7% MONTHLY WEATHER REVIEW Vol. 98, No. 10UDC 551.607.321.2:527.62:528.288:550.386.4:523.165(~)POSITION DETERMINATION METHODS USED TO TRACK SUPERPRESSURE BALLOONS J. E. BLAMONT,* T. F. HEINSHEIMER,t and J.-P. POMMEREAU**Service d'Aeronomie CNRS, Verrieres le Buisson 91, France, and tThe Aerospace Corporation, El Segundo, Calif.ABSTRACTThe method that has been used in the past to determine the position of superpressure balloons employed in long-term meteorological and technological experiments-projects GHOST (Global Horizontal Sounding Technique),EOLE (French, God of the Winds), SOMEX (Solar Monitoring Experiment), etc.-was to measure the solar eleva-tion during the daylight hours and then compute the balloon position at local noon: latitude by the noon solar altitudeand longitude by the time of balloon noon. This method has disadvantages: it requires a lengthy series of measure-ments during the day and has degraded accuracy when the maximum solar angle approaches 90'. Another methodwas therefore devised in which the solar angle data are complemented by data that indicate the local geomagneticrigidity; the correlation of the two data yields an improved position determination algorithm. This method wasapplied to the trajectory determination of the SOMEX balloons with considerable success.1. BALLOON LOCATION BY MEASUREMENT OF SOLAR ELEVATIONThe classic method of nautical navigation is based uponmeasurements of the angular altitude of heavenly bodiesabove the horizon. Each such measurement yields a locusof possible positions having t,he form of a circle on theterrestrial globe whose center is t,he substellar or subsolarpoint and 11-hose angular radius is the complement of themeasured stellar (solar) ele~otion.The position of a ship is determined either by quasi-simultaneous "shooting" of several heavenly bodies or byshooting a single body at, given intervals as it movesacross the sky. The intersection of the resulting minorcircles yields the ship's position.Figure 1 illustrates this concept as used to providetracking data for the superpressure balloon test flights.A simple solar angle sensor mas incorporated in each pay-load, allowing t,he instantaneous solar elevation to be tele-metered. Since only a single heavenly target was used,these measurements had to be made repeatedly during thedaylight hours. This inherent limitation led to the follow-ing dilemma :1. If measurements were too close together in time, theintersection of the loci were nearly parallel resulting inlarge uncertainty in position determination.2. If the measurements were too widely spaced in time,the geographical displacement of the balloon (up to 1000km between sunrise and sunset) caused erroneous inter-section of the loci.This paper discusses the methods adopted by the au-thors to improve location accuracy for the SOMEX 11(Solar Monitoring Experiment) flights (Blamont et al.1969). That flight series included t,hree balloons that werelaunched to the 100-mb level from Pretoria, Republic ofSouth Africa, in May-June 1968 and flew for 4 to 6 mo.These lengthy flights provided a data base for the eval-uation of the navigational techniques to be discussedbelow.This paper mjll shorn how the measuremenk were made,the method by which the data was transformed into esti-mates of balloon position, and the probable errors. Possi-ble improvements of tJhis navigational system through theexecution of other simple measurements will then bediscussed.SOMEX II SUN MEASUREMENT INSTRUMENTATIONThe sensor used was a GHOST (Global HorizontalSounding Technique) solar dtj tude det,ector furnished byV. Lally of NCAR (National Center for Atmospheric Re-search) at Boulder, Colo. (Lichfield and Frykman 1967).As shown in figure 2, a photoconductive cell is mountedinto an opaque cylindrical structure. A filter is placed inSOLARFIGURE 1.-Solar angle method of balloon position determination.October 1970 J. E. Blarnont, T. F. Heinsheirner, and J.-P. Pomrnereau 757--PLASTIC HOUSINGPHOTO RESISTORy 20- IO>W-FIRST DAY'S FLIGHTTYPICAL OUTPUT OF SUN ANGLE SENSORFIGURE 2.-Exploded view of NCAR sun elevation sensor.TYPICAL PAYLOAD IN FLIGHT POSITION"---SUN ANGLESENSORAUXILLARYSOLARCELLSSOLARPANELTHERMALSKIRT(PARTIALLYREMOVED)t.he optical pa.th to restrict optical sensitivity t,o the 6600to 7500 8 band. Three opal glass diffusers inserted be- BOTTOM VlfW (WITH SKIRT RfMOVEDltween the filter and the photoconduct,ive cell compensatefor the inherent anisotropic response of the photocell,eliminating sensitivity to solar aziniuth angle. The re-stricted neck of the instrument improves sensitivity of t.hedetector at high sun angles. The detector is thereforemeasuring the incident solar illumination impinging uponthe horizontal surface of the filter. This illuminabion varies COUNTERSas the sine of the solar elevation angle.GEIGERPAYLOAD CONFIGURATIONThe detector (Heinsheimer and Pommereau 1968) wasmounted into the SOMEX nacelle (fig. 3) so that its axiswas within 5 0.5' of the vertical with the nacelle sus- k. , '.'pended as in flight,. Except for the occasional shadolvsimpinging upon the detector from the supporting tripod, FIGGRE 3.--SOhlES nacelle.or from the balloon itself (which occult the det.ector atvery high solar angles), the detector has a continuousview of the sun during daylight hours. As the solar cellpower supply provides sufficient power to drive tho trans- Iet,ters. For t~he three SOMEX I1 balloons ment.ionedmitt.er when t.he solar elevation is above 3O to So, sun above, the identifying groups n-ere- ELECTRONICCIRCUITS._ ..f .angle data are transmitted continuously for nearly theentire day.Each balloon is equipped with a payload whose designis based upon the NCAR payloads used for the GHOSTprogram. The payloads process onboard measurementsin the following manner. The measuring sensor (solarelevation det,ector, for example) controls the frequencyof an oscillator (fig. 4). The oscillator frequency feeds aMorse code letter forming circuit, such that the repetitionrate of t.he Morse code letter being formed is a monotonicfunction of the input frequency. This signal is then usedt,o key a HF transmitter (generating 100 t,o 1000 nlW RFpower). Since four independent measurements are madein these flights, an onboard programmer. sequences t,hesensors so that each is monitored continuously for 30 sec.Each balloon is identified by its particular group of code~7m712, ZA~XR, ZMYRwhere tho first lct,t,er of each series represented the solarolevat,ion, the second and fourth letters (~14 and R) reprc-sented two Geiger count.ers, and tho t'hird letter identifiedthe measurement, of atmospheric temperature.DATA HANDLINGThe signals are received at stations of the French andNCAR balloon tracking net.works (fig. 5). There, eachreceived signal is processed manually; an operst,or identi-fies a balloon by its code lett,ers and measures the rate ofletter t,rnnsmission with R stopwatch. This process iscontinuously repeated as long as t,herr are balloon signalsbeing received. The tracking rangc of nn individual stat,ion758MONTHLY WEATHER REVIEWvaries from 3,000 to 12,000 mi, depending upon stationequipment, operator acuity, and HF reception conditions.The data are manually recorded on standard formsand tIansmitted to the data reduction center in Paris.SUN ANGLE aSENSORGEIGERCOUNTERVICTOREENGEIGERCOUNTERFORMING MITTERAABBTHIRTYSECONDTIMER-AA BB-90' 120' I50'EAST 180' 150'WEST 120'80'60'40'20'0'2 oo40'Vol. 98, No, 10There, the data are plotted (fig. 6) and, after bad pointsare rejected, the calculation of balloon position isaccomplished.As a single position per day is sufficient to define theballoon's trajectory, the simplest technique is to deter-mine that position at balloon "noon." This can be doneeither manually or automatically (Solot 1968). In eithercase, for a balloon experiencing no simcant accelerationand no large north-south velocity component, the noonposition is determined by the axis of symmetry of theday's solar elevation curve. The GMT time associated withthat axis yields the balloon longitude, and the measuredsolar altitude at that time determines latitude.Should the balloon have either a significant north-southvelocity component or an acceleration during the day,the solar elevation curve will be skewed. Although thiserror can be partially compensated for by a sophisticateddata-processing technique, the procedure is inherentlyiterative and of limited value.Positional accuracy (whether automatically or manu-ally determined) is limited by the quality of the datajudged by three criteria.1. Is the distribution of a day's data sufficientlyextensive to permit tracing of the entire day's solar curveand thereby to determine the noon solar position?2. Is the scatter of data points small enough to precludegross errors?3. Is the knowledge of the calibration curve of thedetector sufficient over all portions of the day's solarcurve?90' 60' 30' WEST 0' EAST 30' 60' 90'FIGURE 5.-Tracking stations participating in SOMEX I1 (ellipses indicate nominal coverage).October 1970J. E. Blamont, T. F. Heinsheimer, and J.-P. Pomrnereau-50--40u- -2 30---0_IWz3 -v)20--IO-00600 0800 1000 1200 1400 1600 1800GMTFIGURE 6.-Typical sun angle curve.Experience has shown that fact.ors (1) and (2) arc wellin hand and do not usually contribute serious errors. Theprecision of the calibrat,ion curve (3) is generally found tobe the greatest source of error.SUN ELEVATION SENSOR CALIBRATIONTho calibration curve of a given balloon is primarily afunction of the solar angle, and a lesser function of balloonaltitude (atmospheric filtering) ~ payload t,emperature(thermal drift of det,ector and electronics), season (solardistance) , and flight duration (sensor aging).Because t,hese error-producing factors are difficult tosimulate, it is necessary to calibrate the sensor in flight..The balloon is launched so that it arrives at ceiling altitudein the early morning. It is thereby able t'o transmit it,sobservations of the entire day's sun angle data. These dataare compared Tvith the actual sun angle seen by the balloonas determined by either radar observation of balloorlmotion or t,rajectory predictions based on quasi-simul-taneous radiosonde wind mensuremcnt's. Comparison ofthe actual sun angle dat,a with measurements made by theballoon establishes a calibration curve of sun anglemeasurements over the range of sun angles seen during thefirst day's flight. This range is bounded at the top by themaximum sun angle seen during t,he day (near 45' for theSOMEX balloon flown from Pretoria in May) and at thebottom by the sun angle at loss of signal (typically 5'to 10").759I I ITEMPERATE ZONE-l3~ " WESTERN HEMISPHERF80 60 40 20 0MAXIMUM SOLAR ELEVATlON ANGLE (degiFIGURE 7.--Nominal latitude errors associat,ed with position determination.Further extension of the calibration curve can beimplemented on subsequent days by using computer-derived standard solar elevation curves. This process isextremely sensitive t,o balloon velocity which dist'orts thrsolar angle curve. As n result,, effect,ire calibrnt'ion is onlymaintained in the range of sun angles seen during the firstday and in t-he linear region immediately adjacent, to it(fig. 2). An addit,ional calibrat,ion point is available forextremely high solar elevation angles. This is t,he angle at.which t,he sun angle sensor begins t,o be occulted by theballoon. Occultation is clearly visible in the data as asudden decrease in the apparent solar eleration angle.The angle is clearly a function of the gcametry of theballoon ensemble (balloon diameter, payload-t,o-balloondistance) and is t.ypicallp of the order of 80'.ACCURACY OF POSITION DETERMINATIONGiven the method previously described, one can det.er-mine tjhe position of the balloon to a certain level ofaccuracy. The nominal errors associat.ed 1vit.h positiondetermination arc shown in figure 7. Note that there areessentially threr zones-those in which the maximumsolar angle is high (tropical zone), intermediate (tem-perate zone), and low (polar zone). The exact geo-graphical location of thcse zones is a function of theseason, or more exact,ly of the solar declination.The curve shows the typical errors associated with thedetermination of a single day's position as a function ofthe maximum solar elevation seen on that day. In somecases, much better accuracy is achieved, particularly ifextensive data is available on days immediately preceding760 MONTHLY WEATHER REVIEW Vol. 98, No. IOand following. The curve is meant to indicate the trend inaccuracy variation with solar angle rather than to give anabsolute value of attainable accuracy.In the polar zone, accuracy is limited by three factors:(1) the solar angle varies very slowly during the day, makin:the determination of balloon noon difficult; (2) the calibra-tion for very low sun angles may be poor; and (3) the strengthof the signals received from the balloon may be too weak topermit reliable tracking. n7eak signal strength is a resultof the low solar angle impinging upon the horizontal solarpanel. This problem is circumventd on some of the NCARflights through the use of nearly vertical panels containingsupplement'ary solar cells. As the SOhlEX balloons merenot expected ho fly in the polar region, the horizontal solarcell array was deemed sufficient.In the intermediate zone where the calibration is goodand the data reception is extensive, position accuracy is atits best. In the case of a balloon for which ext'ensire dataare available over many continuous days, accuracies ofbetter than 1' in both latitude and 1ongit)ude are possible.In the tropical zone, longitude is accurate to 11-ithin lo,but three factors combine to limit t,he accuracy of lati-tude det,ermination. These factors are:1. Poor calibration accuracy (as mentioned above).2. Insensitivity of the sensor output to variation of thesolar angle above 75' (due to the rclatirc insensitivity ofthe cosine of the solar angle in that. range). Note t'hat thisproblem can be partinlly overcome by use of the balloonoccultation data.3. The ambiguity of measurements made by balloonsnearly under the sun. A balloon found 10' north of the sunwill t,ransmit essentially the same solar elevation data as aballoon found 10' south of the sun. It is t,herefore possiblet'o make gross errors in position determination (tens ofdegrees). Such crrors occur when in the course of its flightthe balloon comes close to thc latitudc corresponding tothe solar declination. Subsequent day9 spent mithin=lOOof that latitude are subject to t'his error. If the balloorlsubsequently diverges from this area, the errors can becorrected by measurement of the day's duration (which isradically different 30' north of the sun then it is 30'south). For experiments in thc Tropics, where the balloondoes not subsequently diverge from t,his area, or where itmay cross "under the sun" several times, t,his factor is aserious limitation.2. IMPROVED NAVIGATIONAL METHODSThe positioning of balloons will be improved by up tot1V-o orders of magnitude when the balloon-satellite relays(Fourrjer et al. 1966) such as EOLE (French, God of theWinds) and Ximbus are flying. Even without a satellite,however, some improvements are possible.For eliminating the need for eshaustive data collectionand processing 11011- necessary to arrive at a single balloonposition, two independent measurands could be sampledsimultaneously, and the resulting intersection of locicould be used to give the instantaneous balloon positionorthogonality of intersection of the two loci. Such pairsof measurements could, far example, combine solar eleva-tion with magnetic dip angle, or solar azimuth. Aninstrument for effecting the magnetic measurement hasbeen developed by Lally (1969) for the GHOST flights.Another instrument that measures the solar azimuth(with respect to t,he local magnetic north) mas developedby one of the authors.The SOMEX I1 data showed, however, that an im-provement could be a,ccomplished without suffering thecomplexity (and weight) of an additional instrument andassociated electronics. It was found that the sensorscarried to fulfill the scientific objectives of the missioncould also serve to improve the navigational accuracy.POSITION DETERMINING BY MEASUREMENT OF SECONDARY COSMIC RADIATIONEach of the SOMEX I1 flights \I-as equipped with twoGeiger counters to perform a scientific mission thatrequired monitoring of secondary radiation at an altitudeof 55,000 ft. This esperiment, was primarily designed topermit detection of radiation increases associated withsolar flares. The data served well, however, in reducingthe posit'ion errors. For using the data appropriately,the geographical dependence of the steady-state radiationlevel (cosmic secondaries) must be well known, and errorsdue to the major perturbing phenomena must beeliminated. It mill be shomn below that these twoconditions were satisfied.The radiation level at t'he 100-mb level is n steady-state function of the earth's magnetic field and of themass of air above that level. The magnetic field in-fluences the local radistion level by tending to deflectthe incoming charged particles toward the geomagnetic,poles. To describe quantitatively the relative deflectionexperienced by such particles, one may convenientlydefine the particles' "rigidity" (P) given by the formulaP=- (E2+2?noc2E)"*where e is the elect.ron charge, m.o the particle rest mass,E the particle kinetic energy, and c the velocity of light.An evaluation can then ba made of the rigidity requiredof a particle to reach a given spot over the enrth's surface(above the atmosphere). This leads t30 t,he determinationof a "gcomagnctic cutoff" value that is, for each locationon earth, the minimum rigiditj- required of a chargedparticle (incident from a given direction) to reach thatlocation.Quantitative evaluation of the geomagnetic cutoff wasfirst dona by Stormcr using a magnetic dipole model ofthe earth's field and improved by Quenby and Wenk (1962)who added non-dipole terms. Figure S shon-s the resultingvalues of the geomagnetic cut,off for rcrtically incidentparticles as a function of geographical position. Xoto thatthe relative symmetry of the lines of "isorigidity" in thesouthern regions is distorted by the geomagnetic anomaly1ewith an accuracy limited by instrument error and the in the South Atlantic.vOctober 1970 J. E. Blamont, T. F. Heinsheimer, and J.-P. Pommereau 76150I2: 60 WEST 0 EAST 53 / 17FIGURE 8.--Nominal values of vertical rigidity (in GT').The other fundamental parameter mentioned above,which influences the steady-st,ate radiation level, is thoairmass above the level at which the measurement's aremade. At the top of the atmosphere, the radiation ledis due to incident cosmic primaries. Decending through theatmosphere, the primaries are filtered out by collision, butthe overall radiation level increases due to the augmentedquantity of secondary particles. A maximum int,ensit,y levelis reached, the "Pfot,zer maximum," below which the radia-tion level decreases rapidly. This maximum is found at the80-mb level (17500 m) in the polar areas and at 160 mb(13300 m) in the Tropics (Bet,hery and Legrand 1965).The descent of the maximum level toward the lowxlatitudes (areas of high rigidity) is due to the increasedrigidity of primary part,icles impacting in t!he equatorialregion and t.he resulting greater depth of t,hcir atmosphericpenetration.It is seen therefore that both the geomagnetic cut,off andt,he atmospheric absorption effects depend primarily onthe particle rigidity. For balloon flight,s at constantpressure altitude, the rigidity determines the st,cady-stateradiation level. Although a superpressure balloon float,s ata constant density level rather than constant pressuro, thisdiscrepancy results in an error that, if corrected by usingthe standard atmospheric tables, becomes negligible.The shady-state radiat.ion levo1 has superimposod upnit numerous perturbations. This paper will not discuss t,hegeophysical' and astrophysical phenomena involved, butwill indicate the quantitative effects of position determina-tion. Discussion of the phenomena may be found in t,howork of Dorman (1963) and Hofmann and Sauer (1968).These authors shon- that all perturbations (whethercyclical or occasional), except solar eruptions and gco-magnetic st,orms, have an impact on the st,eady-st,ateradiation level of less t,han 2 percent.The two major perturbations am both associated withsolar flares. Within a few minutes (or hours) of thebeginning of a major flare, a proton shower may bedetected. Such showers of protons having energies1 .. ....9....... .G.. .5! 5":E-!ibi Q!*;,;,T" GbFIGURE 9.-Typical Geiger count ratc vcrsw rigidit,!- (calibra- tion curve).400 MeV (million electron volts) last only afew hours and influence radiation intmsity onlyat high geomagnetic latit,udes. This will rarely affectthe balloon-borne moasurcments. Subsequently, magneticstorms may be experienced when the plasma cject,cd fromt2he flare reaches tho eart,h (1 t,o 2 dsys after flare com-mencement.). For our purposes, the importance of such astorm is measured by the resulting reduction in cosmicrsdiation (Forbush effect) as monit,orcd by a network ofground-level neutron counters. A 5-percent reduction inthe neutron count, ratc is considered sufficient to invalidatea day's balloon-borne measurement of steady-stato radia-tion. This occurs not more frequently than l or 2 daysper m0nt.h.TRAJECTORY DETERMINATION PROCEDUREThe first steps in the trajectory analysis are aimed atestablishing a pair of accurate calibration curves, the762MONTHLY WEATHER REVIEW Vol. 98, No. 10I I I I1200 00 I Oo E 00 1200 1800L 60'FIGURE 10.-Trajectory for balloon ZRXM.first yielding the response of the solar angle sensor to solarelevation and the second yielding the Geiger countingrate as a function of vertical rigidity. This is accomplishedas follo~~: on the first fiight day, the solar sensor is cali-brated using the procedure described earlier (fig. 2).Also measured on the first day of flight is a single pointon the Geiger counter calibration curve (point 1 of fig. 9).In the case of a launch from Pretoria, this point mould fallat a vertical rigidity of =8.5 GV (gigavolts), see figure 8.The errors associated with this point are due to scatteringof the Geiger counter measurements and uncertainty ofballoon position.After establishing the initial calibration points on bothcurves, the procedure is then aimed at both expandingthese curves to their respective limits and on continuouslyreducing the calibration errors. On a subsequent day,when the balloon is found at a latitude such that its noonsun angle is within the calibrated range of sun angles,another calibration point is determined on the Geigercounter curve. For example, if on a given day the solardata showed a position of 30'S., 120' W., the correspond-ing vertical rigidity is found (from fig. 8) to be 12.5.Geiger counter data observed during the interval of noonl hr are then used to give a calibration point for arigidity of 12.5 (point 2 of fig. 9).Note that due to the unsymmetrical nature of the iso-rigidity lines, the range of rigidities that can be calibratedby a balloon at 30 S. ranges widely, from a rigidity of 4.5to 13.1. This asymmetry is critical to the procedure.Having extended the Geiger calibration curve throughuse of the solar angle curve, the reverse process can thenbe accomplished. On a subsequent day when the balloonis found to be at a position such that the solar angle exceedsthe calibrated range and the Geiger counter rate is withinits calibrated range (between points 1 and 2), the positioncan be determined by the intersection of the two knownloci; that is, the relevant isorigidity curve (fig. 8) and themeridional line indicated by the axis of symmetry of theday's solar cuve. This position permits computation of 8new point on the solar calibration curve. This iterativeprocedure is then continued indefinitely, allowing ex-tension and refinement of both calibration curves as well asdetermination of daily balloon position. In the case oflong flights, the continuous reprocessing of positions as thecalibration curves are improved will further upgrade theaccuracy of the trajectory. This also guards againstsudden calibration shifts due to instrumentation or pro-cessing problems that might otherwise go undetected.ACCURACY OF POSITION DETERMINATIONDetermination of ballon position using the sun angleand Geiger counter data results in significant sccuracyimprovements (fig. 7). The improvement will be discussedbelow for the three latitudinal zones previously discussed:1. In the polar region, the Geiger counter method isnot very sensitive to variations in latitude. It does, how-ever, alleviate the necessity of precisely detwmining thetime of balloon noon as 11-ould be required if sun angleonly were rmployed. It thereby allows the measurementof balloon posit.ion with only a, sparse amount of data. Asdata reception from balloons in the polar regions is oftensporadic (due both to weak transmitter power and poorcoverage of the polar region by t,he ground stations), thomethod would permit more frequent positioning of polarballoons.2. In tho temperate region, the use of solar angle dataalone is usually sufficient. The Geiger counter data is ofvalue for the det.ection of unexpected shifts in the solaranglo calibration curves (due to electronic difficulties).The method thereby adds to the credibility of the com-puted trajectories of these balloons.3. In the tropical region, the method of position deter-mination using both types of data is of fundamentalimportance. As seen from figure 8, the variation of theGeiger counting rate is most extreme in areas close to theEquator (where the solar angle sensor is least sensitive).The method therefore allows a dramatic increase in pOSi-LAUNCHED 29 MAY, 1968(THE NUMBERS ARE FLIGHT DAYSFROM DATE OF LAUNCH)(THE NUMBERS ARE FLIGHT DAYSFROM DATE OF LAUNCH)I- 6- ++++2 "" 7 -..-3 -.-4 "" 9 -+.-5 -+-+8 "++9- 12 ""IO " 13 "H-11 -+- 14 ----FIGURE 11.-Trajectory for balloon VMWR.-mL764MONTHLY WEATHER REVIEWVol. 98, No. 10tional accuracy for the tropical balloons. In addition, the de la Recherche Scientifique, Verrieres-le-Buisson, France,use of the Geiger data removes the ambiguity in position Mar. 2oj 1969, 2o PP10" on cither side of the solar declination would have physics, vel. 111, North Holland Publishing co., Amsterdam,nearly equivalcnt solar angle curves (as indicated previ- 1963, 358 pp.different,.determination of balloons nearly under the sun. Balloons Dorman, L. I., "GeoPhYsical and Astrophysical Aspects of Cosmic Radiation,'' Progress in Elementary Particle and Cosmic RayOUs1Y), but tho Geiger count rate bc Fourier, J., Heinsheimer, T. F., Lastcnet, E., and Balazot, G., "The EOLE PACIFIC I Campaign, July/August 1968," Report of the Service d'deronomie No. 104, Centre National de la Recher-RESULTSand balloon VMWR (fig. 11) that flew in the temperate Oct. 28-31, 1968, 10 pp."zone. The meteorological data accumulated bv such bal- Hofmann, David J., and Sauer, Herbert H., "Magnetospheric-loons flying ost,cnsivc]y Over Ocean areas lyheie meteoro- Cosmic-Ray Cutoffs and Their Variations," ' Space Science Re- views, Vol. 8, No. 5/6, D. Reidel Publishing Co., Dordrecht- Holland, 1968, pp. 750-803.logical dat'a is sparse is of considcrablo interest.ACKNOWLEDGMENTThis investigation was supported by the Direction des Recherchesct Moyens D'Essai (IIRME), Contract lOl/68.REFERENCESBethery, M., and Legrand, J. P., "Measure de l'effet de latitude enp6riodc de minimum d'aetirit6 solaire" (Measure of the Effectof Latitude in a Period of Minimum Solar Activity), SPAXMOBulletin, SPARMO (Solar Particles and Radiations MonitoringOrganization) Secretariat, Meudon, France, May 1965, 6 pp.Blamont, Jacques E., Heinsheimer, T. F., Leblanc, J., and Pom-mereau, J.-p., "Expcrience SOXIEX-11, Campagne et premiersrbsultats, Printemps-Et6 1968, lkre partie" (SOMEX-I1 Exper-iment, Flights and Initial Results, Spring-Summer 1968, FirstPart), Report of the Service d'deronmie No. 224, Centre NationalLally, Vincent E., "Superpressure Balloons for Horizontal Soundingsof the Atmosphere, "NCAR TN-28, National Center for Atmos-pheric Research, Boulder, Colo., June 1968, 167 pp.Lally, Vincent E., National Center for Atmospheric Research,Boulder, Colo., 1969 (personal communication).Lichfield, Ernest W., and Frykman, Robert W., "Elect,ronics forAround the World GHOST Balloon Flights," Proceedings of thcAFCRL Scientific Balloon Symposium, New Zealand, April 10,1966, AFCRL 67-0075, Air Force Cambridge Research Labora-tory, Bedford, Mass., Jan. 1967, pp. 59-67.Quenby, J. J., and Wenk, G. J., "Cosmic Ray Threshold Rigiditiesand the Earth's Magnetic Field," Philosophical Magazine, Vol. 7,Solot, Samuel B., "GHOST Balloon Data: I. Introduction,"NCAR TN-34, National Center for Atmospheric Research,Boulder, Colo., Jan. 1968, 21 pp.NO. 81, JU~JT 1962, pp. 1457-1471.[Received January 19, 19?'0]