MONTHLY WEATHER REVIEW Editor, ALFRED J. HENRY, Assistant Editor, BURTON 116. VARNEY,
CLOSED APRIL 3, 1926 IE~UED APRIL 30, 1925 FEBRUARY, 1925 VOl. 53, No. 2 W. B. No.859
CORONAS AND IRIDESCENT CLOUDS
By CHARLES F. BROOK^
[Clark University, Worcester, Mass.]
STh'OPBIS
Published observations and some new ones of coronas and iri-
descence are summarized to show (1) the essential facts about these diffraction phenomena, (2) the explanations for them, espe- cially as to the limited part ice crystals may play in their forma-
tion, and (3) the relation between iridescence and cloud tem- perature.
Observations of coronas.-Coronas are the small colored ring3
frequently to be seen about the sun, moon, and, occasionally. bright stars or planets. The order of colors is such that blue is
on the inside and red on the outside. Repetitions of the spectrum
about the moon are observable, rarely, to a third red ring, but out from the sun at times brokenly to a fifth, and possibly a sixth
or seventh red arc. Brilliant coronas can be observed in breath fog or on a befogged or frosted windowpane. The first red rings
of coronas are most frequently from 1-5" from the sun or moon,
but sizes down to 10' from the edge of the moon, and up to 12' or 13" from the edge of the sun have been observed. Second rings,
when seen beyond the first are usually a t 70-100 per cent of the so-called radius, or distance of the first from the edge of the lumi- nary. Third and fourth rings are commonly a t irregular intervals greater than the radius of the first ring. Recognizable arcs of a fourth or fifth ring have been observed to about 30' from the suu,
and of a second or third red ring to 11 or 12" from the moon. In about an eighth of the observations the corona is either dis- tinctly noncircular or irregular, when cloud parts differing in stage
of condensation or evaporation occur at the same angular distance from the luminary.
Obeeruations of iridescence.-Iridescence on clouds is merely a mixture of portions of coronas of differing radii. Sometimes the
mixture is regular, as in the banding parallel to the edge of a len- ticular cloud. At others the mixture is irregular, as in the interior of a thin cloud of differing age or in wisps of fracto-cumulus. Iri-
descence may be seen about the moon at times to an angular dis- tance of 12O, and about the sun to an angle of more than 57", and even on the portion of the sky more or less opposite the sun. The brightest iridescences rarely extend to more than 30" from the
sun. Lenticular clouds give the best displays, especially when they are forming aud evaporating rapidly at low temperature\.
As many as five spectra in succession, as marked by red band\,
have been observed parallel to the rapidly forming edge of a leii- ticular cloud. Iridescence on fracto-cumulus wisps, while showiiig
a connection with the rapidity of cloud formation and evaporatioii, is distinctly related to the temperature of the cloud. E.ctended
brilliant iridescences occur only when the temperatures of t h i a
type of cloud are below about -5" C. A study of 3s cases shoned that when iridescence extended to 20" or less from the sun the coni- puted average cloud temperature was 3" c. and in no instanr?
below -7" C.; when colors were visible to more than 20" blit
not over 30' the average was 0" C. and the coldest -13' C.; with c_oloornE out more than 30" but not over 40' the average w'nq
. and the coldest -118" C., while with colors to more th:rn
40" the average was -15O C. and the lowest -21" C. Iridescence
on windowpanes covered with ice plates of frozen dew is of a differ- ent nature from the cloud iridescence. Cause of coronas and iridescence.-Diffraction of light by sinall particles is the evident cause of coronas and iridescence. It seems
probable that the particles involved are alniost invari~b1.v water
droplets, for the following reasons: (1) Brilliant coronas,*glories, and iridescence have been observed on clouds known to be coni- posed of water droplets, and, i t seems, not on clouds known t u have been exclusively of ice. (2) The most brilliant and wide-
spread iridescence occurs on clouds forming and evaporating with a rapidity probably not possible for aggregates of ice crystals a t temperatures so low as those where these clouds occur. (3) Liquid
droplets have been observed at temperatures as low as those of most, if not all, iridescent clouds. Under rapid condensation a t
low temperatures such droplets seem more likely to form than ice crystals. Crystals, however, begin to form in such clouds of
liquid droplets very soon, for lenticular iridescent clouds often
This aper was completed and Brst, submitted on July 7.5, 1!120. I t has since been
consideraby shortened and sllghtlg revised.--C. F. U.
40160-25t-1
transform quickly into distinctly different, noniridescent, some- times halo-producing clouds of falling snow. (4). An analysis of the sizes of double, triple, and quadruple coronas gives no evidence that any were formed by ice crystals, and shows definitely that in
all but a very small percentage of cases ice spicules could not have been the cause. Only in the case of the small, nearly colorless annuli almost invariably seen in conjunction with halos on the same cirro-stratus clouds, do ice crystal clouds appear to form
cbronas, except in rare instances. The observation of a colored solar corona on a windowpane covered with small spicules, shows that the lack of positiveevidence of colored coronas due to ice spicule clouds can not definitely ex-
clude the possibility of their occurrence. Furthermore, a small, colored triple corona was once observed on a cloud apparently of tabular crystals. Iridescence and coronas as guides to cloud temperature and, there- fore, cloud height.-If, as seems established, coronas and iridescence indicate water clouds, and if, as appears reasonable, we assume that clouds of liquid droplets do not, except for very brief periods, occur at temperatures lowei than the lowest a t which liquid droplets have been observed, iridescent or corona-forming clouds
inay be correctly considered as having a temperature not lower than -3 5 O C. Since the lower the temperature the more brilliant and estensive the iridescence and coronas, brilliance and angular extent of color may be used as a rough guide to the temperature, on the basis of observations iiiade on fracto-cumulus clouds a t computed temperatures. Once the temperature of the cloud is approximated the height map be estimated with the aid of the
probable lapse rate in temperature at the time. Coronas and iridescence because of this relation to temperature should be observed as regularly as halos, and may be used as a criterion in distinguishing "alto-" from ''cirro "-clouds.
INTRODUCTION
Witli dark glasses or a black mirror nephoscope at hand an observer may become acquainted with the beauties
of the frequently occurring coronas and the less common
iridescence near the sun or moon. Numerous observers have published descriptions of these beautiful phenomena
ttnd not a few have attempted explanations as well. The essentials of some of the published information and some
details of a new series of observations will here be pre-
sented and discussed.
OBSERVATIONS OF CORONAS
Dacripiiori of coronas.-Coronas are most readily seen
about the moon, for here the light is not blinding. Near the moon there is a whitish area often with a ring of bluish tinge a short distance out, which grades into an
ortinge-red ring still farther away. Often there is violet or purple and at times a further run of color through blue,
green, yellow, orang!, and red, with increasing distance from the moon. Still a third spectral series of rings is rarely seen. Around the brighter sun the more intense li ht occasionally gives as many as four spectral series.
Ttere appears to be no record of five series in complete
rings having been seen at once, though five, six, or seven partial ones have been observed. The outermost series
are usually mostly bnly diffuse greens and reds. Small,
generally faint coronas are sometimes observed about a bright 'lanet or star. Beautiful coronas are best seen on
clouds or wisps of fog, fleeting portions of coronas occur. Brilliant coronas can be observed in steam or breath fog or on a befogged or frosted windowpane.
sheet c f ouds or thin fogs. Even with tbe passage of frayed
49
50 MONTHLY WEATHER REVIEW FEBRUARY, 1925
Knmtz ____________.________ BenNevls(summer) .-.... BenNevIs (wlntar) ________ Brooks ___________.________
Allobservations ___________ -
Angular &zee of C O T O ~U ;C B .-~ clouds the radius of the
first red ring of a corona is most frequently about 1-5'.
Occasional1 it is less than N o or more than 10". The
sizes in 132 measurements (mostly at Ben Nevis, Scotland)
presented by Pernter,' and 164 noninstrumental observa- tions of my own at Urbana, IlI., October, 1907-April,
1908, and at Worcester, Mass, November 1921,-June,
1923 :
following ta g le shows the frequency of coronas of different
TABLE 1.-Frequencies of angular sizes of coronas
-
The observations cited by Pernter were made with great care, and stated to the nearest minute of arc. Those
made b Brooks were done roughly by estimation
(Urbanay and by crude angular measurement to the
nearest whole or half degree, with pencil or index finger at
arm's length. The Worcester observations include seven
instances of radii varying by 2, 3, 4, 5, or 6' within the same cloud. In each case these are put with the class
showing the smallest of the observed radii, e. g., the observations of radii varying from 4-6" is classed with
others of 4-5".
TABLE 2.-occurrences of markedly noncircidar coronas
~a d i i ______________ 1 Z-P~ 3-601 3-901{ !$I} 4-601 P:O/ 4-g01 +13(?)"1 T O ~~I
:k
0 0 2 4 4 1 0 0 0 3 2 2 0 3 7 2 3 1
0 1 2 3 3.6 7.5 0 0 2 0
1 1 6 16 1 4 1 1 1 1
4 4 11 22 11.5 19.5 3 4 4 3 9a ---------- -
On l e n t i c u l a r
On fracbcumulus clouds ....
The smallest corona measured at Worcester was 10'
to the first red ring, and the largest (only one observation)
13". At distances of less than so Ironi the edge of the luminary. escept in the case of bright planets or stars, the ring becomes narrow and its coloring faint. Such
very small coronas about the moon are culled lunar annuli, and are almost invariably associated with halos
in the same cloud. When the moon is not full, the corona,
particularly when small, is not circular, but flattened vn
the side where the moon is flattened. The first con-
densation on a windowpane, e. g, when a cold wind springs up outside, may give n, corona of So radius to the first
red ring. One's breath fog in the open gives nearly the
same, 7-9' to the first, and (once accompanying a 9'
inner red ring) 16' to the second, red ring. Anglur distances of second tings boyotd the j&.- In double coronas the second red ring usually occurs at a
slightly less distance from the first red ring than that ring is from the edge of the luminary, as shown in table 3:
TABLE 3.-Angular radius of secoiid red ring i n percentage of an- gular radius of the first
Per cent ___________________
ohserrations include 15 measurements of a third ring,
2 of a fourth ring and 1 of a portion of a fifth, as in Table 4 :
TABLE 4.-Angdar measurements of red rings (i n degrees and minutes from edge of sun or moon) for iriple to quintuple coronas
I Red rings, cited by Pernter I Red rings, Brooks' observations 1
First
___.
0 ,
2 3
1 10
1 2 3
1 43
1 28 2 2 3 1 5
1 2''
0 57
.____.-_--
.
-
Second
-__
0 ,
3 1
3 46
2 52
2 35 2 26
4 30
1 55
2 5
1 48 2 25
?4 M 1X 4
2.5-4
5 5
5-8
4-5 7
4-5
3-5-9 1
K
3? 6
The first three of Brooks' observations are from esti-
mated diameters from which so, the diameter of the
sun or moon, has been subtracted before dividing by
2 to get the radius. Among the others, where more
than one angle is given for a single ring, the ring was of different radii in different directions.
Noncircular or irregular coronns.-Mention has been
made above and Table 3 presented, showing that coronas markedly noncircular have been observed about one
time in seven, on the average. Such coronas are most noticeable on clouds in which the different portions vary
appreciably in age and, therefore, in size of particles.
Sometimes when a smooth cloud with thickness gradually increasing toward the interior drifts in front of the
moon or sun the corona is found to present the appear-
ance of a parabola or hyperbola, open to the ed e of the
cloud is densest. Pf the cloud is not smooth the corona
is observed to have an irregular radius, as on wisps of
fracto-cumulus clouds. The greatest departures from
circular coronas, attaining at times 300 per cent of the
distance of the nearest portion of the first red ring, occur with the most ra idly forming and evaporating clouds,
for on the thin e&es the droplets are so small that the
red is diffracted to a large angle, while toward the interior,
where the droplets have been growing for a short time, the diffraction is to an appreciably smaller angle (cf.
Table 8, below). To observe a noncircular corona it is not necessary to wait till such a cloud is seen, for it
can be viewed on a partially befogged or finely frosted window pane almost any cool evenin . When such
or less irregular coloring out to considerable angular dis-
tances from the center of light.
cloud and standino nearest the sun or moon w fl ere the
noncircular coronas are visible there is o B ten much more
OBSERVATIONS OF IRIDESCENCE
Perhaps the earliest full account of iridescent clouds,
showing their relation to coronas and including a satis-
1 Ibld.
FEBRUARY, 1925 MONTHLY WEATHER REVIEW 51
factory ex lanation of the nature of the color, is that by
Sir John %. W. Herschel, published in 1S62.3 Since Eerschel’s time, iridescence on clouds, usually associated
with bright coronas, has been observed and described by many, but few have undertaken systematic observations.
Hildebrandsson, Mohn, McConnel, Carlheim-Gyllen- skiold, Schips, and Arendt have provided the best series
of observations.‘ Scattered observations are available
in various meteorological books and journals and at times
in general science journal^.^ These all deal with observa-
tions in sunlight. Moonlight is generally too weak to produce this effect. Brooks has seen more or less len-
ticular iridescent clouds in moonlight only twice and to a meximum of 12O from the moon. Three out of seventy-
one observations of iridescent clouds at Upsala, 1S66-
1892, were in moonlight.e In sunlight iridescence may
be brightly developed, usually showing bands of color hemming the edges of the clouds but having a tendency to be more or less concentric about the sun. Near the
am, coronas and this iridescence merge into one another
in such a manner that it is evident they are of the same
nature. Since iridescence usually occurs only on ra idly
change of win on the ap roach of a storm or just after
the wind shifts and the s-y begins t80 clear as a storm passes on. The cold, boisterous winds on the rear of a
cyclone are especially favorable to the develo ment of
&her, lenticular clouds.?
, Some unusual displays qf coronas and iridescence.-In
1922 and 1923, at Worcester, Mass., I observed four
t d y magnificent displays of solar coronas and accom- panying iridescent clouds. In some respects these ex-
celled even the wonderful occurrence described by Her- &el. Brief summaries of these brilliant occurrences
will, therefore, it is hoped, be worthwhile additions to data already published. Again and again I have seen
one ortion or another of such displays almost clupli- catel The essentials of these four displays are: * Those of November 29, 1922, and November 25, 1923,
were noteworthy, because of the rapidit with which the
clouds were forming and the colors clanging, for the
occurrence of five spectral series paralleling the edge of
B cloud, in each case, and for the rapidity of the disinte-
gration of the denser arts of the cloud into falling snow.
the extreme brilliance of the color over large areas, for
the extent of visible colors through six or seven spectra out from the sun, tind for the estent of the iridescence to 40° and
45’ from the sun on clouds of two levels at the saiiie time. The display of February 25, 1923, wtu particularly
interesting because of the successive resentation of
levels; because of the well-inarkecl triple corona with red
rings at successive equal distances from the sun and with
B weak sun pillar and lateral portion of a 22” solar halo
or parhelion in the same (‘1) cloud at almost the same time; because of the marked difference in the radii of
the corona on a lenticular cloud along the edge, vs. to-
forming or eva orating clouds, it is usut$ly seen wit K the
iridescent fracto-cumulus and strato-cumulus c P ouds and
R 9
The iridescence of B eceinber 19, 1922, was notable for
beautiful coronas and iridescence on c P ouds at three
I81r John F. W. Herschel, Meteorology, London, 1863, 2nd ed pp. 224-325.
4 Th. Arendt, Irisirende Wolken, Dad TqtUer, Oct. and, Nov.,”1697, 14: 217-224, 244- l52 A bibliogra hlc summary of data avadable at that time.
I E g Nature ?London) sea “Iridescent clouds” in the indexes for vols. 31 (1885-88), mp. by C. Davison ’in issue for Jan. 28 1886, pp. 28W293 2 flgs. 33 (1888-87) esp. letter by T W. Backhode in issue for Mar. 55 1858 p. 486 ddcribinb iridescent clouds
p y tribgulated at a height of 11-23 miles 55 (1&7-88) ’esp. J. C. McConnel p. 533 0. H. Stone p. 581 the latter descrihing ihliant and kxtended iridascence i t C o b d o Springs, Coio., during chinook. Anunusual display is described in Sdcnce (N. Y.), Mar 10,1922 N 8 55,283
8 drct. Zcidchr’ iwi 1895. 1Cf Arendt I& cit.: p.247-248.
I Tie detail& dekriptfka have bean laced on ffle in thelibrarles of the U. 9. Weather Enrean Waahington D. C the Brltisi Meteorological Offlce, London, England, and the R&lm MeteorhogIcal”Institute, Berlin. Qermany.
ward the center, and because of a similar diversity be- tween the coronal radii in different parts of cumulus wisps. Frequency of iridescence.-Most people being un-
equipped with dark glasses are not accustomed to look- ing up, and so never see these wonderful color effects.
But he who will carry dark lasses can often view the
nomenon in such brilliant form as here described for a
number of instances is rare, nevertheless iridescence on clouds is a common, and, as McConnel agrees,O almost
daily occurrence during some periods in winter. Arendt in his biblio raphic sumnlary of observations of iridescent clouds lo fkewise shows that the phenom-
enon is of frequent occurrence, though the frequency
of observation seems to depend largely on the observer. Thirty-six observations by Carlheim-G llenskiold, 97
(including some of coronas on1 ) by Sc z ips (2 years),
42 by Mohn (32 years), 71 citegby Hildebrandsson (27
years) and 65 by Arendt (5 years) are discussed. Schips’ observations showed 56 days with coronas or iridescent
clouds in two years. From November, 1922, to October,
1923, I observed iridescence in 142 instances on 99 days. Iridescence was observed fourteen times on two or three
well-defined levels of cloud simultaneously. Most of these, or 77 cases on 49 days, were during the cold, snowy months December to March, inclusive:
TABLE 5.-Observations of iridescence by C. F. Brooks, November, 19.99, to October, 1993
unannounced color marvels o f the sky. While the phe-
INov.1 Dec.1 Jan. IFeb.lMar./Apr.lMayl June IJuly IAug.16opt. I 0ct.l Total
I I I I I I I I I l l I 1
February was a particularly cold month, with light snow- fall ractically eve day. Forest &e smoke in May
limited the cases cited. While Schips’ observations l1
show the four warmest (brightest) hours winter and sum- mer to be those with the most frequent occurrence, I hesitate to classify my observations by hours, for the times
I observed were usually between 8 and 9 a. m. and 1 and
3 3. m. I made no attempt to watch the sky at times ot il er than when I happened to be outdoors, which was
an average of about one hour per day. On the whole,
I observed iridescence about two-thirds as often as coronas. Angular extent of iridescence.-One feature not often mentioned in descriptions of iridescent clouds is the observation of these colors t o an les exceedin 23O, the
outermost limit reported by IVfcConnel in %is early article.12 In my observations I have usually detected iri- descence beyond 23’. In fact, it was not uncommon to see
the pinkish or reddish hems of greenish clouds more than 40
from the sun and on several occasions to more than 50’:
TABLE &-A iigitlar extent of (solar) iridescence observed by C. F. Brooks
and s une, as well as T t e spending of less time in observing,
Outer limit---.---- >loo 11-2Qo 21-30’ 3140’ 41-50’ 51-6O0 OG!. Total
Cases ______________ I +%I
27-1 481 S i 1 1 , 11 5 1 136
---------
Red tinge only.
t More consistent observations, with dark glasses, would probably make these flgures far in excess of the others.
In some cases, angles, the
9 J. C. McConnel, on the cawas of iridescence in clouds, Phil. Mop., 6th m., 1887 I
24.423: 1880.29.168-189. io Aiendt loc..cit.
11 Summkized byArendt, ibid., pp. 246-247.
12 LOC. cit., pp. 423-424.
52
Computed cloud temperatures 1
(individual cases)
MONTHLY WEATHER REVIEW
Computed clout1 temperature
FEBRUARY, 1925
---
___-
10-20°.--.
2C-3Oo.---
W O O ----
Over 40'-
red and green tints is over 57' (April 19, 1923), while for
red alone it is 124' from the sun (January 26, 1924). I observed two colors out t,o 55' on July 26, 1920, at Washington, D. C.,'s and on February 23, 1923, at
Worcester, Mass. At least two other American observers
have re orted iridescence out to 45' or more from the sun.
G. H. &one's description of iridescent lenticular c.louds
observed at Colorado Springs, Colo., clurin a chinook,
tion equalled in brilliance anything I have seen." Mabel A. Chase recently reported a brilliant display of iridescent clouds in which the coloring extended to 45' from the
sun,'6 while C. Bonacini has just, ublished an account
of k e iridescence, like a web of si& threads;of different
colors, observed to 70' and more of angular distance from
indicating eat brilliance of color, and distri 5 ut,ion from
5 to more t !T an 45' from the sun, shows that his observa-
'
Aver- age
o c .
OC.
~
--
. 7 7, -6, --4(h), --4(h). -2, 6, 14 .-.--.. 1 83 12 1 1 13, -S(h), -i, -3, -2, U(d), 2(h), 4, , 0 6, 6, 7, 10. 13 -18, -18(h), -17(h), -16(dp), -14, -9 -ll(ph), -10, --S(d), -6, -6, -1 ,
3,3(h). 7 -22(h), -20, -lS(h), -E+), -I I (~), -15
-10, -7(h).
I
the man.l6
McConnel. in the winter following. the Dublication of
his first article, observed iridescence %n f o 6 occasions to more than 23', the farthest being 37O.I' He refers also
to unusually vivid iridescence seen in England and Scotland during the winters 1884-85 and 1885-86, de-
scribed in a number of letters to Nature. Colors were visible on these a t less than 50' and more than 130' from
the sun (coronal and glory colors, respectively, each within 50' of the sun and a point direct1 opposite, i.e.,
been of a s ecial type, ancl that one was found to be at
on occasions of brilliant and extensive iridescence, about the sun, to examine clouds for iridescence at angles
exceeding 130' from the sun.
Extent oj-iridescence on j'racto-cumulus clouds at difereiit temperatures.-Observations of diffraction colors on frac- to-cumulus clouds a t different temperatures me classified
in the following table:
TABLE 7.-Diflraction colors on fracto-cuini~lus clouds at diferent temperatures
180' away). McConnel says these clou K s seem to have
least 11 mi r es high. It would probably be worth while
LOW-
est
"C. '
39 (distributed by uouths as follows: Dec., 3; J i ..3 ; Ft
hIar.,S; hpr., 1; hlny, I; June, 3; July. 1; AUK., 2; Sept.. 4; 1
i I Total--l I Oct., 5 .)
1 The cloud temperature we5 computed from the following approximate formula: t. = t d - 0.2 (t. - t d )/ 0.77, in which t., fd, and f. :ire, respectively, the cloud base tem- $%at?, the dewpoint 8t the ground. and thcl eir tempt'rilture at the ground. 311 in "C.
or t h s formula the dewpoint was. obtained Kith dew-point hygromrter in the case of observations marked (dp) wlth hair hygrorueter (h). or, with sling psychrtimetrr (p or no letter). In one instance the cloud height was tibtamed by trlanguliition with the aid of the cloud shadow at known distnnce. and the temperature was then computed at an assumed lapse rate of 1.0' C. per 100 m.
2 This average is not strictly comparable with the others, as observations at less than
20° were not attempted for July-October.
11 C F Brooks Iridescent Clouds Mo. Wealher RPP. June, 1Q20,48: 333, ft. note.
14 a: H. Stone 'Irldeswnt clouds kature. Apr. 21 IS'S7 353580.
I1 Mabel A. Chase, Iridescent clduds, Science, M&. 30,'1922. N . 8. 55263.
11 C. Bonacini, Nubi iridate, Boll. Bitnemrult Soc. Mcf. Ital., 1924, 3433-74,. The whole paper, pp 71-78 extracts from a longer one. presents some other observations on iridescent clouds, men)tions the scarcity of recorded observations of the sort in Italy, and attempts to place iridescence In a different category from coronas by offering to erplaln iridesosnce 89 the result of supposed fine striations on laminar crystals of which the author fh%umes irldescent clouds must be composed. (I am lndebted to Miss E. C. H. Brooks for translating this paper.-C. F. B.)
11 J. C. McConnel, On diffraction colors, with special reference to coronae and irides-
18 Bid., p. 170. ent clouds, Phil. Mug., 5th ser., 29:168.
The data preeented in this table appear reasonably
three minor sources of error: accurate, notwithstandin
(1) Observin
temperature. The angular observations of iridescence depend not only on distinguishing the colors, but also
on clouds being at distances such that the limit may be
established. In the case of fracto-cumulus clouds, the number of units and their motion allows fairly readily
the determination of the angular limits of iridescence, probably c,orrect to within 5'. Determinin the dew point in c.old weather is always difficult. %uring the
months December, 1922, to October, 1923, when these
observations were made, I took almost daily observa-
tions wit,h the sling psychromet'er, and in the cold
weather of December and January supplemented these witah determinations of the dew point wit'h a table-salt and snow mixture in a silver or bright aluminum cup.
Throughout the period a hair hygrogra h was in opera-
point hygrometer and the psyc.hrometer at wet-bulb
t.emperatures above 0' C. may be considered accurate within 0.5' C., those wit,h the psychrometer below 0' C.,
wit,hin 1' C., a.nd those obtained from the hair hy ro- graph, correct wibhin 1.5' c. The temperature at w E 'ch
air ascending from the round would reach saturation
assuni tion is open to errors which are known to be
ieights of cumulus clouds with those computed by the dewpoint formula and found the computed heights
lower (-) or higher (+) by tlie follouing number of
me.ters in loinstances: $55, -42, - 12, + 104, $42, -4,
+35, - 14, - 133, and -5. The average difference is
9 45 in., which under adiabatic conditions, corresponds t,o an average error of 0.5' C. Thus, it seems safe to
assume t,hat the cloud temperature,s indicated are
probably c0rrec.t to within 1' C!. in 17 instances, to within 1.5' C. in 10 more, and to wifhin 2' C. in 11
more. The remaining case, in which tlie cloud height
was obt,aine,d, wit'h the aid of its shadow, by rough
t.riangulation, is probably not more than 1' C. in error. As the most ext,ended irideswnce is unusually observed
on foiining cumulus wisps, the temperatures of the
cloud bases reasonably apyrosimate those of the clouds
on which the iride,scence IS observed. hc.ce.pting Table 7 as it stands, it is noteworthy that
bob11 the. arerage and extreme temperatures are lowest
for the most extended iridescences and highest for the
h s t est,ended, and that there is an orderly gradation
between. Owing to the large range of cloud tempera-
t,ure, 15 t,o 23' C., wit'liin each group it is evident that
further observat,ions are required before any averages
can be dependable witshin a degree or two.
A closer relat,ionship between extent of iridescence
and computet1 cloud tem ierature seem to be prevented
the clouds me forming, that is, the more intense the
convect,ion, t,he greater are the cloud areas wTith droplets sufficiently alike in size to make the colors stand out
plainly. Therefore, the more intense the convection, cloud temperat,ures being equal, the more extended the iridescence is likely to be. Nevertheless, this factor is
not sufficient, to obscure well marked chan es in the
of cloud tern erat.ure on any day. &ur instances may
the angu 5 ar extent of iridescence, (2)
determining t % e dew point, and (3) obtaining the cloud
tion. The dew-point observations ma c f e with the dew-
has been taken as the pro B able cloud temperature. This
Fenera f ly small. Chyton la has compared observed
by the varying int,ensity o f convection. The more rapidly
extent of iridescence accompanying a preciab 7 e changes
be cited of t K e usual increase in the angular extent of
I* H. H. Clayton, World Weather, Table XI, p. 162. New York, 1923.
FEBRUARY, 1925 MONTHLY WEATHER REVIEW 53
5
than in t B le morning, the comput,ed c.loud tempernture
on Fe B ruary 10 showed an extent of iridescence increas-
robably -4" C., to 22" at 1 5 0 . m., w E en the
iridescence with falling cloud temperature from mornin
to afternoon: December 13, 1922, February 10, 18, an June 27, 1923. The observations on December 13
showed a reater brilliance of iridescence in the afternoon
havin fallen from -14" to -11s" C. The observat,ions
ing from 15' at 1:05 p. m., when the cloud ten1 erature
was clou B temperature had fallen to -7" 8 On February
18, while there seems to have be,en no change in cloud temperature, which was probably - IS" C. at '3 a. m. and
11:40 a. m., the iridescence spread from 30' at the
earlier hour, to 40" with reddish tints visible to 50"
out, at the later observation. Here the spread wns apparently owing to increased intensity of convection.
On June 27 the iride.sce,nce estended to 20" a.t 6 3 0 a. m.,
but to 30' at 1:2O p. m. The. c,onipute,d cloud temper-
tures fell from 14' to 10" C!. in t,his period. At both
obse,rvat,ions there was considerable t,urbulence in the
25 mi./hr. west wind. On August 32 the iridescence
increased from 25" to 28" out from the sun from 12 t,o 4
p. m., during a dry, colder and colder gale. September
30, 1923, there was an instance of decreasing radius of
iridescence with rising cloud temperature as a hurricnne drew near. About 11:30 to 12, anvular limits of 32,
35, 35, 38, and 38' were observed, tge comput,ed cloud temperature being about 3" C., vc-hile at 2 p. m. 25'
was the limit, the comput'ed cloud tem erature bein
into line, it appears that, a change of about 3' C. in cloud temperature is sufficient to change the extent of iricles-
cence by 10".
At low tem eratures cumulus clouds are ephemeral,
temperatures much below -20" C. the gahge is so
rapid that iridescence is too fleeting to be extensively observable at one time, unless new clouds are continually
forming in large numbers. At temperatures above 5" C.
condensation of the more abundant su ply of water
vapor so quickly results in iiumbers of iroplets lqger than t,he very small size (about 0.005 mm. diame.t,er)
which be.st form extensive iridescences, that. these color-
* gs are rarely to be see.11 more than 20" from the sun.
8bviousl , at temperatures between about -20" C. and 5' d the clouds are neither too ephemeral nor com-
posed of droplets too large to provide estensive irid-
escence.
Summary.-The observations presented show t,hat coronas are more or less concentric rin s of color in spec-
sun. Iridescence, however, is a bright to faint coloring from any part of the spectrum, in irmgular splotches or
in bands paralleling the edge of a cJouc1. With movement
toward or from the sun the colors of particular cloudlet,s
change as they move through the more or less fised zones
of potential color for the sizes of the cloud particles. On
near approach of an iridescent cloud to the sun the colors take on a more. or less circular arran ement, with a red
ring at a certain minimum radius.. fridesccnt coloring has been noted on several oc.casions by different observers
to an angular distance of 45" from the sun. The bril-
liance, however, usually f d e s rapidly beyond a cliut8ance
of 20' or 30".
Ti,ndowpa.n.e iiidescen ce.-Color bands suggestire of those on the edges of clouds niay be. seen on w~ndo~vpanes
covered with fr0ze.n fine dew. The,se, however, nre sep-
arated from the corona by a colorless zone, and the ar-
rangement of colors is the reverse of that in the corona.
9 4" C. Though some of the observations s o not fall we1
quickly trans P orming into clouds of fallin snow. At
tral successions a t modernt'e angular 3 ist'ances from the
Unlike iridescence on clouds, such bands are, evidently,
produced by the interference of transmitted light with that trnnsmitted nncl doubly reflected from the opposite
faces of the thin plates of ice, held in parallel orientation on the window ane. One cold morning when there was frozen clew a n f a new deposit of frost on different parts
of talle same. windowpane, solar coronas could be seen on bot8h, but the noncircular color bands on the deposit of
frozen dew only.?O
CAUSE OF CORONAS AND IRIDESCENCE
The ex laiiation 21 for t'liese beautiful colors on clouds,
light,. So far as clouds are concerned, coronas and irides- cence are probably due only to diffraction b . very small
pmticles. In passing through the spaces t etween the partic,les t,he light waves s read as from new sources,
wit,h copves wave fronts.
s aces cross, and, in conse uence, make alternating zones
has its own sets of wave fronts, which on interfering pro-
duce bright and dark bands a t different angular distances
from t8he source of light. Thus the bright phase of one color occurs in the dark phase of another, and this leads
t80 rings of colored light at certain an les from the lum-
inary. t of shorter wave
a dusty or cloudy medium, the color seen nearest the sun
or moon is bluish, and is followed in succession by green,
yellow, orange, and red. Then there is another series marking the second zones of reinforcement of the several
colors-violet, indigo, blue, green, yellow, orange, and red. The size of the innermost red ring, and, of course,
of all the others, too, is dependent on the size of the par- ticles causing the diffraction. The smaller the ring the
larger are the particles making it. This relation between the radius of the first red ring and the diameter of the diffracting spheres is given by the following formula:zz
Sin r = .00082/d,
in which r is the angular distance of the first red ring
from the ed e of the luminary, 0.00052 a constant for the f i s t re$ ring (n [=1]+0.22) 0.000671 [the wave length of red, in mm.], and d the diameter of the droplets (in mm.) res onsible for the red ring of radius T. The following tabfe is derived from this formula:
TABLE 8.-Relation of radius of first red ring of corona (r ) to diame- ter of droplets (d )
is t,h:tt t. K e.y are owing to interference in diffraction of
$ he waves from the different
o P double illumination an 8 darkness. Each wave length
As the wave fronts of the li
lengths, e. g., blue, are least diffracte P in passing through
.
From the radius of the chestnut-brown corona Richard-
son has just shown 23 that
5.3 x lo4 cm.
I (radius of chestnut-brown) (diameter of obstacle) =
corona) minus
56 (radius of source). I
~
20 For detnils concerning interference of light see books on physics, or the Encyclopedia Britanniya, 11th ed., vol. 14, pp. 685-693 (Interference of light), and vol. 8, pp. 238-255 (Diffractinn of liehtl. ..
:I For--&%s tri zilirartion ~J Y water droplets see W. J. Hum breps. Physics of the Air, I'biladellhia, IYZO, pi). SB-Sti nnd onditlraction by icespic&sseeJ. C. McConnel, O n dillractiori rolors, wit ti siwcial ;elerenoe to coronae and iridescent clouds, Phil. Mag., 28: 3iZ-?B, 1#9, ?Y: lei-17.3, 1Y90.
Derived from that presented in Humphreys o cit. p. 534.
93 L. F. Richardson, The brown corona and thb 8ametkr of particlea, war Jorr. Roy. Met'Z. Soc., Jan., 1935, 51: 1-6.
54 MONTHLY WEATHER REVIEW FEBRUARY, 1925
From this formula, applicable for droplet diameters of
ordinary size, 5 to 20 microns, Table S b has been com- puted.
TABLE 8 b.-Relation of radius of chestnut-brown ring (r) to diaine- ter of droplets (d )
~
t ________________________
d (in mm.) ______________
Iridescence on clouds, as already described, seems t.0 be merely a mivture of coronas of different radii resulting from the particles being of different sizes in different
portions of .the cloud at the same angular distance from the sun or moon.
H G t o r y and cl-iticism o th,eories a s t o the cause of
tion of coronas and irridescent colors on clouds, as just explained, can take place only as the result of interference
of light of different wave lengths as it is diffracted in passing minute particles. Snow crystals are une ual
fortuitously. Thus, even when the crystals are suffi- ciently small to make a corona large enough to be seen,
their irregularity of form and orientation should so
diffuse the diffraction pattern as to render it practically, if not uite, invisible. When interference colors are
seen on 9 c ouds, it is presumed, therefore, as must usually
be the case, that the particles involved are liquid droplets. The occurrence of such clouds in the air at te,mperatures
f a r below the usual freezing point of water presents no
difficulty, for clouds or fogs of liquid droplet,s have been observed at tem eratures down to -34.5’ C., as will be
noted below. h y liquid droplets form and do not
freeze a t such low temperatures is not known.
The similarity of iridescence on clouds to iridescent
films led Stoneya4 to suggest that this coloratmion was
caused by interference of light reflected from the op- posite faces of thin, transparent lates of ice. Mc-
command, immediately pointed out ob’ections fatal to
cussion (1SS7) was marred, however, by the unnecessary assum tion that all high, and therefore cold, douds
must Be of ice and, conse uently, that most of the iri-
the result of the diffract,ive effect of ice spicules.a7 Never-
theless, his detailed mathematical discussion of c.oronas includes both spicular and water-dro let kinds. Pern-
the larger and brighter coronas must be c,ausecl by ic.e spicules,. for at the large angles to which iridesce,nce was
observed the intensity of light diffracted by water drop-
lets would not be, nearly so great BS t’hat from ice spicules. Sim )sonae quest,ionecl several of Pernte,r’s esplanations
Simpson’s conclusions were based on observations wit81i the Scott expedition in the Ant,rtrctic. On September
24, 1911, he observed a 3s’ fogbow and it8s double, n.t
temperatures between -1 5 O and -?lo F. (-2 6 .1 O m r l
coronas and kv2escence.- l ’0 far as is known, thq forma-
in their several dimensions, and oriented more or P ess
Connela6 with numerous personal o R servationsZ6 at his
Stoney’s hypothesis. His own general 1 y escellent dis-
descence he had seen on c Y ouds had to be explninecl as
terZ8 accepted McConnel’s theory, anc lp assume,tl that ti11
and i l rought it new point of view to bear on the subject’.
-29.4’ C.):
The observation proves that the fog was composed of water drops having a radius smaller than 0.025 mm., and this with a temperature of -21’ F. (-29’ C.). Support is lent to this con-
clusion by the observation that the hair of sweaters and fur bags became covered with hoarfrost, which is a sure sign of super-
cooled water.
The occurrence of water droplets at such low tempera- t8ures was also noted by J. P. I(oc.11 in the far intenor of
Greenland : 30
Today [June 101 we saw * * * the white rainbow [fogbow]
at a temperature of -31’ [C,]. The rainbow occurs always in water clouds, never in ice clouds, so we knew, therefore, that we here had to do with water at -31’. It is naturally possible that the fog cloud had a higher t,emperature than the air on the ground, though at any rate it could not have been over -18’ [the highest
recent temperature] * * *. [June 11, longitude 42’ 53‘ W.] We have again seen to-day the white rainbow, this time at -33.4’. This record for undercooled water will be hard to break.
12 June, 2 a. m., camp site, 42’ 53’. Again the white rainbow is there, this time at -34.5’. The record has been broken earlier
than I expected. (Trans. by C. F. B.)
While Sim son’s and Koch’s observations give us tern-
observed, clouds of liquid dro lets as low as -20’ to
A. Berson in a balloon.31
On Ben Nevis, fogbows, indicative of liquid droplets, “have been observed at all temperatures.”33 Water has been supercooled to -SO’ C.33
peratures P or liquid droplet’s lower t,han any previously
-22’ C. (-4 t o -So F.) had‘ I! een observed in 1S93 by
Simpson says:
It is now generally admitted that while halos are caused by the refraction and reflection of ice crystals, corons are due to difTrac-
tion effects of either small drops of water or thin ice needles. From certain observations made in the Anarctic I was led to doubt the
possibility of ice crystals ever’forming diffraction effects * * *. Pernter’s reasons for believing that corons are produced by ice crystals may be summed up in the three following statements: (a) Coronse are seen on clouds having temperatures much below the freezink point.
(b) The most beautiful coronae appear on light white cirro- vumulus or fine cirro-stratus clouds; and these clouds are always
com osed of ice crystals.
(cy Halos and corone have been observed at the same time; and as halos are a sure sign of ice crystals the coronz must, therefore, be formed in ice clouds.
Shpson made the following points: (1) It was not necessary to suppose that very cold clouds must be of
ice spicules, for he had observed a fo of liquid droplets
coronas, he claimed, were not to be seen in t.he same cloud
at the same t8ime, notwithstanding entries of both at the
same observat’ion. (3) The tendency of spicules to fall with their long axes horizontal should make any coronas
formed by theni brighter above and below the sun or
moon than on the sides. (4) The diverse arrangement of spicules should so weaken their diffracted light, as to make the coronas caused by them so faint as not to be
noticeable. (5) An explanation of iridescence as a phe-
nomenon not always caused in t’he same nirtnner as
coronas is unnecessar-y. Huniplu-eys 34 a,ccept,s Simpson’s treatment wit,h due
c,aution, and does not) touch on Pernter’s elaborate dis-
c,ussion. Fujiwhara and Nakano, 35 however, review McConnel’s, Pernter’s, and Simpson’s explanations and
concludes t81int while t,he Japa.nese studies are not fatal to the ice-cloud t,lieory, (( the wat,er-cloud the.ory of Doctor
Simpson is correct so far ns supercooling can take place.
at a temperature of -15 to -21’ !6 . (2) Halos and
H a. Johnstone Stoney On the cause of iridescence in ciouds Phil. Map. 5th ser.. ‘24:
87-92, 1887. Repr. from kci. Trans. Rou. Soc., Dublin, 2nd ser.: 3:637 flg, 1&7 (flg. 9).
11 Lac. cit pp. 433-434.
11 Some A n t e d and discussed In Nature Apr. 7, 1887, 35533, before his extended dk-lonfn Phil Mag loc cit 1887,1889, lkW. n M C C O ~~, 10;: cit.”l& p .3 ~
28 J M. Pernter,‘Me&orol&ische Optlk, vol. 3, pp. 449-456. Wien L Leipelg 1808. Q. C. Simpson. Coronae and iridescent clouds, Quar. Jour., Rou. Mel’l. So:., Oct., 1912, 38:291401,3 figs.
10 J. P. Koch, Durch die weisse Wiiste. Die dsnische Forschungsreise quer durch
81 R. Siiring’ Wissenschaftliche Lufifahrten, pol. 2, p. 192, Braunschweig, 1800.
a2 McConnei, op. cit.. ism. p. 1%.
JJ E.W.Washburn, An introduction to the principles of physical chemistry, etc., New
Nordgronland 1912-13. (A. Wegeuer translator.) Ref. to pp. 207-208
Pork. 1915. D. 77.
84 o p . cit.; pp. 534-536.
8. Fujiwhara and H. Nakano. Notes on iridescent clouds, Jour. of the Mdcorologlcal Soc. 01 Japan, June, 1920.39th yr., pp. 14.3 flgs.
FEBRUARY, 1925 MONTHLY WEATHER REVIEW 56
The theory [Simpson's] is proved by means of numerical calculation of some typical cases from achual observations and by a simple experiment. [The] hemming or crossing
natureof iridescent color on clouds, and its reponderating
a pearance on thin clouds such as Ci.- 8 u., A-Cu., or
F!.-Cu. are explained." Esner has recently answered Simpson's criticisms of some of Pernter's nnd McConnel's
ex ran at ion^.^^ Exner replies as follows to all but (5) : (1) Even though
liquid droplets may exist at low temperatures, there &e clouds of ice spicules nevertheless. (2) Observations
in Holland in 1918 reported by Van Everdingen 37 include
10 cases of simultaneous halo and corona on the same cloud. "It is therefore apparently certain that coronas dso form in ice clouds." (Trans. by C. F. B.) (3)
Since halos, which are formed only by ice crystals, occur with equal brightness in all directions from the sun, it is
reasonable to suppose that ice-crystal coronas can do
so, too. (4) Com utations do not sustain Simpson's
noticeable in diffraction by spicules. Here the matter
stands .
Do angular measurements of coronas show ice spicule
origin?'-Among the 132 careful angular measurements of
coronas collected by Ramtz or made at Ben Nevis
Obervatory s8 are the dimensions of 56 double, triple, or
quadruple coronas. Now Pernter shows mathemati-
cdlySB that with nonspicular coronas the second maximum
of light outside the central source comes at a distance of
only 506/610tlq or about 78 per cent, as far beyond the h t maximum as the first maximum is from the luminary,
while with spicular diffraction the successive masima
are at equal angular distances outward from the source
of light. Unfortunately, this does not necessarily allow the identification of ice-spicule coronas, on the basis of
the occurrence of angular radii of second red rings at
just double those of the first, for the clouds are not of droplets all the same size within the ran e of the corona.
known occurrence of horizon tal differences which are
the cause of irregular and noncircular coronas an&of
iridescence. Assmann, making measurements on the
Bracken,@ in a quiet cloud layer near the summit, found
that the droplets ranged from 0.005 mm. in diameter at top of the cloud layer, through 0.008 mm. 10
to 0.013 mm. a t the base (SO m. below the top). Braak found a similar range of sizes in c10uds.'~
Assuming that the effect of the upper half of the cloud
would be as if the diffrnction were by particles 0.009 mm. in diameter, and that of the lower half of the cloud by
particles 0.012 mm. in diameter, hhe following would, in
this case, he the sizes of the saccessive red rings produced:
assumption that t 1 e colors would bc too weak to be
T b is true even if we leave out of consi c f eration the well-
meters the veT elow the top, 0.011 111111. a t 30 m. below the top,
By droplets 0.012 mm. in diamekr. _____..
By droplets 0.009 mm. in diameter. ....... 17
The result would be red rings at.. ..I 4%1 [Extinction) 1 Qk/ 13 1 Near 17
On account of the greens of t,he smaller drops falling at the same clist,ance as the reds of tdie larger ones, t,liese.
M J. M. Pernter and F. M. Eaner. Mcte?rologische Optik, DP. 488-485. 2d ed. partly reworked bv F. M. Exner. Wen and Leiuzirr. 1R22. ._ 11 Ibid p. 462.
u C&d by Pernter, loc. cit., 1906. pp. 4 W 0 S
8) Ibid 1st ed. p. 454 figs. 165 165a' 2d ed. p. 483 flas. 1P2 183.
40 R. Akmann,' Micrdcopischi Beodachtunk der lcolken-Eiemcnte auf dem Bracken,
41 C. Braak, On cloud formation, EL Mag. en Met. Obs. te Bataria, Verh. 10, 1923.
41so in 2d ed., pi. 504-506.
Met. Zeilschr., Feb., 1M5, 2343-45.
pp. lL3-27.
colors would, in this instance, not appear brightly be-
tween about 5% and 8%' from the luminar . It is evident t,hat the corona such as might have { een ob-
served on this cloud would have had it,s second red wing almost ex-
the third and fourth
rings would have been at successive distances smaller
than that between t,he first and second. If there were a greater diversitmy in the sizes of the droplets t,he first
nnd t,hus estinguish the color, or leave but a faint Jhase iffuse mmimn of color would generally fall in opposite
corona, perhaps without a second ring. If the sizes were more nearly alike, say, making first red rings at 4' and
4%"' t,he seconds would be at 7" and Sg", making a good first ring at 49;' and a diffuse second one at 7%',
or 177 per cent of the distance of the first, practically
t,he theoretical 178 per cent. There would be no third ring.
This rough antilysis may explain the peculiarities to be noted in the angular measurements resented in
differ in size within the same cloud, it is normal for the
second ring of double coronas to be between 178 ercent
from the sun or moon. As a considerable diversity in sizes will prevent t,he formation of coronas, it is not
surprising that angular distances of second rings larger
than 305 per cent of those of the first are not often observed. Furthermore, in the cases of triple and quad- ruple coronas it is easy to see why there should be a
considerable diversity in the angular dist,aiices between
the successive red rin s; extiiictioiis and reinforcements
representing the third and fourth maximi for the sm&i
droplets and 6he fourth and fifth for the larger, the second maximum for the larger droplets having been
ext,inguished, as in the e,xample cited. In these measure-
ments there is obviously no assurance that any corona was formed by spicules. *
Ice clouds a .d simdtaneowr occurrence of ha.los and
coTonns.-Even if these observations can not indicate ice-
crystal origin of colored coronas, the almost invariably simultaneous occurrence of lunar annuli and halos con- stitutes practically indisputable proof of the ice-crystal
origin of at least this type of corona. Seventeen times in
27 months I have recorded observations of lunar annuli on cirro-stratus clouds. In 14 of these there was a lunar.
halo at the same time, ap arently on the same cirro-
there ap eared to be a faint halo; on a second, a halo waa
nokd, probably because the sky was too heavily clouded to make it visible, though oss'iblp because the thunder-
Another c,orona was visible at the same time on lower
clouds, and a t.hnnderstorm within he.aring distance
occupied bhe northern sky from northwest to northeast. Thus it seems t,hat annuli are rarely observed except on
cirro-stratus c.louds which have halos, and, therefore,. on clouds unquestionablv formed of ice crystals. Although it is possible that such a corona may be caused b li uid
of falling snow which c.onstitutes cirro-stratus,* ice
droplets all the
Tables 3 and 4. As it is normal for clou g droplets to
and perhaps 205 per cent of the distance of t B. e first
blot out rin s where t. fl ey might be expected and inten- sify others F artlier out, a third and a fourth ring, e
stratus cloud in each case. 8 n one of the remaining three
observe B half an hour later; and on the third no halo was
storm cirrus in which it P ormed did not make a halo.
droplet,s from the ndting of the lower portion of t, XI e s eet
4' C. F. Brooks, in bfo. Weathtr Rm., June, 1920, 48:333, and The Met7 Map., May. 1921,56:95. Also C. K. M. Douglas, The bfet'l Mag., Jan., 1921,55274, and June, 1821' 56:1?&127.
66 MONTHLY WEATHER REVIEW FEBRUARY, 1925
crystals of a size small enough to make droplets of corona- producing size would themselves make a corona, though
of smaller radius. Another possibility, the simultaneous
annuli, for, owing to the
are covered with a deposit of small drops of dew or with
ice plates (frozen dew), flat crystals, or ice spicules, coronas may be seen about the moon, sun, or other light
shining through. Such coronas, in the cases of panes
covered with ice crystals are, ap arentl , invariably small and never double, and their co F l ors, w ile bright, not so
pure as those of water-droplet coronas. Except for their
colors, owing, presumably, to a homogeneity greater than that found in a cirro-stratus cloud, and to parallel orienta-
tions in the same place, these coronas correspond to annuli seen on cirro-stratus clouds.
Direct observation of frosted windowpanes thus up- holds the reasonable surmises from annuli seen on
cirro-stratus clouds, that ice-crystal clouds can produce
small essentially colorless coronas.
Water dro lets a8 the cause of nearly a11 if not dl colored
the cause of near1 all, if not of all, colored coronas, an
cence, a penrs much stronger than that which can be resenhi for ice-spicule or other ice-crystal origins.
gbservation of diffraction colors producecl by fogs or clouds known by direct observation to be of water droplets
indicates that clouds of water droplets do produce
brilliant and extensive coronas and iridescence. Fuji- whara and Nakano experimenting with steam fog
reported that colors were visible to 45' from the source
of light. Lcomotive exhaust fog on a cold sunn day sometimes gives a flash of color immediately gefore
disappearing. On March 8, 1923, I saw this occur at
about 40' from the sun. The same effect may be obtained in cod damp weather or in cold weather, when
breath fo readily forms, by blowing in the direction of tt
light, a n t better still, by blowing onto a cold window- pane from a distance of a foot or two. Experiences of
mountain observers, balloonists, and aviators with
gloriea on clouds of liquid droplets are commonly de-
scribed, but never (?) a colored corona or glory on a cloud known to be wholly of ice.
Iridescence on clouds of water dro lets only.-In view of
in different portions of a cloud in order to produce
iridescence, it is likely that only clouds of uickly con-
geneity to be iridescent. While some coronas may be
spicular in origin and others due to hexagonal ice crystals, iridescence probably occurs only on clouds composed at
least partly of water droplets. Certainl the most bril-
liant phases of iridescence are producelonly by water clouds. Iridescent clouds are frequently seen to yield a
fall of noniridescent snow, which is uite distinct from the mother cloud. Under such con(cttions the corona and iridescence usually do not last long enough for a sheet
of crystals to develop saciently to roduce a halo, if,
indeed, the crystals usually are of a Ea10 making kind.
3
coronas.--?' If e evidence in favor of water droplets bein
eapecially of the z rilliant coronas and associated irides-
the marked differences in sizes of c P oud particles required
densing or evaporating water dro lets coul 8 have suffi-
cient local homogeneity yet su 2 cient general hetero-
48 A phenomenon once observed by Douglas, ]bid., p. 274.
'4 LOC Llt , v 8.
However this may be, I have on three occasions observed
iridescence and on eight others a corona without irides- cence, simultaneously with halo or parhelion, evidently
on a cloud made up largely of liquid dro let,s and a sheet
of simultaneous halo and corona the cirro-stratus cloud
of crystals responsible for the halo seems to have been
above the clouds causing the corona. This occurrence of cirrus clouds so commonly attached to or growing from iridescent clouds 1e.d earlier observers to conclude erro-
neously that in s ite of t,he rapidity of formation and evaporat,ion, whic K pointed to water droplets, the irides-
cent clouds must nevertheless be of ice crystals.
Conclwion. on ca,use of cor0na.s and sridescence.-Since, therefore, (1) iridesceme to the greatest limits observed
occurs on clouds known or resumed, with eatest confi- dence, to be of water drop P ets, (2) liquid c Y ouds occur at,
low temperatures, (3) angular measurements seem to indi- cate that coronas occur general1 if not quite exclusively on water-droplet clouds, and (47 there appears to be no instance of a colored corona on a doud known or reason-
ably presumed to be wholly of ice crystals, it seems justi- fiable almost without reservation to uphold Sim son in
contending that coronas and iridescence occur exc P usively
on clouds of liquid dr0plets.4~ All doubt on this score can probably be removed only through careful direct obser- vations from airplanes or airshps driven into clouds
showing coronas, glorks, or iridescence, though a labo- rious computation or diffraction esperiments with suitable
ast'ificid spicules would be helpful.
The conseque,nces of this conclusion, pending direct observations to establish or overthrow it, are interesting when considered in conjunction with the relations
between temperature and extent of iridescence on fracto-
cumulus clouds.
of snow falling from it.. In number o r otilier instances
IRIDESCENCE AND CORONAS AS GUICES TO TEMPERATURE
AND, THEREFORE, GENERAL CLOUD HEIGHT
The angular distance to which iridescence may be seen is not only a function of the size of the droplets involved
bu€ also one of the brightness of the light and the degree
to ivhich whitelight can be shut out, as with dark glasses. Since the moon, which is one-millionth or less as bright
as the sun, can produce visible iridescent %olors to a third mrt-uimum 11' out, or a second one at 1 2 O from the lighted
edge, iridescent colors, per se, are probably strong enough about the sun to be visible to the theoretical limit of 90'
if reflected light and sky light were not so strong. Homo- geneity in the density of a cloud, usually shown by its
apparent smoothness, and homogeneity in the size of droplets over appreciable areas, is essential to the devel-
opment of the widest angles of visible iridescence. Still
another factor is the contrast in the size of droplets between interior and edge of the cloud. This must be sufficient to yield different diffraction colors at the same
angular distances from the sun. Furthermore, the cloud must be of sifficient density to make bright diffraction colors. A very thin cloud at a large angular distance
diffracts too little of the colored light even to be seen. A thick cloud, however, is not likely to let enough light
t h o u h to give coloring, or else it reflects so much light
The temperature at which condensation takes place
has an important bearing on the sizes of the droplets and
the density of the cloud produced, and therefore, on the
that t % e coloring is obscured.
4sG. C. Simpcon, lor. rit.
FEBRUARY, 1925 MONTHLY WEATHER REVIEW 67
extent and brilliance of diffraction colors in the cloud. The lower the temperature the less is the va or that can
that can be formed. Thus the cloud formed at low temperature is likely to be of smaller droplets and less dense than one formed a t high temperature. Since t,he
extremely small droplets are of little consequence when
other sizes redoinmate, the smaller t8he size of the majority of goplets the more uniform are the sizes with-
in a cloud, and, therefore, the purer are bhe colors pro- duced. A cold cloud, thus, is much more favorable bo widespread diffraction colors t'hnn n warm one. Pernt,er's
conclusion, from obserrat,ion of more beautiful coronas
when t,he temperature was low, that t,he clouds causing them were " surely ice clouds'' does not, seem tenable.4e
With moderate-sized particles, such as give the first red ring at 3' from t,he sun, iridescence from t,he particles
of this size would not be expechecl in brilliant form beyond
1 l 0 , the limit of the fourbh red ring, while wit,h small particles, such as make tjhe first, red ring at 6 O from t,he sun, bright iridescence should occur to 21'. With ex-
ceedingly small particles, such as occur in wisps of
fracto-cumulus on cold wint,er days, on which the first red band occurs 10' or perhaps even 12' from the sun, fairly bright diffractmion colors are visible t.0 35" or 40''
the position of the fourt,h red ring. But port.ions of the
fifth, sixt'h, and perhaps even t'he seventh rings are sometimes bright enough to be visible. This extends
the visible iridescence to more than half again as gre.at a
distance as the bright iridescence. To t,he est,ent., t,here-
fore, that temperature controls the size of t'lie cloud
particles, the angular distance to which diffractmion ccdors are .visible ma be used as a rough index to the
be condensed, and, assumin an equal numler Q of nuclei
at different temperatures, t, B e smaller are the droplets
tan erature of the c T oud, and, t,hrough temperature t.o
the 1 eight of the cloud."
stratus and a 9 to-cumulus clouds from cirro-stratus and
clou C f s, that the observer is Tikely
to veTpidly viewmg fo 7- clou b having temperatures appropriate
Iridescence as an indicator of cloud t,emperatures nnd altitudes thus becomes a factor o value in c1ou.d .nomen-
clature. An examinat,ion of Ta d le 9 indkates at once
the possibilit of generally dist,inguishing thin alto-
cirro-cumulus, on the basis of t,he occurrence of irides-
cence. If we may judge from t,he temperat,ures of iridescent fracto-cumulus clouds (see Table 7), it is only with iridescences ext'endin to 40' or more esce t on
to the average heights of cirro-cumulus and cirro-stratus.
(6 Loc cit lsted p. 449. a The' he&ts of 'hdasesnt clouds waa discussed a t length a few decades ago. H. Mob, in an article, Irisirende Wolken (Met. Zcilschr., lW3, 10: 81-97, 240) sought to rove from the late disappearance of sunset llght (7) on clouds that were iridescent
L Sunset that hidescent clouds were a t great heights u p to 140 km. 0. Jasse, how- ever, said (ibid., pp. 3&1-385) he had seen iridescent clobds upually a t moderate eleva- tlona up to about 7 Mw) m and suggested that the late darkening of the clouds observed by Mohn did not &rea& the actual setting of the sun a t their height, but merely the end of more or less mtense indirect lighting there. Mohn re lied (!bid,. p. 460) that the mddennesa wlth which the light left the clouds preclude8 any lighting less direct than sunlight. ReimaM (ibid 1894 11: ) sustained Jesse in saying that his observa- tions showed iridearant cloud;' to hb a e o great height. Hildebrandsson. however (ibld 1885 12: 71-72), cited another observation of late darkening of a previously iri- deaceh clobd, which under the assum,ption of direc! sun,light,,iqdicated a height of 132 km., conforming to. Mohp's computations. E. Schips cited (lbid., p. 312) an obsema- tion of a very beautiful iridescent cloud t.hat could not have been very higb. C. Pasner in summarizing the discussion (ibid., pp. 379-352) suggested t.hat there Kcre two s0rt.s of iridescent clouds, those a t moderate heights. and those at very great heights. If the clouds observed by Mohn and Hildebrandsson had been lighted .hy the sun dlreetlp up to the time when they darkened,, then they should have continued to he iridescent till that time, instead of losing t.heir iridescence at ahout the t.ime of general sunset. It does not take much light to make a cloud,look hright in a dark sky. and it ia reasonable to suppose that even indirect light might he ,cut ,off rather suddenly when the sun ceased to shine on the distant cloud, or whatever it might he that sent tho indirect light. There seems no reason for believing that iridesrent clouds exist to phenomenal heights.
TABLE 9.-Average heights o j intermediate and upper clouds and the ntwage femperut urcs at lhove heights, summer and winter, in the eastern United States and trestcrn Europe.
United States (Wnshington, D. C., and ,
1 Europe (Trappes, France, and 1 Blue Hill, Mass.) Potsdam, Germany)
I I I I
Sumnier I
h-in.
('i. Ht.~. \V., 1Il.f;
13. H .. 10.1
Ci. Cu.. W., R.R
1 B . H.. 8.4
A. St ... JV., 5.R
B. H .. 13.3
A . C.U.. W., 5.0
B. A.. 3.k
B. i r ., fi.;
Aver- I I >\ver- I I Aver-
OC. 1 A-m. " c. A-m. -4 3 9. F, -49 T,, 7.0
--4n I 2.:) -46 J B ,, 8.1
-3'2 , 1 .4 -37 T.,5.8
-3 1 6.1 -3n I>., 5.0 -17 1 6.2 --M ._.__..___
-11 , 4 ,s -19 T ., 3.S
-14 4. f i -18 P., 3.3
-5 ' 3.R -13 T., 3.7
+z 3.7 -I ? P.. 3.fi
= c. - 30 -31 -1s -16
-3
0
-2
-2
- . - - - - -
I Winter 1
I Aver11
1 Heights from H. H. Clnyton, Discussion of the cloud observations, etc., Ann. Astr.
These cloud heightsire from the tbble In'W. J. Humphreys'. Physics of the air, Phila-
dmperatures for western and central Europe were read fr0.m Humphreys, loc. dt.
Obs Harvard College vol. 30 pt. 1 p. 340 18%.
del hia, 1920, p. 306. copied from J. v. Hann's, Lehrb. d. Meteorologie.
Figure 16. and for the eilstrrn United States, from Oregg, loc. cit.. Figure 13.
Herein lies a justification for recommending the use
of diffraction colors for distinguishing alto-cumulus from
cirro-cumulus and alto-st,ratus from cirro-st,ratus. The
value of coronr, observations for this purpose has lon
in observational prackice only in Austria-Hungary and and very recently in the United Statm.m
Brooks recommendations of three years ago 61 require minor modificat,ions, in the light of the stud of coronas
and iride.scence discussed in this paper. 9 quote the
essentials, with the changes inserted in brackets [ 1:
Alto-curnulus * * *. In the vicinit,y of the Bun or moon diffraction colors are usually visible * * *, Alto-stratus * * *. On thin parts of the other (water- droplet) kind, diffraction colors appear in the vicinity of the sun or moon. Cirro-cumulus * * *. Small white flakes or tenuous plobdar masses which [escept rarely] produce no diffraction colors near the sun or moon * * *. Ci.-Cu. being composed of ice particles
[except in their ephemeral earliest stages], are usually bright, in spite of their tenuity, and do not have the solid appearance char-
acteristic of liquid-droplet, A.-Cu. clouds * * *.
The bracketed modifkations introduced should, at least partially, meet the criticisms raised by British meteorolo-
gists.6a Even in their modified form, however, these sugges- t,ions are not in conformity with the usual practice
abroad, which, as shown in Table 8, is such as to place
cirro-cumulus clouds at an appreciably lower elevation
t,han is customary in America, and which, in conse- cuence of the higher te,mperature and usual water-
cxiniulus being designated twice as often RS alto-cumulus
when an iridescent flocculent cloud is observed. Hilde-
hrandsson's observations at Upsah and Arendt's at
been recogni~ec!,'~ but the.y seem to have been applie %
(
t rople,t coniposition of such clouds, results in the cirro-
' 8 Cf. Bericht des Internationalen Mrteorologischen ComitS und der Internationalen Commkion fiir Woljentorschung. Versammlung zii Upsds. 1894, p. 24.
(0 Tho matter of coronns in the int.ernational cloud definit.ions and national instruc- tions for observers is discussed in det.ail by E. Leyst. in Hofc urn Ronne und hfond in Rnrrsland Bull. drc Nnlurd. dr Moacou 19W No. 182 pp. 9-13.
Jo (1. P.'Brooks. Cloud noiuenclnture,' blo.'ll.mlhcr 'Rca.. SeDtemhor. 19, "0.48: 516.
61 Idem.
6' See Mefrorologieal M a g ., 1921, 56: 158-9. 192-3, 21%?0; 1922, 57:183-4, 211; and Qu. Jour. Ron. MfICor. Soc., January, 1923, 49: 3-4.
58 MONTHLY WEATHER REVIEW FEBRUARY, 1925
Potsdam 63 show, respective1 , 40 and 50 per cent of the
as a ainst 24 and 20 per cent on alt,o-cumuli, while Ci.,
per cent of the cases. & my own observations I have
credited cirro-cumuli with only 7 per cent., and alto- cumuli with 45, while Ci., Ci.St., and Ci.-Cu. together
comprise only 11 per cent of the total occurrences of
iridescence. Cirro-cumulus and cirro-stratus may fea-
ture too seldom in t,hese observations, even though I faithfully .attem ted to keep strictly t,o t8he current
We should not forget the fundamental basis of height
in oiir International cloud forms. Therefore, as ori-
ginally intended,65 all reasonable care should be exercised to reserve the names cirro-stratus and c,irro-cxmulus
for clouds that are distinctly hieher than alto-stratus
and alto-cumulus. Thinness an$ small apparent size
of elements in the hi her clouds are primary criteria, but the thin and sma fl -size initial phases of the lower
ones should not lead the observer to misname them with the names of the higher. The occurrence and
angular extent of iridescence seems to provide a hitherto unused aid in differentiating what mieht be called
pseudo-cirro-stratus and cirro-cumulus, wkch are really alto-stratus and alto cumulus (“* * * finer flakes
(resembling Ci.-Cu.)”)
G’onclwrion.--The apparent value of the extent of
iridescence as a rough index to tem erature, and, there.-
fore, to appr0ximat.e cloud height, s E ould justify, (1) the
regular use of dark glasses by observers, (2) the rou h angular measurement of the radii of coronas and t i e
extent of iridescence, and (3) the entry of such observa-
tions as an essential part of the cloud record. Further-
more, systematic observations of the heights and t,empera-
tures of iridescent clouds should be undertaken at
aerological stations, in order to establish the degree to which angular extent of iridescence on different cloud
ty es forming at different rates may be used as an
iniication of cloud temperature and height.
USING WEATHER FORECASTS FOR PREDICTING FOREST-FI RE DAN Ci ER
By H. T. GISBORNE, U. S. Forest Service’
[Priest River Erperlment Station]
iridescence observations t.o E ave been on cirro-cumuli,
Ci. -4 t., and Ci.-Cu. grou ed together include 58 and 78
International de l nit ion^.^'
from the true and higher types.
Three kinds of weather control the fluctuations of
forest-he- danger-wet weather, cl weather, and windy
weather. fluctuation of fire danger. These are the occurrence of
lightning and the activities of man. But neither of
these fire-starting agencies is fully effective unless the
weather has dried out the forest materials so they are
9 orest fires can not be started and will not spread
unless the forest fuels are dry. Wet weather ma,kes the
fuels wet, dry weather makes them inflammable, windy
weather fans the flames and makes the fires most diffic,ult
to control. If the degree of wetness, dryness, and windi- ness of the weather can be forecast accurately in timc
and place, fire danger c,an likewise he forecast with suffi-
cient accuracy to improve very greatly the efficiency of
forest-fire detection and suppression. The ur ose of the present article is t,o illustrate some of t,{e letailed
rocedures involved in t,he process of translat,in weather
t8ypes of northern Idaho and western Mont,ana.
Two other conditions a P so contribute to the
enough to burn.
forecasts into fie-danger forerasta for the con’ l! er timber
33 Arendt. OD. cit.. p. 2B.
34 International Cloud Atlas, Paris, 1910.
13 Cf Brooks op. cit
M Internatioh Cloid Atlas, Paris, 1910. Part of designation of alto-cumulus.
Investigations of the relation of weather to f i e danger
the Priest River Forest
devot,ed to the compilation and comparison of
and Forest Fires in Montana and Northern Idaho, 1909
to 1919,”’ by Larsen and Delavan, gives s ecific data on bot,h weather and fire fluctuations. ’#he present
object of fire studies, however, is to make available to
the fire-fight,ing organization all possible information concerning present and probable fire danger so that that organizat,ion may expand t.0 meet incre.asing danger and
cont.ract to save unnecessary ex ense whenever ossible.
not always me,an a certain degree of fire danger in this
region. The effect of that hot, dry weather depends on how wet the fuels were to be in with. If it has rained
uired before extreme danger will result. Likewise,
Pollowing a drought, the forecast maybe for aperiod of high humiclit , or rain, and the effect will depend on
high the humidity may be or how much rain may fall.
Before weather forecasts can be used accurately in determining what ‘protective action should be taken, it is nemssary to know the prevailing moisture contents of the various fuels.
Studies a t the Priest River Forest Experiment Station in nort,hern Idaho have shown that the top layer of duff (decaying leaves atid twigs covering the mineral soil)
responds t.o weather chmges about as the average of all t’he coinbustible forest materials, from inoss,, wee Js, and
t,wigs, to slash and the outside wood on windfalls and
snags. The fiuer arid lighter of hhese fuels pick up and
lose moisture rapidly; the heavier fuels, such as branch-
wood, etc., respond more slowly. The top layer of dufT seems to be a reliable criterion of the average response.
revailing moisture content in that to layer of duff, caUe B a duff hygrometer, has been invente B by the U. S. Forest Products Labora-
tory and the Priest River Station. Numerous tests of the inflammability of duff in relat,ion to its moisture content
have permitted the delineation of six z0ne.s of inflamma-
bility-none, very lorn, low, medium, high, and extreme. By this means it is possible to apply weather forecasts to a reliable base and so obtain a translation into t e r n of
fire danger. Past, practice has shown that such a trans- lation can not be made with sufficient accuracy without such a base to build on.
During the past fire season (1921) three duff hygrom-
eters were used t.0 measure prevailing duff moisture contents on thrAe different sites in the vicinity of the
Priest, River Forest Experiment Station in northern Idaho. These three sites may be tormed, (1) moist site, a ftdv t,inil)ered nort,liwest d o e; (2) medium site, a
srt.in’lly cut-ovcr knoll top; (3 P dry site, a clean-cut’, h l y c.spnsed flat,. Figiire 1 shows the fluctuations of
moistiiw content, recorde,d, also the various zones of
inflnaum~thilitS, HS previously described.
As ciigiit be expected, these three sit,es, all within a
circle Iers t,liwn a milc in di:rmeter, general1 exhibited very difforent degrees of fire danger, t8he fu J y t,inihered
st,nt,ion uswlly showing the most niois ture, the clean- cut, mea t,he least,, tmd the partially cut area an inter-
?nediat,e amount. Table 1 shows the ercentage of time during w1iic.h each site esperienced t, B e various degrees
of infianimability.
were initiated in this region b
Experiment Station in 1916. !i! hese first researches were
recor large$ s of weather and forest fires. The report, (‘ Climate
A forecast of several clays o P hot, dry weat R er does
recent,ly, a week or more of a rying weather may be re-
how dry the r uels were to begin with, as well as on how
An instrument for meixmring the
1 Mo. WEATHER REV., Feb.. 1932,50: 6E-@8.