You will require of me that I should explain the centres of radiation so conspicuous here and there on Lowell's map. The meeting of more than two lines at the oases is a phenomenon possibly of the same nature and also requiring explanation. In the first place the curves to which I have but briefly referred actually give rise in most cases to nodal, or crossing points; sometimes on the equator, sometimes off the equator; through which the path of the satellite returns again and again. These nodal points will not, however, afford a general explanation of the many-branched radiants. It is probable that we should refer such an appearance 193 as is shown at the Sinus Titanum to the perturbations of the satellite's path due to the surface features on Mars. Observe that the principal radiants are situated upon the boundary of the dark regions or at the oases. Higher surface levels may be involved in both cases. Some marked difference in topography must characterise both these features. The latter may possibly originate in the destruction of satellites. Or again, they may arise in crustal disturbance of a volcanic nature, primarily induced or localised by the crossing of two canals. Whatever the origin of these features it is only necessary to assume that they represent elevated features of some magnitude to explain the multiplication of crossing lines. We must here recall what observers say of the multiplicity of the canals. According to Lowell, "What their number maybe lies quite beyond the possibility of count at present; for the better our own air, the more of them are visible." Such innumerable canals are just what the present theory requires. An in-falling satellite will, in the course of the last 60 or 80 years of its career, circulate some 100,000 times over Mars' surface. Now what will determine the more conspicuous development of a particular canal? The mass of the satellite; the state of the surface crust; the proximity of the satellite; and the amount of repetition over the same ground. The after effects may be taken as proportional to the primary disturbance. 194