Decametric Radiation from Jupiter
JAMES N. DOUGLAS
Summary-A brief historical summary is followed by a review of
current observations of Jupiter's decametric radiation. Particular
attention is given to the time structure and statistical properties of
the emission, and several important deficiencies in our observational
knowledge are pointed out.
sec recorder time constant; tape recordings showed
them to be 0.1 t o 1.0 sec long, superimposed on a
rather smooth rise and fall of background noise.
Thus, the radiationwhen present is Jluctuating.
3) In only two cases was theflux less than that of the
Crab Sebula, the other
seven being much stronger,
occasionallydriving therecorder off scale. T h e
flux from theCrabNebula,measuredwiththe
llills Cross,
was
5X
wmP2 cps-I. Thiswas
lo8 times the expected thermal flux from Jupiter
at this wavelength. Consequently, the radiation is
extraordinarily intense.
I. INTRODUCTION
T
H E S E R E N D I P I T O U S discovery of decametric
radiation from Jupiter ~7asmade by Burke and
Franklin [l 1, [26] while using the Mills Cross of
the Carnegie Institution of Washington for a survey of
the sky at 22.2 Mc. During the first quarter of 1955, a
strong, fluctuating noise appeared on 10 out of 31 nighttimerecords of a declination strip centered
at +22'.
Nine of these noise events, although resembling terrestrial interference, occurred a t approximately the same
sidereal time and never lasted longer than the time
it
would require for a sidereal source t o pass through the
antenna's 1 ? 6 X 2 ? 4 beam. Theevents couldbeexplained by an intermittently-active fluctuating extraterrestrial noise source having a rightascension approximately equal to the median time
of occurrence of the
noise. The right ascension thus observed coincided with
that of Jupiter, then at a declination near +22". Furthermore,theapparentrightascension
of the noise
source was observed to change over a three month
period,preciselyreflectingJupiter'sgeocentricmotion.
This irrefutable association of an intense and obviously
nonthermal
decametric
noise source
with
Jupiter
marked the beginning of planetary radio astronomy.
Nearly ten years of observation have established the
decametric radiation as a phenomenon of great complexity; decisive proof of any theory is lacking. This review will concentrate on observations,sketchingthe
current(incomplete)picture
of thedecametricradiation.
11. RESULTSOF EARLY
WORK
A . DiscoveryObservations
Threefundamentalcharacteristics
of theradiation
were established in Burke and Franklin's original observations :
1) Recordsobtained on 22 of 31 nightsshowed no
evidence
whatever
of radiation
from
Jupiter;
thus, the radiation is seen to besporadic, occurring
in noise storms with durations of less than a day.
2) A noise stormwascomposed
of a sequence of
bursts, many of which were shorter than the
15Mansuscript received May 4, 1964.
The author is with the Yale University Observatory. New Haven,
Conn.
B. Prediscovery Observations
The intensityof the decameter source makes it noticeable even on records taken for quite different purposes,
where itmight well have been mistakenforterrestrialinterference.Accordingly,availablerecordswere
checked, and a number of workers announced prediscovery observations, suggesting further important characteristics of the radiation.
1) Spectral Limitation: Observations of the occultation of the Crab Nebula (then near Jupiter) by the sun
were madein 1954 at the Department of Terrestrial
Magnetism of the Carnegie Institution of Washington,
withsimultaneousrecordsobtained
at 22.2, 38, and
207 h:Ics. Burke and Franklin [4], [ Z ] report eight periods of apparent Jupiter activity on 22.2 YIc; in no
case was this accompanied by activity on either higher
frequency,althoughthesensitivity
of thehigherfrequencyequipmentwasgreater.Thisresultwasindirectly confirmed by F. G. Smith [SI, who searched 38
and 81.5 Mc records obtained with very high sensitivity
radiointerferometers at theCavendishLaboratories,
Cambridge,England,finding
no instance of Jupiter
activity on either frequency.
2) TypicalDuration:
T h e discoveryobservations
were made with a pencil beam instrument which was
sensitive to Jupiter only fifteen minutes a day; consequently, it could be said only that the duration wasless
than a day.Shain 131, [6] obtainedextensiveprediscovery observations in the courseof an 18.3 Mc cosmic
noise survey in 1950 and 1951. Shain's apparatus was
sensitive to Jupiter for two to eight hours per day;
in
spite of this, noise storms were typically less than two
hours long. Thus,theapparent
flux from thesource
may change over two orders of magnitude in times of
the order of one hour.
3) CorrelationwithRotation:
Opticalobserverscan
measure the rotation period of various long-livedfeatures in the upper cloud layers (e.g., the white spots or
839
._
.
.
840
December
1951I
'0
90'
LONGITUDE $ Y S T C M 0
180'
270'
360. 90.
180.
270'
560'
LONGITUDE (SYSTEM II)
1951
1
180"
270'
360" 90" 160"
Fig. 2-Kumber of occurrences of 18.3 MC radiation per 5" interval
of centralmeridianlongitude.Longitudesystemagreeswith
System I1 on August 14, 1951, androtates withaperiod
of
gh 5Sm 13s (From Shain, [6].)
5 ) Other PrediscoveryObservations: Onseveraloccasions in 1953 and 1954, Reber [31] observed radiation
near 30 Mc which later seemed best interpreted as that
of Jupiter. No otherprediscoveryobservationshave
been reported ; Jansky's original 13-rn cosmic noise records very probably contained many instances of Jupiter radiation, but they are unfortunately lost or deFig. 1-Central meridian longitude during decametric noise storms
stroyed. As a result of inquiries by Smith [GI, and with
as a function of observational data: a)using System I longitudes,
the cooperation of V. Agy, a t NationalBureau
of
b) using System I1 longitudes. (From Shain,I6].)
Standards,Boulder, Colo., 20 years of 24-hour fieldstrength records obtained by A. M. Braaten of RCA's
receiving station at Riverhead, L. I., X. Y . , have been
the GreatRed Spot) by timing the interval between successive passages of the object across the central meridianscanned. Although obtained with sensitive equipment
and beautifully annotated by Braaten, the records
were
of the planet's disk. No two periods obtained in this
unfortunatelyobtainedwith
a logarithmic,discreteway are exactly the same; apparently the features are
floating in the upper atmosphere. However, features in step recorder, and Jupiter radiation is not inevidence.
the equatorial regions tend to rotate with aperiod near
9b50m,and those in temperate and polar regions, with a
C. First Observation Programs
period near 9h55m. TWOlongitudesystemshave been
defined, suitable for equatorial and temperate re,'rJ.lons
Following the discovery of decametricradiation,
respectively.System I (rotatingin 9h50m30.s003),and Jupiter observation programs wereestablished at the
System I1 (rotating in 9h55m405.632).To check for pos- Department of Terrestrial Magnetism of the Carnegie
sible correlation, Shain [3], [6] plotted periods of occur- Institution of IqTashington (Burke and Franklin [ll 1,
rence of radiationagainstlongitude
of thecentral
[I81andFranklinandBurke[7],[25]),Commonmeridian at the time of observation, for both System I
wealth Scientific and Industrial Research Organization
and System 11. Fig. 1 reproduces his original diagram.
(Gardner and Shain [ZO]), National Bureau of StandI t isclearly not a random distribution, the most
fre- ards(Gallet[13],[MI)andOhio
State University
quently occurring longitude drifting with date in
Sys- (Kraus (81, [19], [29]). Observations obtained by these
tem I. This drift is slightly overcorrected in the System groups during the 1956 and 1957 oppositions completed
I1 plot, suggesting the best fit would be with a period the initial description of the radiation characteristics.
slightly shorter than that of System 11. The deviation
1) Ndtifrequency Obserzlations of LongitudeProfile:
is not remarkable, but the distinct correlation with rota-[ f ] , [g], [ I l l , [13], [2O], [ Z ] . Occurrencefrequency
tion certainly is.
histograms, similar t o Fig. 2, were obtained a t 27, 22.2,
4) Directivity of Radiation: Re-expressing Fig. 1(b) in 19.6, 18, and 14 &:IC for the 1955-56 opposition. Their
terms of a period of 9h55m13s,Shain summed the num- generalappearancenear
20 Mcwasconsistentlytrilobed, with one region much more active than the other
ber of occurrencesper 5' region of centralmeridian
longitude, obtaining the occurrence frequency histogram two; in addition,a prominent quiet region occupied 40reproduced here as Fig. 2. T h e region of most frequent 60" of longitude. T h e principle region seemed narrower
50" wide. An
occurrence is only 135" wide, not 180" as would be ex- than in Shain's 1951 observation-about
increase inoccurrenceprobability,
flux, and width of
pected for a n isotropic source localized on the planet.
1964
Douglas: Deca.m&ic Radiation from Jupiter
major peak with decreasing frequency was noted by all
observers; Gardner and Shain [20] noted a shift in the
longitude of the major peak to earlier longitudes a t 27
Mc. The tri-lobed profile was generally interpreted a s
signifying the existence of three or more sources on the
planet.
2) ArmrowBandwidth of Radiation: [ 71, [SI, [I 11,
[13],[20], 1253. Initialindications of sharpspectral
characteristics were abundantly confirmed when it was
found that activity on any frequencywasnot
necessarilyaccompaniedbyactivity
on frequencies 2 XIc
higher or lower. Good correlation over 0.1 Mc was observed, establishing the radiation
as relatively narrow
band,with a bandwidthsomewherebetween
0.1 and
2 Rlc. Center frequencies between 14 and 2 7 Alc were
apparently possible, and on oneoccasion, Kraus [19]
reported radiation as high as 43 nic.
3 ) Time Structwe: High-speedrecordings by Kraus
[8], [19] andbyGallet
[13],[44]
andGalletand
Bowles [9] produced evidence of amplitude variation in
times on the order of milliseconds, as well as the slower
0.1 to 10-sec fluctuationsnotedearlier.Gardnerand
Shain [20] noted only the slower component, and found
that the time structure
was markedlydifferent when
observedwithreceivers
25 km apart, suggesting the
presence of important ionospheric effects.
4) So1a.r Correlation: An inverse correlation of occurrenceprobabilitywithsun-spotnumber
for 1956 and
1957 was found and interpreted by Gallet [9], [I31 in
terms of effects imposed by a Jovian ionosphere.
5) Rotation Period: Both Gallet [13] and Burke
[18]
commented on the persistence of the relative position
of peaks on the longitude profile with time as evidence
that the sources are fixed, perhaps on the solid body of
the planet. Based on a comparison of current data with
the prediscovery data of Shain [ 3 ] , [ 6 ] , Galletcalculated a radio source rotation period of 9h55m29.s5,and
Burke a period of 9h55m28.85.
6) Polarization: Observationswitha
crossed-dipole
interferometer at the Department of Terrestrial Magnetism of the CarnegieInstitution[I1 ], [25]showed
the radiation to be predominantly right-hand circular;
simultaneous observation of one event by Gardner and
Shain [203 showed a similar sense of polarization in the
southernhemisphere,therebyrulingoutaterrestrial
ionospheric effect. This strongly implied the presence of
a magnetic field on Jupiter. Based on a source beneath
Jupiter’sionospherewithonemagnetoionicmodecut
off by the ionosphere, Burke and Franklin [ l l ] gave a
lower limit to the magnetic field of 4 Gauss.
At the conclusion of theseinitialobservations,the
decametric
radiation
was
regarded
as 1) sporadic,
b)directionalandrotatingwithJupiter,
c) narrow
band,d) polarized, and e) affected bothbyaverage
solar activity and by the terrestrial
ionosphere. Observations werewidely interpreted in terms of several sporadicallyemittingsources
on the “solid”surface of
Jupiter, whose radiation wasaffected bytheJovian
ionosphere,
841
111. APPARESTTIME STRUCTURE
No feature of the decametric radiationis better established thanitsintermittentbehavior;however,
one
must be careful to distinguish the apparent time structure of the radiation from the intrinsic behavior of the
source of the radiation. Apparent time variability may
arise partly or entirely from
a changing center frequency
andorientation of anarrow-band,directive
emission
source,togetherwithionosphericscintillation
effects.
The relaxation time of the emission mechanism, important for theoretical discussion, can only be obtained by
a simultaneous study of the apparent time structure,
spectum and directivity of the source, after elimination
of possible time-varying propagation effects.
The decametric radiation is known to be variable on
several times scales: short (rapid flux variation during a
noise storm), moderate (noise storm duration, rotation
correlation), and long (correlation with sun-spot cycle).
This section will summarize present knowledge of time
behavior on theshortest
scale, thesubstructure
of
noise storms.
A . Time Structure of iVoise Storms
Early single-frequencyobservationsestablishedthe
apparentlyhierarchicalsubstructure
of noise storms.
Flux variations, although random in nature, appeared
to becomposed of variations on threewell-separated
time scales: minutes, seconds, and, on rare occasions,
tens of milliseconds. The 22.2 IicJupiterstorm
in
Fig. 3(a) is seen to be composed of clumps of radiation,
or burst groups, each lasting a minute or two and possessing unresolved fine structure. Fig. 3(b), made with a
high-speed photographic recorder and time constant of
30 msec, shows 10 sec of a burst group. The spike-like
structure unresolved in 3(a) is seen in 3(b) as a gentle
rise and fall of asecond’s duration; following Gallet
[MI, wewill call these L pulses. Superimposed on the
L pulses, shorter pulses are seen ; these are presumably
the “clicks” first reported by Kraus
[ 8 ] , [19] in 1956,
andthe S pulsesdiscussed
byGallet
[45]. Some S
pulses have rise times limited by the time constant
of
the recordingpen,or
less than 30 msec. Systematic
studies of the durationsof the S pulses and burst groups
are lacking; however, Douglas [38] has reported
a statistical study of duration of L pulses indicating that on
one night, 99 per cent had durations between 0.”3 and
2.s0 with a most frequent duration of 0 . 5 Since all S
pulses have d~rations<<0.~3, and
all burst groups have
duration~>>2.~0, three separate processes may well be
responsible for this.
Not all noise storms possess all three elements of time
structure discussedabove. The S pulses are relatively
rare; when present,theyusuallycontinuethroughout
the entire storm. Burst groups may merge, giving the
storm the appearance of 10 or 15 minutes of rather continuousburst-likeactivity.
L pulses arealmost continuous emissions with bursts superimposed. Weak continuous emission is frequently seen a s a precursor t o a
. ~.
i=.
843
IEEE TRANSACTIONS O X AhTTELVNAS
AND
PROPAGATION
December
located a t
.heJupiter
period of violent burst activity and remains after burst beunchanged.Gardner’sandShain’s
[20] failureto
activityhas
ceased
(Fig.
4,02:24:00to
02:jO:OO observe correlation of L pulses over a 25-km baseline is
E.S.T.).
thus remarkable, requiring the ionospheric scintillation
rate to beof an order of magnitude faster than hitherto
B. Spaced Receiver Observations
believed. Their conclusion, infact,wasthatthe
L
Diffractioneffects in theterrestrialionosphereare
pulses were real, and that only the longer time fluctuaknown t o produceviolentamplitudefluctuations
on tionscouldbeexplained
by the ionosphere; this also
radiation received from discrete radio sources on a time was pointed out by Gallet
scale of minutes to tensof seconds. Such fluctuations are
Smith, et al.,
have compared records obtained in
uncorrelated on records obtained with receivers spaced
Floridaand
Chile,
finding
occasional
correlated
L
more than 10 km apart. Occasional chance coincidences pulses, together with evidence of ionospheric scintillaof fluctuations will, of course, occur, but even a single tion. In their conclusion, the remarkable fact of even a
long record of good correlation would not be explainable few correlations was notedas evidence of the Jovian oriin terms of the usual ionospheric scintillation theory,
gin of the L pulses.
which assumesdiffractingcloudswith
sizes less than
A systematicspacedreceiverprogram
at theYale
10 km.
Observatory [12], [46],[47], [58], [79] has produced
While the ionosphere may beblamed for slow modula- a number of observations with receivers at 30 and 100tion of the relative amplitudes of the storms’ substruc- km spacings in which perfect correlation of storm subture, the timeof arrival of the rapid components should structure was present€ortheentireduration
of the
[MI,
[MI.
Douglas: Decametric Radiation f r m Jupiter
1964
4
03:m:m
E.S.T.
.+
@Z:SO:X.
E*S,T.
f
02:&0:@0
SeSeT.
1
02:30:09
E.S,T.
843
t
02 :25:00
X.S.T.
02:lO:Oct
E*S.Z.
f
Fig. 4-Lobe-sweeping interferometer records of the Jupiter storm of July 28, 1963, obtained at the BethanyObserving Station of the Yale
Observatory. Traces from top to bottom: 20 >IC phase, 20 3Ic amplitude, 22.2 Mc phase, 22.2 >IC ampltitude.
storm. A portion of one such record is reproduced inFig.
3. Note the good time correlation of burst group structure in Fig. 3(a), even though conventional ionospheric
scintillation modulates the relative amplitudes
of successive burst groups. The perfect envelope correlation
of the L and S pulses received at three sites seenat high
speed in Fig. 3(b) continued throughout the period
of
activity. A simple interpretation is that ordinary ionospheric scintillation acts only as a rather slow modulation on a pre-existing time structure incident on the top
of the ionosphere.
The Yale group has found that L pulses may display
all degrees of correlation from the perfect case shownin
Fig. 3(b) t o a total lack of correlation. In many cases
where envelope correlation is sufficiently goodto permit
recognition of equivalent events on two channels,a systematic time lagis seen. T h e lag, which on different rec-
ords may be anything from Os t o k Is, remains constant
formanyminutesduring
a particularstorm.Fig.
5
illustrates a particularly striking case with a lag of approximately $ sec, the western station receiving the signalfirst. Thisperfectenvelopecorrelationwith
3 sec
lag lasted for 15 minutes. The observed senseof the lag
seems t o reverse at opposition; this led Smith [47] t o
suggest an origin of the L pulses in a diffraction process
byinterplanetaryelectron
clouds,whose
systematic
drift velocity perpendicular to the Earth-Jupiter line
might be expected to change near opposition. Recent
observations at Yale [79] have suggested drift velocities for the hypothetical diffraction pattern cast on the
earth of 200-500 km/sec,and a patternscale of the
order of 1 km. Clouds at 1 astronomical unit capable of
producing such a pattern would subtend an angle much
smaller than that of a discrete radio source, but prob-
844
IEEE TRANSACTIONS O N ANTENNAS AND PROPAGATION
December
ably larger than Jupiter's decametric source
(see Sec- have a periodic structure in frequencyas well, but with
tion V below), thereby explaining the failure
t o see L a tendency to drift to lower frequencies a t 100 kc per
pulses on such objects as the Crab Nebula. This sugsecond. Here also, further observations are required to
determine the origin of the frequency structure.
gestion is being investigated both experimentally and
theoretically; should it prove correct, Jupiter and quasi117. OCCURRENCE
PROBABILITY
stellar sources could serve as valuable tools in studying
interplanetary cloud dynamics.
The sporadic appearanceof noise storms suggests the
Storms containing S pulses are comparatively infreoperation of a random process, whose annual, monthly
quent, and usually display perfect S pulse correlation
ordailyaverageoccurrenceprobabilitymaybeestiover a 100-kmbaseline.A few records have been ob- matedbydividingthetimeintervalin
whichnoise
storms were recorded by the total time interval during
tained,however,in
which nocorrelationwhateveris
which noise storms would have been recorded, if prespresent.Furthermore,onesolaroutbursthasbeen
recordedon the same day as a Jupiter storm with S ent. In a similarway,theaverageoccurrenceprobaof the central meridian
pulses; thesolarradiationwasfoundtocontain
S bility for a given longitude range
of Jupitermaybecalculated,sinceover
pulses, which were also perfectly
correlated. I t is cona period of
cluded that the S-pulses are produced by some kind of months that range is observed many times. Shain's [6]
propagation phenomenon not sensitive to source anguoriginaloccurrencefrequencyhistogram,Fig.
2, may
lar size,whichusually
be transformed t o a n occurrence probability histogram
is perfectlycorrelatedover
100 km.
if one has the additional information of the number of
times a given longitude region was observed. However,
C. Dynamic Spectra of Pulses
the occurrence probabilityat a giveninstant of time may
I n 1960, dynamicspectrumanalyzerswentinto
not be estimated with accuracy; that instant
observed
is
operation at theHighAltitudeObservatory(HAO)
only once and the resultant probability
will be either
(15-34 Mc), University of Florida (4 Mc bandwidth at zero or one.
18 Mc) and Yale University Observatory (1 Mc bandOccurrenceprobabilityisthus
useful indescribing
width at 22 Mc);theYaleandFloridainstruments
aggregates of data, and such descriptions may permit a
were particularly designedfor study of the L and S choice amongseveralsimplebutphysicallyplausible
statistical models.
pulses.
L pulseswerefound
to occursimultaneouslyover
bandwidths up to 9 NIc, and Warwick [6O] reports a A . StatisticalModels
The simplest statisticalmodel of this sporadic source
rough correlation between the durationof the pulse and
the bandwidth on at least one record. The
H A 0 dy- would assume it to be a random process with a probanamic spectrum analyzer can measure position as well bility of occurrence fo(t) =constant. If this hypothesis
as spectrum. Based on its records,Warwick [60] has were true, no matter how one subdivides the data, estireported a correlation between apparent shifts in Jupi- mation of occurrence probability should give the same
the is one of the most signifiter's radio position in the sky and the occurrences of L result. That this is not case
[6]
pulses, suggesting a common, non-Jovianorigin for both cantfactsaboutthedecametricemission;Shain's
finding of a greatly enhanced occurrence probability when
phenomena. Thus,
two
lines of evidence,
spaceda particular 135" region of System I1 longitude faces the
receiverobservationsandposition-fluctuationcorrelation, suggest that the S and L pulses in the decametric Earth necessitates rejection of this simplest model. T h e
occurrence
probability
of the
decametric
radiation
radiationareduetohithertoundetectedpropagation
effects.
is at leastapproximatelyperiodic:
p,(t+nP) =po(t),
The broad-bandL pulses were found on closer exami- n = 1, 2 , 3 , .
. While the periodicity could be intrinsic
nationto possess a fine-structureinfrequency
[ 3 6 ] , to the source and have nothing to do with the rotation
of Jupiter, this does not seem plausible and we there[38], [42], [52], [65]; occasionally a pulse may be invisible at a frequency flanked 100 kc on either side by fore consider only rotating statisticalmodels. While itis
strong activity. The Florida group reports that this
fine hardly possible toenumerate allconceivablemodels,
into three general
structure occasionally takes the form
of a periodic struc- many physicallyplausibleonesfall
ture in frequency, suggestiveof harmonics. Further ob- classes :
1) Isotropic Sources in Polar Regions: Isotropically
servations are needed to distinguish between two possibilities: 1) the propagation effect producing L pulses is emitting sources of radiation are presumed t o be near
refrequency-sensitive, thereby creating the apparent fine thesurface of theplanet,rotatingwithit.Earth
ceives radiationonlywhenseenabovethesource's
frequency structure as well as the time structure, or
2) the radiation leaving Jupiter possesses the fine struc- radio horizon.Since the declination of the earth seen
from Jupiter is different from zero, the period of emisture in frequency with only the time structure being
imposed by propagation effects. Dynamic spectra of the sion may range fromzero t o a complete Jupiter rotation
S pulses have been reported by Smith,
et al., [65] t o depending on the declination of the earth and latitude
1964
Douglas: Deca.nzetric Radiation f r o m Jupiter
845
of the source. Because of ionospheric refraction, the frequency of observation will also be a factor. This typeof
model predicts a very strong correlation of occurrence
probability and length of emission period with the declination of the earth seen from Jupiter.
2) Directive Sources: One or more sources rotate with
Jupiter and emit a relatively narrow beam of radiation
which is observed only when directed toward the earth.
This model is unusually flexible until a detailed emission mechanism is specified; it may be accommodated
to a wide range of possible observational results.
3) StimulatedSources: Sources of emission have an
intrinsicallyhigherprobability
of activity when the
‘3000
60’’
earth is near the meridian of the source. Such a terrestrial influence upon Jovian affairs seems highly unlikely;
/
however, as seen from Jupiter, the earth and sun are
300
always in approximately the same direction ( f 11’) so
00
that the required stimulation may come from the sun.
Fig. 6-1961 22.2 kIc data reduced to occurrence probabilities and
In this case, the periodicity present in the occurrence
plotted as a circular histogram to enhance the longitude symmetries and tri-lobed character of the emission at this frequency.
probability would be Jupiter’s period relative to the sun
Graduations of the principle axes are units of 0.1 in occurrence
rather than to the earth, and the central meridian
longiprobability.Openhistogram
includes weak continuumevents.
(From Douglas and Smith, [69].)
tude of maximum activity (as seen from earth) should
increase systematically by 22’ through the course of an
apparition.
As will be seen, observations generally favor the sec- involves a least-squares fit of the best drift line to a fig1. This method is nonsubjective
ond class of models; this may be due largely to its fiexi- ure such as those in Fig.
but also does not seem to make full use of the available
bilitywhencontrastedwiththeconcretepredictions
information.
made by the other two.
A third technique was introduced by Douglas
1381,
B. Rotation Period
in which the variance of the histogram was calculated
A longitude system for Jupiter may be defined by
for a sequence of assumed periods. T h e peak of the plot
stating the central meridian longitude a t some instant of variance vs assumed period (IVhittaker periodogram,
in time (epoch) and the periodwithwhich
the longi- [83], see Fig. 8) was shown to be an unbiased estimator
tude system rotates. For any assumed period, an occur- of the period present in the data; furthermore, it can
renceprobabilityhistogramfor
a set of observations be shown that this procedure uses all the periodic inmay be calculated; Fig. 6
is such a histogram plotted
formation present in arriving a t a period estimate. T h e
on a polar diagram. If one were to alter the assumed
entire process is quitefree from subjective involvement,
period substantially, the assymetries present in the his- the accuracy of the result may be calculated, and with
togram would be washed o u t ; obviously the more asym- electronic computers, calculations are quite rapid.
All
metric histogram was calculated with a period near the period-determination techniques give equivalent results
period present in the observations.
if the periodicity present is constant;
if the period varies,
Threemethodsfordeterminingaccuraterotation
the estimate may be averaged in different ways.
periods for the radio sources have been used by various
Period determinations have been published by Shain
workers.Onemethod,
used byGallet[g],[13],
[MI, [6], Gallet [IS], Burke
[ l S ] , Gardner and Shain [20],
Burke and Franklin [ l S ] and Franklin and Burke [25], Carr, et al., [22],[42],Douglas
[35], [3S], [53], and
and Carr and Smith, et al. [22], [42] was made possible IVarwick[60]. Themostextensiveworkwasthat
of
by the long-timepersistence of the main peak of the Douglas, who used observations furnished by all previoccurrence probability histogram when plotted in Sysousobservers to obtain a meanperiod and standard
tem I1 longitude. The peak persisted, but its longitude deviation. A subcommission of IXU Commission 40
drifted. The rate of drift is proportional to the error in recommended in 1962 [M] that radio observations be
the period assumed for calculating the histograms; cal- referred tothelongitudesystemdesignatedSystem
culation of the trueperiod is trivial once the slopeof the I11 (1957.0), defined as follows:
drift line has beenestablished.
The histogramdrift
System III(1957.0)
technique is easily applied; drawbacks are its failure t o
Epoch: 1957 January 1.0 U T (JD 243 5839.5)
use all the information present in the histograms and the
Period: 9h55m29s37
difficulty of assessing its accuracy.
Central meridian longitude at epoch: 108?02.
A second technique, used by Gardner and Shain [20],
-
IEEE TRANSACTIONS O N ANTENNAS ,4ND PROPAGATIOrV
846
For any timef expressed in Julian days, the longitudeof
centralmeridian of SystemIII(1957.0), XIII, maybe
obtained from the longitude of the central meridian of
System 11, An (astabulated in thevariousnational
ephemerides) , from the relation
Xm(1957.0) = XU
+ 0:2743(t - 243 5839.5).
C. Changes i n Rotation Period
15--
December
22.2M e
IO -5 --
The System III(1957.0) period discussedin the previous section represents a n average period over the interval 1950-60; smallyearlyfluctuationscertainlycan
havegoneunnoticed.Douglas
[38] reportsanupper
limit to yearly period fluctuations of + 2 sec, and an
upper limit to a slow, monotonic change over the interval of 0.5 sec. Gallet [MI, using thehistogramdrift
technique, found the rotation period between
1956-57
to be 1 seclonger than the periodbetween 1950-56.
T h e significance of thisresultleansheavilyonthe
accuracy with which the observed peak
of the occurrence probability histogram mirrors the location of the
true peak. Fora small number of observations, sampling
fluctuations are clearly possible; this is discussed more
fully below. The 1-sec lengthening reported corresponds
t o a 12' change in longitude of the peak of the histogram; this lengthening, while entirelypossible, does not
seem to be statistically proven.
Douglas and Smith [73]found clearcut evidence fora
change in Jupiter's apparent rotation period sometime
between 1961 and 1962. Fig. 7 shows the occurrence frequency
histograms
obtained,
displaying
a syste300
360
matic drift of the major peak with respect to System
SYSTEM lIL(I957.0) LONGITUDE
III(1957.0). Fig. 8 shows the periodogramsfor
the
Fig. 7-Occurrence frequency histograms of Yale observations in1950-61 and the 1961-63 intervals on the same scale;
dicating change in apparent rotation period of the radio sources.
the Os8 lengthening is visible here as well. I t should be
The ordinate is number of events on all three histograms.
emphasized thatthisdoesnot
necessarily imply a
change in the rotation period
of the solid body of the near 225'; and Region 3, centered 180' away from Replanet. A variety of other explanations might be offered;gion 1 at about 300'. Equally remarkable is the virtual
for example, the longitude of the main source of emis- absence of radiation in the quadrant 350"-80". Smith
sion may have changed.
et al. [52] have suggested that Region 1 is in fact double;
Wanvick's [84] perioddeterminationsbythelandthe fact that its breadth is greater than ofthat
Region 3
mark method (see Section VI below) for the same pe- is certainly striking in most observational material.
riod show no apparent deviation from the System
I11
The overall appearanoe of the longitude profilereperiod;thismustberegarded
as anopportunityto
sembles t h a t of an antenna directivity pattern, with a
learn something about the source rather than as a con- main lobe, side lobes, and unusually good front-to-back
tradiction; the two period determination techniques are ratio. I t should be emphasized at this point that what
basedonobservations
of differentbutpresumably
is plottedisoccurrenceprobability
vs longitude,not
relatedthings.
strength of storms.Furthermore,thefactthatpeak
occurrenceprobabilityoccurs
at approximately 225O,
D . Longitude Profile
System III(1957.0), does not imply that the source
is
The longitude profile plotted in polar coordinates in on the central meridian at that time; nordoes the presFig. 6 was obtained by calculating the occurrence prob- ence of three peaks imply the existenceof three sources.
ability in 5' intervals of System III(1957.0) from 22.2 Several
alternative
suggestions
were
proposed
by
Mcdataobtainedduringthe
1961observation. T h e Douglas and Smith [69], as suggested in Fig. 9.
presence of three regions of enhanced emission is strikI t is a difficult matter to assess the accuracy of pubing: Region 1, centered around 120'; Region 2, centered lished longitude profiles. Eachpoint is essentiallyan
Douglas: Decametric Radiation from Jupiter
1964
1500
c
847
1950-1961
ASSUMED
PERIOD
Fig. 8-Periodogram for the sameobservations as usedin
showing the Oa.8 lengthening in rotation period.
Fig. 7,
estimate of an occurrence probability at that longitude,
given by the ratio of the number of events (fV) to the
number of observations (AT): pest@)=lV/M. If the
observationsareconsidered
tobestatisticallyindependent, the standard deviation of the estimate will be
u = (pest(l -pest)/M)1!2.For pest small, the percentage
error
in f i e s t will be
approximately
Thus,
if the
estimate of a point on the longitude profile is based on
onlyfourobservedstorms,a
+50 percentstandard
deviation in the estimate is expected.
If the observationsarenotstatisticallyindependent,thestandard
deviation will be even greater. Furthermore, since most
Jupiter storms are longer than 1.5' longitude of System
I11 (1957.0) the estimates of occurrence probability at
adjacent longitudes will covary: if one 5'" interval has
an unusually high value as a result of sampling fluctuations, the adjacent intervals will also be raised. Thus,
the standard deviation expression given should be regarded as a minimum, and statements about small features of the longitude profiles must be made with caution.
A systematic bias may creep into longitude
profile
shapes obtained with different instruments if the shape
is a function of the limiting flux t o which Jupiter storms
may be detected. The open histogram in Fig.
6 is the
result of addingasubstantialnumber
of order-ofmagnitude weaker continuum storms to the observations of the main phases of the emission; no systematic
changeinshapeisevident.
If thisinsensitivityto
strength of stormshas heldthroughtheyears,data
from many observers obtained
at the same frequency
may be compared, making up in part for the lack of a
long, homogeneous series of records.
Fig. 9-Several
possibleradiation-coneconfigurationsconsistent
with the occurrence probability histogram of Fig. 6 , all drawn
for the phase when the Region 2 cone is centered on the earth.
The dotted circles represent Jupiter; the heavy points, emission
regions. (From Douglas and Smith, [69].)
E. Correlation of Longitude Prqfile with Frequency
Three kinds of frequencycorrelationsinthe
longitude profiles were noted by early observers: the occurrence probability and widthof the major peak increased
withdecreasingfrequency;andthelongitude
of the
major
peak
increased
slightly
with
decreasing
frequency.Inrecentyears,
low frequencyobservations
in FloridaandChile[42]andthose
of Ellis[49]in
Tasmania have extended these results. Although thorough statistical discussions in the literature are lacking,
the trends, if significant, may be of great importance in
interpretingthe emission mechanism.Ellis'observations at 4.8 I l c show that the shape of the longitude
profile has changed substantially, with only two principlesourcesbeingvisible
a t thisfrequency.Fig.10
illustratesEllisresults;theoccurrenceprobabilityis
practicallyconstant,butthemean
pourer is concentrated intwomajorregionsandvariesalmostsinusoidally.
Observations of the longitude and width
of the Region
2 peak are plotted vs wavelength in Fig. 11. All points
are from Carr, et al., [42] except the points a t 4.8 Mc
( z 6 0 m), which are estimated from Fig. 9 (Ellis [49]).
The shift and broadening is clearly evident (Ellis, [49],
[Xi], [ 6 6 ] , and
Ellis
and
McCullough,
[68], [75],
~
:--
, .
848
December
r
f =4.8 Mc
O T
i
z
1
\
SYSTEM m(1957-0)
Fig. 11-Longitude of peak probability (center line)and $-peak probability (outer lines) of the Region 2 peak as a function of wavelength.
two might reflect any influence of the general activity
of the sun on Jupiter or on propagation conditions in the
interplanetarymedium; a correlationwiththethird
would be expected with models of class 1) or 2) of Section 117, A. Unfortunately, a thorough statistical consideration of the significance of these possible correlaFig. 10-Kumber of occurrences, total power and mean power as a tions is not yet available.
function of System III(1957.0) longitude. (From Ellis and
In addition to the
possible inverse correlation with
McCullough, [68].)
meansolaractivity,Warwick
[37], [40],[45],
[60]
finds evidence for an association of periods of Jupiter
Smith, et aZ. [ 8 O ] ) . Extension of observations to the
activity with periods of solar continuum in 1960 and
vacant intermediate wavelengths as seen in Fig. 11 and
1961. Such a correlation, if well established, could be a
thorough statistical discussion of the results is one
of mostsignificantindicator
of thetype
of emission
the more pressing observational tasks confronting obmechanismresponsiblefor
thedecametricradiation.
servers. Particularly intriguing is the suggestion by EllisDouglas [ 3 8 ] failed t o find a significant correlation in
and
McCullough
that
the
source
mechanism
disthe records on 1960; the effect was present but explainplays itself mostcharacteristically
at very low fre- ableassamplingfluctuationwhenthetendencyof
quencies,thehigh-frequencybehaviorbeingunderJupiteractivitytolastseveraldayswasrecognized.
standable as the result of perturbations.
Carr, et al., [39], [42] suggest a correlation of Jupiter
emission with geomagnetic A index with a n 8-day lag;
F. Correlation of Longitude Profile with Time
onceagainthenumbersaresmallandasignificance
A prominent characteristic of the longitude profile is check is lacking.
itsrelativepersistenceinshapeoverlongperiods
of
G. Intrinsic Duration of Emission
time. The amplitude of the occurrence probability is a
KO definitiveconclusionsregardingtherelaxation
strong long-term function of time, which some observers
have interpreted as a n effect of solar activity. However, time of thesourcemayyetbereached,butseveral
limiting statements may be made based
on statistical
the fortuitous near coincidence of Jupiter’s revolution
[38],
[58]. First, it is
period (1 1.86 years) withthe sun-spot cycle(11.1 years) investigationsbyDouglas
makes it verydifficult t o distinguish effects that are due found that the three emission regions are significantly
correlated, and must be connected by a common directo Jupiter’s position in its orbit and those due to mean
solar activity. Estimates of peakRegion 2 occurrence tivity or stimulation process, or perhaps represent the
probability at 22.2 Mc for each year in which observa- same source viewed from different directions. Second,
tions are available, are plotted in Fig. 12 against three observationof a Jupiterstormgreatlyenhancesthe
probability that activity will be seen at the same longiplausiblepotentially-correlatedfunctions:provisional
tudeonerotationlater.Quantitativeworkismade
sunspot number, solar latitude of the line joining the
sun and Jupiter [52], and the declination of the earth difficult bythesamplingperiodicityimposedbythe
earth’s rotation; however, it is clear that the limitation
as seenfrom Jupiter. If onemakessuitableshiftsin
of storm duration to tens of minutes is not a result of
phase, any of the three functions may be brought into
commencementandcessationofthegeneralactivity
approximatecorrelation.
A correlationwiththefirst
[%I,
1964
Douglas: Decametric Radiation from Jupiter
849
TIME
Fig. 12-22.2 h4c Region 2 occurrence probability (heavy dots with error flags) compared with provisional sun-spot number (open circles).
.Also shown (without ordinate scales) are declination of earth seen from Jupiter (x's) and sub-Jovian latitude on the sun (line). Range
of sub-Jovian latitude on the sun is about +9"; range of declination of earth is about ,3".
timeplanearereproducedrepeatedly;drawing
an
analogy with optical features of a planet, he calls them
landmarks, and. suggests that Jupiter possesses a more
or less permanent set of landmarks, ;.e., a permanent
dynamicspectrum.
The time of occurrence of a distinguishable part of
such a landmark can be determined with considerable
precision. Fig. 14 is a plot of central meridian longitude
corresponding to time of landmark occurrence vs time
in days from opposition. Several remarkable properties
V. ANGULAR SIZEOF SOURCE
ofthisdiagramwerepointedoutbyWarwick
[60].
First, the fact that the landmark
is only observed within
SleeandHiggins[86]successfullyobservedJupiter
a
with an interferometer having a baseline of 1940 wave- about +9' of SystemIII(1957.0)longitudeimplies
of theradiation.Second,the
lengths at 19.7 XIc (32.3 km). The fringe visibility of the verynarrowbeaming
Region 2 landmarkshows
no ssFstematic driftwith
radiationwasnotreduced
a t thisbaseline,givingan
upper limit to its angularsize of one third the diameter time. If thelandmarks were theresult of puresolar
of the planet. This observation confirms the suggestion stimulation (model 3), a systematic drift of 22' would
be expected. -4 repetition of this observed lack of drift
of many observers that the source has small angular size;
we have seenin Section I11 above that this has
rele- in several more apparitions would be sufficient t o rule
vance in unexpected ways to interpretationsof the emis- out this kindof model. Third, it is very curious that the
Regioo 1 source shows just the sort of systematic drift
sion.
that theclass (3) model requires,b u t in the wrongsense.
VI. DTXAMIC
SPECTRA
OF NOISESTORMS
Such a systematic change in longitude would serve t o
The H 4 0 radiospectrograph,
a swept-frequency broaden Region 1 on occurrence-probability histograms;
phaseswitchinginterferometer,hasproduced
an im- such a broadening has already been mentioned. Warpressive collection of dynamic spectra of noise storms wick's observations of the drift of the landmark in this
in the 7.6 to 41 Mc range. Fig. 13 shows the appearance region providesan alternative explanation to the double
of a Jupiter noise storm on this instrument. This storm, source suggested by Smith, et al. [52].
IITarwick has used the rate of drift in longitude of the
which occurred when Region 2 was facing the earth, is
Region
2 landmark to determine a rotation period for
seen to drift lower in frequency with time. Such drifts
the
source
in a methodsimilar to the histogram-drift
arecharacteristic
of emission fromthisregion;the
technique
used by other observers. The smaller width
oppositesenseisobservedwhenthe
emission occurs
of
the
landmark
emissioncone
haspermittedmore
with Region 1 facing the earth. IYarwick 1601 has found
t h a t a number of characteristic shapes in the frequency- accurate period determination in a given length of time.
period. A Jupiter activity 'period generally lasts more
than two days, and frequently more than five days.
It
is interesting40 speculate that even the eventual apparent termination of the activity period after five days
or so is the result of a shift of emission directivity out
of the plane containing the earth; a random tendency
of this sort might also produce a correlation
of occurrence probability with declination of the earth as seen
from Jupiter.
IEEE TRAATSACTIOMS ON AMTENATAS AND PROPAGATION
Fig. 13-Region
December
2 radiation seen on the H A 0 swept-frequency interferometer. Sote the drift of emission from high t o low frequencies.
(From IYarwick, [60].)
Fig. 14-Central meridianlongitude a t landmark occurrence as a
function of days from opposition.The scatterof points isreal, not
a result of inaccuracy in estimation of time of landmark occurrence. (From J. W. IVarwick, [60].)
VI I. POLARIZATION
T h e general predominance of the right-hand sense of
polarization of the decameter radiation near 20 Mc has
beenconfirmed byseveralobservers[42],
[.SI], [69];
multi-frequencypolarizationobservationshavebeen
made by Sherrill at Southwest Research Institute, [74]
and by Barrow at Florida State University [82]; and
observations at 10 Mc by Dowden at the University of
Tasmania [761.
Left-handpolarization is seenwithincreasingfrequency at longer wavelengths; at 16 Mc, one third of
the activity is in the left-hand sense. Dowden reports
that left-hand polarization a t 10 Mc is primarily concentrated in emissions from Region 1; Sherrill and Barrow have both noted a tendency for Region 1 polarization t o differ significantly from that of Region 2.
Interpretation of polarization records is complicated
bymanyequipmentcalibrationproblems,together
with the apparent presence of propagation phenomena
capable of differentially attentuating the two characteristic modes and producing rapid (Iess than one minute)fluctuations
in theapparent axialratio
of the
radiation. Longer term systematic effects are also apparent in the records: several observers have noted
a
change from right hand to left hand and back again in
a ten minute period a t 22.2 &IC.
Even if solutions to the equipment and
ionosphere
problems were known, the interplanetary medium would
be optically active: the electron clouds and interplanetary magnetic fieldwill produce Faraday rotation. A
more important obstacle to straightforward interpretation of polarization records is the mechanism recently
proposed by Lusignan [77], [SI]. Relativisticeffects
in the solar wind will cause it t o possess linear charac-
1964
Douglas: Deca:&ric
Radiaiion from Jupiter
85 1
it is not as yet unambiguously established by the data.
5 ) Rotation period was well established between 1950
and 1961; the change in period or drift in longitude of
major peak, between 1961 and 1963. 6) There are three
obvious peaks of occurrence probability in the longitude
profile; a broadening of the Region 1 peak is probably
VIII. FLUX
due to an elongation effect rather than a fourth peak.
Absolute measurement of the decametric flux is made 7 ) T h e Region 2 peak is later in longitude and broader
at
difficult byseveralcircumstances:onemust
use an lowerfrequencies; theshape of thelongitude profile
antenna of knowngainwith
a power-linearreceiver
changessignificantly a t lowerfrequencies. 8) A perwhose output is integrated over a period long compared manent dynamic spectrum exists. 9) “Landmark” radiato the duration of ionospheric and interplanetary diftion is apparently beamed ina very narrow ( f9”) cone.
fraction shadows. The resulting measurement must be
10) Theradiationispolarized;andispredominantly
corrected for the ionospheric absorption present, and the
righthand near 20 Mc, with an increasing percentage of
wholeprocessmustbedonesimultaneously
in two lefthand a t lower frequencies; there is a suggestion of
orthogonal
polarizations.
ViTith these
precautions,
a correlation of polarizationellipticity,orsense,with
mean flux measurement at one frequency could prob- longitude. 11) Decametric power is probably emitted at
ably be made to an accuracy of 10 per cent or so; such l o 7 t o lo8 w averaged over long periods of time.
observations are not available at a n y frequency. Even
Man~7 authors have proposedtheories of the decamany of the order of magnitude estimates in the litera- metric emission capable of explaining some of the obture are open t o question on one or more of the above servational features listed above; at the present time,
grounds, particularly those made at frequencies below no one theory is generally accepted.
A transition has
15 Mc ( 5 Mc flux reported by different observers covers occurred, however, from one type of theory to another.
a range of 100). At this time, it seems safe to say only I n early work, plasma oscillations
were frequently inthat averageflux near 20 Mc can be on the order
of
voked [ E ] , whichexplained
boththenarrow-band
wm-2 c-l, and that the fluxincreaseswithincreasing
characteristics of the emission and the burst-like strucwavelength down t o 10 Mc. Filling in this picture is
a ture. In theories not dealing specifically with the emisvery important observational problem.
sion mechanism, the sources were generally regarded as
The decametricpower radiated by Jupiter establishes being beneath a Jovian ionosphere, on the surface of the
the radio energy requirement; its calculation requires
planet [6l, 181,[101-[131,1241,
[311, [321, [%I, [451;
knowledge of 1) intrinsic time structure of the source, directionalpropertieswereproducedaspropagation
2) instantaneous bandwidth of radiation, 3) spectrum,
effects in the Jovian ionosphere. Now, stimulated by the
4) flux, and 5) size of directivity cone. T h e results are discovery of Van Allen belts around Jupiter [64] and
particularlysensitivetoitems
3) and 5 ) , andquoted
by the observations and theories of Warwick [37], [42],
figures range from 2 x 1 0 7 t o 7 X 1011w. The author is [43],[60],[65],theoriescenteraroundmechanisms
much inclined to favor the first of these figures, given
operating in the upper atmosphere of the planet at the
by Warwick [6O] and based upon a small
( f 9 ” ) emis- gyrofrequency,producingdirectiveradiation
t h a t is
sioncone, which seemsquitereasonably
well estab- intrinsicallypolarized[43],[52],[53],
[57], [60],[61],
lished by the dynamic spectrum observations, and using [65l,[69], 1701.
average rather than peak
flux values. Residual uncerTheory and observations are unfortunately difficult
tainty must still be of an order of magnitude. Further tocompare in manyareas,primarilybecause
of the
progress must await more information
on each of the great complexity of the observational problem. Perhaps
five contributing points above, particularly time structhe best correlations of the two occur in the theory and
ture and flux.
dynamic spectral work of Warwick, and in the theory
and longitude-profile work of Ellis. Clarification of the
IX. CONCLUSION
observational picture requires a simultaneous study of
The following reasonably well establishedobservaallphenomena,inordertodistinguishintrinsicand
tional characteristics of the decametric radiation from
propagation-inducedproperties of theradiation.ParJupiter may be noted: 1) Decametric radiation is due
ticular attention should be given to the several octaves
to one ormore sources having angular size
less than one below 20 Mc, a region of great potential importance
third the planet’s disk; very probably muchless. 2) The which is not as yet adequately explored.
source is active fora period of days; the hypothesisof a
REFERENCES
continuously active source is not disproved. 3) General
[I] B. F. Burke and K. L. Franklin, “Observations of a variable
probability of source activity isa slow function of time,
radio sourceassociatedwith the planet Jupiter,” J. Geophys.
Res., vol. 60, pp. 213-217; June 1955.
with a cycle on the orderof ten t o twelve years; it is not
[Z] B. F. Burke and K. L. Franklin, “Jupiter as a Radio Source,”
definitely established a s a n inverse correlation with the
presented at I.A.U. Symp. No. 4 on Radio Astronomy, August,
1955. and oublished in “IA\USvmoosium No. 4: Radio Assun-spot cycle. 4) There is a tendency for positive cor-
teristicmodes, which will changetheaxialratioand
sense of polarization of the radiation rather than simply
rotatingit.Thisphenomenoncanproveexceedingly
valuable in permitting a study of the properties of the
continuous solar wind using Jupiter as a probe.
IEEE TR.AArSACTIONS O N BNTE,VT:YAS AND PR0PAGATIO:V
[3] C. A. Shain, ”Location on Jupiter of a Source of Radio Noise,”
presented a t I.A.U. Symp. KO. 4 on Radio Astronomy; August,
1955, a n i published in “IAU SymposiumNo. 4: Radio Astronomy, H. C. vandeHulst,Ed.,CambridgeUniversity
Press, London, England p. 397; 1957.
[4] J. F. Fkor, 1%’. C. Erickson, H. )\;. Wells, B. F. Burke, and K. L.
Franklm, Carnegie I&. Wash.YearBook,RadioAstronomy,
pp. 143-146; December, 1955.
a t 38 and
[SI F. G. Smitb, “A search for radiation from Jupiter
81.5 Mc/s, Observutory, vol. 75, pp. 252-254; December, 1955.
161 C. A . Shain, “18.3 Mc/s radiation from Jupiter, Australian J.
Phys., vol. 9, pp. 61-73; March, 1956.
[7] K. L. Franklinand
B. F. Burke,“RadioObservations
of
Jupiter,” Astron. J . (Abstract), vol. 61, p. 177; 1956.
[SI J. D. Kraus, “Some Observationsof the Impulsive RadioSignals
from Jupiter,” Astron. J. (Abstractj, vol. 61, pp. 182-183; 1956.
[9] R. M. Gallet and K. L. Bowles, “Some Properties of the Radio
Emission from Jupiter,” presented a t Am. Astron. SOC.Meeting,
Columbus, Ohio; March, 1956.
[lo] A. F. O’D. Alexander, “Possible Sources of Jupiter Radio Koise
in 1955,” J . Br. Astron. Assoc., vol. 66, pp. 208-215; May, 1956.
1111 1. F. Firor. W . C. Erickson. H. 1%’. l\Tells. B. F. Burke. and K. L.
Franklin,-Curtzegit!Inst. Miaslt. Year Book, “Radio As’tronomy,“
pp. 74-76; December, 1956.
[12] K. W . Philip, “Dispersion of Short-wave Radio Pulses in Interplanetary Space,” Astron. J . (Abstract), vol. 62, p. 145; July,
1957.
[13] R:M. Gallet, “The Results of Observations of Jupiter’s Radio
IRE
Emissions on 18 and 20 Mc/s in 1956 and 1957,“ TRANS.
(Abstract), vol. AP-5, p. 327; July, 1957.
[14] V.V. Zheleznyakov, “On the theory of sporadic radio emission
from Jupiter,” Astron. Zh., vol. 35, pp.230-240;1958.
(In
Russian.)
[I51 C. H. Barrow and T. D. Carr, “A radio investigation of planetary radiation,” J . Br. Astron. Assoc.vol. 67, pp. 200-204;June,
1957.
[16] C. H. Barrow, T. D. Carr, and A. G. Smith, “Sources of radio
noise on the planet Jupiter,” Nature, vol. 180, p. 381; .August,
1957.
N.Douglas,“Preliminary
Report of a
[I71 H. J. SmithandJ.
Planetary Radio Programof Yale Observatory,” Astron. J., vol.
62, p. 247; October, 1957.
[18] B. F. Burke, CarnegieInst.Wash.
Yeur Book, “Radio Astrononzy,” p. 90; December, 1957.
[19] J. D. If;raus, “Planetary and solar emission a t 11 meters wavelength, PROC. IRE, vol. 46, pp. 266-274; January, 1958.
[20] F. F. Gardner and C. A. Shain, “Further obsen-ations of radio
emission from the Dlanet -lupiter,”
- Azcstralian J . Phys., vol. 11,
p. 55; March, 1958.
C. H. Barrow and T. D. Carr, “18 MC radiation from Jupiter,”
J . Brit. Astron. Assoc., vol. 68, pp. 63-69; February, 1958.
T. D. Carr, 4 . G. Smith,R. Pepple, and C. H. Barrow,“18
&IC observations of Jupiter in 1957,” Astroplzys. J . , vol. 127,
pp. 274-283; March, 1958.
Harlan J. Smith andJ. N.Douglas, “Observations of planetary
non-thermal radiation,“ Proc. Paris Symp. o?t Radio Aslronomy,
Paris, France, Stanford University Press,
Calif., pp. 53-55 ; 1959.
F. Link, “Sur les Ionospheres Planetaires,” Proc. Paris Synp.
on Radio Astronomy, Stanford Cniversity Press, Calif., pp. 5860; 1959.
K. L. Franklin and B. F. Burke,“Radioobservations of the
planet Jupiter,:’ J . Geophys. Res., vol. 63, pp. 807-824; December, 19%
[26] K. L. Franklin, “An account of the discovery of Jupiter as a
radio source,” Astron. J., vol. 64, pp. 37-39; March, 1959.
[27] T. D. Carr, “Radiofrequency emission fromthe planet Jupiter,’
Astron. J . , vol. 64, pp. 3 9 3 1 ; March, 1959.
sources otherthan
[28] H. J.Smith,“Non-thermalsolarsystem
Jupiter,” Astron. J., vol. 64, pp. 4 1 3 3 ; March, 3959.
of Jupiter, PROC.IRE,
1291 J. D.Kraus,“RadioObservations
(Currespondence), vol. 47, p. 82; January, 1959.
[30] B. M. Peek, “The determination of the longitudes of sources of
emission of radiation on Jupiter,” J . Brit. Asfron. Assoc., vol.
69, pp. 70-79; February, 1959.
[31] G. Reber,“Radiointerferpmetry a t threekilometers altitude
J . Geopkys.Res., uol. 6 4 , pp. 287abovethePacificOcean,
303; May, 1959.
Jupiter,”
1321 C. H. Barrow, “The latitudes of radiosourceson
J. Brit. Astron Assoc., vol. 69, p. 211; May, 1959.
[33] A. G. Smith and T. D. Carr, “Radio-frequency observations of
J., vol. 130,pp. 641-647;
the Planets in 1957-58,” Astrophys.
- .
September, 1959.
Bollhagen, X. Chatterton,and
[34] A. G. Smith:( T. D.Carr,H.
X. F. Six, Ionosphericmodification of theradiation
from
Jupiter,” Nature, vol. 187, pp. 568-570; August. 1960.
[35] J. N. Douglas, “-4 Uniform Statistical Analysis of Jovian
.--1
_I
December
Decameter Radiation,1950-1960,” Astron. J. (Abstract), vol. 65,
p. 487 ; Kovember, 1960.
[36] H. J. Smith, B. A I . Laker, and J. N. Douglas, “Fine-structure
of Tupiter’s 20 Mc/s Noise Storms.” Astron. J . (Abstract),
. . vol.
65,-p.-501; Novepber, 1960.
[37] J. lV. It:arn-ick, A comment on Jupiter’s radio emissionat long
wavelength in relation to solar activity,” Arkiv Geofysik, vol. 3,
p. 497; August,~l96O.
[38] J. X. Douglas, A Study of Kon-thermal Radio Emission from
Jupiter,”Ph.D.dissertation,
YaleUniversity,
New Haven,
Conn. ; September, 1960.
[39] T. D. Carr, A. G. Smith, and H. Bollhagen, “Evidence for the
solar corpuscular origin of the decameter wavelength radiation
from Jupiter,” Phys. Rev. U t . , vol. 5, pp. 418420; November,
1960.
[40] J. W . IVanvick, “Relation of Jupiter’s radio emission a t long
wavelengths to solar activity,” Science, vol. 132, pp. 1250-1252;
Ortnher.
- - - - - - ., -1960
.-[41] C. H. Barrow, “The magnetic field of Jupiter,” ,hTatu7e, vol. 188.
pp. 924-925; December, 1960.
.1421- T. D. Carr. 4. G. Smith. H. BollhaKen. X. F. Six. and N.E.
Chatterton;
“Recent
decameter-wa<elength
obseivations
of
Jupiter, Saturn and Venus,”Astrophys. J.,vol. 134, pp.105-125;
July, 1961.
[43] B. F. Burke,“RadioObservations
of Jupiter. I.,” ch.13in
“Planets and Satellites,“ G. P. Kuiper and B. TVI. Middlehurst,
Eds., University of Chicago Press, Illinois; 1961.
[GI R. M. Gallet, “Radio Observations of Jupiter. 11,” Ch. 14, in
“Planets and Satellites,” G. P. Kuiper and B. M. Middlehurst,
Eds.; University of Chicago Press, Illinois; 1961.
[45] J. \V.\Tarwick, “Theory of Jupiter’s decametric radio emission,”
Ann. X . Y. A m d . Sci., vol. 95, p. 39; 1961.
[46] J. N. Douglas and H. J. Smith, “Presence and correlation of
fine-structureinJoviandecametric
radiation,” Nature, vol.
192, p. 741; Xovember, 1961.
[A71 H. J. SmithandJ.
N. Douglas, “Fine-structure inJovian
Decametric Radio Noise,” Astron. J . (Abstract), vol. 67, p. 120;
March, 1962.
(481 I.A. LT. Injnrtnation Bulletin, No. 8, p. 4; March, 1962.
[49] G. R. A. Ellis, “Radiation from Jupiter at 4.8 Mc,” Nature, vol.
191, pp. 667468; May, 1962
[SO] C. H. Barrow, “Recent radio observationsof Jupiter a t decameterwavelengths,” Astroplzys. J . , vol. 135, pp. M7-854; May,
1962.
[51] J. N. Douglas and H. J. Smith, “Obsemations bearing on the
mechanism of Jovian decametric emission,”Proc. 11th Internut’l
Astropltys. Colloq., Liege, Belgium; July, 1962; M e w . Roy. Soc.
Sci. of Liege, vol. 7, pp. 551-562; 1963.
[52] -4.G. Smith, T. D. Carr and X. F. Six, “Results of recent decameter-wavelengthobservations of Jupiter,” Proc. 11thInterJuly, 1962; Mem.
nat’lAstrophys.
Colloq., Liege,Belgium;
Roy. SOC.Sci. Liege, vol. 7, pp. 543-550;1963.
[53] J. X. Douglas “A StatisticalAnalysis of Jupiter’sDecameter
Radiation, 1950-1961,” Astron. J. (Abstract), vol. 67, pp. 574575; Sovember, 1962.
[54] H. J.Smith,“Longitude Effectsin JovianDecametricRadio
Emission, Astrolt. J . (Abstract), vol. 67, pp. 586-587; 1962.
[55] J. P.Rodman and H. J. Smith, “A Photoelectric Spectrophotometer with
Applications
to
Planetary
Aurorae
and
Stellar
Flares,’ Astron. J . (Abstract), vol. 67, p. 585; 196%.
[56] G. R. A. Ellis, “CyclotronradiationfromJupiter,
Australian
J . P h ~ s . ,vol. 15, p. 3 U ; September, 1962.
[57] L. Landovitz and L. Marshall,“Stimulatedelectronspin-flip
transitionsasthe
source of 18 Mc/sradiationonJupiter,”
Xufure, vol. 195, pp. 1186-1187; September, 1962.
[58] J. X. Douglas, “Time Structure and Correlation of Decametric
RadiationfromJupiter,”
presented a t Symp.onThePlanet
Jupiter, KASA GoddardInsitute
for
Space
Studies,
New
York, N. Y.; October, 1962.
[59] T. D. Carr, “The Possible Role of Field-aligned Ducts in the
Escape of DecameterRadiationfrom
Jupiter,” presented at
XASA GoddardInstitute for
Symp. o n ThePlanetJupiter,
SpaceStudies, S e w York, N. Y.; October, 1962.
of Jupiter’sdecametric
[60] J. 11’. Warwick,“Dynamicspectra
emission, 1961,”Astrophy. J . , vol. 137, pp.41-60; January, 1963.
1611 C. H. Barrow, “A brief survey of the decametric-aavelength
radiation from Jupiter,” J . Brit. Astros. Assoc., vol. 73, p. 4 2 4 8 ;
January, 1963.
[62] J. A. Gledhill, G. M. Gruber, and M. C. Bosch, “Radio o b w v a tions of Jupiterduring 1962,” Nature, vol. 197,pp. 47-75;
Februarv. 1963.
B‘&oRrI “38 h k / s radiation from Jupiter,” A’ature, vol.
[63] CI
197, p. 580; Febyary, 1963.
[MI J. 4 . Roberts,Radio
emission from theplanets,” Planetary
Space Sci., vol. 11, p. 221; March, 1963.
[65] A. G. Smith, T. D. Carr, and X. E. Chatterton, “Spectra of the
-H~:
decameterradiobursts
from Jupiter,” Proc. 12th Intwnat’l
d s t r o n . Congress, .Academic Press, New York, N. Y., pp. 689698; 1963.
[66] G. R. A. Ellis, “Theradio emissions from Jupiterandthe
density of Jovian exosphere,” Australian J . Phys., vol. 16, pp.
74-81 ; blarch, 1963.
[67] J. L. Hirschfield and G. Bekefi, “Decameterradiationfrom
Jupiter," ATuture, vol. 198, pp. 20-22; April, 1963.
[68] G. R. X . Ellis a,pd P. IV.hlcCullough, “Decametric radio emission of Jupiter, Xutzrre, vol. 198, p. 275; *April, 1963.
[69] J. K. Douglas and H. J. Smith, ‘‘Decametric radiation from
Jupiter. I., synoptic observations 1957-1961,” Astron. J., vol.
68, pp. 163-180; epril, 1963.
[io] J. \X-. \t’$rwick, The position and sign of Jupiter’smagnetic
moment, Astrophys. J . , vol. 137, pp.1317-1318; May, 1963.
[ill A . G.
.V. Brown. “Occulta- Smith. S . F. Six. T. D. Carr and G. !
tion theory ’of Jovian’radio outbursts,” ,Vature, vol. 199, pp.
267-268; July, 1963.
[72] C. H. Barrow and F. 11.. Hyde, “An experiment to study the
relationshipbetween radio noise from Jupiterand solaractivity,” J . Brit. Astrmz. Assoc., vol. 73, pp. 3 2 i 328; August,
L
-1
~~
1 ah?
AI””.
[i3] J. X. Douglas and H. J. Smith, “Change in rotation period of
Jupiter’s decameter radio source,“ ;Vature, vol. 199, pp. 10801081; September, 1963.
[74] 11.. LI. Sherrill and h.1. P. Castles, “Survey of the polarization
of Jovian radiation at decameter wavelengths,” Astrophys. J.,
1-01.138, pp. 587-598; August, 1963.
[is] G. R. A. Ellis and P. M. hlcCullough, “The decametric radio
emissions of Jupiter,” Australian J . Phys., vol. 16, pp. 380-397;
September, 196:.
I761 R. L. Dowden, Polarization measurements of Jupiter radio outbursts at 10.1 h k / s , ” Au.stialian J . Phys., vol. 16, pp. 398-410;
September, 1963.
[77] B. B. Lusignan, “Detection of solarparticle streams by highfrequencyradio waves,”J . Geophps. Res., vol. 68, pp. 5617-5632;
October, 1963.
[78] D.B. Chang, “Amplified whistlers as the source of Jupiter’s
sporadic decameter radiation,“Astrophys. J . , vol. 138, pp. 12311241; November, 1963.
[i9] J. X . Douglas,“Time Structure of Jovian Decametric Radiation,” Trans. Arne?. Geophys. LTnion (Abstract), col. 44, p. 887;
December, 1963.
[80] A. G. Smith, T. D. Carr, N. F. Six, and G. Leb; “Decametric
Observations of thePlanets in1961 and 1962, presented a t
XGU Meeting, Boulder, Colo.; December, 1963.
(811 B. B. Lusignan,presented at AGU Meeting,Boulder, Colo.;
December, 1963.
[82] C . H. Barrow, ‘Polarization of 16 M c radiation from Jupitcr,”
::hfure, vol. 201, p. 151; January, 1964.
[83] E. Whittakerand G. Robinson,“Calculus of Observations,“
Blackie and Sons Ltd.; London, England; 1921.
[84] J. 11:. \l:anvick, private communication.
[85] A. J. Smith, private communication.
1861 0. B. Slee and C. S. Higgin, Long baseline interferometry of
Jovian radio bursts,” LVatuue, vol. 197, pp. 781-783; February,
1963.
Radio Telescopes
J. IT. FnTDLAY, SENIOR
MEDER, EEE
Summary-A radio telescopeis usedin radio astronomy to measI. THEL\STROKOhlICALREQUIREMENTS
ure the intensity of the radiation received from various parts of the
RADIO TELESCOPE is used to study the radio
sky. Such a telescope mustbe able both to detect and to locate faint
waves received on earth from a variety of astroradio sources of small angular size, and also to measure the brightsky
ness distribution across extended radio sources or over large
nomicalobjects.Someofthesemaybe
well
areas. Ideally the telescope should be capable of making such meas- known from optical observations, but many others have
urements over a wide frequency range and for different types
of
beendiscoveredonly by radio astronomy. IJTithin the
polarization of the incoming waves.
solar
system, the sun,
moon and the nearer planets are
all
The noisepowers availablein radio astronomyare very small,and
some of the radio sources haveangular sizes or angular structureof, well-known radio sources. Beyond the solar system, but
perhaps, only one secondof arc, so that a radiotelescope needs both
within our own galaxy, are many other radio sources,
high gain and good resolving power. The paper describes various
some of which arealreadyidentifiedwithoptically
types of radio telescopes which have been built and tested, and outknown
objects such as the remnants of supernova exlines the astronomical needs which they fulfill.
The parabolic reflector antenna is fist described, with particular plosions or clouds of ionized hydrogen gas. Ilrithin our
A
galaxy, the neutral hydrogen gas radiates and absorbs
reference to the fully steerable 210-foot telescope at the Australian
National Radio Astronomy Observatory and to the 300-foot transit
the 21-cm line, and absorption by the OH radical a t 18
telescope at theU. S. National Radio Astronomy Observatory. Of the cm has been detected. Beyond our galaxy, other galaxies
telescopes which usefixed or partlyh e d reflector surfaces, those at
have been studied by their radio emissions, both over a
the University of Illinois, at the Nangay station of the Paris Observaof thenearer
tory, and at theArecibo Ionospheric Observatory in Puerto Rico are widewavelengthrangeandforsome
described in some detail. Instruments in which the resolution is im- galaxies,byobservingtheirhydrogen-lineemissions.
proved without a corresponding increase of collecting area, such as
Radio waves have been measured coming from objects
the cross-type antennas, are briefly described. The future progress
so distant that in one case the optical red-shift repreof radiotelescopedesign is certainto follow the development of
sents an apparent recession velocity of more than 0.4
parabolic dishes to still greater sizes, and the exploitation of synthetic antenna systems; the article concludes with a survey of both of the velocity of light. hleasurements of the apparent
angular size of radiosourcesindicate
that some may
developments.
have diameters of less than a second of arc.
I t is thus clear that the astronomical requirements
for
hlanuscript received May 20, 1964.
The authoris Deputy Director of the Kational Radio Astronomy a radio telescope may be so varied that no single design
Observatory, Green Bank, W . \.:a. (Operated by Associated Univerof instrument can possibly satisfy all the requirements.
sities, Inc., under contract with the National
Science Foundation.)