With the carets something of the same sort seems to take place—if we may consider as betokening a general law the fact that in 1894 the carets at the mouths of the Phison and Euphrates developed before their affiliated canals. But about them much less is yet known, and we must be content to say that the observations of 1905 made at the opposite season of the canal’s year seem to bear this out. PART IV EXPLANATION CHAPTER XXVIII CONSTITUTION OF THE CANALS AND OASES As rational science does not rest content with raw results, it now becomes obligatory, by marshaling the facts to suitable discussion, to seek to find out what they mean. Now, so soon as we scan these phenomena for some self-interpretation, we perceive one characteristic of the lines which at once appears to direct us to their nature and justifies itself as a signpost with increasing certainty as we read on. This trait is the very simple yet most significant one of showing intrinsic change: the lines alter in visibility with time. This primary proclivity we do not even need the cartouches to establish. That the lines change is palpable to any one who will watch them long enough. Schiaparelli was struck by the fact early in his study of the planet, and it forces itself on the notice of any careful observer who compares his own observations with one another at intervals. But though the cartouches are not needed to a first revelation of mutability, they serve to certify and precise it to much further information on the subject. For, that these changes are not extrinsic, that is, are not caused by varying definition, distance, or illumination, they make patent even to those who have never seen the things themselves by disclosing respective differences of behavior in lines similarly circumstanced optically. The change is therefore intrinsic, and the question arises to what can such intrinsic change be due. In searching for cause, attention is at once attracted by another series of transmutations that manifests itself upon the disk, in the orderly melting of the polar caps. For the existence of the two sets of metamorphoses suggests the possibility of a connection between them. The inference is strengthened when we note that not only are both periodic, but that furthermore the period of the two is the same. Each polar cap runs through its gamut of change in a Martian year; the canals also complete their cycle of growth and decay in a Martian twelvemonth. The only difference between the two is that each polar cap has but one maximum and one minimum in the course of this time, while most of the canals have two of each, though the maxima are not alike nor the minima either.