ELLIS-ON THE DIURNAL V.~RIATIOY OF THE BAROMETEII .u ~~REESIVICII. 467

31r. WHIPPLE said that owing to the great pressure of work and thc small staff at the Iiew Observatory, the sunshine records for last year liad not been discnssed, but he had with him the results of the observations for the first four months of this year, which showed in January, 39 hours of sunshine against 19 liours recorded a t Greenwich ; February, 73 against 36 ; March, 101 against 93 ; April, 75 against 72. The difference between the two, he would suggest, was owing perhaps to Greenwich being more sheltered b London smoke than Kew. He did not think that blackened paper would be getter than unbIackened, as neither would register when the slightest cloud passed before the sun.

Dr. TRIPE did not doubt that sunshine had a great effect on health and v e w tation. I t mould be useful if the relative intensity of the sunshine could be mcasured, as that affects vegetation and health more than duration; and possibly if this mere done it mould throw light on the question why some t o m s are more subject to diarrhea, &c. than others.

blr. SCOTT said he had on the former occasion adverted to the fact that Mr. Campbell had shown, that in the cholera year in London very little wood had beeti burned in his instrument.

Mr. SYXONS said that Roscoe, Marchand and Winstanley had done miich for tlie subject of sunshine, and he would especially refer t o M. Marchand's work, " Eturlc sur la force chimique contenue dnns la. lumiBre du Soleil, la mesure de sa puissance et la dEtermination des climats qu'elle caractbrise. Paris, Gautliier-Villars, Svo, 1874 ''

Mr. ELLIS said that the register of duration had been at once commenced at Greenwich, as being a comparatively simple matter, but any continuous registcr of intensity did not seem to be so easy of accomplishment.

XLV. On the Diurnal Variation of the Barometer at t h ~ Royal ODseruntor!i, By WLLLIAM ELLIS, F.R.A.S., of the Royal Obaerdory. G-reemcich.

(Plate XVIII.)

[Received June 1st.-Read June 20th, 1877.1

IT will be in the remembrance of those Fellows who were present a t the last meeting of the Society that a t the close of the mecting I exhibited some small diagrams of curves showing the diurnal variation of the barometer for months and for years, as deduced from the photographic records taken n i the Royal Observatory during the twenty years ending 1873. A strong wish

for months, I mentioned this wish to the Astronomer Royal, who at once assented to the communication to the Society of the numbers in qnestion. They are contained in the accompanying table, to which I propose to add a few general remarks.

In order better to show the character of the diurnal variation I have pre- pared an illustrative diagram (Plate XVIII.) giving the Greenwich curve for each month, but smoothing the small accidental irregularities. The curvcs apply to the civil day, commencing at midnight, the first six hours being repeated at the end of each curve to show better the form of the curve near to midnight. The thick horizontal lines indicate mean values. The spaces between the vertical lines represent each two hours, and the spaces betveen the horizontal lines each 0.01 inch of barometer reading. The points ni which the dotted lines cut the curves indicate respectively the times of .;iinri-c and sunset.

having been afterwards expressed for the numerical values of the vai i - a t' Ions

A E W JERIES.-TOL. 111. P U

TAB

LE o

f th

e D

iurn

al V

aria

tion

of t

he B

arom

eter

in

each

Mon

th (O

r Ex

cess

of

the

Mea

n R

eadi

ng a

t eac

h ho

ur a

bove

the

Mea

n R

eadi

ng fo

r the

M

onth

), w

ith t

he M

ean

Rea

ding

for

eac

h M

onth

, a8

ded

uced

fro

m a

dis

cuss

ion

of

the

Phot

ogra

phic

Rec

ords

tak

en a

t th

e R

oyal

O

bser

vato

ry, G

reen

wic

h, d

urm

g th

e Tw

enty

yea

rs e

ndln

g 18

73.

Nov

. I

Dec

. ----

Hou

r. 1-F

For

thc

Mid

nigh

t ...

......

...

I a.

m. ...

......

.....

2

......

......

....

29.8293

~ ,

5 ...

......

......

. 6

I, .....

... ...

.I $::::;

Aug

.

In.

+.0069

+003

5 -*

om5

--'me

--.0069

--'oo13

+a00

30

--'0057

5:::;

0109

+008

4 +'0037

--.ow4

--0040

--0087

--.or13

-'0137

-'0134

-*06

01

--'008g

+006

4

$1::;; 29.7990

[Cor

nmun

icat

cd b

y th

e A

stro

nom

er R

oyal

.] Sept

.

In.

+.oo71

+oo

3t

--00

16

-900

62

--0098

--00

63

--.0003

+or1

8 +0

090

+'00

57

f.00

06

-*0036

--00

88

--oil3

--

--.or06

$::;IF;

-*01

16

-so086

--'o

oze

+'00

39

+'0072

$-:

3;

----

29'7873

Feh.

In.

+.ooz9

+000

6

--.0063

--0089

--0090

-*0087

--oo+

+ +.

0028

--'0017

7 ...

......

......

. 8

....

....

....

. 9

......

......

....

:o ..

......

......

.....

:I ...

......

......

do

nthl

y M

eans

....

.. -

I_

Mar

ch.

Apr

il.

In.

In.

+.OIOO

+*004E

+.oo81

+.ooz

e +*0037 --.ooIa

--.ooz6

--.oo

4E

--0060

-.006

8 --.0064

--0066

--00

40

--.oo17

+000

9 +0034

+'004g

+0072

---I_-

May

.

--'a013

-*004z

-*0069

--.0066

--.0040

In.

+0069

+0041

--'0

015

.0028

$ ,0048

+.0072

+.or08

+.0005

'000 I

$::2;

+0045

+O

OIZ

--'0034

--'0

110

--'0073

--.0140

--'ox 56

--'ox45

--'o I 04

-*00

3 I

+*oo

85

+.0098

+'0057

-'0035

-.006c

--.0063

--'0043

--'0

002

June

.

3 1,

...

......

.....

4 ,,

....

....

....

5

,I

......

......

.. 6

,, ...

......

.....

7 ;I

...

......

.....

8 ,,

......

......

.. 9

,I

......

......

.. 10

,, ...

......

....

11

,I

...

......

.....

Noo

n ...

......

......

. I

p.m

. ...

......

....

2

,, ...

......

.....

3 I)

...

......

....

4 ..

....

....

....

In.

+.0070

+0040

--'0

02

8

--.0036

--.0

01

8

+ooo

z

+*0

01 I

+

00

5

I

$::Z

$:;; +.

oo85

-'ooI

I -'0045

--0089

--'0

1 '9

--'or41

--orog

--.0048

+.0058

-'0145

+00

85

+009

9

--'0049

--'0

122

--.00

87

--.or30

-*0

03

2

*OO

# .or04

+*0024

--0060

--.0087

---'

OO

SI

--.0016

--'0099

$. '0112

July

.

+'OX33

+*00

83

--'0014

-*OIIO

--or09

--'0080

-'0081

-'OOI7

In.

+*00

84

+o04

8 +.

oorz

--.0023

-*0

02

8

--'0020

+0002

+oo4

1 +*0076

+.0086

+*oo82

+a0069

+'0039

--'0

00

2

--'0032

-*007

I --'or07

--'o I 40

---or42

--.0108

--'0049

+*007

I +*

@33

8

+-79 --

+'OIOO

+009

7 +.0069

+m5n

-'

OO

O~

+.OOII

-.0124

--'or07

--a150

--or32

--.or@

--or

38

-'0073

--'0043

-'00Q8

--'0

12

2

-*0068

-*0046

--0023

20

.80

87

-*OIO~

--'o

r03

--.0076

Oot

.

In.

+.0089

+.0064

+.oo17

--0041

-*0071

--0081

--.0

023

--'0073

+0054

+*00

88

+m96

+007

9 +wo9

--.0078

--or

16

-90137

--0096

--.003 I

+0

02

3

--'0143

+o

os

s

+*0100

$:z

; .-- 297199

In.

+-00

88

+'0046

+.002

1

-*0023

--'00~~

-*0066

--00

68

--.0

02

8

+-oo41

g;:;

+.

0103

--'0059

-*01

09

--or13

--.0098

--oor6

+.OOIQ

--'0057

+oor

o

+'0043

+-37

$:::;; --

29.7626

The

Hou

rly

Val

ues

cont

aine

d in

this

Tab

le a

re s

uoh

as a

re g

iven

by

obse

rvat

ion,

with

out t

he a

pplic

atio

n of

an

y pr

oces

s, g

raph

ic o

r oth

erw

ise,

The

Mon

thly

Mea

ns a

re r

educ

ed t

o th

e St

anda

rd B

arom

eter

New

man

, an

d co

rrec

ted

to 3

2'. N

o co

rrec

tion

has

been

app

lied

for

elev

atio

n.

for t

he s

moo

thin

g of

sm

all a

ccid

enta

l irr

egul

ariti

es.

The

cis

tern

of t

he B

arom

eter

is c

onsi

dere

d to

be

I j9

feet

abo

ve t

he m

ean

leve

l of

the

sea.

ELLIS-ON THE DIURNAL VARIATION OF THE BAROMETER AT GREENFPICH. 469

It will be seen'that there are the usual double maxima and minima, indicating semi-diurnal inequality, or inequality with period of 1 2 hours. There is the morning minimum, the forenoon maximum, the afternoon mini- mum, and the evening maximum. There is, also, in most months, evident indication of a diurnal inequality, or inequality of 24-hours period, as shown by the greater depression of the second or afternoon minimum as compared with the morning minimum, and a tendency to greater depression in the second or evening maximum as compared with the forenoon maximum.

I have marked on the curves by short lines the positions of the maxima and minima, They are simply estimated positions, but will serve quite well for pointing out certain peculiarities. Comparing together the different months of the year, it will be seen that the morning minimum and the fore- noon maximum both occur earlier a s the year advances: they are emliest in summer, and become later again on the approach of winter. The afternoon minimum and the evening maximum, on the contrary, occur Iater as the year advances; are latest in summer, and become earlier again towards the end of the year.* They all change in a certain degree with the change in the times of sunrise and sunset. As a consequence of this the intervals between the morning minimum and the forenoon maximum, and between the afternoon minimum and evening maximum, do not change very much through the year, whilst that betmeen the forenoon maximum and afternoon minimum is much shorter in winter than in summer, and that between the evening maximum and morning minimum is much longer in minter than in summer. An examination of the volumes of the Oxford observations shows that the same thing takes place at Oxford.

It is seen that although the diurnai or 24-hourly inequality is well shown in the Greenwich curves in nearly every month of the year, in one or two months it is less perfectly shown. But still it is a real, and not an acci- dental inequdity. To illustrate this point I have prepared another diagram (Plate XVIII.), in which, using the same scale as before, I have given mean diurnal curves for various places at which good determinations have been made. For Greenwich (latitude 51;)' north) for the average of the whole 20 years; for Oxford (latitude 51%' north) for the average of 1 7 years' observations (1858 to 1874); for Washington (latitude 59' north) for the average of 5 years' observations (1862 to 1866); for the Cape of Good Hope (latitude 34' south) from observations made between 1841 and 1846 (about 5 years); and for Ascension (latitude 8' south) from observations made between 1863 and 1865 (about 2 years). The Oxford, Washington and Cape curves depend on observations made at the observatories at these several places ; that for Ascension depends on results contained in a volume recently published by the Neteorological Office.

* A slight inaccuracy is introduced into the comparison together of different months of the year, in consequence of the employment (as usual) of mean solar time instead of apparent time. I t is, however, sufficient to remark the existence of the inequality, as its effect is practically unimportant in considering the broadly mnrkocl changes spoken of above.

470 QUARTEBLY JOURNAL OF T E E METEOROLOGICAL SOCIETY.

The general similarity of the curves at these five different places is suffi- ciently apparent, and also the increase in magnitude of the variation as the equator is approached; the variations at Washington and the Cape being greater than those at Greenwich and Oxford, whilst that at Ascension is still greater. The approximate positions of the points of maxima and minima are marked by short lines as before. It will be seen that in these mean curves they occur at nearly the same times at all the five places. But with regard to variation in the times of occurrence of the maxima and minima in different parts of the year, pointed out as existing at Greenwich, it may be remarked as follows :-

At Ascension (latitude 8') there appears to be no strongly marked effect of the kind. At the Cape (latitude 34O) there is some such tendency, especially as respects the afternoon minimum, which occurs much later in the Cape summer than in the Cape winter. But at Washington (latitude 39') and Oxford (latitude 614') the variations in the times are strongly marked, and are altogether similar to those existing at Greenwich. They thus appear to be greater in the higher latitudes; but they are not apparently accompanied by change in the amount of the diurnal variation; for taking in each month of the year, the difference between the principal or forenoon maximum and the principal or afternoon minimum for Greenwich and for Washington (at both of which places the variations during the year in the times of occurrence of the maxima and minima are large), it is found that the mean difference for the six months from April to September at Greenwich is 0.024 inch, and for the six months from October to March is also 0,024 inch; and that at Washington the corresponding mean differences are : from April to September, 0.067 inch, and from October to March also 0.067 inch.

Reverting now again to the diurnal or 24 hourly inequality, it Foil1 be seen that it is distinctly indicated in each of the five mean curves by the unequal deviation of the two maxima and two minima from the zero line, and it has the same general character in each curve. It is much more strongly marked at Washington, the Cape, and at Ascension, than at Greenwich or Oxford. It is especially apparent in the Washington curve. And on examining the numbers for the different months at Washington and Ascension, it is found to be as strongly marked and to possess the same general characteristics at each of these places in dl months of the year. A similar remwk applies with nearly equal force to the Cape. But at Greenwich, where the whole diurnal variation is SO much smaller, the general character of the diurnal or 24-hourly inequality appears to be a little modi6ed in some months of the year, as is seen in the diagram showing the monthly curves for Greenwich,

The result of this is that in the daily variation of the barometer the diurnal and semi-diurnal inequalities are both real inequalities, representing real physical effects. Had methods of calcdation been applied, we should probably have deduced other smaller inequalities of shorter period. But lesser inequalities may be in part accidental, and it was rather my wish on the present occasion to point out the connection of the diurnal anJ semi-

ELLIS-ON THE DIURNAL VARIATION OF THE BAROXETER AT GREENWICH. 471

diurnal inequalities in a graphical way, because the general effect seems to be thus more readily perceived.

It seemed to me, however, that it might still be interesting to calcnlate the constants for the ihe mean curves, because they enable us to study more completely the peculiarities of each curve. The whole diurnal variation at the several places may be nearly represented by the following formulq in which s denotes the interval of time, expressed in degrees (15 to each hour), elapsed since midnight :-

In. 0 In. 0 In. 0

Greenwich . . . . . . = + '0029. sin ( z5'4+s) + *oogo. sin (142*1+22) + '0012. sin ( 5.0 + 3s) Oxford . . . . . . . . .. = + '0034. sin ( 51'2fs) + .oog8. sin (157'6+zs) t '0010. ein (350.5 + 3s) Washinqton . . .. = + .or86. sin (348.64-2) + '0175. sin (164'3+22) + *oor6. sin (351.8 + 32) Cape of Good Hope = + .0048. sin (344'7+2) + '0195. sin (161-3+2s) + 'oor5. sin ( 17-6 + 3s) Ascension . . . . . . = + 'OIIZ. ain ( 17'1+2) + *0282. sin (1~8.zfzs) + .0006. sin (126.7 f 32)

I mentioned at the last meeting of the Society that an arrangement of the Greenwich hourly barometer values for the whole 20 years, with reference to the time of the moon's meridian passage, results in giving no certain indi- cation of lunar diurnal variation, although a small lunar diurnal variation would appear to have been detected at some places. We have seen, however, that the solar diurnal variation is, in comparison, large; and this points to the action of heat as its principal producing cause. The changes, in the course of a year, in the times of occurrence of the maxima and minima of the solar variation (spoken of in an earlier part of the paper) also point to the influence of heat. It would appear that the order of these changes may be stated generally as follows :-Near the equator there is not much change in the course of the year in the times of sunrise and sunset; neither does there seem to be much change in the times of maxima and minima in the diurnal barometric variation. But in high latitudes there is great difference in the times of sunrise and sunset in different parts of the year, and (as appears by the observations at Greenwich, Oxford and Washington, and in some degree by the observations at the Cape) corres- ponding great difference between the times of occurrence of the maxima and minima in the diurnal barometric variation. That is to say, in high latitudes the morning minimum and forenoon maximum appear to be both accelerated when the sun rises early, and the afternoon minimum and evening maximum both retarded when the sun sets late; whilst in low latitudes, where there is little change in the times of sunrise and sunset, there appears to be also little change in the times of the barometric maxima and minima.

DISCUSSION. The PRESIDENT (Mr. Eaton) remarked that the pressure of the atmosphere at

Greenwich for the period embraced in the paper was the same as the average from a much longer series of years. On a former occasion he had communicated to the Society a statement of the average height of the barometer in London for S3 years." €Ie had now iiicolporated the results of the 14 years since elapsed,

a Ytde Proceedmp of the BntiJh Xcteorologio&l Society: Vol. I. p. 5 3 .

472 QUARTEaLY JOUBNAL OF T H E METEOROLOGICAL SOCIETY.

and the results were shown in the accompanying table, which also included the values given by Mr. Ellis, reduced to the same standard, viz. sea level.

Average height of the barometer, reduced to sea level, in London, for 97 ears,

years from the photographic records.3 comprising the Greenwich observations since 1843, and at Greenwich P or 20

1774-80 Difference

In. In. Month. 1787-1876. from average.

January .. 29.947 -woo7 February.. *962 + ~008 March , . . . '957 + .003 April .... May ...... June .... Jnly . . . . . . August .. September October .. November December ..

'943 -968 '999 '969 '983 '97' '905 ,

'901 29'938 --

Year .... 299.54

- '011 + .or4 + '045 + *ox5 + .0=9 + '017 - '049 - '0.53 -0.016

18.54-73. In.

29'907 30'010

,980 29'951 3 o * w o 29'979

.969 '959 '895 '939

29,969

29'955

Zg*gOO

-

DifTerence from average.

In. -0'048 + '05.5

- '055 + '02.5 - '004 + '04.5 + *024 + * 0 I 4 + '004 - *060 - 'or6 +w014

Notwithstanding this accidental accordance in the average annnal reading of the barometer, it would be seen that the period of 20 years was not nearly long enough to determine the true month1 averages. Moreover, the sudden altera- tion in the progression of the diurnaf curve of pressure between 11 p.m. and midnight, noticeable in several instances, especially November, showed that the distribution of pressure due to the change of season sensibly affected the curves of monthly diurnal range.

Mr. HARVEY SIJIMONDS said that the time and extent of the morning maximum and of the afternoon minimum at any place varied with the sun's xcnith distance ; that the phenomenon was most marked within the tropics, and was hardly dis- tinguishable within the polar circle.

Mr. C. HARDING said that the connection between the times of occurrence of the various maxima and minima and the times of sunrise and sunset was clearly pointed out by Kaemtz, and was also alluded to in numerous other works. The difference in times of occurrence of the extremes though less marked in the tropics was yet clearly shown: he quoted the diurnal range worked from ship's observations for 0" to 10" N. lat. and 20" to 30" W. long., where the morning maximum ranges from 9h. 10m. in February to 9h. 52m. in June, and at Ascen- sion the morning maximum ranges from 8h. 48m. in January to 9h. 47m. in July. H e referred to the apparent influence of land and sea on the amplitude of oscil- lation in the diurnal variations, remarking that the amplitude at Bombay was in excess of that at Ascension, or that just north of the Equator in the Atlantic.

Colonel KNIGHT had found from observations taken at Allahabad, in India, similar diurnal variations of the barometer to those described b Mr. Ellis ; they

matter because he knew there were some persons who believed these variations of the barometer only took place near the ocean, and accounted for them by the condensation of sea vapour, but his observations had been taken at many hundreds of miles from any sea.

Mr. ELLIS said that the omissions from loss of the photographic register probably amounted to about 2 per cent. There were also a few periods during which alterations were being made in the registering apparatus. I t has been assumed that as regards diurnal variation these omissions would exert no sensible influence on the general results, indeed no means of making corrections exists.

sometimes were nearly so great as the tenth of an inch. d e mentioned the

average height of the barometer for tio:Treduced to sea level, is 20.054 incheu.

these years from the volumes of 'I Greenwich 0bsen.a-

MARCET-CONTRU3UTION TO T H E METEOROLOGY OF CANNES. 473

But in calculating the monthly means the omissions occurring through alteration of the apparatus have been tilled up from the eye observations. He did not lay claim to any thing new in what had been brought forward, but in exhibiting the Greenwich results, it seemed to him that it would be interesting to point out their various peculiarities, and also make some comparison with results found at other places. With regard to any supposed difference between the diurnal variation on the coast and in the interior of a country, he might dram attention to ? papcr by Mr. Buchm, in a recent volume of the ‘‘ Edinburgh Transactions,” in which is given a map showing lines of equal barometric variation ovcr a large extent of the earth’s surface.

XLVI. Coxtribut ion to the ilileteolology of Cnunes (Deppnl%ment of the Alpes By WILLIAM MARCET, M.D., F.R.S., F.M.S. JIicritinies, Fmizce) .

(Plates XIX. -XXII.)

[Received June 18th.-Read June 19th, 1877.1

THE small town of Cannes, which has become such a favourite resort for invalids, is situated in lat, 43’ 33‘ N., or nearly the same latitude as Florence, 43O 41‘, and Pau, 43’ 17’. WhiIe at both these latter places ice often attains some degree of thickness in winter, and snow is of common occurrence, water freezes but on rare occasions at Cannes, and a fall of snow is quite an exceptional event at that place.

The town is built on a charming bay of the Mediterranean coast, and a succession of hills intervenes between it and the chain of the Maritime Alps, the nearest place of which is situated at a distance of about 46 miles north of the town. These mountains rise to a height of 10,433 feet (Xount Clapier ‘ 6 Ball ”), and are covered with snow throughout the winter. The melted snow and rainfall find their way to the sea mainly by the rivers Var and Paglione, both at the east of Cannes; the Paglione crossing the town of Nice 20 miles from Cannes.

A glance at the map will show that the country east and north-east of the “Corniche” is flatter than that which extends on its western side, while further east the plains of Lombardy are reached. On the western side are the mountains of Auvergne, and further south the Pyrenees. Quite in the north, beyond Sardinia, the Swiss mountains are met. I t is, therefore, apparent at the outset that the wet winds will be the south from the Medi- terranean, the easterly and north-easterly; while the north-west and westerly winds, having deposited their mcjisture in the western part of France, and reaching a warm climate, will be comparatively dry.

The high winter temperature of the north coast of the Mediterranean is due to a combination of circumstances; these are: the dry and clear state of the air, the high temperature of the sea, the rare occurrence of northerly minds, and in a great measure, I believe, the heat given out on the condensa- tion of atmospheric hnmidity. There is an immense evaporation going on from the surface of the Mediterranean on clear days. This moisture requires a very large quantity of heat to preserve its form of invisible vapour, as according to Hcrschcl (JIctcorology, p. Bl), erwg p a i n of w t c r evnpuratcd