28 A, W. Roberts.— Variable Star observing and [Oct. 28, 
a long period variable approaches a maximum its red colour becomes 
less intense, the remarkable variable R Carine being cited as an 
instance of this change. 
If the increase of light be due to the increase of heat this is just 
what we would expect. 
A strange connection exists between the intensity of the minimum, 
the duration of minimum and the whole period of variation, especially 
with regard to Algol variables. The connection, also, between the 
spectra of stars, their colour and their variation will no doubt be a 
future means of new discoveries. 
But into these interesting questions we cannot go; nor, although 
-we would fain do so, into the now celebrated meteoric theory of Mr. 
Lockyer, as an explanation of the phenomenon of stellar variation. 
Notwithstanding all Dr. Huggins has to say in favour of spectroscopic 
results the time has not yet come when one can build up a theory 
of the universe upon them. 
But the search after new variables though most interesting and 
engrossing as far as the observer is concerned is not the most 
important section of variable star work. The systematic careful 
observation of known variables is, as far as the southern variables 
-are concerned, a great desideratum. 
The number of Cape cireumpolar variables—or rather of variables 
south of 30° south declination is 35. 
Twenty-one of these have been discovered by Dr. Gould while 
engaged on his great catalogue of the Southern Heavens. Six of 
these are invisible in ordinary binoculars. Five are visible all through 
their variations, four are visible at maximum but not at minimum, 
-and twenty-one vary between the 6th and 9th magnitudes. 
For the observation of known variables Argelander’s method is 
perhaps the best. Call the variable star under examination “a” and 
-eall two other stars in the neighbourhood of “a,” one of which is 
slightly brighter than “a” at max. and the other slightly less than 
“qa” at min., “b” and “ce.” Consider the distance between “6” and 
“e” as 10—the absolute value of this “ten” can afterwards be 
obtained. Then the various ratios of @ to 6 and ¢ can be thus set down: 
b—1la—9ec (From 6 to a=1 step: from a to c=e9 steps) 
or b—2a—8e (From 6 to a=2 steps: from a to c=8 steps) 
or b—3a—7e (From 6 to a=3 steps: from a to e=7 steps) 
always making the sum of the co-efficients (if we may ao call them) 
==10. . 
