ments through clouds, just long enough to get imperfect observations for position, but nothing more. It was noticed, however, that the eccentric arcs had disappeared. On October 2d the comet was observed for more than an hour before daybreak with the great telescope. The most notable features were a single bright cap or envelope at a distance of about half a minute from the nucleus, and the nucleus itself, which, instead of being round, was considerably elongated. There were, however, no jets, or starfish-like projections such as the comet of 1881 presented so often. There was not much of structural detail to be made out in the head of the comet, but the dark stripe behind the nucleus was very conspicuous. This dark stripe, by-the-way, is a very remarkable phenomenon, not yet explained, so far as we know, though observed in most large comets. The common impression is, that it is merely a space behind the nucleus, screened as it were by the nucleus itself, from the rush of luminous matter which is being driven backward by the sun's repulsion. But if this be so, then, as Mr. Proctor has pointed out, in a recent article, there is no reason why it should appear so well defined and so dark. The cross-section of the tail, a little way behind the nucleus, was, in the present case at least, 100,000 miles in diameter: now, merely taking away the luminous matter from a tunnel 6,000 or 8,000 miles in diameter along the axis of the tail, could make but little difference with the amount of light received by the eye at a distance. If there were no tunnel, we should get from the central line of the tail the brightness corresponding to a thickness of 100,000 miles of luminous matter. Boring the tunnel would only reduce it to some 90,000 miles, and the difference would be hardly perceptible.
It seems more likely, if the writer may venture the suggestion, that the stripe is a stream or beam of non-luminous, cooler vapor or gas, which is nearly opaque to the radiation emitted by the same kind of gas when luminous, and therefore cuts off all the light from whatever portions of the comet's luminous drapery is behind it; in the same way that cool sodium-vapor, for instance, is relatively opaque to the light of a sodium-flame. If this is correct, the dark stripe ought not to be black, but just about half as bright as the neighboring nebulosity; which corresponds to the actual fact. If one could catch a star passing behind the stripe, it would perhaps be easy to settle the question. At any rate, if the star shone more brightly when in the stripe, we might be sure that the hypothesis is wrong. The star should be dimmed a little, if anything, though, of course, star-light would not be so much affected as the light from cometary matter. Mr. Proctor has suggested a different hypothesis, which seems to the writer rather less probable, but there is no time to discuss it here.
On October 4th the nucleus had become much more elongated, so as to be shaped something like an Indian club. The envelope, which was conspicuous on the 2d, had disappeared, or degenerated