Table III. The mean of observations of three days, when the wind was in the SE quarter.

Table IV. The mean of observations of three days, when the wind was in the NE quarter.

Table V. The mean of observations of three days, when the wind was SW by W, or NW.

Table VI. The experiments with the Fort gun, arranged as those in Table I., with the Mount gun.

Tables VII, VIII, IX, and X. The experiments with the Fort gun arranged according to the state of the wind, as in the former Tables of experiments with the Mount gun.

Table XI. Shows the mean motion of sound for each month at the Madras Observatory, as found by the experiments, at the mean height of the thermometer, barometer, and hygrometer, given in the table.

Upon a cursory inspection of Tables I. and IV. it will be seen that the motion of sound varies under different states of the atmosphere and weather: that, according to the first table, sound at one time has been as long as 27-6 seconds in travelling from the Mount to the Observatory station; and at another time only 24.8 seconds; the distance being 29547 feet. In the first case, therefore, the velocity of sound was only about 1078 feet in a second; while, in the other, its velocity was nearly 1191 feet. The extremes in Table VI. show a still greater difference. This proves the necessity for making experiments during a long interval, in order to obtain an accurate general result.

In Tables II. and VII. we find, as the thermometer rose, the atmospherse at the same time decreasing in density and increasing in its elasticity, that the sound moved with greater rapidity.

That with the wind in the SE quarter the velocity was considerably increased, both from the Mount and Fort; more, however, in proportion, as might be expected, from the former than

the latter.

That with the wind at NE. the sound from the Fort gun travelled with a greater, and from the Mount gun with a less velocity than when the wind was in any other direction; that wind being favourable for increasing the velocity from the Fort, and unfavourable from the Mount: the full effect of the wind, however, is not to be ascertained by this table alone, as the thermometer during the time the NE wind prevails is comparatively low, and the barometer high; both which, as will have been seen by inspection of the tables, occasion the sound to travel slower than ordinary.

The wind SW. W. and NW. the velocity from the Mount was accelerated, and that from the Fort retarded; but not in the degree that would have taken place had the thermometer, baro

meter, and hygrometer, remained the same as in the NE monsoon; but having been different, the velocity was accelerated from both guns on this account, in like manner as it was retarded in the NE monsoon.

The following are the results deduced from the experiments in the different tables. I shall first give the general results from Table I. and VI.

Or almost precisely the same as the velocity by the theory.

Here we find a difference from the former general result by the observations with the Mount_gun; the reason of which appears to be, that I could not, as I have before stated, carry on the observations during at least a complete revolution of the changes in the atmosphere; and that this is the reason I shall now endeavour to show. The interval wanting is between the 28th of March and the 16th of July. Had this interval been wanting in the experiments with the Mount gun, there would have been a difference of 0.237 seconds in the mean result; for the mean of the experiments in this interval is 25.632", and the mean of the whole 25-869′′, making the difference just mentioned.

Now 25-869" + 0.237" = 26-106", which would have been the mean number of seconds had the observations with the Mount gun been continued during the same interval only as the experiments with the Fort gun. Then 26-106": 0-237" :: 12-306" (the mean of the Fort observations) : 0.112". Now 12.3060-112" 12-194", which would have been the general mean of the experiments with the Fort gun, had the same been continued as long as the experiments with the Mount gun. Then the distance 13932-3 feet, divided by 12-194, will give 1142.5 for the motion of sound by the experiments with the Fort gun thus brought on; and this also agrees, within a fraction of a foot, with the velocity according to Sir Isaac Newton; and with the results by the two other celebrated philosophers before named (Halley and Flamsteed).

Feet.

We then have by the Mount gun 1142-18 for the velocity. And by the Fort 1142.5.

gun

The mean is 1142-34, or very nearly the velocity above alluded to. Nothing could be more satisfactory than this general result; and it may be presumed, that the other results in different states of the atmosphere are equally to be depended

upon.

The velocity also by the Fort gun, which, it will be recollected, is little more than half the distance of the Mount gun from the station, shows that sound travels equally during its progress.

In the NE monsoon, the sound was very indistinct at times: this however does not appear to have sensibly affected its motion. The French academicians indeed proved, that this made no difference in the velocity.

I shall now proceed to the conclusions from the other Tables; and first, those of the experiments with the Mount gun.

The results in these Tables, like the separate observations, show the necessity of making a series of experiments long continued, in order to obtain the correct general rate at which sound travels; and this may afford a clue, as I observed in the first part of this paper, for discovering the cause of the differences in the results by the authorities there named: it is difficult, undoubtedly, to ascertain the distance of two stations, one far from the other, to the nearest foot; but errors of many feet in this respect, would make but a small difference in the velocity in a second found by experiment, when the gun and station were even at a moderate distance ;t we must, therefore, be led to conclude, that these differences have chiefly arisen from the

The results by the Mount gun may however be taken as the standard.

+ For example, a difference of about 26 feet in the distance, between the Observatory station and the Mount gun, would make only about a foot difference in the velocity in a second.

experiments having been made during a limited period only, and at unfavourable times for obtaining a mean result, instead of the interval which appears by these experiments to be necessary.

A particular examination of the Tables and results will show the difficulty of ascertaining what proportion of the differences should be allowed to each of the instruments used for finding the state of atmosphere, exclusive of the effects of the wind.

During the calms, we might expect that the proportional parts to be allowed for the difference in the thermometer, barometer, and hygrometer, might be found with some degree of accuracy; the discrepancies, however, are very considerable. Comparing the results of Tables II. and VII. we find the barometer 0.121 lower, the thermometer 4.6° higher, and the air about 8 more dry by the former Table than by the latter, while the velocity in a second is only 177 feet greater by one Table than the other.

We give, however, in addition the following results taken from the Tables of calms, and arranged according to the different heights of the thermometer and barometer. These results may assist us in coming to some conclusion upon this part of the subject.

Experiments with the Mount gun.*

Barometer. Thermometer. Hygrometer Seconds.

Distance.

Velocity in a second.

Where the changes are so numerous and so frequent as in the atmosphere of the earth, we cannot expect that our imperfect instruments will be of a construction sufficiently delicate to show accurately every alteration that may affect the motion of the pulses of the air; but by various comparisons and combinations of the results, we may hope to arrive at general conclusions, somewhat approaching the truth.

These are deduced from 100 observations.

Now, by numerous combinations of the observations just given, when the air was calm, we are led to conclude: first, that for each degree of the thermometer 1-2 feet may be allowed in the velocity of sound for a second; for each degree of the hygrometer 1:4; and for one tenth of an inch of the barometer* 9-2 feet. Then taking these numbers as the basis of the comparison, we find the mean difference of the velocity between a calm, and in a moderate breeze of wind, to be nearly 10 feet in a second. And by comparing other results together, a difference of about 214 feet in a second, or 1275 in a minute is found between, the wind being in the direction of the motion of sound, or opposed to it.

Before I conclude these introductory observations, and explanations of the experiments, it may be proper to refer more particularly to Table XI. containing the mean motion of sound for each month of the year, by the experiments with the Mount gun, according to the state of the atmosphere indicated by the different instruments; and to the prevailing monsoons, which may be considered to be the same, during the same months, every year; full information respecting which is given in the former Tables. On examining this Table, it is rather curious to observe how regularly the mean velocity proceeds to a maximum about the middle of the year, and afterwards retraces its steps; giving us a velocity in one case 1164 feet in a second, and in the other of only 1099 feet. This regularity would, no doubt, be still greater with the mean of the observations of several years.

TABLE XI.

Mean Motion of Sound for each Month, according to the Experiments with the Mount Gun.

* The rise and fall of the barometer is very limited in this country, as will be seen by an examination of the Tables. A sudden fall of 0.3 inch indicates a gale of wind.