[[Categoria:Pagine che usano RigaIntestazione|Scientia - Vol. VII.djvu{{padleft:257|3|0]]few years give many instances of streamers leaving the head in directions making a considerable angle with this line. It seems clear that the forces producing these are not wholly solar; some expulsive force, for whose nature I can only suggest electrical repulsion, shoots particles violently from the head in various directions; perhaps there is a slight favouring of the direction towards the sun, as we seem thus to get an explanation of the paraboloid envelopes so often seen on the sunward side of the nucleus, resembling the jet of a fountain. Morehouse’s comet of 1908 showed these very distinctly, and they were discussed by Mr Eddington at the Royal Institution on 1909 March 26. He deduced a velocity of projection from the nucleus of 70,000 miles per hour, and a solar repulsive force 800 times gravity, a startling result, being far in excess of the values previously found by Bredichin or Jaegermann. Another point that shows that the tail-producing forces reside partly in the nucleus is the cycle of changes in the aspect of the tail that several comets, and notably Morehouse’s, have shown. In the latter comet the Greenwich photographs revealed a fairly regular cycle that was many times repeated, so that it was even found possible, at the beginning of a cycle to predict the subsequent behaviour. Since we cannot plausibly assign these short-period variations to any change in the solar action, we must suppose the source is in the nucleus.

Allied to the last are the numerous instances of apparent rotation of the tail, it appearing alternately broad and narrow, like a sword seen broadside and edgewise. Halley’s comet showed features of this kind in 1835, which were attributed by Bessel to rotation of the tail, in a period of about 5 days. There is a difficulty in supposing the tail to rotate, since this implies either a rigid body or a powerful central force to control the rotating particles, neither of which is present in the tail. A rotating head would however produce a semblance of rotation in the tail emitted by it. The conception of a rotating head involves the assumption that its separate meteoric masses are concentrated pretty densely, since they would not otherwise have sufficient mutual gravitation to control the rotation. The other cyclical variations of the tail, referred to above, also seem to involve concentration in the head, for if it was scattered over a wide