formerly Assistant in the Heidelberg Observatory and now of Princeton University, throughout the entire period of work, from unpacking to repacking. Professors SVANTE ARRHENIUS and GUSTAF KOBB, of Stockholm University, were with us. for ten days following August 20th. Mr. C. M. OLMSTED, an American student in Bonn University, was with us from August 25th to September 1st, and Mr. FREDERICK PALMER, Jr., of Haverford College, from August 27th to September 3d. To these six gentlemen the expedition is under special obligations. Even with their skillful assistance the work of preparation was exceedingly strenuous. We were able to obtain excellent unskilled laborers, but we found it expedient to be our own carpenters and machinists.

The available weather records for past years gave promise of very light rainfalls in the months of July and August. This promise was realized in August, but not in July. From July 21st to August 2d there were six heavy rainstorms, and the fall probably amounted to seven or eight inches. These storms were reported to be of wide extent in northeastern Spain, causing serious floods in many quarters. During the period August 2d to September 1st the weather conditions were excellent for erecting and testing the apparatus. There were sprinkles or slight showers on August 4th, 10th, 23d, and 25th, which caused no damage and little inconvenience. Nearly all the nights and forenoons were clear. The majority of the afternoons were clear, but there was a marked tendency for thin clouds to form in the early afternoon. On many days. the sky was beautifully blue and markedly free from glare around the Sun. The Milky Way frequently shone as we have seldom seen it except on Mt. Hamilton.

The other volunteer assistants who were to take part in the observations arrived at Alhama on August 27th and in the morning of August 28th. They were: Professor Dr. J. HARTMANN, Astrophysical Observatory, Potsdam; Sr. D. ERNESTO GREVE, National Observatory, Santiago, Chile; Dr. VIGGO STRÖYBERG, formerly in the observatory, Copenhagen; Sr. D. ARTURO CUYÁS, Madrid; Professor JOSÉ CASARES, Central University, Madrid; Professor HILARIÓN GIMENO, University of Zaragoza; Professor ANTONIO ROCASOLANO, University of Zaragoza; Professor FELIPE LAVILLA, University

of Valencia; Lieutenant MANUEL HERNANDEZ, Geodetic Survey, Madrid; Sr. D. ESTEBAN TERRADAS, University of Madrid; Sr. D. ENRIQUE IBAÑEZ, Secretario de Municipalidad, Alhama; Sr. D. FELIPE HERREROS, Telegraphico, Alhama; Sergeant ESTEBAN BARBAJOSSA, Guardia Civil de Espagna; Sr. JUAN BLASCO, Guardia Civil de Espagna.

The remaining days were devoted to training the twentyfour observers to their duties at the instruments, to testing the final adjustments, to preparing the photographic plates and their holders, and to arranging the multitude of details which affect success.

Threatening weather conditions arose on the evening of the 28th. From early morning of the 29th until late in the afternoon the wind blew a veritable gale and the sky was thickly clouded. Although the ground at the station was thoroughly sprinkled, the wind brought clouds of dust from a distance. The rehearsals of programme were made very difficult, and the situation was discouraging. However, the wind ceased and the sky cleared in the early evening.

The night of the 29th and the forenoon of the 30th were as perfect as any during our stay in Alhama. About noon of the 30th clouds began to form here and there; by 12:30 they were numerous, especially in the northwest; and before 1:00 a sheet of light mackerel clouds was streaming southeasterly across the Sun. During totality,-from III to 1:15, Greenwich mean time,-the principal prominences and the general features of the corona were visible through the thin clouds. The clouds probably permitted from twenty to thirty per cent of the photographic rays to pass. The lower atmosphere was perfectly calm; not a breath of air was stirring.

66

66

All preparations were completed, and the observers were in their places, several minutes before totality began. Signals "5 minutes," "I minute," and 20 seconds" were given by Mr. OLMSTED at the timepiece. Certain of the spectrographic exposures were to begin at 12 seconds" before the computed time of total eclipse. Totality began about 17 seconds earlier than the computed time, on which account the few affected spectroscopic exposures were necessarily omitted; but all the exposures after the beginning of totality were secured as planned.

Much has been tried for at this eclipse, and the presence of clouds was a severe discouragement. Nevertheless, considerable confidence was felt that sufficient light had penetrated the clouds to give useful results with most and perhaps all of the eighteen instruments. Development of the plates on succeeding days showed that the coronal photographs were most excellent, the "seeing" had been good; that nearly all the spectrograms were of splendid quality; but that the intramercurial plates and those spectrograms which demanded all the light from an unclouded sky were only partially successful.

Mathematical astronomers have made comparatively little use of eclipse contact-time observations for improving our knowledge of the Moon's motion. The expense and time involved in transporting, erecting, and using meridian instruments for the purposes of the three expeditions seemed unjus tified, and it was accordingly decided to depend upon sextant observations of the Sun, not only for time, but for latitude determinations also. In Labrador the eclipse-path lay due east and west, and a knowledge of the longitude was an unimportant consideration so far as selecting a station was concerned. It was known that His Excellency, Governor MCGREGOR of Newfoundland, a geodesist of considerable experience, was to make a latitude and longitude expedition to the eclipse stations in the month of August. From the large-scale maps of the Aswan district, kindly sent to the Lick Observatory by Captain H. G. LYONS, R. E., Director-General of the Egyptian Survey, it was also evident that the latitude and longitude could with confidence be taken directly from the charts. The transportation of chronometers on long railway journeys is attended with considerable risk, and it was thought that good watches would be satisfactory substitutes. Through the kind offices of Mr. F. H. McCONNELL, of San Francisco, eleven watches were sent to Mt. Hamilton late in 1904 for trial. Three Elgin watches were selected, whose performance compared very favorably with that of standard chronometers when they were handled with the same care that chronometers demand. One of these watches was sent with each expedition.

It is probable that observations of the Sun, secured with

1 For an account of his work at the Lick Observatory-Crocker station in Labrador, see these Publications, No. 105, p. 178.

sextant and watch, before and after the eclipse, furnish time for the eclipse as accurately as observations with transit and chronometer made on the stars during the preceding and following night, especially if there are considerable changes in temperature from night to day and day to night.

When Mr. CAMPBELL was in Madrid he arranged with Director IÑIGUEZ to send electric time-signals from the standard clock of the Madrid Observatory to the eclipse station. Through the kindness of the telegraphic service these signals were received satisfactorily on several dates between August 3d and August 16th, inclusive. The longitude as obtained from these signals and the local time observations, the latitude obtained from circum-meridian altitudes of the Sun, and the altitude above sea based upon the figures published for the railway at Alhama, are

Longitude 7m 368 1° 54' 00" W. of Greenwich.

Latitude

=

=

=

41° 17′ 40′′.

+41°

=

Altitude 2,200 feet 670 meters.

The station was almost exactly on the central line of the shadow.

The computed times of beginning and ending of totality, based upon American Ephemeris data, determined by Messrs. KOBB, OLMSTED, and CAMPBELL, were

II, 1h 11m 238, Greenwich M. T.

The signal that totality was complete was given by Mr. PERRINE, based upon naked-eye observation, at 1h 11m 078, Greenwich mean time. Dr. DUGAN, who was observing the diminishing crescent on the plate-holder of the 40-foot camera, estimated that totality occurred between one and two seconds earlier. We think it is entirely possible that the very bright prominence near the point of second contact may have influenced the naked-eye observation, and 1h 11m 06s is perhaps the time to be adopted. The Sun reappeared at 1h 14m 45s. The duration was therefore 3m 398, or six seconds less than that given by the American Ephemeris. Totality began 17 seconds earlier and ended 23 seconds earlier than the predicted times. The middle of the eclipse thus occurred 20 seconds earlier than expected. There seems to be no escaping the fact that the

Moon was far ahead of its predicted right ascension. Timesignals received on six days directly from the Madrid Observatory made a serious discrepancy in the adopted longitude impossible, and it agrees exactly with that read off from the excellent large scale map of the Province of Zaragoza prepared by the Geographical Engineers, and presented to us by the Government of Zaragoza. Very accordant sextant observations for time had been secured at 8:10 A. M., 10:40 A. M., and 3:10 P. M. of the eclipse day, and this element could not be in error by as much as one second.1

This was a "dark" eclipse, notwithstanding the light diffused by the clouds. It was very much darker than those of India, 1898, and Georgia, 1900, and darker than that of 1901 in Sumatra. The amount of cloudiness at the two stations in 1901 and 1905 was not very different. At the former eclipse Mr. PERRINE Could read the figures in a table of logarithms easily, while at the latter he had to look closely to distinguish them.

The eighteen instruments of observation were mounted in six groups, each group depending upon one driving-clock. The six groups were so situated that all the observers could hear the time-counter, who occupied a central position with reference to them.

TIME SIGNALS.

In charge of Mr. CHARLES M. Olmsted.

Each of the three expeditions was provided with a pendu lum, consisting of a rod and a heavy flat lead disk, which could be adjusted to one-second period. The Spanish pendulum, supported by a stout bracket on a heavy timber post, was set in motion shortly before totality. Beginning with the swing following the signal for totality, Mr. OLMSTED counted "One, two, three, four, . . ." until, after the reappearance of the Sun, the pendulum counts had been compared twice with the watch and recorded. This method was entirely satisfactory.

FORTY-FOOT CAMERA.

In charge of Dr. DUGAN, assisted by Professor LAVILLA.

The outside general features of this instrument, designed by Professor SCHAEBERLE for observing the eclipse of 1893,

1 After this was written, a number of preliminary eclipse accounts of other expeditions have come to hand, and many of them noted that totality occurred some twenty seconds earlier than predicted.