ASTRONOMY: E. DOOLITTLE 
137 
A STUDY OF THE MOTIONS OF FORTY-EIGHT DOUBLE STARS 
By Eric Doolittle 
Flower Observatory, University of Pennsylvania 
Communicated by H. H. Donaldson. Read before the Academy, November 21, 1917 
The double stars examined in the following investigation will all be recog- 
nized as old and well-known pairs, on each of which a large number of measures 
have accumulated. Each of these has shown a considerable motion since 
discovery, but although orbits have been computed for several of them, there 
is no case in which certain knowledge of the true form of the orbit can be had 
at this time, nor with the greater number of these pairs is it even certain that 
the motion is orbital at all. It was, in fact, in view of this uncertainty that 
the following computations were undertaken. The complete list of double 
stars at present known to us was examined, and all of those in which, not- 
withstanding the large number of measures made upon them, it seemed uncer- 
tain whether the motion is rectilinear or orbital were selected for study. The 
primary object was to ascertain, when possible, in which pairs rapid motion 
during the next few years might be reasonably expected and those in which 
the distance is now near a maximum value. It is upon such pairs that observa- 
tions are most urgently needed, while upon those for which the analysis gives 
no indication that the companion is now near a critical part of the orbit the 
multiplication of observations is at the present time unnecessary. 
It is well known that w^ith the type of pairs here considered, in which less 
than two quadrants have been described by the companion, a series of widely 
different orbits may be obtained, any one of which may satisfy the observa- 
tions within the limits of error naturally to be expected; and this, even al- 
though it may afterward appear that the motion is only rectilinear and the 
pair is not a true binary system. It was therefore thought best not to begin 
each computation by so adjusting the six constants of the orbit that the 
observations should be represented by them, but rather to assume that the 
motion is rectilinear, and having found the straight line which best represents 
all the observations to examine in each case whether there is definite evidence 
that the motion of the companion in the straight line is, or is not a uniform 
motion. If a continuous progressive change is found, greater than can be 
assumed due to errors of observation, it is evident that the motion is not 
rectilinear, while in the many cases in which the variations are less, or but 
little greater, than might be attributed to errors of observation the evidence 
for the binary character of the system must be regarded as inconclusive. 
The first step in each case was to form suitable means of the numerous 
observations. Among the earlier measures, those of Maedler and also those 
of the few occasional observers were rejected, especially since in no case were 
there wanting accurate observations by 2, OS and A during the same intervals. 
Practically all modern observations were, however, included, and these were 
