for one observer. His results are compared with a similar average derived by Major Tupman and the writer from observations in 1869-'71 and 1877-'80 respectively:
| DATE. | FALLING STARS IN ONE HOUR. | |||
| Schmidt. | Tupman. | Denning. | ||
| 6 | August | 6 | 36 | 13 |
| 7 | " | 11 | 37 | 23 |
| 8 | " | 15 | 45 | 26 |
| 9 | " | 29 | . . | 44 |
| (max.) 10 | " | 31 | 59 | 71 |
| 11 | " | 19 | 53 | 38 |
| 12 | " | 7 | 27 | 24 |
Schmidt's figures are very small and much below the numbers found in recent years. But the averages in the table are not thoroughly reliable, inasmuch as they are based upon only a few years' observations. A longer series might give a closer comparison, but it is seldom that the results of independent observers agree within small limits. There are differences in vision, modes of observation, and in position, which must obviously affect the numbers to no small degree; and the intermittent character of the meteor-shower itself must give rise to discrepancies which can not at first sight be accounted for. The horary number of meteors on August 10th may vary, according to Heis, from 160 (in 1839) to 24 (in 1867). During the last ten years the writer has found little variation in the intensity of the annual returns when the conditions of weather and moonlight are fully taken into account; and there is no question that some of the variations ascribed to the shower have no real existence, but are to be explained by the differences referred to above.
A fair comparison can not be instituted between the horary numbers found by observers, unless the observations, from which the values are deduced, are made, in each case, at similar hours of the night; for shooting-stars, though often plentiful after midnight, are comparatively scarce in the evening hours. This is readily explained by the fact that the principal radiant points of the showers are massed together in the eastern region of the sky where the earth's orbital motion is directed, and it is obvious that in the evening hours, when the altitude of many of them is very low, and when others have scarcely appeared above the horizon, their operation is in a great measure restricted, so that only a feeble indication of their displays is perceptible at such a time. The case is entirely different at a later period of the night, when the constellations in which the several radiant points are situated have ascended high in the sky, and are in fact so placed that they may be seen to the greatest advantage. The August Perseids are always best observable in the morning hours, for the radiant point is very low on the
