THE VALLF.Y OF RAU. 415 



day, we found the bottom of the valley abounding 

 in rich, vegetation, though that was three hundred 

 feet lower than this place, because that valley is so 

 short that the air has no room to become heated to a 

 dry simoom, which can mther the vegetation as it 

 sweeps along. It is, therefore, in this valley that the 

 simoom is formed, not on the high mountains that 

 border it or on the adjacent ocean. 



2£arch 1st. — Left Rau at 6 a. m., for we have an- 

 other long day's journey before us. As yesterday, 

 the road led along the bottom of the valley, but soon 

 a range of mountains appeared before us, and we 

 began to ascend along the side of a deep ravine. The 

 rock here was exposed, and proved to be a soft sand- 

 stone covered with clay. Here we came to a third 

 water-shed two thousand one hundred and fifty feet 

 high, and could look back do\vn the valley of Rau to 

 the southeast. Its lenoih in a rio-ht line, from this 

 water-shed to that at the gorge near Lubu Sikeping, 

 is thirty geographical miles, but, instead of being 

 straight, it curves to the northeast, and is of a 

 crescent form, A\ddest in the middle, and gradually 

 narrowing toward the extremities. In its broadest 

 part it is not more than six or eight miles wide. We 

 now turned to the northwest, and began to descend 

 into another valley, that of Mandeling. Here the 

 mountains are quite devoid of forests, and only cov- 

 ered with a tall, rank, useless grass, the Andropogon 

 caricosum. 



At Marisipongi, the first village we came to in 

 this valley, Ave found we were among an entii'ely new 

 people, the Battas or Bataks. They also belong to 



