fii('-\v()()(l, and llie lra\ cllcr, wlu'ii iiiHlerlakiii}^ a journey licro, will iiave to carry with 

 liim (he necessary fuel. Already previous travellers through the steppe regions of Asia 

 have ol)servcd llial the tree vegetation occurring at all on the stcjipcs. is especially to 

 be found on the hilly ranges. ■1'ani-ii..ii.\v and others, who have .studied this more closely, 

 are. as is known, of opinion that one reason of the steppes wanting trees is the 

 salinity of the soil, and liial the forest will advance in ])roportion as the .soil is being 

 washed out. .\s this washing out of salt must be supposed to be going on first on the 

 height.s, at the watershed, the pioneers of the forest will accordingly find reasonable terms 

 and appear first in these places. .Mthough this supposition may possibly agree in some 

 cases, it cannot, however, be adopted for all, and for instance not for the yVbakan Steppe, 

 where, as mentioned above, formerly, in a moistcr period, the steppe was wooded. On 

 the contrary, from my own experience in other regions, I consider lliese lonely larches 

 on the ridges as the last remains of those forests in former times also extending over the 

 lower parts. 



There are, on the other hand, indications that the Ireelessness of the steppe is not 

 always due to the long-continued drought only, for trees are also wanting in places on 

 the steppe where the soil is moist or even swampy all the year round, and where the 

 water — to judge from the general character of the vegetation — is ([uite fresh. 



During my stay on the Abakan Sleppc, I collected a number of tree-samples in the 

 ground, and after my return home I have subjected them to an investigation with the 

 microscope. As the structure of these samples, thousands of years old, was, as. might also 

 be expected, somewhat decayed, I have not been able to settle definitely w'hether they 

 originate from spruce or larch, the microscopic wood structures of which, as is well- 

 known, resemble each other so much that it is difficult to distinguish them, even in fresh 

 material. There are, however, features indicative of the samples consisting of larch 

 wood. 



'r^^r ^' 



Fig. 3 I-'rom the .\bakan Sleppc near Ust .\bakansk. 



13 



