can be found by subtracting the parts on initial circumstances, but, for example, a of the overall variances that depend on correla- set of Perseids leaving their comet in the tion with the distance from the earth's orbit. vicinity of the earth with root-mean-square For the Perseids and Orionids, it is found that relative velocities of 30 meteors per second the observable quantities are highly correlated were spread all along the orbit four revolutions with the distance from the earth's orbit; thus later. Consequently, it has not been necessary the variance in the observable sample is much to take account of the distribution in the smaller than in the stream. From another stream along its orbit. point of view, this means that perturbations Once the meteors are well separated, they tend to disperse meteors of both streams undergo independent perturbations which are mostly within the orbit planes. (effectively) a random walk in the elements. It is not a true random walk because, if only Perseids one perturbing planet is considered, the per- The Perseid reference orbit is very similar to turbations have a complicated periodic char- the orbit of the parent Comet 1862 III, except METEOR SYMPOSIUM—SOUTH WORTH 301 TABLE 1.—Differences within the Perseid stream Va Date Observed standard deviation 4?7 1!7 1.0 km/sec 3.4 days as 21.9 1.2 0.5 15.0 Observed variances S B 2.8 -1.3 1.1 Va 1.0 0.5 Date 11.6 Computed increase in as .08 .08 -.04 .08 variances per revolu- SR . 11 -.04 .09 tion (per 100 years) Va .02 -.04 Date .09 Deduced age (years) 27,000 2,600 5,300 13, 200 for a shorter period. The reference orbit passes of the meteors would be somewhat less than 1.3 a.u. from Jupiter's orbit, and only 0.3 a.u. these ages, but the results agree no better. from Saturn's. It follows from table 1 that planetary per- OBBITAL ELEMENTS (1950.0) turbations alone cannot explain the present Reference 1862 III Perseid distribution. Since other processes a 21.56 24.28 can scatter the meteors, but can hardly bring e .956 .960 them back together, the ages in line 10 are q .946 .963 upper limits. If we take the smallest of those, P 100. 1 119.6 and allow an additional 400 years for the initial i 113?2 113!6 w 149?8 152?8 spread along the orbit, the conclusion is that D, 137?9 138?7 the average age of the Perseids does not exceed 3,000 years, and that the overall age is not The first line of table 1 shows the standard likely to exceed 6,000 years. deviation of Perseids from the mean. aR and To explain the remaining variances, there 8R are the right ascension and declination of must be another force that must be nongravita- the radiant, and Va is the geocentric speed. tional and must have large components per- The radiant deviations are from a fixed mean; pendicular to the orbit plane. It is impossible standard deviations from a moving mean to determine the nature of this force definitely radiant are 1?6 in both coordinates. The because only a sample of the stream is observed; next four lines show observed values of the nonetheless it is possible to guess. Guigay four variances and six covariances in the same (1947, 1948) pointed out that the observed units as the standard deviations. Perseid orbits tend to diverge from a point on The computed increase of the variances per the comet orbit before perihelion. Accordingly, revolution (i.e., per century) is shown in lines the hypothesis was made that an explosion 6 through 9 of table 1. On the simplest model took place there in the past, and several Monte for the stream history—all the meteors leaving Carlo computations were performed to study the comet at one time—the age of the stream the subsequent perturbations. By adjusting would be found by dividing the present vari- the date and place of the explosion, and the ances by their rates of increase. line 10 shows relative speed, the best fit to the observed these ages; they are quite discordant. If one variances was found to be with an explosion considers models in which the meteors leave 1.5 a.u. from the sun, 1.3 a.u. north of the the comet at different times, the average age ecliptic plane, roughly a thousand years ago. 638S05—63 -24 302 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS Table 2 shows the computed variances, which TABLE 2.—Computed Perseid variances on the explosion are a passable fit. The root-mean-square hypothesis relative velocity found is 2.6 km/sec, which is SR Vo Date very high, but is actually observed in dust <*B 23. 1 4. 1 2. 5 11.6 outbursts from comets (Whitney, 1955). These SB 2. 0 -0. 1 2.4 outbursts are not well understood, but it is Va 1. 1 0.4 clear that a large quantity of material is ejected Date 10.5 at high velocity; hence some connection with meteors seems likely. the case of the Perseids, line 10 contains upper limits to the average time since the meteors Orionids were spread along the orbit. The observed There has long been a doubtful association dispersions in the declination of the radiant between the Orionids and Comet Halley. and in the speed are both so small that they Although the reference orbit is quite similar are certainly affected by observational errors, in size and shape to the comet orbit, the and should in fact be even smaller. Further- mutual inclination is eight degrees. All the more, it is not difficult to produce small dis- observed Orionid orbits lie in nearly the same persions—for example, by collision with dust plane and are scattered in that plane. The particles (see Whipple, this symposium, p. 239). comet orbit is well outside the apparent Thus one can accept the results from the right distribution of meteor orbits. ascensions and dates without being troubled ORBITAL ELEMENTS (1950.0) by the declinations and speeds. The conclusion is that the Orionids have been spread along Reference Halley the orbit for only 400 or 500 years, and are a 17.80 17.95 e .968 .967 perhaps 700 years old in all. q .569 .587 P 75. 1 76.0 Conclusions i 164T7 162?2 The perturbations of individual meteors in a H 82?6 111?7 29?1 56?7 stream can be much more important than the secular perturbations of the stream as a whole. Table 3 shows the observed and computed The dust outbursts from comets deserve more differences within the Orionid stream. As in attention as a possible source of meteor streams. TABLE 3.—Differences within the Orionid stream «B Vo Date Observed standard deviation l?03 0?48 . 87 km/sec 1.23 days a« 1.06 -.03 .57 .72 Observed variances 8& .23 .02 .25 Vo .75 .51 Date 1.52 «B .170 .006 -.019 .228 Computed increase in 8S .0004 -.001 .008 variances per revolu- Va .002 -.025 tion (per 75 years) Date .304 Deduced age (years) 470 44, 000 26, 000 380 METEOR SYMPOSIUM—SOUTHWORTH 303 References JACCHIA, L. G., AND WHIPPLE, F. L. GUIGAT, M. G. 1961. Precision orbits of 413 photographic me- 1947. Recherches sur la constitution du courant teors. Smithsonian Contr. Astrophys., d'etoiles filantes des Perseides. Journ. vol. 4, no. 4, pp. 97-129. Observateurs, vol. 30, pp. 33-48, 69-87. SMART, W. M. 1948. Recherches sur la constitution du courant 1953. Celestial mechanics. Longmans, Green d'etoiles filantes des Perseides (suite). and Co., New York. Journ. Observateurs, vol. 31, pp. 1-32, WHIPPLE, F. L. 49-64, 77-94. 1954. Photographic meteor orbits and their HAMID, S. E. distribution in space. Astron. Journ., 1950. The formation and evolution of the Perseid vol. 59, pp. 201-217. meteor stream. Doctoral thesis, Har- WHITNEY, C. A. vard University. 1955. Comet outbursts. Astrophys. Journ., vol. HAWKINS, G. S., and SOUTHWORTH, R. B. 122, 190-195. 1961. Orbital elements of meteors. Smithsonian WRIGHT, F. W., AND WHIPPLE, F. L. Contr. Astrophys., vol. 4, no. 3, pp. 1953. The photographic Perseid meteors. Har- 85-95. vard Coll. Obs. Reprints, ser. 2, no. 47. A Note on the Cometary Nature of the Tungus Meteorites By V. Fessenkov1 The Tungus meteorite appeared in the early Northern Hemisphere, its total amount could morning of June 30, 1908, at 0b17m UT near be roughly estimated to be 106 tons. Never- Vanovara, Central Siberia (X=6h47m28, A Short Note on the Origin and Age of the Quadrantids By S. E. Hamid * and Mary N. Youssefl The Quadrantid shower is distinguished a minimum value «0.1 a.u.) around 1500 among the periodic meteor streams by its very years ago (time of minimum inclination), and short duration ( TABLE 1.—Orbital elements of Quadrantid meteors Trail no. o (a.u.) e i (degrees) to (degrees) Q> (degrees) x (degrees) Reference 9945 3.046 0.682 6& 6 165.2 282.5 87.7 (1) 9985 3.074 0. 683 70.8 16a 3 282.5 90.8 (1) 9953 2.906 0.664 72. 7 170.3 282.5 92.8 (1) 9983 3.002 0.675 72.4 169.6 282.5 92. 1 (1) 9974 2.999 0.674 72.5 170.4 282.5 92.9 (1) 9907 2.880 0.659 72. 1 177.2 281.4 9a 6 (2) (1) L. O. Jacchia, and F. L. Whipple, "Precision Orbits of 413 Photographic Meteors," Smithsonian Contribution to Astrophysics, vol. 4, no. 4, pp. 97-129,1961. (2) O. 8. Hawkins, and R. B. Southworth, "Orbital Elements of Meteors," Smithsonian Contribution to Astrophysics, vol. 4, no. 3, pp. 85-95,1961. 309 310 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS TABLE 2.—Angular orbital elements (in degrees) of the Quadrantids in the past Trail no. Time (years) 9945 9985 9953 9983 9974 9907 i CO i CO a i CM i eo a » (0 a i a 0 69 165 282 71 168 282 73 170 282 72 170 282 73 170 282 72 177 281 -400 65 162 284 68 164 284 71 165 284 70 164 284 70 165 284 71 169 283 -800 58 160 288 60 163 287 67 162 286 64 163 286 65 163 286 68 164 284 -1200 42 157 295 40 159 295 58 161 289 50 161 292 51 161 291 62 162 287 -1600 15 107 351 14 66 394 36 153 302 17 128 331 17 128 331 46 158 294 -2000 37 24 440 47 18 449 22 42 420 41 21 445 41 20 445 14 94 365 -2400 57 18 450 63 16 455 53 19 449 62 16 454 62 16 454 46 20 446 -2800 66 17 454 71 14 457 67 17 454 71 15 456 71 15 457 64 16 452 -3200 71 14 456 74 9 458 73 15 456 74 12 458 75 11 458 71 14 456 -3600 73 8 457 75 — 1 459 75 11 458 76 4 459 76 3 459 74 9 457 -4000 73 -2 459 74 -10 461 76 3 459 76 -6 460 76 -7 460 75 1 458 -4400 72 -11 460 71 -16 463 76 -6 460 75 -13 461 74 -14 462 75 -8 459 -4800 70 -17 462 65 -17 465 75 -13 461 71 -17 463 70 -17 463 73 -14 461 rA Radius vector at ascending node r_ Radius vector at decending node (au.) — r. 5.0 rD - \ \ / N \ 4.0- \ \ / \ I | 3.0- 1 / I / 1 / 2.0- 1 / — 1 / * / 1.0 _ 1 / ^ — — ^ >v 1 1 • N, • N \ . 1 ^' >^ 1 1 0 -1000 -2000 -3000 -4000 -5000 -1000 -2000 -3000 -4000 -5000 T (years) T(years) FIGURE 1.—Quadrantid meteor no. 9983. Radius vectors FIGURE 2.—Quadrantid meteor no. 9983. Orbital inclina- at the nodes plotted against time in the past. tion t plotted against time in the past. METEOR SYMPOSIUM—HAMID AND YOUSSEF 311 180 160 - O -1000 -2000 -3000 -4000 -5000 T (years) FIGURE 3.—Quadrantid meteor no. 9983. Longitude of perihelion x plotted against time in the past. -20 0 -1000 -2000 -3000 -4000 -5000 T (years) FIGURE 5.—Argument of perihelion w plotted against time in the past. -1000 -2000 -3000 -4000 -5000 T (years) FIGURE 4.—Quadrantid meteor no. 9983. Perihelion dis- tance q plotted against time in the past. Abstracts of Other Papers Southern Hemisphere Vela-Puppid Meteor the importance of short-range echoes is empha- Activity: C. Ellyett and K. W. Roth 1.—It has sized. Comparison of results obtained from gradually become established from radar surveys both hemispheres reveals a considerable meas- at Christchurch, New Zealand, in 1953, 1956, ure of agreement between various surveys. and 1960, that intense meteor shower activity The Seasonal Variation oj the Twilight Sodium occurs annually in early December, with radi- Air glow in the Southern Hemisphere:3 B. A. ants in both the Velid and the Puppid constel- Tinsley4 and A. Vallance Jones.8—The plateau lations. The Velid group has a mean right brightness of the sodium twilight airglow has ascension of about 133°, and the Puppid group been measured near Christchurch, New Zea- is somewhat widely spread around 110°. The land, by photographic spectroscopy between declination of both showers is about —45°. July 1960 and September 1961. A seasonal The Absence oj Detectable Infrared Radiation variation with a maximum in mid-winter and from Meteor Trails: B. A. Tinsley, J. J. Tait, a minimum in mid-summer was observed. and C. Ellyett1.—A lead sulphide cell immersed This is similar to the seasonal variation in the in dry ice was used in conjunction with an northern hemisphere, indicating a six-month arsenic trisulphide lens to look from ground phase difference between the two hemispheres. level for transient infrared bursts radiated from The annual distribution of meteors at high the trails produced by large incoming meteors rates has just been determined at Christchurch, in the upper atmosphere. No bursts were New Zealand, for a complete year. The south- detected. A theoretical analysis was then ern hemisphere meteor distribution is out of made of the probability of detection. The phase with the northern hemisphere, and both mechanism of infrared radiation from meteor are out of phase with their own sodium distri- trails is largely unknown, so assumptions must butions. Thus, unless the concept of long-term be made in order to give a figure for radiated storage of sodium in the upper atmosphere is power. It is then possible to calculate the in- introduced, it is unlikely that upper atmospheric frared flux at the ground. The conclusion sodium is of meteor origin, at least from meteors reached was that, with the equipment available, of a size detectable by radar methods. an average of 200 hours of clear night-time All-Sky Meteor Rates in the Southern Hemi- observing would be required to detect a single sphere:6 C. Ellyett and C. S. L. Keay.1—An meteor. extensive experiment has been carried out The Latiivde Dependence oj Radar Meteor at Christchurch, New Zealand (43°37'S, Shower Observations:2 C. S. L. Keay and C. 172°24'E), to determine meteor rates over the Ellyett.1—A simplified method of assessing the visible sky, using radar equipment with all relative significance of meteor showers is dis- parameters maintained constant throughout a cussed. Most of the tables of meteor showers complete year. The effect of varying back- quoted in the literature give a false impression ground noise was eliminated. of the actual influx of meteoric shower particles From a total of 724,681 meteors, a seasonal into the earth's atmosphere, particularly when variation has been established, with a peak rate measurements are quoted. The bias in- rate in local summer. The meteor rate at herent in single-station rate measurements is 69.5 Mc/s was found to be unaffected by iono- largely eliminated by the present method, and » Published hi Journ. Atmosph. Terr. Phys. vol. 24, pp. 345-351, 1962. i The authors are associated with the University of Canterbury, Christ- • University of Canterbury, Christchurch, New Zealand. church, New Zealand. • University of Saskatchewan, Canada. > Published in Journ. Oeopbys. Res., vol. 66, pp. 2337-2343,1961. • Published in Journ. Geophys. Res., vol. 66, pp. 2590-2591,1961. 313 314 SMITHSONIAN CONTRIBUTIONS TO ASTROPHYSICS spheric behavior, such as increased absorption, An additional 9 to 10 radiants, however, are magnetic storms, or aurorae. found by each survey alone. The Identification of Meteor Showers with The errors inherent in the identification of Application to Southern Hemisphere Results:7 meteor showers from radar data are considered. C. Ellyett, C. S. L. Keay, K. W. Roth, and Different techniques have been used at Christ- K. G. T. Bennett.8—Further meteor shower church and at Adelaide. It is concluded that results observed by radar at Christchurch, New no single method of analysis is fundamentally Zealand, are now given, including results at better than any other. high rates, and all Christchurch data are The technique of shower identification compared with visual observations, and radar depends partly on the nature of the back- results from Adelaide, for the declination range ground rate, which may bo due either to from +26°.5 to about —50°. Seven meteor sporadic meteors, or to the resultant of many showers appear in all three surveys, and can minor radiants. The resolution of this un- be regarded as certain, and another 6 to 9, certainty is necessary before the occurrence of which are each reported twice, are probable. less active showers can be confirmed. The rate-count method has been extended to ' Published In Monthly Notices, Roy. Astron. Soc., vol. 123, pp. 37- multiple range bands, resulting in a useful 60, 1961. additional means of sorting out the significance 1 The authors are associated with the University of Canterbury, Christchurch, New Zealand. of peaks in the rate curves.