yet received their store of light and heat, while there are others which have entirely lost it. The stars are probably only those stellar masses which having recently had an encounter have become possessed of light and heat. They have gained in light and heat what they have lost in motion, but they have gained a possession which they cannot retain, and when it is lost they become again what they originally were -dark bodies.

2nd. "We have no instances of stellar motions comparable with those demanded by the theory." A little consideration will show that this is an objection which, like the former, can hardly be admitted. No body of course moving at the rate of 400 miles per second could remain a member of our solar system; and beyond our system the only bodies visible are the nebulæ and fixed stars; and they are according to the theory visible because like the sun they have lost their motion-the lost motion being the origin of their light and heat. Their comparatively small velocities are in reality evidence in favour of the theory than otherwise; for had the stars been moving with excessive velocities this would have been adduced as proof that their light and heat could not have been derived from motion lost, as the theory assumes.

3rd. "If suns or stars have been formed by collision of bodies moving in space, proper motion can be none other than the unused and unconverted energy of the original components. And as stellar bodies are likely of all sizes and moving with all manner of velocities, it must often happen, from the unequal force of the impinging masses, that a large proportion of the original motion must remain unconverted into heat. Consequently some of the stars ought, according to the theory, to possess great velocities-which is not the case, as none of the stars have a motion of more than 30 or 40 miles per second."

I freely admit that, if it could be proved that none of the stars have a proper motion of more than 30 or 40 miles per second, it would at least be a formidable difficulty in the way of accepting the theory. For it would indeed be strange that amidst all the diversity of dimensions of heavenly bodies, it should invariably happen that the resultant movement of the combined masses should be reduced to such comparatively insignificant figures. But something more definite must yet be known in reference to the motion of the stars before this objection can be urged.

All that we are at present warranted to assume is simply that, of the comparatively few stars whose rate of motion has been properly measured, none have a greater velocity than 30 or 40 miles per second, while nothing whatever is known

with certainty as to the rate of motion of the greater number of the stars.

There seems to be a somewhat prevailing misapprehension regarding the extent of our knowledge of stellar motions. Before we can ascertain the rate of motion of a star from its angular displacement of position in a given time, we must know its absolute distance. But it is only of the few stars which show a well-marked parallax that we can estimate the distance; for it is now generally admitted that there is no relation between the apparent magnitude and the real distance of a star. All that we know in regard to the distances of the greater mass of the stars is little else than mere conjecture. Even supposing we knew the absolute distance of a star and could measure its amount of displacement in a given time, still we could not be certain of its rate of motion unless we knew that it was moving directly at right angles to the line of vision, and not at the same time receding or advancing towards us; and this we could not determine by mere observation. The rate of motion, as determined from its observed change of position, may be, say, only twenty miles a second, while its actual velocity may be ten times that amount.

By spectrum-analysis it is true we can determine the rate at which a star may be advancing or receding along the line of sight independently of any knowledge of its distance. But this again does not give us the actual rate of motion, unless we are certain that it is moving directly to or from us. If it is at the same time moving transversely to the observer, its actual motion may be more than a hundred miles per second, while the rate at which it is receding or advancing, as determined by spectrum-analysis, may not be 20 miles a second. But in many cases it would be difficult to ascertain whether the star had a transverse motion or not. A star, for example, 1000 times more remote than a Centauri (that is, twenty thousand billion miles), though moving transversely to the observer at the enormous rate of 100 miles per second, would take upwards of 30 years to change its position so much as 1", and 1800 years to change its position 1'; in fact we should have to watch the star for a generation or two before we could be certain whether it was changing its position or not. And even after we had found with certainty that the star was shifting, and this at the rate of 1' in 1800 years, we could not, without a knowledge of its distance, express the angle of displacement in miles. But from the apparent magnitude or brilliancy of the star, we could not determine whether its distance was 10 times, 100 times, or 1000 times that of a Centauri; and consequently we could form no con

jecture as to the actual velocity of the star. If we assumed its distance to be 10 times that of a Centauri, this would give a transverse velocity of one mile per second. If we assumed its distance to be 100 times that of a Centauri, this would give 10 miles a second as the velocity, and if 1000 times, the velocity of course would be 100 miles per second.

As there are but few of the stars which show a measurable parallax, and we have no other reliable method of estimating their distances*, it follows that in reference to the greater number of the stars, neither by spectrum-analysis nor by observation of their change of position can we determine their velocities. There does not, therefore, appear to be the shadow of a reason for believing that none of the stars has a motion of over 30 or 40 miles per second: for any thing that at present is known to the contrary, many of them may possess a proper motion enormously greater than that.

There is, however, an important point which seems to be overlooked in this objection, viz. that, unless the greater part of the motion of translation be transformed into heat, the chances are that no sun star will be formed. It is necessary to the formation of a sun which is to endure for millions of years, and to form the centre of a planetary system like our own, that the masses coming into collision should be converted into an incandescent nebulous mass. But the greater the amount of motion left unconverted into heat, the less is the chance of this condition being attained. A concussion which would leave the greater part of the motion of translation untransformed, would be likely as a general rule to produce merely a temporary star, which would blaze forth for a few years, or a few hundred years, or perhaps a few thousand years and then die out. In fact we have had several good examples of such since the time of Hipparchus. Now, although it may be true that, according to the law of chances, collisions producing temporary stars must be far more numerous than those resulting in the formation of permanent stars, nevertheless the number of those temporary stars observable in the heavens may be perfectly insignificant in comparison with the number of permanent stars. Suppose there were as many as one hundred temporary stars formed for one permanent, and that on an average each should continue visible for 1000 years, there would not at the present moment be over half-a-dozen of such stars visible in the heavens.

4th. "Such collisions as the theory assumes are wholly

It is true that we may one day be able to determine by spectrumanalysis the distance of some of the binary stars; but as yet this method has not been applied with success.

hypothetical; it is extremely improbable that two cosmical bodies should move in the same straight line; and of two moving in different lines, it is improbable that either should impinge against the other." In reply, if there are stellar masses moving in all directions, collisions are unavoidable. It is true they will be of rare occurrence: but it is well that it is so; for if they had been frequent the universe would be in a blaze, and its store of energy soon converted into heat.

II. On the Analysis of Alloys containing Copper, Zinc, and Nickel. By THOMAS BAYLEY, Assoc. R.C.Sc.1.*

THE

HE analysis of these alloys can be very rapidly effected by a combination of colorimetric and volumetric methods. The alloy is dissolved in nitric acid, and the solution then evaporated to dryness with excess of sulphuric acid to expel nitric acid, which must not be left in the solution.

Determination of the Copper.-The solution is mixed with excess of potassic iodide, which causes the formation of cuprous iodide, according to the following well-known reaction:

2 CuSO4+4KI=2K2 SO4 + Cu2 I2 + I2.

The solution containing the precipitate is then titrated with a standard solution of sodic thiosulphate. The free iodine present is an exact measure of the copper, each gram of copper being equal to two grams of iodine. The following is the result of a series of determinations of copper made under various circumstances by this method:

1408 Cu (free H2SO4 present)

1408 Cu (NiSO, present)

Iodine.

•1275

•2500

•5100

•2740

•2755

1408 Cu (ZnSO, and free H, SO, present) 2760

The solution, after the titration, is filtered and the precipitate washed. The filtrate is free from copper and contains the nickel and zinc. The absence of copper was proved in several experiments by evaporating to dryness and gently heating the residue, after which it was dissolved in a little dilute sulphuric acid and excess of ammonia added. In no instance was any blue colour perceptible.

Determination of the Nickel.-The fact will have been observed by chemists, that solutions of nickel and cobalt salts

Communicated by the Author.

are so far complementary in colour, that when they are mixed together the resulting liquid, if moderately dilute, is hardly to be distinguished from pure water. I conceived this fact might be made the basis of a method for estimating nickel and cobalt, and therefore undertook the following experiments.

A large hollow prism filled with a moderately strong solution of a nickel or cobalt salt was placed immediately in front of the slit of the spectroscope; and the thickness of the liquid traversed by the light was regulated by moving the prism until the eye could most clearly determine the dark absorption-band caused by the metal in solution. On referring to the accompanying diagrams, which show the absorption-spectra of the two metals, it will be seen that cobalt and nickel are almost exactly complementary in their relations to light. The black band of cobalt is well defined at the edges, especially at the end nearest to the red; while the absorption-bands of nickel are not so sharply defined, but fade away at each end. If the Spectrum of light passed through Co.

Spectrum of light passed through Ni and Co.

Spectrum of light passed through Ni.

The black parts here represent the bright parts of the spectrum, and vice versâ. spectra were exactly complementary, on superimposing the nickel spectrum upon the cobalt spectrum, the dark part on the one would exactly cover the light part on the other. This, however, though nearly the case, is not exactly so; for the bright band in the nickel spectrum overlaps the dark cobalt band at the end nearest to the red, although with diminished brilliancy. Consequently, when we employ a mixture of nickel and cobalt salts in solution, we do not get a uniformly dark spectrum, but an excess of light coming through at the part where the overlapping occurs, as seen in the diagram. This is why the solution obtained by mixing strong solutions of nickel and cobalt is not grey, but reddish brown in colour.

Having so far demonstrated the complementary character of the two metals, I next endeavoured to find in what proportions they must be mixed in order to neutralize each other. For this purpose a tall glass cylinder (150 cubic centims. capacity),