'90 



KN OWLEDGE. 



[May 1, 1898. 



Should the reader object that these electrical effects 

 would not be likely to occur on a rainy night, he may be 

 reminded that every shower is an electrical phenomenon, 

 and that the rain itself is sometimes luminous. Thus, 

 an officer of the Algerian army states that during a storm 

 on the 25th of September, 1840, the raindrops that fell on 

 the beards and moustachios of the men were luminous. 



The author also refers to a luminous appearance which 

 he witnessed one night in the Isle of Ely. It advanced 

 rapidly, and distinctly touched his cheek, when he per- 

 ceived a sensible glow, but his breath seemed to make it 

 bound away again. This appears to have been a case of 

 marsh gas, but the author thinks that it — namely, the 

 ignis fatuus — is simply " a luminous air, a phosphoric light, 

 arising from the oily particles of decomposed aquatic 

 plants ... or from the gas of the many resmous 

 pine trees which once grew thei'e, and whose remains still 

 lie buried in the peat." 



But the most singular case remains to be told. A brig 

 off New Zealand, having been visited by some electrical 

 brush discharges, the author accounts for them in the 

 following manner : — " These marine lights most probably 

 arise from the decomposed remains of fish, raised to the 

 surface by the violent motion of the water in storms." 



Yours faithfully, 



C. ToMLINSON. 



WHAT IS A STAR CLUSTER? 



By A. C. Eanyard. 



IT is hardly possible to conceive that a star cluster such 

 as that shown in our plate can form a permanent 

 system, in which the stars move on from age to age 

 without collision. The facts that we are already in 

 possession of, with regard to the motion of binary 

 stars, seem to render it very probable that the reign of 

 gravity extends to stellar distances, though it must be 

 admitted that we have no indisputable evidence that 

 gravity acts across the interstellar spaces, or even that the 

 apparently elUptic paths, which observation shows that the 

 components of a binary star describe about one another, 

 are swept out under the action of a force varying as the 



inverse square of the distance from their common centre of 

 gravity. But in view of the facts with which we are 

 familiar in our solar system, this is the simplest assumption 

 to make, and the observed motions of such binary systems 

 are what gravitation easily accounts for. 



More that fifty years ago Sir .John Herschel saw the 

 extreme difficulty of imagining the " conditions of con- 

 servation of such a system as that of u> Centauri, or 47 



Fio. 1. — Uiitouolicd block made b_v a pliotograpliie procesa from a 

 photograph of the Groat, Cluster in Hercules, taken bv tlic Brothers 

 Henrv. 23ril .Tunc, 1886. 



Fig. 2. — LTntouched block made by a photographic process from a 

 photograph of the Great Cluster in Hercules, taken witli the 

 36-in. refractor of the Lick Observatory, 24th Septemljer, 1892. 



Toucani, &c., without admitting repulsive forces on the one 

 hand, or an interposed medium on the other to keep the 

 stars asunder."* But if we put on one side assumptions 

 as to repulsive forces, and endeavour to consider the 

 changes which wiU probably take place under the action of 

 attractive forces such as we are familiar with in our region 

 of space, we shall be forced to admit that a spherical star 

 cluster cannot be a permanent system, and that the 

 individual stars of the cluster can only escape collisions 

 for a limited period of time. Next let us proceed to 

 consider the results that will probably follow such cata- 

 clasms. At every such collision, motion of translation 

 will be turned into heat, and a change in the physical 

 condition of the colliding bodies, or a part of them, may 

 be expected to follow. 



Prof. George Darwin has pointed out that if two solid 

 bodies were to collide with planetary velocities, there would 

 be a very rapid development of gas between them at the 

 region of contact, which would cause them to rebound from 

 one another almost as if they were perfectly elastic bodies. 

 If the moving bodies were liquid or gaseous, no doubt a 

 similar evolution of heat would take place at the region of 

 contact with an explosive development or expansion of 

 gas, that would drive the masses away from one another, 

 causing them to rebound with velocities almost equivalent 

 to their velocities of approach ; and it seems probable that 

 within a short period after such a collision, the gaseous 

 matter evolved at the region of contact would be distributed 

 in space between the rebounding bodies. 



Now let us pass to what is actually observed. It has 

 long been remarked that globular clusters frequently 

 present a partially radiated appearance in their outer parts, 



* " Cape Observations," p. 13'J. 



