344 THE EMPTY QUARTER damp and clammy about us and the north-east wind, which had blown throughout the night, chilled us as we marched. We were still in the gravel plain of Abu Bahr, no longer however a dead flat wilderness, but rather a low gently undu- lating steppe, bare and desolate with streaks of sand and here and there an outcrop of rocky ground, which was a welcome relief from the monotony of gravel. We had scarcely been marching ten minutes when we came upon a strip of withered Eadh bushes, which would have given us coffee the previous night had we marched but a little further. Now, however, it did not tempt us to stop. It is better, said Zayid, to go on yet a bit that, perchance, we may find fodder for the camels as well as fuel for our coffee. So we marched on past two or three mounds of the outcropping rock, some 20 feet above the plain level, and noticed signs of a recent flood. Beyond them we came to a light spread of sand that marked the transition from Abu Bahr into Rumaila, whose first range of dunes lay upon our right hand, while the gravel plain spread out on our left, gently undulating, dotted with patches of rock and streaked here and there with tongues of sand. Almost immediately we came upon a few green bushes of' Arrad, the first edible vegetation encountered in 48 hours, and no power on earth could have got our camels past them. Their hunger was terrible to watch, and we gave them their heads to do as they willed until they had made an end of the scanty herbs. Then we passed on into the real sands of Rumaila, whose gently-billowing bosom was scarred here and there by exposed patches of the underlying bed-rock. Here was much Eadh, though withered and dead, and Al Aqfa went off on her own, searching hungrily for tracks of hares. Rumaila, a tract of broad sandy valleys separated each from the other by long waves of dunes, is generally regarded as a southerly continuation of the Dahna, from which it is said to be separated at a day's journey north- ward of our line of march by a MarbaJch or flat, sandy plain .-besprinkled with Hadh, Abed and other herbage. In Ru-' madia itself there appears to be no Abal but we came upon dry tufte of BirJcan and Sahat stubble.