Consider, now, the disposition of these seven pairs of lines. All of them lie in one semicircle about the Luci, beginning with the Protonilus on the east and ending with the Deuteronilus on the west. Furthermore, all follow approximately arcs of great circles, except the Djihoun, and all send one of their twin lines to one Lucus, one to the other. The data are enough to make this statement possible. Although the west line of the Naarmalcha has not been caught entering its oasis, the east one has been seen to enter the other, and the width of the double shows that the west one must enter the corresponding spot. In the case of the Astaboras the double has only been observed as far as the Vexillum, but the south line has continued on to the west Ismenius, and here again the width makes it certain that were the canal double throughout, the other line must enter the east Ismenius. From the base line of the Proto-Deuteronilus the inclinations of the seven pairs are as follows:— Protonilus 0° Due East Astaboras 40° North of East Naarmalcha 70° North of East Euphrates 80° North of West Hiddekel 55° North of West Djihoun 0° North of West Deuteronilus 0° Due West Now, the width between the two lines of the four canals to the east increases regularly from the Protonilus round; the Protonilus being the narrowest double, the Astaboras the next, the Naarmalcha the next, and the Euphrates the widest. And from the width between the twin oases, it would seem that they severally enter the centres of them. What takes place in the case of the Hiddekel, which is wider than its tilt would imply, and in the Djihoun, which is narrower, is not so clear. But that they enter the oases in some place is certain. [Illustration: _Lucus Ismenius. March 1903._] The spots make common termini for all the canals of a given neighborhood. In other words, canals converge to the places occupied by the spots and do not cross haphazard according to the laws of chance. Only one instance exists where a spot fails to gather to itself the whole sheaf of canals and even there it collects all but two. This anomaly is the Pseboas Lucus. The peculiarity of this oasis is that it lies not on, but just off, the Protonilus. That it does so is exceeding curious, considering that it is the sole example of such extra-canaline position. Strictly speaking, it is not the Protonilus but the point where the Protonilus turns into the Nilosyrtis to which it stands thus neighboringly aloof. And this may explain the anomaly. For the Nilosyrtis has not the full geometric regularity of the normal canal, and seems to have been a more or less fundamental feature of the region.