The article indicates how three facts concerning the long-period comets, which hitherto were not well understood, namely the random distribution of orbital planes and of perihelions, and the preponderance of nearly-parabolic orbits, may be considered as necessary consequences of the perturbations acting on the comets.