25-year sets of solar x-ray observations, measurements of plasma and magnetic field parameters in the solar wind and Dst index variations are analyzed with the purpose of revealing the factors rendering the greatest influence on development of magnetospheric storms. Value of correlation between solar flares and magnetic storms practically does not exceed a level of correlation of random processes. Furthermore it was not possible to find out any dependence between importance of solar flares and value of magnetic storms. Coronal mass ejections (CME) only in half of cases result in storms with Dst < -60 nT. The most geoeffective interplanetary phenomena are magnetic clouds (MC) which, as many believe, are interplanetary manifestations of CMEs and compressions in the region of interaction of slow and fast streams in the solar wind (so-called Corotating Interaction Region, CIR): about all 2/3 observed magnetic storms, and for medium storms with -100 < Dst < -60 nT numbers of storms from MC and CIR are approximately equal, and for strong storms with Dst < -100 nT the part of storms from MC considerably is higher.