378 p. S. MARTIN 



zones of New England and Michigan, is not found in the Arctic. 

 Both Ambrosia and Ephedra, another steppe species in the Late- 

 glacial of Europe and America, present the problem of how we might 

 distinguish cool prairie from tundra in a pollen diagram. Today these 

 vegetation types are separated by a belt of woodland and forest. 

 Is it possible that they were in contact during the glacial periods? 



Perhaps the periglacial landscape was not entirely treeless. If 

 scattered spruce, larch, or jack pine grew near the ice margin, they 

 would have formed a taiga or boreal savanna. Presently the taiga 

 lies between boreal forest and treeless tundra (Rousseau, 1952 ; Hare, 

 1954). Occasionally pieces of coniferous wood are found in glacial 

 drift (Flint, 1957, p. 323). Rather than indicating that forest was 

 overridden by ice, they may mean that the glacier swept through a 

 taiga type woodland, a more plausible ice-margin environment. The 

 relatively well-known and widely discussed Two Creeks "forest" 

 bed, silted and covered by Valders ice (Wilson, 1932, 1936) is not an 

 exception. In stump diameter, taper, and growth rate the Two 

 Creeks trees resemble spruce woodland in central Ungava (see 

 Hustich, 1954, for comparative data). In brief, fossil wood is not 

 proof of forest ! 



We may expect that Full-glacial tundra, boreal forest, and 

 deciduous forest formations were not identical in species composition 

 or even in vegetational structure with their present bioclimatic 

 analogues. Nevertheless, if there is an adaptive relationship between 

 vegetation and climatic zones, it seems unreasonable to postulate an 

 azonal system during the glacial period, as Drury has done (1956, 

 pp. 80-90). The model proposed by Dansereau (1957), with nar- 

 rowed tundra bordering the ice at one point and maple or oak forest 

 at others, also does not agree with either the concept of bioclimatic 

 gradients or with Late-glacial pollen diagrams. In general the suc- 

 cession of pollen zones, tundra -^ boreal forest — » deciduous forest 

 in New England and boreal forest — > mixed deciduous forest -^ oak- 

 pine forest in North Carolina (Frey, 1953)^ shows, I believe, the 



2 My interpretation of boreal forest in North Carolina is based on Frey's pollen zone 

 M2 in Singletary Lake and Jones Lake J-1. This reveals dominance of pine, including 

 many small grains, with up to 9% spruce, 7% oak, 1% birch, and 1% hickory. In fair- 

 ness to Frey (1953, 1955) it should be noted that, although he regarded his results as 

 evidence of climatic change, he does not advance the hypothesis of Boreal Forest in the 

 Carolinas. The case I would make for Boreal Forest rests chiefly on the small but crucial 

 percentage of spruce and the scarcity or absence of broad-leaved species. 



Another authority on this region, D. R. Whitehead (personal correspondence) takes 

 strong exception to such an interpretation of Frey's work, noting: (1) Size-frequency 



