Ambartsumyan's assumption that during the formation of the various structural elements the galactic centers eject fragments in various planes may be to a certain extent substantiated by a number of specific examples. There occur structures in which the major axes are variously orientated in the same plane, as well as some systems, like M82, for instance, in which there are streams of matter directed perpendicular to the principal plane of symmetry.
We consider as more probable the hypothesis that the components of multiple galaxies originate jointly in close proximity and proceed to draw apart in the course of their formation. Some kind of repulsion sometimes grows up between them, causing their mutual recession in spite of gravitation. The material experiencing the strongest repulsion forms tails. The high viscosity of the stellar systems, which must be assumed, leads to the formation of bars. The stability of the tails and bars is the same as that of spiral arms, and is thus considerable. The nearness of a galaxy disturbs for some reason the development of a spiral structure and leads to its "becoming disfigured". The phenomena responsible for the repulsion of interacting galaxies must be related to the phenomena causing the mutual recession of galaxies in groups.
It should be recalled that while the Metagalaxy is made up of galaxies, it represents at the same time a unique type of continuous medium in which the galaxies are condensations. Experimental physics has never had to deal with anything even remotely similar to it. Phenomena may therefore be discovered on a galactic scale which are utterly unexpected and as yet incomprehensible.
It is not impossible that paired galaxies should have properties quite different from those associated only with the attraction of their combined masses, similarly to the way in which the properties of molecules differ from the properties of their constituent atoms.
Certain pairs of living organisms are capable of reproduction, certain pairs of atoms exhibit new properties when combined, etc., and these properties cannot be deduced from a knowledge of the properties of the individual components. There may be something similar in the galactic world as well. This seems to be borne out by the fact that, according to Zwicky, galactic clusters apparently do not interact gravitationally with each other.