« AnteriorContinuar »
ay, the obsing to the hills who with the azimuthor
of observation at the hours corresponding with the azimuth of the hill. Thus, owing to the hills which break the horizon at Rothesay, the observed temperatures fall short of the calculated temperatures to the extent of 0°.38 at 10 A.M., and 0°.43 at 5 P.M. ; whereas at Leith and Inverness, where the horizon is open, the observed and calculated temperatures do not differ so much as 0°.25 at any hour.
147. It is a singular coincidence that the means of observations of hours of the same name—8 A.M. and 8 P.M., 9 A.M. and 9 P.M., 10 A.M. and 10 P.M., &c. — do not differ very much from the mean of the day. The hours which come nearest the mean are the following: 9 A.M. and 9 P.M., 10 A.M. and 10 P.M., 3 A.M. and 3 P.m., and 4 A.M. and 4 P.M. The mean of four hours at equal intervals gives a result still nearer the true mean.
148. Best Hours for Observation.—Hence the best hours for thermometric and barometric observations are the following, the convenience of observers being considered : For two observations, 9 A.M. and 9 P.M. ; for four observations, 3 A.M., 9 A.M., 3 P.M., and 9 P.M., or an hour later in all these cases is nearly as suitable. For certain practical purposes, when only two or three observations can be made daily, and where there are no maximum and minimum thermometers, it is most desirable to include an observation at 3 P.M., when the temperature is near the maximum of the day.
149. Daily Extremes.-A comparison of the mean temperature with the mean of the daily extremes—that is, the mean warmest and mean coldest hour of the day—gives some interesting results. From about thirty places on the globe where those have been ascertained, if an average of six months be taken, the difference between the two does not exceed the third of a degree; the difference for any month seldom exceeds a degree, and the mean annual difference seldom more than half a degree. At Rio Janeiro the difference for any month does not amount to 0°.3, whereas at Catherinenburg, in the Ural Mountains, it exceeds this amount in every month but one. In some places the differences are all
in excess, in others they are all in defect; in some places an excess occurs in winter and a defect in summer, and in other places vice versa. In most places the greatest difference is in October and November, but in a few places the reverse holds good. Comparing the Leith and Greenwich observations, we find that the mean annual deviation at Leith is 0°.2, at Greenwich, 0°.7; the lowest monthly deviation at Leith, 0o.1, and at Greenwich, 09.2; and the highest at Leith, 0°.6, and at Greenwich, 19.1.
150. Mean Temperature deduced from Maximum and Minimum Temperatures.—Of late years, since the invention of selfregistering thermometers, the mean temperature has been more commonly deduced from observations of the highest and lowest daily temperatures. Owing to the general use of these thermometers and the general adoption of 9 A.M. and 9 P.M. as the hours of observations, an interesting question may be raised—viz., Whether should the mean temperature of a place be deduced from observations with the common thermometer (usually the dry-bulb thermometer), or from the daily maximum and minimum temperatures ? I have examined the maximum and minimum temperatures for a considerable number of places in Scotland, and, adopting the simple arithmetic mean of these temperatures as the mean temperature, I have found at places on the same side of the island, and in the same district, that this mean does not differ at any place from the mean of all the stations more than half a degree, generally only 0°.1 or 0°.2. At one or two places where the difference amounted to, or exceeded, half a degree, it turned out to have arisen from an error in one of the thermometers, which was discovered by this mode of examination. Hence, then, the indications of maximum and minimum thermometers give great uniformity of results, and so far are admirably adapted for investigations of mean temperature. I have compared the mean of the 9 A.M. and 9 P.M. temperatures with the mean of the maximum and minimum temperature for four years, but find considerable contrariety of results. Comparing the means of 55 places, which may be regarded as representing the whole of Scotland, the mean annual difference is only 0°.1. In 15 out of the 48 months the difference did not exceed this small amount. In October 1861 the registering thermometers were 1°.0 higher, and in November 1862 they were 0°.9 lower than the other mean. These were the extreme differences, but the average difference was small, the one being sometimes above and sometimes below the other. On examining particular places great differences are apparent among those stations, the local situations of which are peculiar—such as being in close proximity to the sea, or being surrounded by hills. Thus, in open situations, the 9 A.M. and 9 P.M. means are in most cases about 0°.7 lower than those of the registering thermometers ; at Braemar, in Aberdeenshire, the difference is 1°.0, whereas at Thirlestane Castle, in Berwickshire, it is only 0°.2, both places being in deep valleys surrounded by hills. On the other hand, at Aberdeen, previous to 1864, the 9 A.M. and 9 P.M. means were 1°.1 higher than those of the registering thermometors; but since that date, the situation of the observatory being changed, the two means have been nearly the same. There is little uniformity in the differences, even in the same month, from year to year. Thus, at Braemar, a station established by the late Prince Consort, and in every way a well-equipped and well-appointed station, the difference in July 1861 was 30.0, whereas in July 1863 it was only 0°.3. At some places the greatest difference is in summer, whilst at other places it occurs in winter. It follows from this that the data supplied by observations of temperature taken at the above hours—and the remark may be extended to the observations of any two or three of the twenty-four hours—are not so trustworthy guides to a knowledge of the mean temperature, as are observations made by maximum and minimum thermometers.
151. The chief source of these different results of the 9 A.M. and 9 P.M. observations, is the modification of the law of the daily march of temperature, first stated by Sir David Brewster, and already referred to in par. 146. Thus it is evident that, if the shadow of a hill, or even of a tree, house, or wall, does not leave the place where the thermometers are situated till near to or after the hour of observation, the temperature will from this cause be too low. Similarly, objects obstructing the skyview at the observatory will retard the effects of radiation, and the thermometer will read higher at 9 P.m. than it would otherwise do. Again, if the instruments are placed in a low situation, with rising ground surrounding it wholly or in part, then during calm nights cold air will flow down upon the thermometers, thus hastening the period of the lowest temperature at the place. These considerations do not apply, or but slightly, to the observations of maximum and minimum thermometers. It need scarcely be added here, that if the dry bulb be not read immediately on approaching the instruments, but is allowed some time to be affected by the temperature of surrounding objects, and if care be not taken to have the bulb of the dry-bulb thermometer on the same level with the bulb of the minimum thermometer, the observations will be vitiated by errors which it will be impossible to rectify.
152. How far does the mean of the daily maximum and minimum temperatures represent the true mean temperature ? This is a question of first importance in meteorology. The closest approximation we need make to the mean temperature is, to observe the temperature each hour, and the mean of these 24 observations may be accepted as the mean temperature. Hence the above question will be answered, if daily observations of the maximum and minimum temperatures be conjoined with the above. I have examined seven places where such observations have been made for a number of years-viz., St Petersburg, Tiflis, Catherinenburg, Barnaul, Nertchinsk, Pekin, and Calcutta ; and have added to the tables the means for Brussels, as given in M. Ad. Quetelet's excellent · Meteorology of Belgium, compared with that of the Globe, 1867, and the means for Greenwich, prepared by Mr Glaisher, and published in the Transactions of the Royal Society' for 1848.
COMPARISON OF THE MEANS OF MAXIMUM AND MINIMUM DAILY
TEMPERATURES, with OBSERVATIONS OF A COMMON THERMOMETER, OBSERVED TWENTY-FOUR TIMES A-DAY. The former exceed the latter, except in one case, which is marked with a minus sign.
153. All these places, with the exception of Greenwich, show a remarkable uniformity—the mean temperature deduced from the mean of maximum and minimum observations being about half a degree above the true mean. So regular, on the whole, is this difference, that it may be accepted as the rule, unless where it is modified by peculiar changes of climate in different seasons and localities.
154. The following is suggested as the cause of the higher temperature which self-registering thermometers give: It will be seen further on that the descent of the temperature during night is arrested near the dew-point—the heat which is given off in the condensation of the vapour into dew being sufficient to maintain the temperature near that point. Hence in the curve of the daily march of the temperature, the part representing the lowest temperatures presents a flattened appearance when much dew falls, as compared with the part of the curve showing the highest temperatures. Thus, since the minimum temperature does not fall so low as