244 NEW STARS IN ANDROMEDA AND ORION. 
the star had fallen to the tenth magnitude. It has now 
entirely disappeared. 
If we compare recent astronomical catalogues with 
early records, there would appear to exist many cases 
of disappearance of stars. In the majority of such 
cases, however, it is safe to assume that the want of 
accuracy in recording positions among early observers 
will explain the apparent vanishing of the star. But 
some such disappearances are undoubtedly authentic. 
For instance, there are a few stars recorded by so recent 
and careful an observer as Sir William Herschel, which 
are no longer found in the given places. 
Variable stars are divided into various classes. Prof. 
Pickering gives five, placing temporaries in the first. 
The second class includes long period variables, whose 
changes run through several months. A third, short 
period variables, the duration of change not exceeding 
afew days; a fourth, stars undergoing slight changes 
so irregular that no period can be suggested, such as 
Alpha Orionis. Into the fifth division fall the few stars 
of the Algol type (the eighth quite recently discovered) 
whose variations are remarkable, and yet of such regu- 
larity that the periods can be determined within a few 
seconds of time. The Algol class probably bears no re- 
lation to our subject of new stars. 
Ceti, called Mira Ceti, is one of the most interesting 
variables of long period. At its maximum, it has at- 
tained to the third and even to the second magnitude 
and at its minimum it passes beyond visibility to the 
eighth or ninth magnitude. Its period is about eleven 
months, though quite irregular in its duration. About 
forty days after it becomes visible it attains its maxi- 
mum and about sixty days following maximum it again 
becomes invisible ; that is, longer in losing light than 
gaining it. At its maximum in 1886 it reached only the 
4.9 magnitude. This is almost the lowest recorded, 
128 
