SCHUMACHER ON KJOKKENMODDINGS OF OREGON. 29 



north bank of the river. Over the main rancheria, marked by a thick 

 ] ayer of kjokkenmoddings, we find the usual obstruction, an orchard ; while 

 across the rivulet (the efflux of a spring issuing but little over 150 yards 

 farther up on the rocky rise), the house-sites remained well defined ; which 

 we also notice 150 yards farther up the river, in an indentation of the 

 steep shore, and still in another similar nook at a distance of 100 yards 

 farther on. These places were still inhabited at the time of the Rogue 

 Eiver war in 1856, when here, on the left bank of the river, just across 

 from the main rancheria, peace was accepted by the leader of the United 

 States Army, and the Indians were accordiogly removed to reservations. 

 The present owner of the land and ferry, a " squaw-man ", liberally gave 

 us permission to dig in his orchard, where all signs of former houses 

 were obliterated by the plow and obstructed by high weeds and trees. 

 Although we made a careful search for graves, the many test-holes we 

 dug revealed ouly sites of houses; the kitchen-refuse consisting of all 

 kinds of shells (see Smithsonian Report of 1873) and a great many bones 

 of elk and of deer, averaged about 8 feet in depth at the main station, 

 while none were found across the rivulet on the rocky ascending bot- 

 tom, where it is likely the rains had washed them into the river, and 

 very few, not enough even to form a layer, at the two upper town-sites. 

 The houses we excavated were square ; that is to say, the subterranean 

 part reached to a depth of about 4 feet below the surface, and measur- 

 ing variously from 6 to 10 feet square. The casing of the excavation 

 consisted of boards arranged horizontally, contrary to the vertical iDosi- 

 tion in the houses of the present Klaujath Indians, and was kept in its 

 place by jjosts along the front. The general impression which the traces 

 of an old aboriginal town-site makes is that of a group of huge mole- 

 hills inverted or sunk to a small rim at its base (Sketch A) [Plate 4]. 

 Although the excavation was found to be square, the remaining con- 

 cavities, always shallow, and hardly evermore than 3 feet deep, were circu- 

 lar, which is attributed to the circular embankment that still surrounds 

 it, and to the natural action of the elements in filling up a depression in 

 loose ground. No doubt, the superstructure of the hut was of a circular 

 shape, corresponding to the remaining embankment, and was probably 

 placed in such a manner as to meet conically, and was covered with 

 earth, &c. The fireplace we find on one side of the floor in a small ex- 

 cavation, and the smoke escaped through a draft-passage, as shown in 

 section sketch (B) and the plan (G) [Plate 4]. We find among these 

 house-sites a few well-preserved ones, exceptionally with square em- 

 bankments (compare the sites of the first branch settlement. Map 2 

 [Plate 3]), but they are no doubt of recent date, and a modification be- 

 tween an aboriginal hut and a white man's shanty, such as we had occa- 

 sion to witness among the i^resent Klamatbs at the mouth of the Kla- 

 math River, one of which I show in sketch (D), as also an inner view (E), 

 a plan (F), and a section (G) [all on Plate 5J. The inner view shows the 

 depression, which is in this case pentagonal, incased by boards j)laced 



