Some double, triple or quadruple stars could be the final stage of the dynamical evolution of small galactic clusters. This evolution is nothing but a redistribution of the kinetic and mutual potential energy of the stars. It has two main effects: the escape of stars from the cluster and the formation of tight subsystems.

The dynamical decays of random, unstable, negative-energy, n-body systems have been followed numerically, to investigate the statistics of stable decay products. From initial quadruple systems, about 20% of the products were stable triples while 30% resulted from initial quintuple systems. These numbers are consistent with the observed ratio of triples to binaries, especially for the higher-multiplicity initial systems.

It is a great honor to be asked to summarize a conference such as this one, but it is also an impossible task. For a summary I refer you to the table of contents; what I will do here instead is to offer some personal impressions of what we have been doing.

My initial feeling was that of an outsider; I have never worked in the field of double and multiple stars, to which many of you have devoted your careers. Yet this meeting has impressed me with the interrelationship of our areas. In this room, are astronomers who represent and have discussed celestial mechanics, stellar dynamics, stellar spectroscopy, and the astrophysics of star formation. I have not counted heads, but we outsiders might even outnumber you hard-core binary-fanciers. And this is, in some sense, a measure of the value of this meeting, and, even more pointedly, of the value of your work: how much does it matter to the rest of us? Every astronomer should ask himself—particularly when the technicalities get thickest— “What am I doing this for?” The answer is emphatically not merely that this particular work is interesting to do. The appeal of astronomy is not in the bricks and mortar that each of us prepares, but rather in the architecture of the structure that we build with those materials. The questions that we really pursue are not the orbits of binaries, nor the structure of star clusters, but rather the basic problems of the universe: how and why are stars made, and why do they develop as they do?