294 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1926 
VERTICAL CLEAVAGE OF LOESS 
One of the puzzling problems is the explanation of the cleavage 
responsible for the vertical walls of the canyons and cliffs which 
form such a feature of loess districts. Many are dissatisfied with the 
theory of v. Richthofen that this is due to the lines of weakness 
produced by plant roots. 
It is true that to-day in almost any loess gully plant roots or their 
remains, often encased in a tubelike sheath of carbonate or other 
cement, may be seen piercing to several feet below the surface. A 
photograph taken near Kuo T’s’un (Huan-hua) shows such a root 
exposed by a collapsing cliff to a vertical depth of 13 feet without 
reaching the tip. 
Among the objections raised to the “rootlet theory” is the fact 
that it is curious that plants which could thrive prolifically under 
such a variety of climatic conditions fairly recently have no real 
living counterpart to-day. Such root tubes equally might be the 
result and not the cause of vertical weakness. 
As an alternative Willis (12-253) suggests the following explana- 
tion. Owing to the lightness of the grains, the dust on first falling 
was very loosely packed. Under the weight of further deposits on 
top and alternate drying and soaking by surface water sinking into 
the ground, the material settled slowly and became compressed ver- 
tically, though no corresponding lateral force existed to lessen the 
distance between grains on the same horizontal plane. Moreover, 
the closer the grains the stronger the bond between them made by 
the weak cementing action of salt-charged percolating waters. Hence 
the direction of least resistance, which moving air and water would 
tend to follow, would be the up and down direction in which the 
material now splits to-day. 
A much more convincing explanation was recently suggested to the 
writer by Dr. C. P. Berkey, geologist of the third Asiatic expedition, 
on his return from Mongolia. As far as can be ascertained, this 
particular point of view has not been emphasized by other students 
of the problem, and Doctor Berkey has been so kind as to permit it 
to be offered here for the first time.t° He points out, however, that 
the explanation is based primarily upon observation of conditions at 
present ruling in parts of the Gobi and upon certain considerations 
regarding climatic variations in recent times established by re- 
searches in Mongolia, rather than upon any extended study of the 
loess formation itself, which is not extensively developed in the 
areas explored by the expedition. 
15 Since this paper was first printed, the writer finds that a closely similar explanation 
was arrived at independently by Dr. Bailey Willis for the vertical cleavage in adobe 
deposits in Patagonia in the neighborhood of springs where the moisture was enough to 
allow vegetation to take hold. (See “ Northern Patagonia,’ Scribners Sons, New York.) 
