408 
ASTRONOMY: H. N. RUSSELL 
sity. But the problem is capable of investigation by observation. 
There are many close eclipsing pairs in which the components are ellip- 
soidal in form, as is proved by variability of the Beta Lyrae type. In 
such systems the lines of apsides of the orbits should advance, at a 
rate depending on the masses, dimensions, and internal constitution 
of the components. If the last is like that of Jupiter or Saturn, the 
advance of periastron should be rapid. What little evidence there is 
indicates a slower motion, and hence a very strong central condensation; 
but more intensive studies are necessary before definite conclusions 
can be drawn. There are several systems for which the necessary data 
concerning the dimensions and forms of the orbits and the stars are 
accessible to suitably planned observations, — notably a Virginis and 
U Herculis. A careful study of such stars, by means of simultane- 
ous photometric and spectroscopic observations, would be remunerative. 
The singular and so far inexplicable changes which occur in the 
periods of most eclipsing variables, and so far have defied prediction, 
also deserve extended study; and Eddington has recently called atten- 
tion to the fact that secular changes in the periods of Cepheid vari- 
ables are likely to give a clue to the rate of stellar evolution.^^ The 
first scanty evidence points to a very extended time scale. 
12. In the investigation of star-clusters, measures of position, for the 
purpose of detecting future proper motions, are obviously a duty to pos- 
terity. There is little chance that anything more than the motion of the 
clusters as a whole will be perceptible in our generation, and only meas- 
ures of the utmost attainable accuracy and freedom from systematic error 
are likely to be of use to the astronomers of the future. Of far more 
promise are studies of the distribution of the stars within the clusters, 
their magnitudes, and, above all, their color indices. Such investiga- 
tions, in Shapley's hands, have given us for the first time a true con- 
ception of the distances and magnitudes of the globular clusters. Stu- 
dents of the subject are eagerly awaiting the detailed publication of the 
evidence on which he bases his conclusion that the apparent avoidance 
by these clusters of the region within 1500 parsecs of the galactic plane 
is due to a real absence of clusters from this region, and not to obscura- 
tion by absorbing matter. 
The variable stars in clusters also deserve further attention. Those so 
far discovered appear to belong to the Cepheid type, which is natural, 
as these seem to be actually the brightest of all variables. Long period 
and eclipsing variables may yet be discovered among the fainter stars. 
Good work can still be done also upon the irregular clusters, — as is 
shown by Triimpler's^^ study of the outlying members of the Pleiades. 
