AMERICAN

JOURNAL OF SCIENCE.

EDITORS

JAMES D. and E. S. DANA, and B. SILLIMAN.

Professors ASA GRAY, JOSIAH P. COOKE, i JOHN TROWBRIDGE, of Cambridge,

GEORGE F. BARKER, of Philadelphia. THIRD SERIES.

VOL. XXVIII. -[WHOLE NUMBER, CXXVIIL] Nos. 163-168.

JULY TO . DECEMBER, 1884.

NEW HAVEN, CONN.: J. D. & E. S. DANA. 1884.

CONTENTS OF VOLUME XXVIII.

NUMBER CLXIII.

Art. I. — Contributions to Meteorology; by Elias Loomis.

With Plates I and II, 1

II. —Light of Comparison Stars for Vesta ; by E. C. Picker-

III. — Mineralogical Notes from the Laboratory of the U. S.

Geol. Survey ; by F. W. Clarke and T. M. Chatard,. . 20 IV.— Occurrence of Alkalies in Beryl ; by S. L. Penfield,. 25 V.— The Niagara River and the Glacial Period ; by G. F.

Wright, 32

VI. — Discovery or" Primoi\li;il Fossils in Columbia County,

N. Y. ; by S. W. Ford, 35

VII— Some apparently nndescribed forms of Fresh-water

Infusoria ; by A. C. Stokes, 38

VIII.— The Causes of Variation ; by R. Hitchcock, 49

IX. — Crustacea of the Albatross Dredgings in 1883 ; by

S. I. Smith, 53

X.— Crystallized <;«.ld in Prismatic Forms; by W. P. Blake, 57 XI. — M*o,lr of Action of Shell- and Rock-boring Mollusks ;

by F. H. Storer, 58

XII. — Memorials of George Engelmann and of Oswald Ileer ;

by Asa Gray, 61

SCIENTIFIC INTELLIGENCE.

Physics and Chemistry.— -The diffusion of Gases and Vapors, A. Wn

Atmospheric Currents: Results of the Electrical Con-

gress of 1884, 71.— Manual of Chemistry, Physical and Inorganic, H. Watts, 72.

Genb,,v and Natural History.— Origin of Crystalline Rocks, T. S. Hunt, 72.—

.-■:■■■

. ■

i ,. • ..., ^ ' ■: ■' ■ ■ i: : ■' :■' - ■:"'■ '■:•-

I n the genus iimgenunt of the Annual

: Course of Instruction in

/.not. miy (Vt'rtcl.rau). T. • I. I'akkkk. 7C — 1 landi.ook < .f V-n, l.r.-.n I>i.-s.vU,<t:.

medal of the U. S. National Academy 7 7 r the Advancement of Science: Inter-

Session: Life History Album, F. Galton,

Obituary.— Robert Angus Smith, 79.

NUMBER CLXIV.

Art. XIII. — Contributions to Meteorology ; by E. Loomis

(concluded), . 81

XIV. — Notes on the Bocks ami Ore-Deposits in the vicinity of Notre Dame Bay, Newfoundland ; by M. E. Wads- worth, 94

XV.— Origin of Bitumens ; by S. F. Peckham, 105

XVI. — Measurement of rapidly alternating Electric Currents

with the Galvanometer; by L. M. Cheesman, 117

XVII.— Specimens of Nickel Ore from Nevada; by S. B.

Newberry, 122

XVIII.— Gorges and Waterfall- bv W. M. Davis, 123

XIX.— The Influence of Light on the Electrical Resistances

of Metals ; by A. E. Bostwick, 133

XX.— Vanadinite in Pinal County, Arizona ; by F. H.

Blake, 145

XXL— United Metatarsal Bones of Ceratosaurus ; by O. C.

Marsh, 161

SCIF.XTIFIC INTELLIGENCE. Physics and Chemistry.— General Law of Solidification of Solvents, Raoult, 146. - Production of Cvanides and Ferrocyanides from Trimethylamine, Ortlieb .md Mri.LEH, 147.- Preparation of Marsh Gas, Gladstone and Tribe: Crystal- line Form of Beryllium. Brogiif.u and Fink. 1 4>.— Vapor Density of Beryllium

mile. Xaimn: Lactosin. a new Carliohydrt

hydrocarbon from (

The boiling point of Oxygen, of Air, and of Carbonic oxide under atmosphi pressure, Wroblewski, 150. — The Scientific Papers of James Prescott Joule: Whirlwinds, Cyclones and Tornadoes, W. M. Davis, 151.

■.'!'• Survey of Minnesota. X. li WixniKLi.: Origin of Tnseau Granite. B. LOTTI : Mosses of North America, L. LESyuEREtix and T. P. James, 155.— Synopsis of the Genera of Vascular Plants in tl vieinity of .- m Krai . isco, H. H. Beiir. 156.— Flora < »i i lis lloissir.i Flora I Xoi , Pa i.-on i. J. Ball, 157.— Change of the generic name Solenotus, A. C. Stokes, 158.

■ ■ 158.— New England Met. val Society of Canada, 159.—

Royal Society of New South Wales: Census Reports on Cotton: German Devo-

Obituary— Ferdinand von Hochstetter^lCO.

NUMBER CLXV.

Art. XXII. — Amount of the Atmospheric Absorption ; by S. P. Langley,

XXIII. — Tornadoes ; by H. A. Hazen,....

XXIV. — Absorption of Radiant Heat by Carbon Dioxide ; by J. E. Kibbler -

XXV.— Triassic Insects from the Rocky Mountains ; by S.

H. SciTDDKR,

XXVI.— Flexibility of haed. unite; by O. A. Derby,

XXVII.— Age of the Glazed and Contorted Slaty Rocks in the vicinity of Schodack Landing, Rensselaer County, N. Y.; by S. W. Ford, ..

XXVIII.— Relations of the Mineral Belts of the Pacific Slope to the Great l/pheavals ; by G. F. Becker,

XXIX.— Notice of the remarkable Marine Fauna occupying the outer bank- oil' the Southern Coast of New England, No. 9; by A. E. Verrill,

XXX— Geology of the Blue Ridge near Balcony Falls,

modified view; by J. L. Ca

SCIENTIFIC INTELLIGENCE. Photographing Colored Objects in their Natural

:,'•_•■-•::. Pi- •! ■ ,.,/.. \\. \.

: of temperature due to Magnetization and Demagnetiz

- ir tea Mai

■tie ..f t

NUMBER CLXVI.

Art. XXXI. —Duration of Color Impressions upon the

Retina; by E. L. Nichols,

XXXII.— Fulgurite from Mount Thielson, Oregon ; by j."s~.

DlLLKE, ........

XXXIII— Paramorphosis of Pyroxene to Hornblende in Rocks; by G. H. Williams, __ .

XXXIV.— Southward ending of a great Synclinal in the Taconic Range ; by J. D. Dana. With a map (Plate

XXXV— Supposed Glaciation in Pennsylvania south" of" the Terminal Moraine; by H. C. Lewis. With a map (Plate IV, but not numbered), P

XXXVI. —Mass of Meteoric Iron from Wichita County Trx.is ; by J. W. Mallet,

XXXVII.— Jean-Baptist e-Andiv Dumas; bv j. P Cookf

XXXVIII.-New Meteorite; by I. R. Eastman, . . . .'

SCIENTIFIC INTELLIGENCE.

Q at the meeting o

'■»■ Q":0IJCC :it :

309.— Report on th.

Modern' High Expl-s' '

Geology and Natural History.— Professor James Hnll on the "Hudson River age of the Taconic slates. 311.— Earr

Cla.vs of' Chicago and vicinity, H.'

«gre*e: Text-book of Des

''• i,UM,1[AV: Ib-nl.-nto. A. Wiisi; \cii. 318.— Deep-sea

n : btonsan Society of New York: George Bentham, 31

• Annals of the Astronomical Ohservi

Scientific

NUMBER CLXVII.

Art. XXXIX. — Characteristics of the North American

Flora ; by Asa Gray, !

XL.— Columbite in the Black Hills of Dakota ; by Wm. P.

XLI. — Spectro-photometric study of Pigments ; by Edward

XLII. — Criticism of Becker's Theory of Faulting ; by Ross E. Browne, ;

XLIII. — Difference between Sea and Continental Climate with regard to Vegetation; by M. Buysman, . !

XLIV. — Chemical Affinity; by John W. Langley, :

XLV. — Relation between the Electromotive force of a Daniell Cell and the strength of the Zinc Sulphate Solution ; by H. S. Carhart, . - ;

XL VI. — Notice of the remarkable Marine Fauna occupying the outer banks off the Southern Coast of New England, No. 10 ; by A. E. Vkrrill, i

XLVIL— Note on the Cortlandt and Stony Point Horn- blendic and Augitic rocks; by James D. Dana, J

SCIENTIFIC INTELLIGENCE.

P v * X tions.l Coiil'oretK of Klcetri. -» T> ■ t nation of Yapor

Infra-red portion of the solar spectrum, 391.

linen Schiefergesteine. J. Id i the origin of bedding in so-

E. V. d'Intilliehs. 396. — Note on the nu di, J. D.Dana,

398. — Glacial Studies, F.-A. Forel, 400.— Sketch of the Gecu>u

of Lake Lahontan. I. C. RrssEi.i . <.■■..<■-• ■ • . .-

Eicellenz Herrn Dr. H. von Dechen : On the Composition of Horderite, J. B.

Flora Brasilie

> . 403.

..-• I) .;ili ml \Y. I ! I K'S i 'I ['li M ( riiiis ot Engineering, R. [|. Thi-rmviv; I r.

.' of the National Academy of -ci n - - Newport 11. I., October 14, 1884: French Academy of Sciences: The Young Mineralogist and Antiquarian, 106.

NUMBER CLXVIII.

Art. XL VIII. — Distribution and Origin of Drumlins ; by

W. M. Davis, .*. .. ' 407

XLIX. — Geological Relations and Genesis of the Specular

Iron-Ores of Santiago de Cuba; by J. P. Kimball, 41G

L.— A New Tantalite Locality ; bv C. A. Schaeffer, 430

LI.— Paleozoic Hocks of Central Texas; !>v C. D. Walcott, 431 LIL— Sufficiency of Terrestrial dotation for the Deflection

of Streams; by A. C. Bainks, 434

LIIL— Chemical Affinity; l.y J. \V. [,v.V(;i.eV, 437

LIV.— Peculiar Modes of Occurrence <»f (iold m Brazil- bv

O. A. Derby, [_/_ 440

LV.— Colemanite, a new Borate of Lime ; by A. \V. Jackson, 147 LVL— Decay of Quartzyte, and the formation of sand, kaolin

and crystallized quartz ; bv J. D. Dana, 448

SCIENTIFIC INTELLI

'iu/.-if.s.— Cause o motive force of a Battery and tl

diminished Pressure I, Mkvk;

maline for Heat, 45C— Ilelmholtz's dispersion theory, A. Wi

nation of tin- ndpal lines and bands in tl

- a and with the development of perlitic structure, F. Rutley, 461. 7eology and Mineralogy.— The Copper- bearing; Rocks of Lake Superior. R D rsmro, f;-'.— Note on the Paramorphic origin of the "

Geological Railway Cuide: Compn'-dti..n .>f Hen!, rite, !•' \.{\>'\\\ Botany and Zoology.— V^uAo^no ot'tiie Flora of Minnesota. Warrk.v

ische Practicum, Prof. Sth ,!s of Botany, (

Oraithorhynchus and Echidna. 47.V— Orsran isn is , Astronomy and Geodesy.— Observations made on the Expedition to Cai W. Upton, 477.— New form of Primary Base Apparar -

igence.— Extracts of a letter to J. D. Dam Gould: Sir William Thomson's Lectur,

AMERICAN JOURNAL OF SCIENCE.

[THIRD SERIES.]

Art. I. — Contributions to M'>teoroh>/,/ : bv Ki.i \s Loomis, Pro- fessor of Natural Philosophy in Yale College. Twentieth paper. With Plates I and II. [Read before the National Academy of Sciences, Nov. 13, 1883, and April 15, 1884.] Reduction of Barometric observations to sea-level. In my fifteenth paper of contributions to Meteorology, I endeavored to determine the proper reduction to sea-level of barometric observations made at elevated stations. For this purpose I made an extensive comparison of observations at five mountain stations ; three in the United States and two in Europe. The reductions deduced from these observations were compared with the reductions computed from the theories of Laplace and Plantamour and exhibited very great discrepancies. For each of the stations the discrepancies appear at the lowest pressures, and also at the highest pressures, while for a medium pressure the observed reductions agree with the computed reductions. In order to discover if possible the cause of these discrepancies I selected for examination all the cases in which during a period of three years there was a decided area of high or low pressure near either of these five stations. Table I shows the most remarkable cases of barometric minima on Mt. Washington during a period of three years. Column first gives the number of reference ; column second the date of the barome- tric minimum ; and column third gives the height of the barometer at sea-level deduced from the observations at Bur- lington, Vt. and Portland, Me. The distance of Mt. Washing- Am. Jour. Scl— Thibd Sebies, Vol. XXVIII. No. 163.— July, 1884.

2 E. Loomis — Reduction of Barometric Observations.

ton from Portland is almost exactly two-thirds of its distance from Burlington, ami T have assumed that the pressure at sea- level under Mt. Washington is given by the formula |(2B + 3P). The temperature at sea-level given in column fourth is deduced from the observations at Burlington and Portland by the same formula. Column fifth shows the pressure observed on Mt. Washington, column sixth the temperature on Mt. Washington and column seventh gives the half sum of the temperatures in columns four and six.

The numbers in this table were arranged in the order of the mean temperatures given in column seventh, and were divided into four equal groups and the average of the numbers in the several columns for each group was computed. The results are given in the first five lines of table III, under the heading barometric minima; and these results may be regarded as nor- mal values deduced from the barometric minima of three years.

I next proceeded to compare these results with theory. Column sixth of table III shows the Mt. Washington obser- vations reduced to sea-level by the use of Guyot's tables as contained in his collection of tables, series D, page 33, employ- ing the mean temperatures given in column fifth. Column seventh shows the difference between the numbers in columns

It has been a common opinion among meteorologists that anomalies like those shown in column seven result from the erroneous assumption that the mean temperature of the air col- umn between the upper and lower stations is equal to the half sum of the temperatures observed at the two stations. In order to test this explanation I computed what must be assumed as the true temperature of the air column in order that the Mt. Washington observations, reduced to sea-level by the Laplace formula, might agree with the barometric heights observed at sea-level, and the results are shown in column eighth. It will be seen that the differences between the numbers in columns five and eight range from 5° to 10°.

I next computed the reduction of the Mt. Washington observations to sea-level by Ferrel's tables, published in his Meteorological Eesearches, Part III, pages 40-41, and the results are shown in column ninth. Column tenth shows the difference between the numbers in columns one and nine; col- umn eleventh shows what must be assumed as the true temper- ature of the air-column in order that the Mt. Washington observations reduced to sea-level by Ferrel's tables may agree with the barometric heights observed at sea-level. It will be seen that the differences between the numbers in columns five and eleven range from 8° to 15°. Some remarks upon these results will be found on a subsequent page.

E. Zoomis — Red >/et >'<>// of Barometric Olservatio Table I. — Barometric minima on Jit. Was/tin</to

47-1 55 Jan 16.1 29-6

40-4 -01 16 28"2 77 Oct. L.l '65 0-022-96 - 4 - 2'0 78 11.1 "63

16-2J -92| 4J 10-1 81 Novll'.3 -86

	'•]		■15	20 :;i\>			-20 - 46
	-64	41-2	.1 ;	l'j; 30-1	89! Jan. 3.1 29-80		- 22 - 7-3
				40 47-2	901 4.1 -81
	■54	79-S			91 7.2 -18
		694		44 56-7
		54 0		27 40-5
		62-8		ll\ HI	91 Febi;u :;n-ii	1:5 6 23-04
				34 38-9
iHl			•11		99Mar'l5:3 -'72		_2g	361
	«fl
		ill ~		i-i ^			39 nl
-=" ■'■! '	•42	M-0!		- 5 10-5	107 Mr LI -64
54 2:G	'"■	16 1 '		■■

4 E. Loomis — Reduction of 11*, <>,,,, i ,■■',,*. observations.

I next selected for examination all the cases in which, during n period of three vears there w < a ilcc led area of high pressure over Mt. Washington. The results are shown in table II. which is constructed upon the same principles as table I. These numbers were then arranged in the order of the mean temperatures given in column seventh and were divided into four equal groups, and the average of the numbers in the several columns for each group was computed. The results are given in the first five "lines of table III under the heading

I then proceeded to compute the remaining columns of the e same n

We see from table III that if we undertake to explain the differences between the pressures observed at sea-level and the Mt. Washington observations when reduced to seadevel in accordance with theory, we must admit errors of ten degrees in the assumed temperature of the air-column when the compu- tations are made by Guyot's tallies, and errors of fifteen degrees when the computations are made by Fen-el's" tables ; and this is true not simply for exceptional cases but for the average of all the barometric minima during a period of three years. The question to be decided is, can errors of such magnitude be admitted? In order to decide this question, I have consulted all the available observations of the thermometer made in the ■:on. In the Annual Report of the Chief Signal Officer for 1873 is given a series of hourly observations for May 1873 on the summit of Mt Washington and also at a station situated on the side of the mom feet below the summit; and for the month of June 1873 is given a series of hourly observations on the summit of Mt. Washington and at three stations on the side of the mountain, one situated 732 feet below the summit, a second 2227 feet below the summit, and a third 3387 feet below the summit. These observations can be compared with the tri-dailv obser- vations made at Burlington and Portland, and thus for the month of May we have the temperature given at one station intermediate between sea-level and the summit of Mt. Wash- ington, ami for the month of June we have the temperature at lions. As the evidence for the month of June is much fuller than for the month of May, I have selected h for a careful comparison and the details are given in table IV, where column first shows the day of the month and the four following columns show the temperatures observed at 7.35 A. M. at the elevations indicated at the top of the col- umns. The sixth column shows the temperature deduced from the observations at Burlington and Portland by the formula

E. Zoomis — Reduction <>/ H> Table II. — Baroinftrh' ma.<'im

/•/,- Ohsrr cations. 5

Mt. Washington.

6 E. Loomis — Reduction of Barometric Observ Table III.— Mean Results : Barometric Mir

. Bernard, 6, 7 ±2 feet above Geneva.

12 -115 -057 29-71 -078 -094 27"08 091 +"0«

E. Loomis — Reduction of Barometric Observations. 7

In order to deduce from these observations the most probable value of the mean temperature of the air column I proceeded as follows :

I multiplied the half sum of the temperatures observed at the first and second stations by the difference of elevation between the two stations ; I multiplied the half sum of the temperatures observed at the second and third stations by their difference of elevation. I proceeded in like manner with the third and fourth stations, and also with the fourth and fifth. I then divided the sum of these four products by the difference of level between the first and fifth stations and regarded the quo- tient as representing the mean temperature of the air column. The difference between this result and the half sum of the tem- peratures at the first and fifth stations is given in column seventh, where the + sign indicates that the mean temperature of the air column is greater than the half sum of the tempera- tures at the upper and lower stations. I proceeded in the same manner with th ■ - rvati<»ns t k:*>.> r. m. and 11 P.M.

tn

: n

ha»

7e of te,

iperature with ele

Itn

735 X.M.

u,m.

-"

62855553

,05S

2S9S

L«

Mil.

^

,nr,.s

..*„

V.U

KM Dill.

i

£ I

p

.'f.

!

:H

I

':•■

g'

<; ! ■: -1-4

8 E. Loomis — Bedtcction of Barometric Observations.

The average of the differences shown in this table (paying- attention to the algebraic signs) for 7.35 A. M. is + 0°-54, for 4.35 p. m. it is +2°-88 and for 11 p. m. it is +0°'05. In six cases we find differences as great as five degrees. These large differences occur only at 4.35 P. m. and appear to depend mainly upon other circumstances than the height of the barom- eter. The average oi all the differences when the barometer on Mt. Washington was considerably below the mean is + 0°-84; and the average of all the differences when the barometer was considerably above the mean is + 1°22, showing a difference

-38 depending upon the height of the barometer ton, and this is only five per cent of tl wn in table III for the warmest season of the

: the difference . __ .he year. We conclude that, the error arising from assuming that the mean temperature of the air column is equal to the half sum of the temperatures at the upper and lower stations, is quite inap- preciable if the observations embrace a considerable period of tune and are made at all hours of the day. The observations mndeon Mt. Washington in Mav, 1873. and those made in May, ^'-^i;;/1 P'»blished_ in the Eeport of the Chief Signal Officer

I consider it to be proved that the differences shown in table

III between the pressures observed at sea-level and the Mt. Washington observations reduced to sea-level by the tables of - ■ Plantamour or Ferrel, are not due in'anv consider:-

of the

column The question then remains unanswered, what is the cause of these discrepancies? In mv fifteenth paper, on page 6 i nave iMven a table of forty cases in which the observed reduc- tion of the Mt. Washington observations to sea-level was O30 men greater than that given by the formula of Laplace. Inese forty cases are derived from 3285 observations viz: three daily observations for a period of 36 months, and 'there- tore constitute but little more than one per cent of the entire number of observations. They represent the cases in which the cause or causes which gave rise to the anomalies shown in taoie 111 acted with their greatest energy, and they are there- lore well adapted to indicate what tins cause was. These cases are enumerated m table V, and I have endeavored to free them trom certain errors which may have affected the results as puh- nsned in my fifteenth paper. In each case I have deduced the :ure at sea level from the observations at UurJington and Portland by the formula 4(2B + 3P) and the results are given in columns three and four. The four suc- ceeding columns are obtained in the manner explained in my fifteenth paper, and the discrepancies are shown in the ninth column under the heading O-C. Itwill be perceived that these

E. Loomis — Reduction of Barometric Observations. 9

numbers differ a little from th The average of the differences and the average of the differer 035 inch, a change which is unexpectedly small.

Table V. — Cases in which the reduction to sea-level u

n in my fifteenth paper, that paper was 0*36 inch.

6'SE. 4: SW. 38; iW.

10 E. Loomis — Reduction of Barometric Observations.

In some cases, an appreciable error may arise from assuming that the pressure at sea-level under Mt. Washington is given bv the formula |(2B+3P) and a more accurate value may be derived from the isobars drawn to represent all the observations in the vicinity of Mt. Washington. Column tenth shows the height of the barometer at sea-level under Mt. Washington obtained by this method. These numbers were generally obtained from the isobars drawn on the Signal Service maps. but in a few of the cases I have made a slight change in the position of the isobars when the curves on the Signal Service maps did not appear to have been, drawn with sufficient care. Column eleventh shows the differences between the numbers in columns three and ten. The average of the numbers in column eleventh is 0-34 inch. The error in the assumed mean temper- ature of the air column produ the preceding refinements are materially changed.

In order to have a graphic representation of the relation of these quantities to the state of the barometer, I have drawn the curves representing the barometric observations on Mt. Wash- ington, and also the curves of pressure at sea level, for the entire

tab! si m'd n" mi'd'on 'these em'vcs Mum'' 'f.i'li It.Tbv small black circles, the position of each of the fortv cases incbi- d< d in table V. The accompanying chart, plate [, shows such

spaces between the vertical lines represent intervals of eight hours, and the spaces between the horizontal lines represent tenths of an inch of barometric pressure.

From an inspection of these curves we see that each of these forty cases occurred during the prevalence of an area of low pressure about Mt. Washington. Twenty-one of them were coincident with the lowest pressure observed during the preva- lence of the low area ; nine occurred eight hours after the lowest recorded pressure ; and five others followed soon after the preceding, when the barometer on Mt. Washington was rising but had not yet reached the point o! mean pressure. Of the

Were these cases of low pressure on Mt. Washington accom- tnied bv a cyclonic movement of the winds? In order to ■cide this question I have examined the direction and veloc-

K Loomis — Reduction of Barometric Observations. 11

ity of the winds on Mt. Washington not only at the dates of the forty cases in table V, but also at several of the preceding observations, and the principal results are exhibited in a c6n- densed form in the last column of table V. When only one wind direction is given, it is to be understood as corresponding to the date in column second. When two wind directions are given, the first corresponds to a date eight hours previous to that in column second; when three wind directions are given, the first corresponds to a date sixteen hours previous to that in column second, with the exception of the cases in which the direction of the wind is included in brackets. In these cases the date of the observation immediately precedes the wind's direction. The forty cases shown in table V correspond to twenty-one differ- ent areas of low pressure, as was indicated in my fifteenth paper. Two of these storm areas began with a N.E. wind on Mt. Washington ; one began with an E. wind ; five began with a S.E. wind ; four with a S. wind ; and six with a S.W. wind. One of the remaining cases was preceded by a calm, and one of those which began with a S.W. wind was also preceded by a calm. In the two remaining cases, the wind, as reported in the tri-daily observations, blew continually from the W., N.W. or X., but its velocity at one of the early observations was consid- erably less than at the observations immediately preceding and following.

The twelve storms which began with a wind from the N.E.— E. — S.E — or S. are reiranled as unquestionably evclonic; and the six which be<:an with n S.W. wind are regarded as prohably cyclonic. In the three remaining cases, one of which began with a calm, and the other two with a comparatively feeble wind from the N. or N.W. the evidence of a cyclonic motion is unsatisfactory unless it is confirmed by other circumstances. I have endeavored to obtain additional evidence bearing upon this question, from observations of the upper clouds made at neighboring stations on the E. and S.E. sides of Mt. Wash- ington ; but during the periods in question the upper clouds were generally concealed from view by the lower clouds. The following are the only observations I have been able to find which bear directly upon this question.

12 E. Looinis — R> <]<<<■! ion of Barometric Observations..

There can be no doubt that the upper clouds here reported! had a considerable elevation, which was probably at least equal to that of Mt. Washington, and they indicate a cyclonic motion of the winds about a center but a few miles E. of Mt. Washing- ton. I think then • i \ safeh » nde that in a majority of the cases enumerated in table V, the fall of the barometer on Mt. Washington was due to a cyclonic movement of the winds which prevailed at that elevation. In the remaining cases the observations seem to leave it doubtful whether the fall of the barometer on Mt. Washington was due to a cyclonic movement of the winds which extended to the height of 6000 feet, or was due to a cyclonic movement of the winds which was confined to a lower stratum of the atmosphere. From Plate II accompanying my tenth paper we see that the fluctua- tions of the barometer on the top of Mt. Washington are often quite unlike those at the base (which is 3387 feet below the summit, and 2898 feet above sea-level), and the fluctuations are sometimes greater in amount. From the plate accompanying this paper we also see similar differences between the barome-

whieh alieet the result being supposed to be the same), in the ratio of the mean pressures at the two stations, i. e. as 29'98 to 23-63 or nearly as five to four. Hence we see that on the sum- mit of Mt. Washington there are frequently evclonic move- ments of the wind which are more violent than those at inferior elevi ons md this ex| bus it pai t m< mites shown in table V. But these anomalies are partly due to the fact that the barometric minima on Me. rally occur

later than they do at sea level. In several" of my former papers 1 have dwelt upon this subject, and the same fact is clearly indicated by the accompanying plate. The average date of minimum pressure on Mi. Washington is more than eight hours later than it is at sea-level ; and there are frequently secondary minima on Mt. Washington which do not occur at sea-level or only in an inferior degree. Hence it results that when the barometer on Mt. Washington stands at its lowest point, the barometer at sea-level has generally risen above the preceding minimum one or two tenths of an inch, and occa- sionally four tenths of an inch ; and at the time of a secondary minimum on Mt. Washington, the barometer at sea-level may have risen three or four tenths of an inch, or even five tenths of an inch more than it has risen on Mt. Washington. Thus on Jan. 15.1, 1875, when the barometer on Mt. Washington was- at its lowest point, the barometer at sea-level had already risen

K Zoomis — deduction of Barometric Observations. 13

four tenths of an inch : and at the secondar}^ minimum of Dec. LM-.3, 1S72, the harometer at sea-level had already risen above the preceding minimum a half inch more than the Mt. Wash- ington barometer had risen A similar e;tse occurred Dec. 30.2. 1874. The discrepancies shown in table V are due partly to the causes here stated, and partly to the violence of the winds on Mt. Washington, for according to theory, the velocity of the wind is the ino-t important factor which determines the depres-

Tie—.,	ess of the barometric pressure at sea-level above th< ington observations reduced to sea-level as shown by
It. Wa<h
!ble III i	n the case of barometric minima, is ascribed to the

The differences between the observed and computed reduction to seadevel in the case of barometric maxima shown in table III are explained in a somewhat similar manner. From the accompanying plate, as well as from the i late accompanying my tenth paper, it is seen that generally the date of a barometric maximum is not as distinctly marked as that of a barometric minimum, but the average date of maximum pressure on Mt. Washington is more than eimht hour- later than it is at sea- level; and when the barometer on Mt. Washington stands at its highest point, the barometer at seadevel has frequently fallen below the preceding maximum, a tenth of an inch or

If we attempt -to represent the reduction of the Mt. Washing- ton observations to "sea-level by the Laplace formula with modified coefficients, we find that the high pressures require a larger value of the pressure coefficient than the low pressures. Also that the low temperatures require a larger value of the temperature coefficient than the high temperatures. It is not possible therefore to find values for these coefficients which shall represent the observed reduction for all pressures and temperatures. I have sought to obi, in v lues which shall best represent all the observations, and in doing this I have given the observations made near the time of barometric maxima, twice the weight of those made near the time of barometric minima, for the reason that in the former case the winds are more feeble, and the atmosphere probablv approaches nearer to the condition of statical equilibrium. With this assumption [ have found that the value of the pressure coefficient which best represents the Mt. Washington observations is 60372, and that of the temperature coefficient is -g^-g-; indicating that the pres- sure coefficient employed by Laplace (60158-6) is too small.

I next undertook a »n of the observations

on Pike's Peak and Denver. Table VI shows the principal 'barometric minima on Pike's Peak during a period of three

14 E. Looinh—Uxl Uit'nni of Barometric Observation. Table Xl.— Itarom.trh minima on Pike's Peak.

1 -ill	H	in!
	H	290 240
"; -.v.		•i::-n
4 -27
i •:::
9 -J		0-0

E. Loomis — R>ihi<ii<»> <>f £<> rum< t /•>'<.' ()l*< rcathm Table VII.— Barometric maxima on Pike's Peak.

57 Jun 16.S 24-77

5 58 21.3 24-83

6 .V.Uuh :,.i 24 92

16 E. Loomis — Reduction of Barometric Observations.

years, and the observi in the same manner as

in table I. Table VII shows the principal barometric maxima on Pike's Peak during a period of three years.

The numbers in each of these tallies were divided into four equal groups, and the average of the numbers in each group was taken. The results are given in the first live columns of table III, and the numbers in the other columns were computed in the same manner as has been explained for Mt. Washington.

We see that for barometric minima, the differences between the observed reductions to seadevel and those computed by the Laplace formula are quite small, while for barometric maxima they are quite large ; but when the computations are made by Ferrel's tables the average difference between theory and observation is almost exactly the same in both cases, with the exception of the algebraic signs. These results accord with those for Mt. Washington in indicating that the pressure coeffi- cient in the is too small. The value of the pressure coefficient which best represents the Pike's Peak observations is 60357, and that of the temperature coefficient is ToVt,- 'For barometric minima, the average difference between the observed reduction and that computed by Ferrel's formula is less than half of that found for the Mt. Washington observa- tions, a result which may be ascribed, at least in part, to the feebler winds which prevail on Pike's Peak.

For barometric maxima, the differences for Pike's Peak com- puted by Ferrel's formula are greater than for Mt. Washington, and for the lowest temperatures they are more than twice as great. We have found for Mt. Washington, three causes for these discrepancies : 1, an erroneous assumption with regard to the mean temperature of the air column ; 2, a retardation in the date of barometric maximum at the upper station ; and 3, iges of pressure at one of the stations which are not felt (or in a less degree) at the other station. I know of no observation- tvhich indicate whai • t iscribed to the

first of these causes in the present case, but probably the effect is small. The effect due to the second cause is often quite large. The barometer at Denver (after a maximum) generally begins to fall from eight to sixteen hours sooner than on Pike's Peak, and it sometimes falls a quarter of an inch or more before the descent begins at Pike's Peak. Thus Dec. 13.3, 1873, after an uncommonly high pressure, the barometer at Denver fell -28 inch teter at Pike's Peak began to fall.

Again Jan. 14.1, 1874, the barometer at Denver was at a maxi- mum, from which time it fell for four days uninterruptedly with the exception of two slight reactions, one amounting to '01 inch and the other to '04 inch. The barometer at Pike's Peak did not begin to fall until after Jan. 16.1, when the

E. C. Pickering—Light of Comparison Stars for Vesta. 17

"barometer at Denver had already descended 43 inch. A sim- ilar case occurred between April 8 and April 10, 1874. This cause of the discrepancies between the observed and computed reductions, seems to be more efficient for Pike's Peak khan for Mt. Washington. The third cause above mentioned affects the observations on Pike's Peak, but in general the curve of pres- sure for Pike's Peak is much less jairued than for Denver. The barometer on Pike's Peak frequently remains above its mean height for several days — sometimes a week or ten days — with only small fluctuations, while during the same period at Denver there have been numerous maxima and minima of con- siderable magnitude. Thus it sometimes happens that a baro- metric maximum on Pike's Peak occurs nearly, if not exactly, at the time of a barometric minimum at Denver. [To be continued.]

Aet. II.-	-Light	of Compa,	• for Vesta PlCKEEING.	/ by	Edward
IN Professor Harrington's appeared in this Journal, DM. +22° 2163 and 2164	Ill, XXVI. The o'bser	'Study o 461, the >ns with	f Vesta,"which light of the the two stars rere made with

tiiewed«rl? photometer, and were aceordingl v duic rential, so that

magnitudes of rhe stars, which were taken from the Durchmust- erung. It therefore appeared desirable that the stars should be observed with the large meridian photometer of the Harvard College Observatory, with the object of providing means for the reduction of Professor Harrington's results to absolute measures. The meridian photometer has been described in the Monthly Notices of the R. Astron. Society, xlii, 305.

The following table exhibits the results respectively obtained for the two comparison stars. The first column contains the numbers of the series to which the observations belong, the second the dates, and the third the Initials of the observers. E. C Pickering and O. C. Wendell. The fourth and fifth columns contain residuals expressed in tenths of a magnitude. The mean results, from which these residuals are derived, when cor- rected for atmospheric absorption, are it-i>6 for DM.-f 22° 2163 and ;V4s for DM.+22° 2164. The fifth observation of DM.+ 22° 2163 was rejected because it appeared that an error of 30° in reading the graduated circle of the photometer had probably occurred in one of the four comparisons which constitute a complete observation with the meridian photometer. The resi- dual corresponding to the rejected observation is placed in brack-

Am. Jour. Sci.— Third Skimks. Vm,. XXVIII. No. 163.— July, 1884.

18 E. C. Pickering— Light of Comparison Stars for Vesta.

ets. If the presumed error of 30° is left without correctio this residual would become — 09 instead of — 0-2. T separate reduction of the four comparisons gives the residue — 24, 0-0, —0*5, +0-1. Correcting the first reading by 30°, i residual is reduced to — 0-3.

March 31 P. [-0-2] -02

April 14 P. +0-1 +0-2

be applied to the DM. magnitudes of the stars appear from these observations to be +-28 for DM.+ 22* 2163 and+ -18 for DM. 4 22° 2164. From these corrections may be derived the formula M— m='023m + -058, in which M denotes the photometric magnitude of Vesta corresponding to the magnitude m given by Professor Harrington.

In the following table the first column is repeated from Pro- fessor Harrington's table in the article above mentioned. The second column contains the corresponding magnitudes of Vesta computed for mean opposition, after correction by the formula just obtained. By mean opposition is understood, as usual, the situation in which a planet is in exact opposition to the Sun, while both the planet and the Earth are at their mean distances from the Sun. The third column contains the residuals from the mean, 6-64, of the corrected magnitudes thus found. The last column contains the residuals showing the accordance of Professor Harrington*s observations of the two comparison stars. Taking the differences between the two columns of his table headed 2164 and 2163, we have a series of quantities expressed in seconds of time, the mean of which is 206 ; it corresponds to the photometric difference in magnitude resulting from the observations made here with the meridian photometer. This photometric difference is 9-06-548 = 3-58. These data show that in Professor Harrington's observations one second of time may be expressed in terms of magnitude by -174. The final column of the table here given accordingly contains the products by "174 of the differences between Professor Harrington's columns 2164 and 2163, diminished by the photometric differ- ence 3-58. If reduced to the equator, the quantity "174 becomes '16, which furnishes a determination of the constant of reduction required by the particular wedixe employed. The last line of the table contains the numerical means of the quantities in the last three columns. It may be observed that ill the first and third lines of the table the large residuals in