590 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1959 
Nevertheless the artifacts in the two layers are alike; even the types 
of the vessels are identical. Further, it was proved that a potsherd 
found in the upper culture layer is part of a sherd found in the lower 
layer, the two surfaces of the fracture matching exactly. There is 
no doubt, therefore, that the culture remains in the two layers are 
contemporaneous, though they are separated by a deposit about 40 
em. thick. Another significant fact is that there are cracks in the 
peat layer separating the two culture layers, and in some of the 
sections it can be seen how things from the upper culture layer have 
fallen through these cracks and have spread in the lower layer. To 
this we can add another observation throwing light on the problem. 
The pollen analysis shows that the upper part of the substratum (the 
former lake bottom) has exactly the same age as the lower part of 
the peat and gyttja layer that separates the two culture layers. Pol- 
len analysis from the lower culture layer shows, however, that this 
is considerably younger than the layers surrounding it; 1.e., younger 
than the layer it rests on and also younger than the one directly above. 
The only possible explanation is that the upper culture layer orig- 
inates from a dwelling place on a floating island. The inhabitants 
of the island threw their rubbish into the lake, and the waves washed 
a part of these culture remains underneath the floating island, together 
with leaves and branches, and finally the interspace between the float- 
ing island and the former bottom of the lake was filled with washed- 
in plant and culture remains. 
But how is such a floating island created? As mentioned above, 
the pollen analysis shows that the upper part of the substratum and 
the lower part of the floating peat island are of the same age, which 
means that the peat island at some previous time must have lain 
directly upon the layers which formed the lake bottom; namely, 
before the peat rose as an island. The reason for the peat breaking 
away from its base must be a rise of the water level in the lake. It is 
characteristic of bog plants that their roots are hollow and filled 
with air. These cavities are in reality a kind of oxidation system 
that enables the plants to grow in peat poor in oxygen. When such 
peat formations are covered with one-half to 1 meter of water, the 
buoyancy will be sufficient to make the peat layers break away from 
the bottom and rise like a ball which has been held under water (fig. 
10). Such floating peat islands are still being formed; e.g., when mill 
ponds that have been drained for some time are dammed (pl. 2, fig. 2). 
THE LOCAL VEGETATION 
The plants that grew on the floating island and along its shores 
have left traces such as seeds and fruits (fig. 9). On the dwelling place 
itself the vegetation probably was rather poor, and only some grass, 
perhaps some moss, was able to resist the tramping feet. A few larger 
