temperature through the North Atlantic states was above normal. On the 23rd, the cyclone had deepened to 982 mb. and moved 1000 kilometres to Newfoundland, while the anti-cyclone had greatly increased in extent and intensity. (1040 mb.) and moved to the Great Lakes. The temperature had fallen from 10° to 18° C. in the North Atlantic states, and in Connecticut ranged between -18° and -10°; a strong northwesterly gale prevailed throughout the day at Middletown and nearby places. On the 24th, at sunrise, the anti-cyclone was central south of New York, the temperature ranged between -19° and -22°, the wind continued northwesterly with greatly diminished velocity, and except between 6 and 8 A. M., when a large patch of A-Cu passed, the sky was clear. Altogether, at the time of first contact (8.09 A. M.) the conditions at Middletown and in Connecticut generally were very favorable and remained so. Summary of Meteorological Records. Figure 1, which is self-explanatory, contains the meteorological data described, plotted for the period after 9 A. M. with reference to the position of the shadow and the time of occurrence and for that before 9 A. M. according to a uniform time-scale without reference to the shadow. The relative positions of New York, Middletown and Blue Hill, and of the nearest cyclone and anti-cyclone are indicated on the same scale. The normal pressure and temperature on days with conditions similar to those on the day of the eclipse are indicated by (A), the normals for all conditions by (B) and the actual pressures and temperatures by the solid lines (C). These normals are from records of a wide-scale mercurial barograph and a thermograph at Blue Hill Observatory, (170 km. northeast of Middletown); any small uncertainties due to difference of position are probably more than offset by the small errors of the instruments and their exposures. Apparently, the morning maximum pressure for conditions similar to those prevailing during the eclipse of 1925 occurs about 40 minutes earlier than the normal maximum for all conditions; the temperature on days with similar conditions lags slightly at first, but, as might be expected, rises more rapidly than does the normal for all conditions. Pressure. The effect of the eclipse of 1925 may be submerged in irregular variations of pressure such as prevailed during the preceding 20 hours; but there is a slight general resemblance of the record at Middletown to that obtained at Washington, Georgia, during the eclipse of 1900 (which occurred near the same time of day) the maxima and minima showing the same tendency to lag behind the shadow. Temperature. The changes of temperature, also, at Middletown and Westerly, resemble those recorded at Washington, Georgia, in 1900, but are smaller, the maximum fall being 3.1° at Washington, and 1.9° at Middletown. The return to normal at Middletown was much slower, the temperature remaining near the minimum for about 20 minutes, then rising slowly, although, as stated, the normal rise of temperature on cold days is more rapid than the normal rise during all conditions. The irregularity of the changes of temperature may be due in part to local influences unavoidable during calm weather when the temperature is low and the small differences between the records at Middletown and Westerly may be accounted for by the nearness of Westerly to the ocean. Both records were made with great care under most favorable circumstances and show very clearly the time and amplitude of the eclipse-effect. Wind. The decrease of velocity during the passing of the shadow was observed at Middletown and Westerly and recorded at New York. At Blue Hill and other places where the average velocity is high, the eclipse-wind, which, in this instance, could hardly have exceeded 1 metre a second, could not be separated from the ever-present irregular variations of velocity, although, at this time these were smaller than usual. The effect on the direction of the wind was very conspicuous, particularly at Middletown and Westerly, where the velocity throughout the morning was low. The changes at these stations agree very well although the actual directions are different. At New York and Blue Hill where the velocity was higher there was a tendency to blow toward the region of lowest temperature. The data from the four stations, limited as they are to the region near the centre and toward the rear of the shadow, are insufficient to show a definite cyclonic circulalation. Clouds. The observations of A-Cu. show an inblowing of the air toward the region of lowest temperature at a height between 2000 and 3000 metres as well as at the earth's surface, although the velocity aloft was much the higher; the changes in velocity, although irregular, were very definite. Ordinarily, the heights of clouds increase toward noon, but there is little or no change in velocity and no regular change could have been expected during the period occupied by the eclipse of 1925, consequently, the variations in velocity and direction following totality are chiefly due to cooling by the shadow. These results confirm the first observations of the effect of the shadow on the upper air obtained at all heights during the eclipse of 1918. Apparently the temporary disturbance caused by the shadow is at least as effective in the upper atmosphere as it is near sea-level and more easily measured when clouds are present. Observers in Connecticut and Rhode Island are in good agreement regarding the changes in amount and appearance of clouds. At Middletown, Dr. Milham reported the sky entirely clear at 5.45 A. M. on the 24th, but clouds appeared soon afterward, and by dawn covered seventenths of the sky. My first observation, at 7 A. M., indicated about nine-tenths of dark, broken clouds moving rapidly from the W.N.W. Realizing, that through lack of illumination near sunrise and sunset, clouds invariably appear to be lower than they really are, and that, at this time of year in the forenoon and under the conditions prevailing, these clouds were probably a low A-Cu., they were assigned to Level 3, at a height of 2000 metres, following Clayton's detailed classification, in which clouds are conveniently placed in five levels: Level 1, Cirrus, Observations of Clouds at Middletown Connecticut, During the Total Solar Eclipse of 1925 Note When two kinds of clouds are mixed or in the same level the prevailing kind and its density are entered first. The column headed "Position" indicates the position of clouds with reference to the observer or that of greatest density or number, indicated by the exponent d. 84 17 78 16 66 13 64 13 90? 18? 74 15 80 17 78 16 82 17 80 17 74 15 of evaporation and of changes in illumination as the eclipse progressed and had other characteristics of a low A-Cu. The changes of outline or form impaired the accuracy of the individual measurements and doubtless account for the differences between consecutive measurements at 10.05 to 10.07 A. M. and elsewhere. The actual velocities in Figure 1 were computed from observed relative velocities, assuming that the Ci. were 6500 and the A-Cu. 2000 metres above sea-level. Relative velocity is the distance in millimetres traveled by the cloud-image in one minute when the observer's eye is 167 mm above the reflecting surface of the nephoscope, and is convertible into actual velocity in metres a second by multiplying it by the height of the cloud, in metres and dividing by 10,000—an operation performed mentally. Mr. Clayton, who was at Westerly, suggests that the apparent clearing of the sky, observed at Middletown, New London and Westerly, was probably due to the shadow. Shadow-bands. The methods of observation followed were suggested by Upton and Rotch for the eclipse of 1889 and described in the Annals of Harvard College Observatory, XXIX, 1. The "lie" and direction of motion at each appearance were marked by colored sticks laid on the snowcovered ground and afterward measured by means of the alt-azimuth. The conditions for observations were excellent although better results might have been secured by two observers, one measuring the "lie" and the other the direction of motion. On first appearance the bands were a faint mass of interlacing bright lines, rapidly and continually changing in length and having an indefinite rapid lengthwise motion as well as a slower general movement from the W.N.W. The visual effect was that of a layer of transparent liquid some 15 centimetres deep into which has been poured another liquid having a different density. Precise measurements were impossible. The bright lines apparently were about one centimetre wide and 3 to 4 centimetres apart. During the second appearance these bands were more definite, and better measurements were secured, the "lie" being about 45°—225°, the azimuth of motion from about 110°, and the velocity 1 to 2 metres a second. As nearly as could be determined, the width and distance apart were about 3 to 4 centimetres. Azimuths are from south (0°), through west (90°), north (180°), east (270°) to south. By reason of their irregularity and indefiniteness the bands observed by the writer at Middletown (and by others) were very unlike those accompanying earlier eclipses, particularly those observed at Washington, Georgia, in 1900, which were sharply outlined and retained a definite form and dimensions throughout the two periods of visibility. The most probable reason for the difference is that the atmosphere, on the 24th of January, 1925, was unusually quiescent or homogeneous, a state to be expected at so early an hour in midwinter in this region, but, in this instance, intensified by anti-cyclonic conditions, and by clouds. The attempt to photograph the bands failed. General agreement regarding the cause of shadow-bands has not yet been reached. Neither convection near the earth's surface, nor changes in refraction when sunlight passes through strata differing in temperature are considered sufficient, themselves, for shadow-bands have been observed on a sheet suspended from a balloon at a height of 3800 metres, far above convection from the ground, and stratification is a persistent if not a normal state of the atmosphere, at all heights. Information accumulated since 1900, indicates that these bands probably occur chiefly, perhaps only, during the mixing of masses of air having different densities or temperatures, as do the familiar artificial shadowbands or waves that form over chimneys, heated roofs or when air from a warm room escapes through a window. These contrasts of density or temperature in the free atmosphere are maintained to an appreciable degree only during the rapid decrease and increase of temperature before and following totality, and the motions of the bands or waves probably are more closely related to those of the eclipse wind than to the normal wind prevailing at the time. At Middletown, the velocity of the shadow-bands was 1 to 2 metres a second, and that of the A-Cu., 15, while a calm prevailed at the ground. THE METEOROLOGY OF FUTURE ECLIPSES If worth-while contributions to meteorology are accomplished by the collection of data during future eclipses, methods and equipment must be given first consideration. The need of this was clearly evident from an examination of many records obtained during the eclipse of 1925, some of which, although made by conscientious observers, were of small value because of the lack of information concerning the conditions or requirements of such work. Studies of former eclipses show clearly that the methods of the ordinary climatological station (where temperatures are read to whole degrees, pressures to 0.3 mb and the directions of wind and clouds to eight points) are hopelessly inadequate, for the probability of a calm day when eclipse-effects are large is usually too small to be depended upon. Sensitive recording instruments are preferable to instruments that must be read frequently, for not only can readings of record-sheets be made much oftener and pressures and winds determined more accurately, but the valuable time of observers can be devoted to observations of clouds and other phenomena. A ventilated meteorograph of the Assmann type, having a time-scale of at least 10 mm an hour, and whose pressure-pen moves 3 mm for a change in pressure of 1mb, exposed in an open shelter of the French type should be very satisfactory; pressures recorded on such a wide scale are more accurate than readings of the ordinary portable mercurial barometer, and the temperatures will not be affected by the nearness of the observer as might be the case if frequent readings of thermometers were depended upon. For data of the wind, the portable Draper anemoscope and anemometer, employed during eclipses since 1889, is most satisfactory; the direction should be read to degrees and velocity to half-metres a |