500 
PACIFIC SCIENCE, VoL VII, October, 1953 
Fig. 5. Diagrammatic profile of part of a typical 
"slumped” stream bed and underground channel. Such 
stream tunnels, A-c, are fairly common in the Aleutians 
and can be dangerous should a person fall through the 
ceiling B. The lower outlet A and the upper entrance 
C are usually choked with ferns and other plant growth. 
The tunnel ceiling usually possesses a meadow vege- 
tation, consisting of grasses, umbellifers, Streptopus, 
Claytonia, etc. At d a series of terraces are formed from 
a former tunnel ceiling which slumped gradually and 
finally caved in. The vegetation here is usually bog- 
like, with Sphagnum, Carex, Juncus, liverworts, and 
Saxifraga predominating. 
the ditch so that the stream flows through an 
underground tunnel (Fig. 5). These places 
can be quite dangerous. Hulten (1937^) men- 
tions falling through the ceiling of one such 
tunnel. Hidden streams may persist this way 
for a considerable time, but usually the over- 
hanging mat is penetrated by surface drainage, 
holes reappear, the overhanging mat slumps, 
and the ditch becomes a series of depressions 
or wet terraces composed of sphagnum and 
typical bog plants, a much different vegeta- 
tion than that which formerly occurred in the 
ravine. 
In the Aleutians, this instability of sub- 
stratum and plant cover does not have precisely 
the same causes as on the mainland. Perma- 
nent subsurface ice (permafrost), for example, 
is not the very important factor that it is over 
widespread areas of Siberia, mainland Alaska, 
and Canada. Rather, inst;ability is probably a 
result of the interaction of the edaphic and 
climatic factors which make the Aleutians 
more or less unique: namely, a comparatively 
thin soil mantle largely of volcanic ash, more 
or less continuous ash deposition, temperate 
conditions with heavy precipitation, cool 
summers, and unusually high, rather constant 
winds. 
Kellogg and Nygard (1951) have charac- 
terized the Aleutian substratum as "tundra 
without permafrost” and composed of silty 
or sandy loam with ash. There are, however, 
small areas of permanent subsurface ice in the 
Aleutians which may be fairly extensive albeit 
sporadic and seemingly of little influence on 
surface vegetation. I observed one such area 
on Great Sitkin Island where Glacier Creek 
has cut through an ancient mud flow. Ice is 
exposed in the profile about 20-30 feet below 
the surface of the old flow. Thick sheets and 
occasional lenses of ice can be traced along 
the floor of the stream bed for distances of 25 
to 50 yards. Strictly speaking this is not per- 
mafrost but relict glacial ice. The surface 
topography and vegetation are unmodified, 
and it is presumed that the depth at which 
the ice occurs is too great to affect plant 
growth above. Melting of the ice masses, 
however, will undoubtedly cause local land 
slumps leading to the formation of kettles 
and, thus, different microhabitats which will 
have their peculiar plant communities. 
RECENT PLANT INVASIONS 
The foregoing examples of fluctuating veg- 
etation in the Aleutians illustrate the spread- 
ing ability, i.e., plasticity, of some plants of 
this area. This ability has been noted by 
Hulten (1937^) who writes that on Bogoslof, 
the "jack-in-the-box” island that has appeared 
and disappeared a number of times in the last 
150 years, a pioneering vegetation is already 
well established. It includes Lathyrus mari- 
timus, Elymus arenarius subsp. mollis, Honckenya 
peploides major, Euccinellia pumila, and Merten- 
sia maritima, all of which are widespread lit- 
toral plants. Senecio psetido-arnica also occurs 
on the still- warm shores. Our own field par- 
ties studied briefly several pioneer plant com- 
munities near a small glacier on the upper 
